Original article

The search for TCP analogues binding to the low affinity PCP receptor sitesin the rat cerebellum

Jacques Hamona, Florence Espazea, Jacques Vignonb, Jean-Marc Kamenkaa*aCRBM, CNRS – UPR 1086, Ecole Nationale Supérieure de Chimie, 8, rue de l’Ecole Normale, 34296 Montpellier cedex 5, France

bINSERM U 336, Ecole Nationale Supérieure de Chimie, 8, rue de l’Ecole Normale, 34296 Montpellier cedex 5, France

(Received 18 December 1997; revised 25 September 1998; accepted 8 October 1998)

Abstract – With the aim of obtaining selective ligands of the low affinity binding sites of [3H]-1-[1-(2-thienyl)cyclohexyl]piperidine([3H]TCP) in the rat cerebellum, oxygen and sulfur atoms were introduced in the TCP structure and derivatives to obtain analogues with alowered lipophilicity. These compounds, and others already obtained, were assayed comparatively to determine their affinities for three siteslabeled with [3H]TCP: one in the forebrain, the originally described PCP receptor, and two in the rat cerebellum. Lowering the lipophilicityand modifying the hetero-aromatic moiety yielded some ligands with increased affinity for the low affinity sites in the rat cerebellum anddecreased affinity for the high affinity sites in the forebrain. Particularly, two compounds displaying both a high affinity and a good selectivitymight be valuable tools to elucidate the pharmacology of the low affinity PCP sites labeled with [3H]TCP in the rat cerebellum. © Elsevier,Paris

TCP / TCP analogues / PCP receptor subsites / rat cerebellum

1. Introduction

[3H]-5-methyl-10,11-dihydro-5H-dibenzo[a, d]cyclo-hepten-5,10-imine ([3H]MK-801) and [3H]-1-[1-(2-thienyl)cyclohexyl]piperidine ([3H]TCP) are the mostfrequently used ligands for the labeling of the non-competitive antagonists binding site within the N-methyl-D-aspartate (NMDA) receptor associated Ca2+ channel(the PCP receptor) [1, 2]. However, competition dataanalysis using a two-site model revealed that both ligandslabeled at least two different binding sites in the rat brain:(i) high affinity sites, well represented in the forebrainand corresponding to the initially discovered PCP recep-tor and (ii) lower affinity sites, more abundant in thehindbrain and particularly in the cerebellum [3–8]. Thenature and the role of the low affinity binding sites arepoorly documented mostly because selective ligands arenot available. Particularly, their possible role in neuronalprotection is unknown whereas the high affinity sites is awell-established target for neuroprotective agents [6].Thus, the finding of ligands binding selectively and

potently to the low affinity sites might be crucial for theirpharmacological characterization. For clarity, the differ-ent binding sites discussed here will be marked PCP1 (the[3H]TCP high affinity binding sites in the forebrain, i.e.,the PCP receptor within the NMDA receptor associatedionic channel), PCP2 (the [3H]TCP high affinity bindingsites in the cerebellum), and PCP3 ([3H]TCP low affinitybinding sites in the cerebellum).

We have previously introduced oxygen and sulfuratoms in the cyclohexyl and piperidinyl moieties of theTCP structure to obtain analogues with a lowered lipo-philicity. In rat forebrain membranes, the affinity of theseanalogues for the PCP1 sites labeled with [3H]TCP wasdecreased as a function of lipophilicity: the lower lipo-philicity was, the lower affinity was [7]. However, theiraffinity for PCP2 and PCP3 sites in rat cerebellummembranes was not studied. The possibility that loweringthe lipophilicity decreased the affinity for PCP1 in theforebrain while increasing the affinity for PCP3 in thecerebellum was attractive since the lipophilic TCP andMK-801 (log P = 4.56 ± 0.30 and 3.71± 0.39 respec-tively) displayed very low affinities for PCP3 (table I).Thus, we have decided to investigate this hypothesis by:*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 125−135 125© Elsevier, Paris

(i) preparing TCP analogues with varied lipophilicities bymeans of O, S or hydroxyl substitution; (ii) using thesenew TCP analogues and some of those previously ob-tained for competition measurements in rat forebrain andcerebellum membranes labeled with [3H]TCP; (iii) deter-mining the more probable model of interaction (single- ortwo-site) in the cerebellum and comparing the affinities ofcompounds for PCP1, PCP2, and PCP3.

2. Chemistry

Two different synthetic strategies were used to obtainTCP analogues according to the presence or absence of amethyl substitution in the cyclohexyl or hetero-cyclohexyl moiety (X, Y-substituted ring, seetable II).The unsubstituted compounds were easily obtained bymeans of a Bruylants reaction [9]. It consists in thereplacement of a cyano group by an aryl- or hetero-arylgroup by the reaction of anα-aminonitrile with an aryl-or hetero-arylmagnesium halide(figure 1). The suitableα-aminonitrile resulted from a Strecker-like synthesis inan organic or aqueous medium [10, 11]. When the hetero-aryl moiety was a 2-furyl (8, 9), 4-methyl-2-thienyl (15)or a 5-methyl-2-thienyl (14) group, the Grignard reagentwas best obtained by means of a magnesium/Lithiumexchange reaction [12, 13] between MgBr2 and the suit-able 2-Li derivative. The pure (e.e.> 99%) enantiomersof the 3-methyl-piperidine derivatives10 and 11 wereobtained according to the same pathway(figure 1) butstarting from optically active 3-methyl-piperidines. Theoptical resolution of the 3-methyl-piperidine racematewas achieved by means of a crystallization procedurewith (+)- and (–)-mandelic acid in ethyl acetate resultingin enantiomeric purities up to 98–99% [14, 15]. It shouldbe noticed that we have previously described10-(–) and10-(+) obtained by a different strategy [15] but with verysimilar enantiomeric purities.

The new compounds obtained according tofigure 1(6–9, 11–15) were all checked by13C-NMR spectroscopyof the hydrochloride salts(table III). In these series

indeed, the hydrochlorides solutions are stabilized inalmost homogeneous conformations: the aromatic orhetero-aromatic rings are essentially restrained to theaxial position. Consequently, the conformational trappinginduced by the protonation allows for structural compari-sons (see below) between similar (axial aromatic rings)conformations [16–19].

Diastereomers16and17,bearing a methyl substitutionat the hetero-cyclohexyl ring(table IV), were preparedusing the azide synthesis shown infigure 2[15, 20]. Thecis/trans configurations were attributed from the13C-NMR spectra of their HCl salts. The chemical shifts wereattributed by comparison with the spectra of GK-11 andGK-12 hydrochlorides, two analogues whose diastereo-meric configurations have been previously character-ized [20, 21]. Thecis (Me/Pip) configuration was attrib-uted to compound16where the specificγ-interaction dueto the axial methyl substitution causes a clear upfield shiftof carbon 5(table IV).

The synthesis of non-commercial ketonic or piperi-dinic starting materials was required. Briefly, accordingto reference [22], alkylation of ethyl 2-sulfanylacetatewith ethyl 4-chlorobutanoate gave a 76% yield of ethyl4-[(2-ethoxy-2-oxoethyl)sulfanyl]butanoate which wassubmitted to a Dieckmann cyclization to afford a 54%

Table I. Inhibition of [3H]TCP binding in rat forebrain and cerebellum membranes by TCP and MK-801. The mean of at least threeindependent determinations was analyzed according to a single-site model in the forebrain and cerebellum (IC50

a, nM, Hill’s number) or atwo-site model in the cerebellum. The two-site model was statistically more probable in the cerebellum: the proportion of PCP2 and PCP3 sites(%) and affinities (IC50, nM) are given. SEM are in brackets.

Compound Forebrain single-site model Cerebellum single-site model Cerebellum two-site model

IC50 (PCP1) nH IC50 nH IC50 (PCP2) % (PCP2) IC50 (PCP3) % (PCP3)

TCP 9.3b 1.00 b 188 (63) 0.56 (0.05) 59 (17) 72.6 (3.6) 3716 (1195) 29.8 (4.5)MK-801 3.67 (0.65) 0.95 (0.04) 995 (458) 0.33 (0.04) 9.3 (4.6) 45.3 (4.9) 11125 (3179) 57.2 (5.8)

a IC50: concentration of unlabeled drug that inhibited 50% of specific [3H]TCP binding on specified sites;b from [20].

Figure 1.

126

yield of ethyl 3-oxotetrahydro-2H-thiopyran-2-car-boxylate. A decarboxylation in a 10% sulfuric acidsolution gave 54% of pure dihydro-2H-thiopyran-3(4H)-one (figure 3). 3-Methyltetrahydro-4H-thiopyran-4-one

was obtained by reacting tetrahydro-4H-thiopyran-4-onewith methyl-iodide in THF in the presence of oneequivalent of LDA at the temperature of –80 °C.3-Methyl-4-piperidinol was obtained essentially as its

Table II. TCP derivatives and analogues unsubstituted at the cyclohexyl or heterocyclohexyl ring and their calculated logP.

Compound X Y Z K R1 R2 R3 R4 log P a

TCP CH2 CH2 S CH2 H H H H 4.56 ± 0.301 S CH2 S CH2 H H H H 3.66 ± 0.502 O CH2 S CH2 H H H H 2.85 ± 0.393 CH2 CH2 S O H H H H 3.02 ± 0.394 S CH2 S O H H H H 2.11± 0.565 O CH2 S O H H H H 1.31 ± 0.416 SO2 CH2 S CH2 H H H H 1.99 ± 0.407 CH2 S S CH2 H H H H 3.52 ± 0.538 CH2 CH2 O CH2 H H H H 4.05 ± 0.299 S CH2 O CH2 H H H H 3.14 ± 0.5210 CH2 CH2 S CH2 H H CH3 H 5.06 ± 0.3011 S CH2 S CH2 H H CH3 H 4.15 ± 0.5312 CH2 CH2 S CH2 H H CH2OH H 3.35± 0.3713 CH2 CH2 S CH2 H H CH3 OH 3.21± 0.3714 S CH2 S CH2 CH3 H H H 4.12 ± 0.5315 S CH2 S CH2 H CH3 H H 4.12 ± 0.53

a Calculated with the ACD/LogP program (ACD, Inc.) (95% confidence, octanol/water).

Table III. 13C-NMR a chemical shiftsb (hydrochloride in CDCl3, δ ppm from TMS).

Carbon 6 c 7 8 9 11 12 13 14 15

1 – 22.7 23.4 – – 22.2 22.5 – –2 48.0 – 22.5 24.5 24.9 23.0 30.8 24.7 25.23 30.5 33.4 29.8 30.2 33.4 33.4 32.8 32.8 33.64 66.6 68.0 67.6 66.8 68.9 64.2 69.0 68.8 69.45 30.5 32.4 29.8 30.2 33.2 33.0 32.8 32.8 33.66 48.0 26.8 22.5 24.5 24.9 23.0 30.8 24.7 25.2α 47.3 47.0 47.0 46.7 52.3 49.7 50.6 46.5 47.2α’ 47.3 47.0 47.0 46.7 46.2 46.8 45.3 46.5 47.2â 22.8 25.5 22.3 22.0 28.2 36.6 35.4 22.3 22.8â’ 22.8 25.5 22.3 22.0 30.5 25.3 32.8 22.3 22.8γ 21.5 23.9 21.7 21.3 22.1 24.0 70.0 21.7 21.1R – – – – 18.9 69.7 15.1 14.7 15.6CAr 132.9–128.4 134.6–128.6 146.6–110.7 145.2–110.7 134.5–127.9 135.8–128.0 135.1–127.5 142.9–125.8 138.9–123.9

a For carbon atom numbering seetable IV; b italicized chemical shifts may be exchanged;c DMSO-d6.

127

trans isomer by a 4-step synthesis starting from benza-mide and ethyl acrylate(figure 4)[23]. The configurationwas attributed by comparison with the13C-NMR spectraof cis- and trans-2-methylcyclohexanol [24].

3. Pharmacology

The binding assays in rat forebrain and cerebellummembranes and the data analysis are described in experi-mental protocols (Section 5.2). We have first checked thata two-site model was more probable than a single-sitemodel to describe the competition of TCP and MK-801 inthe rat cerebellum membranes labeled with [3H]TCP(table I). The results were consistent with those previ-ously reported in membranes [3, 4] as well as in culturedcerebellum cells [5]. The same treatment was applied to21 compounds derived from the TCP structure; the resultsare presented intable V.

4. Results and discussion

In forebrain homogenates, [3H]TCP inhibition curveswere better fitted to a single-site model (PCP1-sites)(table V). This result was likely since Hill numbers weremost generally close to unity. However, in cerebellumhomogenates, Hill numbers were mostly lower than unityand the inhibition curves were better fitted to a two-site

Table IV. Comparative13C- (up) and1H-NMR (down) chemicalshifts of GK11, GK12, and compounds16, 17 (hydrochloride inCDCl3, δ ppm from TMS).

Carbon GK11 (cis) GK12 (trans) 17 (trans) 16 (cis)

1 17.8 22.5 – –2 30.2 30.0 a 29.8 a 33.23 35.4 36.3 34.7 34.84 72.9 74.5 73.3 72.45 26.5 30.7 a 29.6 a 27.06 22.6 22.0 22.7 24.9CH3 15.8 17.1 16.8 15.1α 48.9 48.2 48.6 48.9α’ 46.7 47.9 47.2 46.3â 22.3 21.72 21.9 22.0â’ 22.1 21.69 21.6 22.0γ 22.5 21.9 21.2 22.52’ 137.1 136.4 135.8 134.83’ 127.6 126.8 127.3 128.34’ 127.2 126.4 127.0 127.85’ 130.1 130.6 130.3 131.1

δ-CH3 1.6 1.1 1.2 1.8J (Hz) 6.9 6.6 6.2 6.7

a Italicized chemical shifts may be exchanged.

Figure 2.

Figure 3.

Figure 4.

128

model (PCP2- and PCP3-sites) which produced a signifi-cant reduction in the sum of squares (P < 0.05, Student’st-test) and a Durbin–Watson coefficient between 1.5 and2.5 (see Experimental protocols). The results presented intable Vconfirmed that [3H]TCP labeled two sites in thecerebellum in a 70:30 mean relative proportion(PCP2/PCP3). Interestingly, most compounds were moreor less able to interact with these binding sites. Howevercompounds4, 5, 6,and17 displayed a very low affinitywhatever the binding sites and thus their affinity andselectivity could not be evaluated. Compounds10-(–),11-(–), and13 displayed Hill numbers close to unity andtheir interaction was best described by a single-sitemodel.

As previously shown, the affinity for PCP1-sites wasreduced with the decrease of lipophilicity [7] in homoge-neous groups of structures. Indeed, TCP and derivativesdisplayed greater affinities for PCP1 than their thio- oroxa-analogues: TCP had a higher affinity for the PCP1-sites than1 or 2 (tables I and V), similarly 10 was ahigher affinity ligand than11 whatever the chirality is(table V). The same behavior was revealed when compar-ing TCP and8. Interestingly,1 and 7 were equipotent

ligands for the PCP1-sites although their sulfur atoms (inthe six atoms ring) are located in two different positions.

Affinity for PCP2-sites could be determined for 16compounds only(table V). Among them,3, 11-(±), 11-(+), 13, and14 were unable to discriminate significantlyPCP1- and PCP2-sites since they displayed statisticallyclose affinities for both sites. At the contrary, the 12remaining compounds had affinities significantly differ-ent at PCP1- and PCP2-sites. These affinities were appar-ently linearly correlated (P < 0.004, r = 0.76, seefigure 5). The lower affinity of these compounds forPCP2-sites might be related to the distribution of NMDANR2 subunits. Indeed NR2A and NR2B subunits are highlyexpressed in the forebrain while NR2C subunit is mainlyexpressed in the cerebellum [25, 26]. NMDA receptorsinvolving NR2C subunits are less sensitive to MK-801blockade than those comprising NR2A and/or NR2B

subunits [27]. These results are confirmed by competitionexperiments with [3H]MK-801 [4, 5] and consistent withMK-801 affinities in table I. Thus PCP1- and PCP2-sitesare likely to represent two different states of the NMDAreceptor discriminated by some of the new molecules.

Table V. Inhibition of [3H]TCP binding in rat forebrain and cerebellum membranes. The mean of at least three independent determinationswas analyzed according to a single-site model in the forebrain and cerebellum (IC50

a, nM, Hill’s number) or a two-site model in thecerebellum (SEM in brackets). When the two-site model was more probable, the proportion of PCP2 and PCP3 sites (%) and affinities (IC50a, nM) are given. Very low affinities precluded the two-site computation (n.d.).

Compound Forebrain single-site model Cerebellum single-site model Cerebellum two-site model

IC50 (PCP1) nH IC50 nH IC50 (PCP2) % (PCP2) IC50 (PCP3) % (PCP3)

1 71.6 (10.5) 1.05 (0.05) 178 (40) 0.68 (0.10) 586 (96) 75.2 (8.8) 8.5 (4.5) 25.2 (7.2)2 1223 (143) 0.98 (0.05) 2396 (475) 0.58 (0.12) 5840 (1690) 80.3 (6.5) 35 (18) 15.5 (5.7)3 294 (48) 0.77 (0.02) 1001 (408) 0.69 (0.09) 183 (65) 64.3 (3.4) 2127 (1273) 33.0 (1.1)4 8253 (320) 0.78 (0.06) 29700 (8022) 0.72 (0.02) n.d. – n.d.5 > 100 µM _ 63800 (10400) 0.60 (0.08) n.d. – n.d. –6 97700 (19066) 0.96 (0.18) 198500 (8500) 0.77 (0.22) n.d. – n.d. –7 73.4 (1.4) 1.00 (0.05) 228 (43) 0.65 (0.02) 659 (206) 77.8 (3.7) 8.0 (4.8) 23.4 (5.0)8 47.8 (1.1) 0.83 (0.06) 150 (78) 0.70 (0.11) 187 (42) 66.3 (12.0) 5.6 (3.5) 30.7 (10.0)9 133 (20) 0.87 (0.06) 0.55 (0.03) 0.55 (0.03) 1015 (262) 68.3 (2.8) 13.8 (3.6) 32.3 (4.4)10-(±) 5.5 (1.3) 1.09 (0.07) 122 (38) 0.80 (0.13) 47.3 (5.2) 68.5 (9.6) 1450 (590) 33.7 (8.1)10-(+) 5.2 (1.0) 1.04 (0.06) 155 (10) 0.62 (0.08) 28.3 (17.1) 52.5 (10.5) 808 (47) 51.0 (11.8)10-(–) 158 (20) 1.08 (0.09) 452 (53) 0.90 (0.08) –b – – b

11-(±) 132 (39) 1.14 (0.11) 265 (80) 0.69 (0.07) 265 (101) 69.1 (9.8) 2850 (433) 31.6 (11.1)11-(+) 24 (8.2) 0.84 (0.03) 231 (71) 0.59 (0.04) 53.2 (15.8) 66.3 (1.1) 2076 (747) 34.6 (2.5)11-(–) 529 (50) 0.93 (0.05) 996 (314) 0.85 (0.13) –b – – b –12 28.4 (6.6) 0.96 (0.09) 467 (158) 0.61 (0.06) 2240 (330) 69.0 (4.0) 12.4 (5.7) 31.0 (5.0)13 2406 (218) 1.07 (0.04) 3362 (1013) 1.02 (0.09) –b – – b –14 462 (88) 0.97 (0.07) 1170 (242) 0.63 (0.06) 532 (20) 83 (3.5) > 100 µM 15.3 (6.9)15 77.8 (9.3) 0.93 (0.06) 762 (335) 0.62 (0.06) 772 (202) 76.4 (2.6) 5.4 (3.4) 21.8 (3.8)16 26.8 (4.5) 0.89 (0.01) 206 (30) 0.71 (0.07) 1355 (733) 62.8 (7.3) 17.6 (6.6) 34.7 (9.5)17 17233 (1417) 0.94 (0.09) 50467 (2663) 0.60 (0.06) n.d. – n.d. –

a IC50: concentration of unlabeled drug that inhibited 50% of specific [3H]TCP binding on specified sites;b a one-site model was moreprobable.

129

The affinities for the PCP3-sites could not be correlatedwith those for PCP1- or PCP2-sites and they appearedvery sensitive to structure. Indeed, molecules testedappeared grossly separated into two groups: (i) moleculesdiffering from TCP only by substitution of one heteroa-tom in the cyclohexyl ring and/or in the hetero-aromaticmoiety, (ii) molecules differing from TCP by introductionof a methyl substitution in any of the constitutive rings orby specific modifications of the piperidine ring. Com-paratively to the TCP model structure, lowering thelipophilicity (table II) clearly directed the first group ofmolecules (1, 2, 7, 8, 9) toward the PCP3-sites since theydisplayed higher affinities for these sites (5.6 to 35 nM)than for PCP1- (47.8 to 1223 nM) or PCP2-sites (187 to5840 nM) (table V).Moreover, plotting IC50 PCP3/IC50

PCP1 against logP confirmed this tendency: the lowerlog P was, the higher the selectivity for PCP3 was(figure 6), in line with our hypothesis. A similar tendencywas found with regard to PCP2 (not shown). Finally, thisgroup of compounds revealed interesting selectivities forPCP3 when compared to PCP2 and PCP1. 2, the lesspotent (35 nM) and the less lipophilic (logP = 2.85) inthis group, displayed a high selectivity for the PCP3-siteswhen compared to the PCP1- and PCP2-sites (PCP3/PCP1

< 0.026; PCP3/PCP2 < 0.006).In the second group, changing the piperidine for a

morpholine ring decreased considerably the affinities (3)or precluded the two-site computation given very lowaffinities (4, 5). Substitution of a methyl group in thepiperidine ring gave compounds with low affinities for

PCP3-sites:13, 10-(±), 10-(+), and their thio-analogues11-(±), 11-(+). Interestingly, the affinities of10-(–), 11-(–), and13were better described by a single-site model inthe cerebellum. Compound12 bearing an hydroxymethylsubstitution in the piperidine ring exhibited a high affinityfor both PCP3- and PCP1-sites with a high selectivitywith regard to PCP2-sites. The methyl substitution in thethiopyranyl ring gave a high affinitycis-compound (16)equipotent at the PCP1- and PCP3-sites and an inactivetrans-compound (17). The difference betweencis andtrans diastereomers in the forebrain was higher thanpreviously observed in the cyclohexyl homologue se-ries [15, 20] although affinities were lower. Finally, theposition of a methyl group in the hetero-aromatic moietywas crucial. In theα position from the sulfur atom (14),it lowered the affinity for PCP3 and at the contrary, in theâ position (15), it increased the affinity for these sites.Moreover15 displayed a good selectivity with regard toboth PCP1- and PCP2-sites (PCP3/PCP1 < 0.07;PCP3/PCP2 < 0.007). This is a confirmation of the veryimportant role played by the aromatic or hetero-aromaticring in the arylcyclohexylamines selectivity of bind-ing [28]. In this second group of molecule steric interac-tions due to substitutions are likely to influence moreselctivity than lipophilicity.

It is now well admitted that the NMDA receptor is anhetero-oligomeric protein composed with 5 sub-unitsdifferently expressed in forebrain, cerebellum, and spinalcord [29–31]. The resulting diversity of NMDA receptors

Figure 5. Linear relationships between affinities for PCP1- andPCP2-sites (r = 0.76,p < 0.004) for 12 compounds discrimina-ting significantly both sites.

Figure 6. Relationships between selectivity (PCP3/PCP1) andlog P (r = 0.84,p < 0.08) for molecules differing from the TCPmodel only by substitution of one heteroatom in the cyclohexylring and/or in the hetero-aromatic moiety.

130

in the CNS, and consequently the diversity of the PCPreceptors, is responsible for the heterogeneity of pharma-cological responses [32]. The present study confirms thebinding sites heterogeneity since TCP analogues wereable to interact differently with subsites labeled with[3H]TCP in the rat cerebellum. In this region interestinglyNMDA receptors display particular properties when com-pared with NMDA receptors found in the forebrain [26].Some of the structures we have prepared and tested mightbe valuable tools to clarify the pharmacological role ofthe low affinity sites in the rat cerebellum (PCP3). Indeed,compounds2 and 15 possess both a high affinity and agood selectivity for these sites. Such molecules mightalso be leads for new compounds able to interact with thisspecific target.

5. Experimental protocols

5.1. Chemistry

Melting points (uncorrected) were determined with aBüchi–Tottoli apparatus. Yields were not optimized. El-emental analysis was performed at the CNRS Microana-lytical Section in Montpellier on the hydrochloride saltsand were within±0.4% of theoretical values.1H- and13C-NMR spectra were obtained on a Brucker AC 200spectrometer at 200.13 and 50.32 MHz respectively in5-mm sample tubes in the FT mode. For some13C-signalassignments, a spin-echo sequence (Jmod) was used.Chemical shifts are reported in (δ) ppm downfield fromTMS. Enantiomeric purities were determined on a Shi-madzu HPLC equipment (LC-1O AD pump, SPD-6A UVspectrometer), computer-controlled by the Class LC-10program. Analysis were made on a Chiralcel-OD column(10 mm, 4.6× 250 mm) (Daicel Chemical Industries) inheptane (0.6 mL /min) at 36 °C. UV-detection was madeat 240 nm. Typical injection volumes were 5 mL of a 30µM solution of base compound in heptane. Opticalrotations were obtained in methanol with a Perkin-Elmer241 polarimeter in a 1-dm microcell at 20 °C. For NMRand in vitro experiments, compounds were used as theirhydrochloride salts; salts were precipitated by adding adry HCl ethereal solution in ether to a solution of base inether. After filtration, the solids collected were dried invacuum.

5.1.1. Synthesis of dihydro-2H-thiopyran-3(4H)-one

5.1.1.1. Ethyl 4-[(2-ethoxy-2-oxoethyl)sulfanyl]buta-noate

Sodium (4.6 g, 200 mmol, 1 eq.) was added cautiouslyby portion in a nitrogen atmosphere to ethanol (100 mL).

After the solid was consumed, the mixture was cooled to0 °C and ethyl 2-sulfanylacetate (22 mL, 200 mmol, 1eq.) then ethyl 4-chlorobutanoate (30.1 g, 200 mmol, 1eq.) was added slowly. The resulting mixture was stirredfor 20 h at room temperature. The NaCl precipitate wasfiltered, the filtrate concentrated in vacuum, the oilobtained was diluted in water (100 mL) and extractedwith ether (3 × 80 mL). The combined organic layerswere dried over MgSO4 and concentrated in vacuum. Theresulting yellow oil was distilled to yield 35.6 g (76%) ofa colorless oil.

5.1.1.2. Ethyl 3-oxotetrahydro-2H-thiopyran-2-carbo-xylate

Ethyl 4-[(2-ethoxy-2-oxoethyl)sulfanyl]butanoate inanhydrous ether (200 mL) was added dropwise in anitrogen atmosphere to a solution of sodium ethanolate(20.7 g, 0.3 mol, 2 eq.) at 0 °C, stirred at 0 °C for 45 minand at room temperature for 3 h. The mixture washydrolyzed with a water/acetic acid (80:20) solution, theaqueous phase was separated and extracted with ether (3× 50 mL), the combined organic layers were dried overMgSO4 and concentrated in vacuum. The resulting yel-low oil obtained was distilled under reduced pressure toyield 15.3 g (54%) of a colorless liquid.

5.1.1.3. Dihydro-2H-thiopyran-3(4H)-one

Ethyl 3-oxotetrahydro-2H-thiopyran-2-carboxylate(15.2 g, 80.9 mmol) was refluxed in a 15% sulfuric acidsolution for 18 h then cooled to room temperature. A 10%NaOH solution in water was then added dropwise toreach pH 6. The mixture was extracted with ether (3×50 mL), the organic phases washed with water, dried overMgSO4, and concentrated in vacuum. The resulting oilwas distilled under reduced pressure to yield 5.1 g (54%)of a colorless oil.

5.1.2. Synthesis of 1,1-dioxo-tetrahydro-1λ6-thiopy-ran-4-one

A solution of hydrogen peroxide (11.4 mL, 0.1 mol, 2eq.) was added dropwise to a mixture of tetrahydro-4H-thiopyran-4-one(5.6 g, 48 mmol, 1 eq.) and acetic acid(25 mL) keeping the temperature below 30 °C. Themixture was stirred for 4 h, the acetic acid distilled underreduced pressure, and the crystallized yellow residue wasfiltered and washed with ether to yield 4.7 g (67%) ofwhite crystals.

131

5.1.3. Synthesis of 3-methyl-4-piperidinol

5.1.3.1. Ethyl 1-benzoyl-4-oxo-3-piperidinecarboxy-late

Sodium hydride (4 g, 0.1 mol, 1 eq.) was added in anitrogen atmosphere to a solution of benzamide (12.1 g,0.1 mol, 1 eq.) in toluene (200 mL), the mixture wasrefluxed for 1 h, cooled to 0 °C and ethyl acrylate(32.6 mL, 0.3mol, 3 eq.) was then rapidly added. Thesolution was stirred at 60 °C for 24 h, cooled to 0 °C,diluted with ice-cold water (100 mL) and stirred for 0.5 h.The aqueous phase was separated, washed with ether(50 mL), acidified until pH 3 and extracted with CH2Cl2(3 × 50 mL). The combined organic layers were driedover Na2SO4, filtered and concentrated in vacuum. Theyellow oil obtained was purified by column chromato-graphy (SDS Chromagel 70–100µ) in ether to yield 9.4 g(34%) of a red oil.

5.1.3.2. Ethyl 1-benzoyl-3-methyl-4-oxo-3-piperidine-carboxylate

A mixture of ethyl 1-benzoyl-4-oxo-3-piperi-dinecarboxylate (9.3 g, 33.8 mmol, 1 eq.) and sodiumhydride (1.35 g, 34 mmol, 1 eq.) in dimethoxyethane(50 mL) was refluxed for 2 h then cooled to 0 °C. ICH3

(5.2 mL, 84 mmol, 2.5 eq.) was added and the mixtureheated at 60 °C for 40 h. The mixture, after concentrationin vacuum, was diluted with water (100 mL) and ex-tracted with CH2Cl2 (3 × 50 mL). The combined organiclayers were washed successively with a 5% NaOH watersolution (50 mL), a 5% HCl water solution (50 mL), withwater (50 mL), then dried over Na2SO4, and concentratedin vacuum. The resulting brown oil (9.4 g) was purifiedby column chromatography (SDS Chromagel 70–200µ)in ether to yield 7.7 g (79%) of a slightly yellow oil.

5.1.3.3. 3-Methyl-piperidin-4-one hydrochlorideA solution of ethyl 1-benzoyl-3-methyl-4-oxo-3-

piperidinecarboxylate (7.7 g, 26.6 mmol, 1 eq.) in a 6 NHCl aqueous solution was refluxed for 72 h, then thebenzoic acid precipitate was filtered, the filtrate washedwith ether (3× 50 mL) and concentrated in vacuum. Thebrown solid obtained was crystallized in ethanol to getwhite crystals (2.85 g, 72%).

5.1.3.4. 3-Methyl-4-piperidinolA solution of 5% NaOH in water (7.6 mL) was added

dropwise to a solution of 3-methyl-piperidin-4-one hy-drochloride (2.85 g, 19 mmol, 1 eq.) in methanol (30 mL)and the mixture was stirred for 0.5 h at room temperature.A solution of NaBH4 (2.46 g, 6.5 mmol, 1.4 eq.) inmethanol (30 mL) was added, the mixture was stirred for4 h at room temperature, cooled to 0 °C, made acidic with

few drops of a 5% HCl aqueous solution, and concen-trated in vacuum. The resulting yellow solid was dis-solved in hot ethanol and precipitated at room tempera-ture by the addition of drops of ether to yield a white solid(1.9 g, 87%).

5.1.4. Synthesis ofα-aminonitrilesα-aminonitriles were obtained by means of two differ-

ent synthetic methods. Since the same method wasapplied to various compounds we describe only oneexample in each case.

5.1.4.1. Method AAcetone cyanohydrine (2.2 g, 25.8 mmol, 1 eq.) was

added dropwise to a stirred mixture of tetrahydro-4H-thiopyran-4-one (3 g, 25.8 mmol, 1 eq.), anhydrousMgSO4 (9.3 g, 77.4 mmol, 3 eq.), dimethylacetamide(2.25 g, 25.8 mmol, 1 eq.) and piperidine (4.4 g,51.6 mmol, 2 eq.). The pasty mixture was heated at 45 °Cfor 48 h, cooled to room temperature, poured onto ice andstirred for 30 min. The aqueous mixture obtained wasextracted with ether and the organic layer was washedwith water until neutrality, dried over Na2SO4, filtered,and concentrated in vacuum to yield 5.3 g (98%) of anorange oil of 4-piperidinotetrahydro-2H-thiopyran-4-carbonitrile.

The following α-aminonitriles were similarly synthe-sized: 4-piperidinotetrahydro-2H-pyran-4-carbonitrile(79%,F = 46–47 °C) from piperidine and tetrahydro-4H-pyran-4-one; 4-(3-methylpiperidino)tetrahydro-2H-thio-pyran-4-carbonitrile (> 98%, oil) from 3-methyl-piperidine and tetrahydro-4H-thiopyran-4-one; R-4-(3-methylpiperidino)tetrahydro-2H-thiopyran-4-carbonitrile(> 98%, oil) and S-4-(3-methylpiperidino)tetrahydro-2H-thiopyran-4-carbonitrile (> 98%, oil) from R-(–)-3-methylpiperidine and S-(+)-3-methylpiperidine [7, 8, 21],and tetrahydro-4H-thiopyran-4-one; 4-(3-methylpipe-ridino)tetrahydro-2H-thiopyran-4-carbonitrile (81%,solid) from 3-hydroxymethylpiperidine and cyclohex-anone.

5.1.4.2. Method B5% HCl (a few drops) was added to a stirred mixture of

3-methyl-4-piperidinol (0.8 g, 7 mmol, 2 eq.) in 10 mL ofwater and cyclohexanone (2.7 g, 23.6 mmol, 1 eq.) toreach pH 3. KCN (0.47 g, 7.3 mmol, 1.05 eq.) was added(pH reached 11). The mixture was stirred at roomtemperature for 24 h then extracted with CH2Cl2, driedover Na2SO4, filtered, and concentrated in vacuum toyield 1.43 g (93%) of 1-(4-hydroxy-3-methylpiperi-dino)cyclohexanecarbonitrile as a colorless oil.

132

1,1-Dioxo-4-piperidinohexahydro-1λ6-thiopyran-4-carbonitrile (64%,F = 148 °C) was similarly preparedfrom piperidine and 1,1-dioxo-tetrahydro-1λ6-thiopyran-4-one.

5.1.5. Synthesis of TCP analogues fromα-amino-nitriles

Compounds1 [7], 2 [7, 33], 3–5 [7] as well as com-pounds10-(±), 10-(–), 10-(+) [15] have been previouslydescribed. TCP analogues were obtained by means of twodifferent synthetic methods. Since the same method wasapplied to various compounds we describe only oneexample to illustrate each strategy. Compounds purifica-tions and properties are given intable VI.

5.1.5.1. Method 1: TCP analogues with a thiophenylhetero-aromatic ring (6, 7, 11–13)

1-[3-(2-thienyl)tetrahydro-2H-thiopyran-3-yl]piperidi-ne 7: An ethereal solution of 3-piperidinotetrahydro-2H-thiopyran-3-carbonitrile (1 g, 4.76 mmol, 1 eq.) wasadded dropwise at room temperature to a well stirredsolution containing a Grignard reagent prepared from2-bromo-thiophene (2.3 g, 14.3 mmol, 3 eq.) and Mgturnings (0.35 g, 14.3 mmol, 3 eq.). The mixture wasrefluxed for 20 h, cooled to room temperature and treatedas follows: the mixture was poured carefully on anice-cold saturated solution of NH4Cl, stirred for 30 min,extracted with ether, the combined ether layers wereextracted 3 times with 10% HCl, 20% NH4OH was addedto the aqueous phase until neutrality. The aqueous phasewas extracted with ether, the organic phase washed withwater, dried over Na2SO4, filtered, and concentrated invacuum. The crude product obtained was purified asdescribed intable VI to yield 1.05 g (83%) of7 as awhite solid.

5.1.5.2. Method 2: TCP analogues with a furanyl (8, 9)or a substituted thiophenyl (14, 15) hetero-aromatic ring

1-[4-(2-furyl)tetrahydro-2H-thiopyran-4-yl]piperidine9: Firstly, a MgBr2 solution was prepared from 1,2-dibromo-ethane (6.76 g, 36 mmol, 4 eq.) and magnesiumturnings (0.88 g, 36 mmol, 4 eq.) in ether (80 mL) in anitrogen atmosphere. Secondly, a solution of 2-furyl-Lithium was prepared in a nitrogen atmosphere by thedropwise addition of a n-butyl-lithium solution 1.6 M inhexane (28 mL, 45 mmol, 5 eq.) to a mixture of furane(3.1 g, 45 mmol, 5 eq.) and TMEDA (5.2 g, 45 mmol, 5eq.) in anhydrous ether (100 mL) at –20 °C (withoutTMEDA in the synthesis of15). The mixture was thenrefluxed for 2 h, cooled to room temperature, and addeddropwise to the MgBr2 solution in ether. A solution of4-piperidinotetrahydro-2H-thiopyran-4-carbonitrile (1.9 g,9 mmol, 1 eq.) in ether was added dropwise at roomtemperature, the mixture refluxed for 16 h, cooled toroom temperature, and treated as described above. Thecrude product was purified as described intable VI toyield 1.6 g (71%) of9 as a white solid.

5.1.6. Preparation of compounds16 and17

5.1.6.1. 3-Methyltetrahydro-4H-thiopyran-4-onen-Butyl-lithium (37.5 mL, 60 mmol, 1 eq.) was added

dropwise in a nitrogen atmosphere to a stirred solution ofdiisopropylamine (8.4 mL, 60 mmol, 1 eq.) in THF(74 mL) and the mixture was stirred 0.5 h at roomtemperature, then cooled to –80 °C. After slow additionof tetrahydro-4H-thiopyran-4-one (6.96 g, 60 mmol, 1eq.) and stirring for 0.5 h, ICH3 (5.6 mL, 90 mmol, 1.5eq.) was added, the mixture allowed to warm up to roomtemperature while stirred for 5 h. The mixture was dilutedwith a NaCl saturated 5% solution of sodium bicarbonate

Table VI. Purification and properties of new compounds.

Purification Solvent (v/v) Fbase(°C) FHCl (°C) Yield (%)

6 Crystallization AcOEt 167 179–183 487 Column chromatograpy (Al2O3)

c Petroleum ether/ether (40:60) 58–60 184–186 838 Column chromatograpy (Al2O3) CH2Cl2 Oil 166–167 769 Column chromatograpy (Al2O3) Petroleum ether/ether (98:2) 79–81 171–173 7111-(±) Column chromatograpy (Al2O3) Petroleum ether/ether (95:5) Oil 168–169 5711-(+) a Column chromatograpy (Al2O3) Petroleum ether/ether (95:5) Oil 170–172 5911-(–) b Column chromatograpy (Al2O3) Petroleum ether/ether (95:5) Oil 169–171 6412 Column chromatograpy (Al2O3) Petroleum ether/ether (50:50) Oil 176–178 5613 Column chromatograpy (Al2O3) Petroleum ether/ether (20:80) 120–122 170–172 7114 Column chromatograpy (Al2O3) Petroleum ether/ether (98:2) 93–95 160–162 6715 Column chromatograpy (Al2O3) Petroleum ether/ether (90:10) 89–91 144–146 55

a HPLC (chiral phase):Rt = 29.1 min, e.e.> 99%; [αD20]base= +12° (c 1, CH3OH); b HPLC (chiral phase):Rt = 32.3 min, e.e.> 99%;

[αD20]base= –11° (c 1, CH3OH); c aluminium oxide 90, 2–3 Merck.

133

in water and the organic phase separated, dried overNa2SO4 and concentrated in vacuum. The orange oilobtained was purified by chromatography (SDS Chroma-gel 70–200µ) in petroleum ether/ethyl acetate (90:10) toyield 3.2 g (41%) of a colorless oil.

5.1.6.2. Cis- and trans-3-methyl-4-(2-thienyl)tetra-hydro-2H-thiopyran-4-ol

2-Thienyl-magnesium bromide was prepared by add-ing dropwise in a nitrogen atmosphere a solution of2-bromo-thiophene (6.1 g, 37.2 mmol, 1.1 eq.) in ether(100 mL) to Mg turnings (0.9 g, 37.2 mmol, 1.1 eq.).After the mixture was gently refluxed for 3 h, a solutionof 3-methyltetrahydro-4H-thiopyran-4-one (4.2 g,32 mmol, 1 eq.) dissolved in ether (50 mL) was added atroom temperature. The mixture was refluxed for 16 h,cooled to room temperature, poured onto a saturatedNH4Cl solution in water, stirred 0.5 h. The aqueous phasewas separated, extracted with ether (3× 50 mL) and thecombined organic layers extracted with 15% HCl (3×50 mL). 25% NH4OH was added to the aqueous phaseuntil neutrality and the solution extracted with ether. Thefinal organic phase was washed with water, dried overNa2SO4, filtered, and concentrated in vacuum. The greenoil obtained was purified by chromatography (SDS Chro-magel 70–200µ) in petroleum ether/ether (90:10) to yield6.8 g (98%) of a slightly blue oil containing the diaste-reomeric alcohols.

5.1.6.3. Cis- and trans-3-methyl-4-(2-thienyl)tetra-hydro-2H-thiopyran-4-azide

Sodium azide (4 g, 61.7 mmol, 2 eq.) was cautiouslyadded to a solution of trichloro-acetic acid (15.1 g,92.4 mmol, 3 eq.) in CHCl3 (100 mL), the mixture wascooled to 10 °C, and a solution of diastereomeric alcohols3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-ol(6.6g,30.8 mmol, 1 eq.) in CHCl3 (50 mL) was added. Themixture was stirred for 72 h at 10–12 °C, then a 10%NH4OH solution was added until neutrality and theaqueous phase extracted with CH2Cl2 (3 × 100 mL).The pooledorganic phases were washed with water,dried over Na2SO4, filtered, and concentrated invacuum. The oil obtained (6.8 g, 92%) was used in thenext step without further purification.

5.1.6.4. Cis- and trans-3-methyl-4-(2-thienyl)tetra-hydro-2H-thiopyran-4-amine

A solution of cis- and trans-3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-azide (5.5 g, 23 mmol, 1 eq.)in THF (30 mL) was added dropwise at 0 °C to a solutionof lithium aluminium hydride (0.87 g, 23 mmol, 4 eq.) inTHF in a nitrogen atmosphere stirred for 24 h at roomtemperature. The strict minimum amount of NH4OH

necessary to destroy lithium aluminium hydride wasadded slowly, the precipitate filtered, washed withCH2Cl2 (300 mL), and the filtrate concentrated invacuum. The brown oil obtained was diluted in ether,extracted with a 10% HCl solution (3× 100 mL) and 20%NH4OH added to the aqueous phase until neutrality. Theaqueous phase was extracted with ether (3× 100 mL) andthe combined organic layers washed with water, driedover MgSO4, filtered, and concentrated in vacuum. Thecrude diastereomeric oily mixture was purified by chro-matography (SDS Chromagel 70–200µ) in petroleumether/ether (50:50) to yield 3.7 g (75%) of acis/transprimary amines mixture as a colorless oil.

5.1.6.5. Cis- and trans-1-[3-methyl-4-(2-thienyl)tetra-hydro-2H-thiopyran-4-yl]piperidine16, 17

Potassium carbonate (5.8 g, 42 mmol, 2 eq.) and 1,5-dibromopentane (3.6 mL, 26.2 mmol, 1.25 eq.) wasadded in a nitrogen atmosphere to cis- and trans3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-amine (4.5 g,21 mmol, 1 eq.) dissolved in HMPT (50 mL) and theresulting mixture stirred at 60 °C for 48 h then cooled toroom temperature. The mixture was poured onto water(200 mL) and extracted with ether (3× 100 mL). Thecombined organic layers were extracted 3 times with 10%HCl and 20% NH4OH was added to the aqueous phaseuntil neutrality. The aqueous phase was extracted withether and the final organic phase was washed with water,dried over MgSO4, filtered, and concentrated in vacuum.The resulting yellow oil was chromatographied (SDSChromagel 70–200µ) in petroleum ether/ether (90:10) toyield the pure diastereomers as white solids:17 (2 g,34%) and16 (0.7 g, 12%).

5.2. Biochemistry

5.2.1. Binding assays[3H]TCP (Amersham, 48 Ci/mmol) binding to the PCP

receptor subtypes was measured as previously de-scribed [3]. Briefly, the rat (wistar) brain (minus thecerebellum) or the cerebellum was homogenized with anUltraturax (Ika Werke, maximum setting) in a 50 mMTris/HCl, pH 7.7 buffer for 20 s at 4 °C. The homogenatewas then centrifuged at 49000g for 20 min. The pelletwas resuspended in the same buffer and the homogeni-zation–centrifugation steps performed a second time. Thefinal pellet was resuspended in 10 volumes of a 50 mMTris/Hepes, pH 7.7 buffer and used without furtherpurification.

The forebrain or cerebellum homogenate (0.5–0.8 mgprotein/mL) was incubated with [3H]TCP (1 nM or 2.5nM respectively) in a 5 mM Tris/Hepes, pH 7.7 buffer(0.5 mL or 1 mL respectively) in the absence (total

134

binding) or in the presence of the competing drug for30 min at 25 °C. The incubation was terminated byfiltration over GF/B (Whatman) glass fibre presoaked in0.05% polyethyleneimine (Aldrich) with an MR24 Bran-del cell harvester. The filters were rinsed three times with5 mL 50 mM NaCl, Tris HCl 10 mM, pH 7.7 buffer andthe radioactivity retained was counted in 3.5 mL ACS(Amersham) with an Excel 1410 (LKB) liquid scintilla-tion spectrophotometer. The non-specific binding wasdetermined in parallel experiments in the presence of 100µM unlabelled TCP.

5.2.2. Data analysisIn each experiment, values are the mean of three

independent determinations. Each experiment was per-formed 3–5 times. The data from competition experi-ments were first analyzed by Hill’s representation accord-ing to a single-site model, then by a non linear regressionmethod (Marquardt–Levenberg algorithm) according to atwo-site model using the Sigmaplott 4 software (Jandel).The two-site interaction was represented by:

[LB] = (B1 + B2) – {([I] × B1/(IC501 + [I]))+ ([I] × B2/(IC502 + [I]))}

where [LB] was the percentage of radioligand concen-tration specifically bound, [I] the competitor concentra-tion, B1 and B2 the percentage of each binding site, IC501and IC502 the concentrations of unlabeled competitor thatinhibited 50% of specific [3H]TCP binding on specifiedsites. Two constraints were fixed: (i) IC501, IC502, B1, B2> 0; (ii) 95% < (B1 + B2) < 105% because of theuncertaincy on the total binding (10%). Experimentalresults were submitted to an ANOVA test followed by aDurbin–Watson test. The two-site model was preferredwhen it produced a significant reduction in the sum ofsquares (P < 0.05, Student’st-test) and when the Durbin-–Watson coefficient was closer to 2 than in the single-sitemodel and comprised between 1.5 and 2.5.

Acknowledgements

This work was supported by D.R.E.T. (grant 94/141).Thanks are due to M. Michaud for her excellent technicalassistance.

References

[1] Vignon J., Chicheportiche R., Chicheportiche M., Kamenka J.M.,Geneste P., Lazdunski M., Brain Res. 280 (1983) 194–197.

[2] Wong E.H.F., Kemp J.A., Priestley T., Knight A.R., WoodruffG.N., Iversen L.L., Proc. Natl. Acad. Sci. USA 83 (1986) 7104–7108.

[3] Vignon J., Privat A., Chaudieu I., Thierry A., Kamenka J.M.,Chicheportiche R., Brain Res. 378 (1986) 133–141.

[4] Ebert B., Wong E.H.F., Krogsgaard-Larsen P., Eur. J. Pharmacol.208 (1991) 49–52.

[5] Berman F.W., Murray T.F., J. Biochem. Toxicol. 11 (1996)217–226.

[6] Choi D.W., Cerebrovasc. Brain Metab. Rev. 2 (1990) 105–147.

[7] Hamon J., Vignon J., Kamenka J.M., Eur. J. Med. Chem. 31 (1996)489–495.

[8] Vignon J., Chaudieu I., Allaoua H., Journod L., Javoy-Agid F.,Agid Y., Chicheportiche R., Life Sci. 45 (1989) 2547–2555.

[9] Bruylants P., Bruylants P., Bull. Soc. Chim. Belg. 33 (1924)467–478; Bull. Soc. Chim. Belg. 35 (1926) 139–154.

[10] Geneste P., Kamenka J.M., Dessapt P., Bull. Soc. Chim. Fr. 3& 4II (1980) 186–191.

[11] Ilagouma A.T., Dornand J., Liu C.F., Zénone F., Mani J.C.,Kamenka J.M., Eur. J. Med. Chem. 25 (1990) 609–615.

[12] Minato A., Tarrao K., Hayashi T., Susuki K., Kumada M.,Tetrahedron Lett. 52 (1981) 5319–5322.

[13] Coderc E., Martin-Fardon R., Vignon J., Kamenka J.M., Eur. J.Med. Chem. 28 (1993) 893–898.

[14] Marwaha J., Palmer M., Hoffer B., Freedman R., Rice K.C., PaulS., Skolnick P., Arch. Pharmacol. 315 (1981) 203–209.

[15] Coderc E., Cerruti P., Vignon J., Rouayrenc J.F., Kamenka J.M.,Eur. J. Med. Chem. 30 (1995) 463–470.

[16] Geneste P., Kamenka J.M., Org. Magn. Res. 7 (1975) 579–584.

[17] Kamenka J.M., Geneste P., In: Domino E.F. (Ed.), PCP (Phencyc-lidine): Historical and Current Perspectives, NPP Books, Ann Arbor, 1981,pp. 47–82.

[18] Manoharan M., Eliel E.L., Carrol F.I., Tetrahedron Lett. 24 (1983)1855–1858.

[19] Manoharan M., Eliel E.L., Kanan W., Tetrahedron Lett. 25 (1984)3267–3268.

[20] Michaud M., Warren H., Drian M.J., Rambaud J., Cerruti P.,Nicolas J.P., Vignon J., Privat A., Kamenka J.M., Eur. J. Med. Chem. 29(1994) 869–876.

[21] Geneste P., Kamenka J.M., Ung S.N., Herrmann P., Goudal R.,Trouiller G., Eur. J. Med. Chem. 14 (1979) 301–308.

[22] Young T.E., Heitz L.J., J. Org. Chem. 38 (1973) 1562–1566.

[23] Baty J.D., Jones G., Moore C., J. Chem. Soc. (C) (1967)2645–2647.

[24] Pouchert C.J., Behnke J. (Eds.), The Aldrich library of 13C and 1HFT NMR Spectra, Vol. 1, Aldrich Chemical Company, 1993, pp. 235–236.

[25] Conley E.C., In: Conley E.C. (Ed.), The Ion Channel Facts BookI, Academic Press, New York, 1996, pp. 140–233.

[26] Sekiguchi M., Okamoto K., Sakai Y., J. Neurosci. 10 (1990)2148–2155.

[27] Yamakura T., Mori Y., Masaki H., Shimoji K., Mishima M.,Neuroreport 4 (1993) 687–690.

[28] Chaudieu I., Vignon J., Chicheportiche M., Kamenka J.M., Trouil-ler G., Chicheportiche R., Pharmacol. Biochem. Behav. 32 (1989) 699–705.

[29] Moriyoshi K., Masu M., Ishii T., Shigemoto R., Mizuno N.,Nakanishi S.,, Nature 354 (1991) 31–37.

[30] Durand G.M., Bennett M.V.L., Zukin R.S., Proc. Natl. Acad. Sci.USA 90 (1993) 6731–6735.

[31] Widdowson P.S., Trainor A., Lock E.A., J. Neurochem. 64 (1995)651–661.

[32] Lynch D.R., Lawrence J.J., Lenz S., Anegawa N.J., Dichter M.,Pritchett D.B., J. Neurochem. 64 (1995) 1462–1467.

[33] Eiden F., Schmidt M., Buchborn H., Arch. Pharmacol. 320 (1987)348–361.

135

Invited review

Inhibitors of 5-lipoxygenase: a therapeutic potential yet to be fully realized?

Robert N. Young*

Department of Medicinal Chemistry, Merck Frosst Centre for Therapeutic Research,P.O. Box 1005, Pointe Claire - Dorval, Québec H9R 4P8, Canada

(Received 21 June 1999; accepted 21 June 1999)

Abstract – Inhibition of leukotriene biosynthesis has been extensively studied as a potential for the development of novel therapies forinflammation and respiratory diseases and, in particular, for asthma. Many compounds have been identified which inhibit the key enzyme,5-lipoxygenase. Four distinct classes of compounds have been identified, namely, (1) redox inhibitors (alternative substrates), (2) ironchelating inhibitors, (3) competitive reversible inhibitors, and (4) inhibitors of the 5-lipoxygenase activating protein. Experience over the pasttwo decades with redox inhibitors has been disappointing and although a number of potent compounds have been identified, they have oftenbeen associated with ancillary toxicity and non-specificity due to their redox activity. Iron chelating inhibitors have been more successful andone compound, Zileutont, has reached the market. However, more potent analogues have often encountered toxicity problems. Competitiveinhibitors have been identified by a number of research groups but, as yet, none has been successful. Inhibitors of the 5-lipoxygenase activatingprotein (FLAP) have been identified and compounds such as MK-0591 and BaY-X-1005 have shown efficacy in asthma trials. To date,however, no clear advantage for inhibitors of lipoxygenase has been demonstrated relative to the leukotriene D4 receptor antagonists such asSingulairt and Accolatet. © 1999 Éditions scientifiques et médicales Elsevier SAS

5-lipoxygenase / inhibitor / 5-lipoxygenase activating protein / asthma / inflammation

1. Introduction

After the initial characterization of the slow-reactingsubstance of anaphylaxis in 1940 [1], nearly four decadespassed before the characterization by Professor BengtSamuelsson of the Karolinska Institute [2, 3] of thesepotent contractile substances as novel peptido-lipid hy-brids, which became known as leukotrienes C4, D4 andE4. These substances and a companion lipid whichbecame known as leukotriene B4 were proposed bySamuelsson to be derived from a common epoxideintermediate which he named leukotriene A4. Samuels-son’s biosynthetic detective work, which led to theproposal of the leukotriene biosynthetic pathway(fig-ure 1) was a scientific triumph. Considering the potentpro-inflammatory properties of leukotriene B4 [4] and themultiple activities of the peptido-lipid leukotrienes LTC4,D4 and E4, it became apparent that modulation of thispathway could have important implications in the devel-opment of novel therapeutics for diseases such as asthma,

allergy and a host of other inflammatory diseases. It wasrecognized in the course of this work that a key enzymein the process, 5-lipoxygenase, could transform arachi-donic acid (AA) in a two-step process to first,5-hydroperoxyeicosatetraenoic acid (5-HPETE), andthence through a dehydration step to leukotriene A4. Thiscommon unstable intermediate is then taken on to leu-kotriene B4 via leukotriene B4 synthase in certain cells,such as neutrophils, or could be converted via a specificglutathione-transferase enzyme (leukotriene C4 synthase)to provide the peptido-lipid leukotrienes C4, D4 and E4 incells such as eosinophils [5].

In the early 1980s, a number of pharmaceutical re-search laboratories throughout the world recognized thatnovel therapeutics could be derived from the develop-ment of inhibitors of 5-lipoxygenase (which would po-tentially modulate the entire pathway) or from receptorantagonists at the specific leukotriene receptors. At theonset, it was not clear which of these alternative ap-proaches would prove most successful, although, as ofthis date, it would appear that the latter approach, namely*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 671−685 671© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Figure 1. Leukotriene biosynthetic pathway.

672

the development of leukotriene D4 receptor antagonists,has been most successful. Nonetheless, the efforts thathave gone into the development of specific inhibitors ofthe 5-lipoxygenase enzyme have been extensive and haveyielded at least one novel therapeutic agent, namelyZileuton®. A number of reviews on the development of5-lipoxygenase inhibitors have been published in thepast [6–8]. This review will attempt to provide an over-view of what has been learned in the course of theseefforts and, hopefully, provide a useful perspective forfuture drug development.

2. 5-Lipoxygenase mechanism

A detailed review on the 5-lipoxygenase mechanismhas been published [7] and doesn’t bear detailed repeat-ing here. However, a number of key factors are pertinentin understanding the efforts to discover a novel and safeinhibitor of 5-lipoxygenase. 5-lipoxygenase was found tobe a cytosolic enzyme which contained a non-haem ironatom. For maximum activity, it is necessary that theenzyme be converted from an inactive reduced state(Fe(II) state (Er)) to the active Fe(III) state (Eo) throughinteraction with an oxidizing agent such as fatty acidhydroperoxide, AOOH.

The oxidized enzyme then interacts with the fatty acidsubstrate (AH) to yield the Fe(II) enzyme with a boundpentadienyl radical (A•), which then reacts with molecu-lar oxygen to yield a hydropyroxy radical (A00•), whichthen picks up a hydrogen atom to yield the fatty acidhydroperoxide and to regenerate the oxidized enzyme.The exact mechanism whereby the 5-HPETE is thenconverted to leukotriene A4 is not fully understood but iscatalysed in a second rapid step by the same enzyme(figure 2). The enzyme itself is subject to turnover inac-tivation, presumably through the generation of reactiveradical biproducts. The arachidonic acid substrate isgenerally limiting in the system and is generated near thecell wall surface by the enzyme phospholipase A2.Processes are in place to rapidly reacylate liberatedarachidonic acid, such that free levels of arachidonic acidare normally very low in the cells. Thus, in order to haveoptimal activity, the 5-lipoxygenase enzyme translocatesunder mediation of a variety of factors, such as calciumlevels in the cell, to interact with the cell membrane.Thus, a variety of approaches can be considered for thedevelopment of inhibitors of 5-lipoxygenase. Consider-ing the redox cycling nature of the enzyme, it should bepossible to inhibit the enzyme by competition with analternative substrate which would itself be oxidizedthrough radical transfer, and thus, divert the enzyme fromits normal task. Such a substrate could itself produce

reactive intermediates which could facilitate turnoverinactivation of the enzyme. The inhibitor radicals thusproduced, if stable, could be cycled in their own right orgo on to decompose. Alternatively, inhibitors which arenot substrates for the enzyme could interact with eitherthe reduced state or the oxidized state of the enzyme (orboth) and form reversible dead-end inhibitor-enzymecomplexes. Finally, it could be possible to inhibit thetranslocation of 5-lipoxygenase to the membrane where itobtains substrate or to inhibit the transfer of the substrateto the enzyme. Although the details of these possibilitieswere not known as research began in the early 1980s tofind inhibitors of 5-lipoxygenase, inhibitors in all three ofthese classes have been discovered and, in some cases,shown to be useful drugs.

3. Redox inhibitors of 5-lipoxygenase(figure 3)

Many small organic compounds such as phenols,quinones, dihydroquinones, etc., can interact in redoxmechanisms. Early screening to find lipoxygenase inhibi-tors employed cell models such as human or rat polymor-phonuclear cells (PMNs) stimulated with calcium iono-phores, measuring the production of leukotriene B4.These initial efforts were rewarded by the discovery of awide variety of inhibitors, some of which showed quitepotent and apparently selective inhibition (selectivity wasgenerally determined relative to other oxygenase en-zymes such as cyclooxygenase or other lipoxygenasessuch as 12-lipoxygenase). Considering the lipophilicnature of the substrate, these inhibitors were generallysmall lipophilic molecules such as mono- and polycyclicaromatics.

In our own laboratories, we were initially elated todiscover the tricyclic benzothiazinone, L-615,919, whichshowed nanomolar potency for inhibition of leukotrienesynthesis in PMN cells stimulated with calcium iono-phore. The compound also showed bioavailability via theoral route and biochemical activity ex vivo. Our initialelation, however, was rapidly quenched when it wasfound that the compound promoted methaemoglobininformation in the blood of dogs treated with the drug. Itwas apparent that the drug not only interfered with theredox cycling of 5-lipoxygenase but also could serve toconvert iron in haemoglobin to the oxidized Fe(III) stateproducing profound toxicity. This was the first indicationof problems of non-specificity and ancillary activity dueto redox cycling that were to plague the development ofredox inhibitors of 5-lipoxygenase for many years tocome. Our initial feelings were that L-615,919 was tooredox-active and the product of its reduction (a dihy-droaminoquinone) would be too powerful an oxidizing

673

agent, and thus shows the observed toxicity. Examinationof the literature indicated a number of redox-activesubstances such as menadione [9] which were known tocause methaemoglobinaemia and also to produce super-oxide anions which led to other manifestations of toxicitysuch as Heinz body formation and haemolysis of bloodcells. Our efforts were then directed to produce com-pounds with less potent redox activity hoping that thesetypes of toxicities could be differentiated from the inhi-bition of the lipoxygenase enzyme itself. Modulation ofredox potential via substitution led to the discovery ofL-651,392 [10] which did not show methaemoglobin

formation in dog blood. The compound was, however, apotent 5-lipoxygenase inhibitor (IC50 = 60 nM (ratPMN)). Unfortunately, the compound was only poorlysoluble and poorly absorbed. It also showed a variety oftoxicities including genotoxicity and therefore the com-pound was abandoned. Further studies on phenothiazi-none analogues including the naphthyl analogue, such asL-649,927, provided compounds which were free of thetendency to form methaemoglobin, although, again,poorly absorbed. A prodrug carbonate, L-654,623, withbetter solubility and bioavailability which showed goodin vitro activity was identified (Y. Girard, unpublished

Figure 2. A kinetic mechanism of 5-lipoxygenase: oxygenase activity (right hand side), results from the reaction of the Fe(II) enzyme(ER) with the arachidonic acid substrate (AH) to yield the Fe(II) enzyme (ER) with a bound arachidonyl radical (A•), followed bybinding of O2 and reaction to yield hydroperoxyl radical (AOO•) bound to ER and then regeneration of EO and release of5-hydroperoxyeicosatetraenoic acid (AOOH). Free ER can result from dissociation of A• from ER(A•) or from reduction of EO byvarious reducing agents (RH). The product of one electron oxidation of RH (R•) can be reduced by thiols. Reduced enzyme (ER) canbe reoxidized by fatty acid hydroperoxide (AOOH) to yield a hydroxide ion and an alkoxy radical (AO•). Non-redox reversible deadend inhibitors (I) can bind to ER, EO and possibly other states of the enzyme as well.

674

results). Unfortunately, again, problems were encoun-tered in toxicological evaluation and L-654,623 wasshown to cause Heinz body formation in blood cells indogs. As this kind of toxicity could be attributed to theredox activity of this series, further work on phenothiazi-

nones was stopped. In later work, a number of hydroxy-benzofurans were identified through screening. Furtherstudies and optimization led to the identification ofL-656,224 [11, 12]. Although the compound showedgood in vitro and in vivo activity, toxicological evaluation

Figure 3. Redox inhibitors of 5-lipoxygenase.

675

revealed indications of hypersensitivity reactions, appar-ently due to metabolic conversion to quinone-type me-tabolites [13], and development was suspended. Furtherefforts at Merck Frosst derived a second dihydrobenzo-furanol, L-670,630 [14]. Unfortunately, again, signs oftoxicity similar to that observed for L-656,224 wereobserved and investigation of redox inhibitors was aban-doned at Merck Frosst in preference for a search forcompetitive non-redox inhibitors.

Other companies have investigated a variety of redoxinhibitors over the years, although none to date have beenbrought to market. The quinone AA-861 has been re-ported to be in clinical development but little recentinformation is available [15, 16]. Other phenolic5-lipoxygenase inhibitors have entered development andclinical trials, such as, TMK-688 (linazolast) [17], DuP-654 [18], R-68151 [19] and E-6080 [20]. None, however,have proceeded to market to this date and presumablyhave encountered a variety of problems. Many redox-active heterocyclic compounds such as BW-755C [21],ICI-207968 [22] and A-53612 [23] have been reported as5-lipoxygenase inhibitors over the years. However, stud-ies have shown that compounds of this type are prone toone electron oxidation [24] and can cause methaemoglo-bin formation in blood [22] and, thus, they apparentlyhave not been developed further.

Recently, a study on a series of tetrahydro-1,2,4-triazien-3-ols has been published by workers at Ab-bott [25] which are reported to be free of methaemoglo-binaemia in rats. This suggests that it may be possible toderive such inhibitors free of the toxicity problemsobserved in related phenothiazinone and phenol series of5-lipoxygenase inhibitors. The heterocyclic phenol fromJansen, R-68151, has been reported to be in clinical trialsfor psoriasis by the topical route and has apparentlyshown mild to moderate therapeutic effect [26]. A morepotent analogue, R-85355, was also investigated forpsoriasis but, unfortunately, was found not to showsignificant clinical activity [27].

4. Iron chelator inhibitors (figure 4)

The most successful efforts to derive non-toxic redoxtype inhibitors of 5-lipoxygenase have been in the area ofhydroxamic acids and related N-hydroxy ureas. Thesecompounds were designed basically with the expectationthat the functional groups might chelate iron and there-fore inhibit the enzyme. Studies in this area have beenpursued by researchers from Glaxo/Burroughs Wellcome,Abbott Laboratories, SmithKline Beecham, Wyeth Ay-erst, Ciba Geigy and many others.

Early studies of N-acylhydroxylamine compounds ledto the discovery of BW-A4C [28] and A-63162 [29] aspotent and selective 5-lipoxygenase inhibitors. However,BW-A4C was found to be rapidly metabolized in hu-mans [30] and shown to be oxidized by 5-lipoxygenase toform nitroxide radicals [31]. It was also found thatO-glucuronidization was a problem in this series, as wellas the hydrolysis of the hydroxamic acid which impairedin vivo activity [30]. Researchers at Abbott continuedextensive studies in this area concentrating on an analo-gous series of N-hydroxyureas which were more hydro-lytically stable, had reduced glucuronidization and there-fore superior in vivo properties. These efforts ultimatelyyielded A-64077 (Zileuton) which has undergone exten-sive clinical evaluation [32]. Zileuton has shown efficacyin chronic asthma where it provided some degree ofbronchodilation and anti-inflammatory and steroid spar-ing effects [33]. The compound was brought to market in1996 as the first of the new class of anti-leukotriene drugsas a 600 mg q.i.d. dose. The compound, however, hasshown a variety of adverse effects including elevatedliver enzymes and other hepatic toxicities as well assignificant drug interactions [32]. Zileuton has beenshown to induce a variety of liver enzymes, includingP450-2B and P450-4A, and induces hepatomegaly onchronic treatment in rats [34]. Recent reports have indi-cated that Zileuton showed efficacy in a rat model ofulcerative colitis [35]. However, a subsequent clinicaltrial at a dose of 600 mg q.i.d. for six months failed toshow activity significantly better than placebo, andshowed less efficacy when compared with mesala-zine [36]. Efforts by Abbott researchers have derived anumber of more potent analogues of Zileuton with betterpharmacokinetics which promise the potential for a q.d.drug with doses lower than required for Zileuton. Studiesderived an optimized compound, A78773, and its morepotent R(+) enantiomer, A79175 [37] which was reportedto have entered clinical trials [38, 39]. A phenol analogue,A-76745 (fenleuton) has been evaluated as potentialtreatment for allergic and inflammatory disorders inhorses [40]. The compound has been reported to havesome efficacy in a horse model of chronic obstructivepulmonary disease [41]. A thiophene analogue (ABT-761) (atreleuton) has apparently proceeded to Phase IIItrials as a second generation hydroxyurea 5-lipoxygenaseinhibitor and is reported to show potent biochemicalefficacy for nine hours after a single oral dose of 200 mg.The compound showed efficacy in blocking bronchocon-striction in asthmatics following a challenge by exer-cise [42]. ABT-761 has a long elimination half-life ofabout 15 hours consistent with once-a-day dosing [43].However, the compound has been reported to show some

676

drug interaction with oral contraceptive steroids [44]. Itremains to be seen whether atreleuton will reach themarket. Structure activity studies on the research whichderived atreleuton have been published [45].

SmithKline Beecham have studied a series ofdihydrobenzofuran-N-hydroxy ureas (which have derivedSB-210661 and SB-202235) [46, 47]. These compoundshave not yet been reported to have reached clinical trials.Ciba-Geigy/Novartis have reported on a series of hetero-cyclic hydroxy ureas including CGS-26529 [48] andCGS 23885 [49, 50], the latter compound has been

reported to have been in Phase I clinical trials. WyethAyerst researchers have reported structure-activity studieson a series of azophenoxyhydroxy ureas [51], however,the developmental stages of this series is unknown.Although the N-hydroxy ureas have been designed asiron chelators, it has been shown that these compoundscan be turned over by the 5-lipoxygenase enzyme toderive radical by-products. In particular, Zileuton hasbeen shown to be metabolized to form nitroxides and, assuch, serves as a reducing substrate for 5-lipo-xygenase [52]. If the degree of turnover by the enzyme

Figure 4. Iron chelator inhibitors of 5-lipoxygenase.

677

correlates with potency of these compounds, it is possiblethat potent analogues of Zileuton will continue to beplagued by toxicities derived from production of radicalsand radical by-products.

5. 5-Lipoxygenase inhibitors with dual activities(figure 5)

Many companies have reported over the last decadeson compounds that inhibit both 5-lipoxygenase and otherenzymes involved in inflammation as approaches to treata variety of inflammatory diseases. BW-B70C has beenreported to have dual 5- and 15-lipoxygenase inhibitoractivity [53]. The compound showed some efficacy in

allergic models in guinea-pigs. The phenolic benzothiaz-ole analogue, E-3040, from Eisai, was reported to havepotent dual 5-lipoxygenase and thromboxane A2 synthaseinhibitory activity. The compound has been evaluated inmodels of experimental ulcerative colitis [54] and hasbeen reported to be in Phase II clinical trials. Anotherquinonoid dual inhibitor of 5-lipoxygenase and throm-boxane A2 synthase (CV-6504) has been reported to showanti-tumour activity in murine models of adenocarci-noma [55].

Most efforts, however, have been directed towardsdevelopment of dual 5-lipoxygenase/cyclooxygenase in-hibitors. The compound tepoxaline, or RWJ-20485, hasreceived extensive biochemical and clinical evaluation by

Figure 5. Inhibitors of 5-lipoxygenase with dual activities.

678

Johnson & Johnson as an anti-inflammatory agent. How-ever, clinical trials indicated potent cyclooxygenase inhi-bition at doses from 25–800 mg p.o., whereas only weakinhibition of leukotriene synthesis was observed at themaximum 800 mg dose [56]. Development of the com-pound has been reported to be discontinued (companycommunication, March 1995). More recently, Johnson &Johnson have reported a cyclooxygenase 2/5-lipoxy-genase dual inhibitor, RWJ-63556 [57]. The companyMerckle has reported on a series of perazolene analoguesas dual cyclooxygenase/5-lipoxygenase inhibitors fromwhich they derived the clinical candidate ML-3000. Thecompound has shown some oral activity in rat models ofinflammation [58] and is reported to work by a non-redoxmechanism [59] and to show activity in a sheep model ofasthma [60] by aerosol administration. The compound isreported to be in Phase II clinical trials. Another dualcyclooxygenase/5-lipoxygenase inhibitor, VUFB-16066(flobufen), has reportedly been evaluated as inflammatoryin clinical trials for rheumatoid arthritis [61]. Researchersfrom Cytomed have recently reported on the hybridizatonof N-hydroxyurea functionality to diaryltetrahydrofuranto derive a novel series of compounds showing dual5-lipoxygenase and platelet activating factor receptor(PAF) antagonism. An optimized compound CMI-392was reported [62].

Although many companies continue to try and developdual inhibitors, it is clear that the difficulty to balance theinhibitory activity toward two different targets in vivo hasmade it difficult to optimize compounds in this area.

6. Competitive (non-redox) inhibitorsof 5-lipoxygenase(figure 6)

The multiple toxicities and difficulties encountered indeveloping redox inhibitors of 5-lipoxygenase has ledmany research groups to strive to find competitive non-redox inhibitors of the enzyme. Only as a better under-standing of the mechanism of the enzyme became avail-able, has it become possible to develop reliable criteriafor the evaluation and identification of non-redox inhibi-tors [52]. The observation by ICI (Zeneca) that a series ofmethoxyalkylthiazoles [63] and methoxytetrahydropyr-ans [64] were potent inhibitors of 5-lipoxygenase which,in some cases, showed enantioselective activity [63, 65]suggested they might act by stereospecific binding at theactive site of 5-LO. Structural features suggested theywere unlikely to enter into redox reactions. Subsequentstudies showed that compounds in this series, such asZM-211965, did not act as reducing substrates for5-lipoxygenase, while they did inhibit turnover-dependent inactivation of the enzyme and therefore could

be considered as true non-redox inhibitors [52]. Optimi-zation of these compounds led to the discovery ofZD-2138 [65, 66]. This compound was evaluated exten-sively by Zeneca in clinical trials for arthritis and asthma,and in normal volunteers. For example, a 350 mg p.o.dose was shown to maximally inhibit leukotriene synthe-sis for at least 24 hours [67]. Unfortunately, in spite ofthis potent activity, Zeneca has recently reported to havediscontinued development of this compound due tomixed and unconvincing results as an anti-arthritisagent [68]. It may still be in development for asthma. Ananalogue, ZM-230487, has also been evaluated preclini-cally by Zeneca [69, 70] and more recently structure-activity relationships on a series of thiene analogues havebeen published [71].

Recent studies have suggested that compounds such asZM-230487 are potent inhibitors of 5-lipoxygenase underconditions of low peroxide tone and are less efficientunder conditions of oxidative stress [72]. This couldexplain the disappointing activity observed for thesecompounds as anti-inflammatory agents.

A related series of non-redox inhibitors of 5-LO hasbeen discovered by Merck Frosst scientists. Screening ofthe Merck sample collection identified a class of lignansderived from a natural product, Justicidin E, as a moder-ately potent non-redox inhibitor 5-lipoxygenase [73].Based on similarities between these compounds andZD-2138, hybrid molecules such as L-697,198 wereprepared which were not only markedly more potent thanthe initial lignans, but also by virtue of the possibility ofdosing as a prodrug dihydroxyacid form, showed excel-lent bioavailability and oral activity [74]. However, ex-tensive metabolism of the pyran ring in these com-pounds [75] complicated their development. Structure-activity studies were carried out with the intention ofderiving a metabolically stable analogue. These effortsresulted in the identification of L-708,780 [76],L-739,010 [77] and L-746,530 [78]. These compoundsshowed only traces of metabolism and had excellentbioavailability and duration in a number of animalspecies. Unfortunately, it was noted that recovery of drugunder in vitro metabolic conditions was poor for thesecompounds and subsequent studies showed that both thebicyclo moiety and the furan moiety were metabolized toderive reactive metabolites which labelled plasma andhepatic proteins [79, 80]. Toxicological evaluation ofL-739,010 showed a variety of toxicities [81] and, thus,development of these compounds was discontinued.Other studies in the Merck Frosst Laboratories haveidentified a series of thiopyranylindoles [82] structurallyrelated to series which had been shown to indirectlyinhibit lipoxygenase activity in cells (vide infra). These

679

Figure 6. Competitive (non-redox) inhibitors of 5-lipoxygenase.

680

compounds, exemplified by L-689,065 [83], showed po-tent in vitro activity against human 5-lipoxygenase andoptimized compounds L-691,816 [84] andL-699,333 [85] were subsequently identified with excel-lent in vivo activity in a variety of models of asthmawhen dosed by the oral route. Development of this serieswas, however, inhibited by the modulation of ex vivoactivity through binding to plasma proteins in blood andby competing development of a series of indirect inhibi-tors in the same laboratory (see following discussion).

7. Inhibitors of the 5-lipoxygenase activating protein(FLAP) (figure 7)

Early screening studies in PMN cell assays at MerckFrosst looking for novel inhibitors of leukotriene biosyn-thesis identified a new class of indole inhibitors whichshowed potent inhibitory activity in intact PMN cells.Optimization of this series of indoles led to the discoveryof L-663,536 (or MK-886) [86]. This was a specificinhibitor of leukotriene biosynthesis in a variety of intact

Figure 7. Inhibitors of the 5-lipoxygenase activating protein (FLAP).

681

cell preparations but had no effect on either cyclooxyge-nase or 15- or 12-lipoxygenase derived products. Inbroken cell preparations or on purified 5-lipoxygenase,the compound had no activity. In addition, the compoundhad no activity on phospholipase A2 and did not inhibitarachidonic acid release. The compound also showedexcellent activity in a variety of animal models of asthma.Attempts to find the molecular target for MK-886 led tothe preparation of a photoaffinity label125I-L-669,083which was shown to label an 18 Kd protein in cellmembrane preparations [87]. In addition, MK-886 and avariety of analogues inhibited the binding of the photoprobe to this protein in a dose-dependent and potency-dependent manner. An analogue of MK-886 was used toprepare an affinity column from which it was possible toisolate this 18 Kd protein and based on sequence infor-mation attained from the rat protein, the cDNA for boththe rat and human protein were isolated, cloned andexpressed [87, 88]. Subsequent studies indicated that thisnovel protein (termed 5-lipoxygenase activating proteinor FLAP) is a necessary factor which must be present incells to facilitate the transfer of arachidonic acid to5-lipoxygenase [89, 90]. Cells which contain 5-lipoxy-genase and not FLAP are unable to biosynthesize leuko-triene products unless provided with a large excess ofexogenous arachidonic acid substrate. It was thus pro-posed that MK-886 inhibited the biosynthesis of leuko-trienes in PMNs and other cells by blocking the FLAP-mediated transfer of the arachidonic acid substrate to theenzyme. Other photo-affinity studies showed that arachi-donic acid binds to FLAP and this binding is competed byMK-886 and related analogues [91]. The discovery ofthis protein led to the possibility of setting up a bindingassay whereby inhibitors of arachidonic acid binding toFLAP could be evaluated directly [92]. MK-886 itselfentered clinical trials for asthma and while it showedsome efficacy, the results were somewhat disappoint-ing [93]. At maximally tolerated doses, however, inhibi-tion of leukotriene biosynthesis was not complete [94]and it was felt that a more potent compound would benecessary.

After the discovery of MK-886, a survey of theliterature identified a weak inhibitor, REV-5901 [95], aquinoline structure, which appeared to have similaractivity to MK-886 in that it was active in whole cells butnot in broken cells. Evolution of structures related toREV-5901 identified a series of quinoline inhibitorscharacterized by L-674,636 [96]. These compounds,though relatively potent and orally absorbed, wereplagued by significant modulation of their activitythrough binding to plasma proteins. Observation of simi-larities between the quinoline series and the indole series

of MK-886 led to the investigation of hybrid struc-tures [97]. The utilization of the FLAP binding assayallowed optimization of analogues of MK-886 and thediscovery of a significantly more potent FLAP inhibitor,L-686,708 (MK-0591) [98–100]. This compound inhib-ited FLAP binding with a potency at least 10-fold greaterthan MK-886, and showed potent inhibition of LT bio-synthesis in stimulated whole blood and also showedgood oral absorption and pharmacokinetics and was lessattenuated by plasma proteins. MK-0591 has been exten-sively evaluated in clinical trials where it showed efficacyboth for biochemical inhibition of leukotriene biosynthe-sis in vivo, as measured by excretion of LTE4 in urine inasthmatic patients, and in ex vivo assays as measured byinhibition of leukotriene B4 biosynthesis in stimulatedwhole blood [101]. Subsequent trials in chronic asthmashowed efficacy by a variety of parameters includingincrease of FEV1, decrease inâ-agonist usage andsymptom scores, although a small incidence of rash wasobserved as well [102]. The compound was administeredunder a b.i.d. regimen at 125 mg dose. Although MK-0591 showed clinical efficacy and oral activity, its develop-ment was suspended by Merck in favour of the leuko-triene D4 receptor antagonist MK-0476 (montelukast),which had superior activity and could be given by aonce-a-day regimen. MK-0591 has also been evaluated inclinical trials in ulcerative colitis. Again, MK-0591showed excellent biochemical efficacy [103, 104]. How-ever, similar to trials previously described for Zileuton,clinical efficacy in the disease was marginal and notstatistically significantly different than placebo [104].

Other laboratories have pursued the quinoline series ofleukotriene biosynthesis inhibitors and, in particular,Bayer have derived compounds which have been evalu-ated in clinical trials. BAY-X-1005 has been extensivelyevaluated in allergen and cold air challenged asthmaticsin clinical trials where it has shown efficacy [105, 106].In spite of this activity, the compound has been reportedto show rather limited anti-inflammatory potential [107].The compound has been reported to be in Phase II clinicaltrials and with Phase III trials planned, no recent infor-mation has been available as to its status. A relatedcompound from Bayer, BaY-Y-1015, has been investi-gated in animal models of IBD where it has shownefficacy better than olsalazine [108] but, to date, noreports of human clinical evaluation have been published.

8. Summary

Intensive efforts to develop clinically useful drugsfrom inhibitors of the 5-lipoxygenase enzyme or ofleukotriene biosynthesis have been carried out for almost

682

two decades now. As only one marketed drug, Zileuton,has emerged from this massive effort, this speaks stronglyto the high risk and great difficulty of developing newtherapies today. It is apparent from clinical results onMK-0591, BaY-X-1005 and ABT-761, that a variety ofleukotriene biosynthesis inhibitors could provide usefultherapy in asthma [109]. However, other issues of toxic-ity, pharmacokinetics or tolerability have frequentlyblocked bringing such compounds to the market. Of greatimportance has been the success of two leukotrienereceptor antagonists, namely, zafraleukast and mon-telukast which have also shown good efficacy in thetreatment of asthma [110–113].

To date, the leukotriene receptor antagonists appear tohave superior properties of safety, pharmacokinetics and,to some degree, efficacy. It remains to be seen whetherthere are disease states or situations where the morecomplete blockade of leukotriene biosynthesis, includingthe elimination of leukotriene B4, may prove to be anadvantage.

Acknowledgements

The author wishes to acknowledge the efforts of DianeSauvé and Mary Lynn Gaal for their help in preparing thismanuscript.

References

[1] Kellaway C.H., Trethewie E.R., Q. J. Exp. Physiol. 30 (1940)121–145.

[2] Murphy R.C., Hammarström S., Samuelsson B., Proc. Natl. Acad.Sci. USA 76 (1979) 4275–4279.

[3] Samuelsson B., Borgeat P., Hammarström S., Murphy R.C.,Prostaglandins 17 (1979) 785–787.

[4] Ford-Hutchinson A.W., Bray M.A., Doig M.V., Shipley M.E., SmithM.J.H., Nature (London) 286 (1980) 264–265.

[5] Ford-Hutchinson A.W., in: Rokach J. (Ed.), Leukotrienes andLipoxygenases, Elsevier, Amsterdam, 1989, pp. 405–425.

[6] Fitzsimmons B., Rokach J., in: Rokach J. (Ed.), Leukotrienes andLipoxygenases, Elsevier, Amsterdam, 1989, pp. 427–502.

[7] Ford-Hutchinson A.W., Gresser M., Young R.N., Annu. Rev. Bio-chem. 63 (1994) 383–417.

[8] Bell R.L., Summers J.B., Harris R.R., Annu. Rep. Med. Chem. 32(1997) 91–100.

[9] Goldberg B., Stern A., Biochim. Biophys. Acta 437 (1976) 628–632.

[10] Guindon Y., Girard Y., Maycock A., Ford-Hutchinson A.W., Atkin-son J.G. et al., Adv. Prostaglandins Thromboxane Leuk. Res. 17A(1987) 554–557.

[11] Bélanger P., Maycock A., Guindon Y., Back T., Dollob A.L. et al.,Can. J. Physiol. Pharmacol. 65 (1987) 2441–2448.

[12] Lau C.K., Bélanger P.C., Scheigetz J., Dufresne C., WilliamsH.W.R. et al., J. Med. Chem. 32 (1989) 1190–1197.

[13] Rouzer C.A., Riendeau D., Falgueyret J.-P., Lau C.K., Gresser M.J.,Biochem. Pharmacol. 41 (1991) 1365–1373.

[14] Lau C.K., Bélanger P.C., Dufresne C., Scheigetz J., Thérien M. etal., J. Med. Chem. 35 (1992) 1299–1318.

[15] Ancill R.J., Takahashi Y., Kibune Y., Campbell R., Smith J.R., J. Int.Med. Res. 18 (1990) 75–88.

[16] Fujimura M., Sasaki F., Nakatsumi Y., Takahashi Y., Hifumi S. et al.,Thorax 41 (1986) 955–959.

[17] Scrip No. 1306 (1988) p. 28.

[18] Harris R.R., Batt D.G., Galbraith W., Ackerman N.R., AgentsActions 27 (1989) 297–299.

[19] Marien K., Morren M., Degreef H., DeDoncker P., Rooman R.P.,Cauwenbergh G., Arch. Dermatol. 128 (1992) 993–994.

[20] Beetens J.R., Somers Y., Goossens J., Vanwauwe J., De Clerck F.,Arch. Int. Pharmacodyn. Ther. 290 (1987) 151.

[21] Fort F.L., Pratt M.C., Carter G.W., Lewkowski J.P., Heyman I.A. etal., Fundam. Appl. Toxicol. 4 (1984) 216–220.

[22] Bruneau P., Delvare C., Edwards M.P., McMillan R.M., J. Med.Chem. 34 (1991) 1028–1036.

[23] Albert D.H., Machinist J., Young P.R., Dyer R., Bouska J. et al.,FASEB J. 3 (1989) A735.

[24] Duniec Z., Robak J., Gryglewski R., Biochem. Pharmacol. 32(1983) 2283–2286.

[25] Bhatia P.A., Brooks C.D., Basha A., Ratajczyk J.D., Gunn B.P. et al.,J. Med. Chem. 39 (1996) 3938–3950.

[26] Degreef H., Dockx P., DeDoncker P., De-Beule K., CauwenberghG., J. Am. Acad. Dermatol. 22 (1990) 751–755.

[27] van de Kerkhof P.C., van Pelt H., Lucker G.P., Steijlen P.M.,Heremans A., Skin Pharmacol. 9 (1996) 307–311.

[28] Tateson J.E., Randall R.W., Reynolds C.H., Jackson W.P., Bhattach-erjee P. et al., Br. J. Pharmacol. 94 (1988) 528–539.

[29] Summers J.B., Gunn B.P., Martin J.G., Mattin M.B., Mazdiyasni H.et al., J. Med. Chem. 31 (1988) 1960–1964.

[30] Payne A.N., Jackson W.P., Salmon J.A., Nicholls A., Yeadon M.,Garland L.G., Agents Actions Suppl. 34 (1991) 189–199.

[31] Chamulitrat W., Mason R.P., Riendeau D., J. Biol. Chem. 267 (1992)9574–9579.

[32] Wenzel S.E., Kamada A.K., Ann. Pharmacother. 30 (1996) 858–864.

[33] Liu M.C., Dube L.M., Lancaster J., J. Allergy Clin. Immunol. 98(1996) 859–871.

[34] Rodrigues A.D., Machinist J.M., Toxicol. Appl. Pharmacol. 137(1996) 193–201.

[35] Zarif A., Eiznhamer D., Callaghan C., Doria M.I., Broutman L.,Keshavarzian A., Inflammation 20 (1996) 217–227.

[36] Hawkey C.J., Dube L.M., Rountree L.V., Linnen P.J., Lancaster J.F.,Gastroenterology 112 (1997) 718–724.

[37] Bell R.L., Bouska J.B., Malo P.E., Lanni C., Harris R.R. et al., J.Pharmacol. Exp. Ther. 272 (1995) 724–731.

[38] Bouska J. et al., 6th Int. Conf. Inflamm. Res. Assoc., Sept. 20–24,White Haven (1992) Abst. P119.

[39] Brooks D.W., Stewart A.O., Basha A., Bhatia P., Ratajczyk J.D. etal., 206th ACS Nat. Meeting, Aug. 22–27, Chicago (1993) Abst.Medi 41.

[40] Annual Drug Data Report, 17 (No. 8) (1995) 711–712.

[41] Marr K.A., Lees P., Page C.P., Cunningham F.M., Inflamm. Res. 44(Suppl. 1) Abst. W20/07, (1995) S289.

[42] Lehnigk B., Rabe K.F., Dent G., Herst R.S., Carpentier P.J.,Magnussen H., Eur. Resp. J. 11 (1998) 617–623.

[43] Wong S.L., Drajesk J., Chang M.S., Witt G., Awni W.M., Biopharm.Drug. Disp. 19 (1998) 159–162.

[44] Wong S.L., O’Dea R., Dube L., Awni W.M., 99th Meeting Am. Soc.Clin. Pharm. Ther., New Orleans (1998) Abst. PI-106.

683

[45] Brooks C.D., Stewart A.O., Basha A., Bhatia P., Ratajczyk J.D. etal., J. Med. Chem. 38 (1995) 4768–4775.

[46] Adams J.L., Garigipati R.S., Sorenson M., Ross S., Schmidt S.J. etal., 204th ACS Nat. Meeting, Aug. 23–28, Washington (1992) Abst.Medi 121.

[47] Chabot-Fletcher M.C., Underwood D.C., Breton J.J., Adams J.L.,Kagey-Sobotka A. et al., J. Pharmacol. Exp. Ther. 273 (1995)1147–1155.

[48] Kimble E., Kowalski T., Peters P., Lee W., Koehler J. et al., Inflamm.Res. 44 (Suppl. 2) (1995) S147–S148.

[49] Satu Y., Stanton J.L., Hutchison A.J., Libby A.H., Kowalski T.J. etal., 205th ACS Nat. Meeting, March 28–April 2; Denver (1993)Abst. Medi 143.

[50] Raychaudhuri A., Chertock H., Chovan J., Jones L.S., Kimble E.F.et al., Naunyn-Schmiedeberg’s Arch. Pharmacol. 355 (1997)470–474.

[51] Malamas M.S., Carlson R.P., Grimes D., Howell R., Glaser K. et al.,J. Med. Chem. 39 (1996) 237–245.

[52] Falgueyret J.-P., Hutchinson J.H., Riendeau D., Biochem. Pharma-col. 45 (1993) 978–981.

[53] Yeadon M., Price R.C., Payne A.N., Pulmon. Pharmacol. 8 (1995)53–63.

[54] Oketani K., Murakami M., Inoue T., Yamanaka T., Yamatsu I., Jpn.J. Pharmacol. 64 (Suppl. I) Abst. (1994) P351.

[55] Hussey H.J., Bibby M.C., Tisdale M.J., Br. J. Cancer 73 (1996)1187–1192.

[56] Depre M., VanHecken A., Verbesselt R., Verpooten G.A., Arnout J.et al., Int. J. Clin. Pharm. Res. 16 (1996) 1–8.

[57] Parnham M.J., Drug News Perspect. 9 (1996) 631–639.

[58] Laufer S.A., Augustin J., Dannhardt G., Kiefer W., J. Med. Chem. 37(1994) 1894–1897.

[59] Laufer S.A., Tries S., Augustin J., 13th Int. Symp. Med. Chem.,Sept. 19–23, Paris (1994) Abst. P181.

[60] Abraham W.M., Laufer S., Tries S., Pulmon. Pharmacol. Ther. 10(1997) 167–173.

[61] Panajotova V., Anderova E., Jandera A., Kuchar M., Arzneim.-Forsch 47 (1997) 648–652.

[62] Cai X., Scannell R.T., Yaeger D., Hussoin M.S., Killian D.B. et al.,J. Med. Chem. 41 (1998) 1970–1979.

[63] Bird T.G.C., Bruneau P., Crawley G.C., Edwards M.P., Foster S.J. etal., J. Med. Chem. 34 (1991) 2176–2186.

[64] Crawley G.C., Dowell R.I., Edwards P.N., Foster S.J., McMillanR.M. et al., J. Med. Chem. 35 (1992) 2600–2609.

[65] Lambert-van der Brempt C., Bruneau P., Lamorlette M.A., ForsterS.J., J. Med. Chem. 37 (1994) 113–124.

[66] Bird T.G.C., 12th Int. Symp. Med. Chem., Sept. 13–17 Basel (1992)Abst. OC-05. 2.

[67] Walker E.R.H., 13th Int. Symp. Med. Chem., Paris, (1994) Abst.ML22.

[68] Scrip No. 2050 (1995) p. 7.

[69] Kusner E.J., Buckner C.K., Dea D.M., DeHaas C.J., Marks R.L. etal., Eur. J. Pharmacol. 257 (1994) 285–292.

[70] Teixeira M.M., Hellewell P.G., Brit. J. Pharmacol. 111 (1994)1205–1211.

[71] Crawley G.C., Briggs M.T., J. Med. Chem. 38 (1995) 3951–3956.

[72] Werz O., Szellas D., Henseler M., Steinhilber D., Mol. Pharmacol.54 (1998) 445–451.

[73] Thérien M., Fitzsimmons B.J., Sheigetz J., Macdonald D., ChooL.Y. et al., Bioorg. Med. Chem. Lett. 3 (1993) 2063–2066.

[74] Ducharme Y., Brideau C., Dubé D., Chan C.-C., Falgueyret J.-P. etal., J. Med. Chem. 37 (1994) 512–518.

[75] Chauret N., Li C., Ducharme Y., Trimble L.A., Yergey J.A. et al.,Drug Metab. Disp. 23 (1995) 65–71.

[76] Delorme D., Ducharme Y., Brideau C., Chan C.C., Chauret N. et al.,J. Med. Chem. 39 (1996) 3951–3970.

[77] Hamel P., Riendeau D., Brideau C., Chan C.C., Desmarais S. et al.,J. Med. Chem. 40 (1997) 2866–2875.

[78] Dubé D., Blouin M., Brideau C., Chan C.C., Desmarais S. et al.,Bioorg. Med. Chem. Lett. 8 (1998) 1255–1260.

[79] Zhang K.E., Naue J.A., Arison B., Vyas K.P., Chem. Res. Toxicol. 9(1996) 547–554.

[80] Chauret N., Nicoll-Griffith D., Friesen R., Li C., Trimble L. et al.,Drug Metab. Disp. 23 (1995) 1325–1334.

[81] Molon-Noblot S., Gillet J.P., Durand-Cavagna G., Huber A.C.,Patrick D.H., Duprat P., Toxicol. Pathol. 24 (1996) 231–237.

[82] Hutchinson J.H., Charleson S., Evans J.F., Falgueyret J.P. et al., J.Med. Chem. 38 (1995) 4538–4547.

[83] Hutchinson J.H., Prasit P., Choo L.Y., Riendeau D., Charleson S. etal., BioMed. Chem. Lett. 45 (1993) 978–981.

[84] Hutchinson J.H., Riendeau D., Brideau C., Chan C., Delorme D. etal., J. Med. Chem. 36 (1993) 2771–2787.

[85] Hutchinson J.H., Riendeau D., Brideau C., Chan C., Falgueyret J.P.et al., J. Med. Chem. 37 (1994) 1153–1164.

[86] Gillard J., Ford-Hutchinson A.W., Chan C.C., Charleson S., DenisD. et al., Can. J. Physiol. Pharmacol. 67 (1989) 456–464.

[87] Miller D.K., Gillard J.W., Vickers P.J., Sadowski S., Léveillé C. etal., Nature 343 (1990) 278–281.

[88] Dixon R.A.F., Diehl R.E., Opas E., Rands E., Vickers P.J. et al.,Nature 343 (1990) 282–284.

[89] Kargman S., Vickers P.J., Evans J.F., J. Cell. Biol. 119 (1992)1701–1709.

[90] Kargman S., Rousseau P., Reid G.K., Rouzer C.A., Mancini J.A. etal., J. Lipid Mediators 7 (1993) 31–45.

[91] Mancini J.A., Abramovitz M., Cox M.E., Wong E., Charleson S. etal., FEBS Lett. 318 (1993) 277–281.

[92] Charleson S., Prasit P., Léger S., Gillard J.W., Vickers J.P. et al.,Mol. Pharmacol. 41 (1992) 873–879.

[93] Friedman B.S., Bel E.H., Buntinx A., Tanaka W., Han Y.-H.R. et al.,Am. Rev. Respir. Dis. 147 (1993) 839–844.

[94] Depre M., Friedman B., Tanaka W., VanHecken A., Buntinx A.,DeSchepper P.J., Clin. Pharmacol. Ther. 53 (1993) 602–607.

[95] Musser J.H., Chakraborty U.R., Sciortino S., Gordon R.J., Khand-wala A. et al., J. Med. Chem. 30 (1987) 96–104.

[96] Prasit P., Belley M., Evans J.F., Gauthier J.Y., Léveillé C. et al.,Bioorg. Med. Chem. Lett. 1 (1991) 645–648.

[97] Young R.N., Gillard J.W., Hutchinson J.H., Léger S., Prasit P., J.Lipid Mediators 6 (1993) 233–238.

[98] Prasit P., Belley M., Blouin M., Brideau C., Chan C., Charleson S.et al., J. Lipid Mediators 6 (1993) 239–244.

[99] Mancini J.A., Prasit P., Coppolino M.G., Charleson S., Léger S. etal., Mol. Pharmacol. 41 (1992) 267–272.

[100] Brideau C., Chan C., Charleson S., Denis D., Evans J.F. et al., Can.J. Physiol. Pharmacol. 70 (1992) 799–807.

[101] Diamant Z., Timmers M.C., van der Veen H., Friedman B.S. et al.,J. Allergy Clin. Immunol. 95 (1995) 42–51.

[102] Chapman K.R., Friedman B.S., Shingo S., Heyse J., Reiss T.,Spector R., ALA/ATS Meeting, Boston, May 21–25 (1994) A215.

[103] Hillingso J., Kjeldsen J., Laursen L.S., Lauritsen K., von Spreck-elsen S. et al., Clin. Pharmacol. Ther. 57 (1995) 335–341.

684

[104] Roberts W.G., Simon T.J., Berlin R.G., Haggitt R.C., Snyder E.S. etal., Gastroenterology 112 (1997) 725–732.

[105] Hamilton A.L., Watson R.M., Wyile G., O’Byrne P.M., Thorax 52(1997) 348–354.

[106] Fischer A.R., Rosenberg M.A., Roth M., Loper M., Jungerwirth S.,Isreal E., Thorax 52 (1997) 1074–1077.

[107] Muller-Peddinghaus R., J. Physiol. Pharmacol. 48 (1997) 529–536.

[108] Murthy S., Murthy N.S., Coppola D., Wood D.L., Inflamm. Res. 46(1997) 224–233.

[109] Negro J.M., Miralles J.C., Ortiz J.L., Funes E., Garcia A., Allergol.Immunopathol. 25 (1997) 209–216.

[110] O’Byrne P.M., Chest III (2 Suppl.) (1997) 27S–34S.

[111] Larson J.S., Jackson S.K., Am. J. Health Syst. Pharm. 53 (1996)2821–2830.

[112] Sorkness C.A., Pharmacotherapy 17 (1997) 50S–54S.

[113] Renzi P.M., Can. Med. Assoc. J. 160 (1999) 217–223.

685

Original article

The COREPA approach to lead generation: an application to ACE-inhibitors

Verginia Kamenska, Julian Ivanov, Ovanes Mekenyan*

Bourgas University “As. Zlatarov”, Department Of Physical Chemistry, 8010 Bourgas, Bulgaria

(Received 17 August 1998; revised 19 January 1999; accepted 21 January 1999)

Abstract – The recently derived algorithm for identifying the COmmon REactivity PAttern (COREPA) of structurally diverse chemicalshaving similar biological behaviour was employed to recognize the structural requirements for high ACE inhibition. COREPA is based on anassessment of all energetically-reasonable conformations of the compounds under study. It is not dependent upon a predetermined andspecified pharmocophore or an alignment of active conformers. The approach describes the common reactivity pattern in terms of global andlocal stereoelectronic parameter ranges. These ranges are associated with compounds having extreme (highest and lowest) biological activitythrough a comparison of conformer distributions for specific descriptors. The defined parameter ranges were combined into Booleanexpressions providing flexible screening of chemicals by making use of a new chemical rule interpreter allowing a simultaneous searchaccording to all available 2-D and 3-D information. The structural combinations, i.e., COREPAs, derived on a training set of ACE inhibitorswere used for screening chemicals in an external test series for identifying those with high ACE inhibition activity. The obtained resultsshowed that the structural requirements derived in the present work can be useful for screening chemicals from databases as potential leadACE inhibitors. © 1999 E´ditions scientifiques et médicales Elsevier SAS

quantitative structure-activity relationships / active analogues / stereoelectronic structure / conformational flexibility / ACE-inhibitors/ lead generation / COREPA

1. Introduction

Angiotensin-converting enzyme (ACE) is a Zn-containing metallopeptidase which catalyses the hydroly-sis of the terminal dipeptide from the angiotensin I(decapeptide) to produce the octapeptide angiotensin II.The latter appears to be one of the most potent vasocon-strictors. Although the primary amino acid sequence ofACE is known, [1–3] its 3-D structure is still undeter-mined. Inhibitors of ACE are widely prescribed to controlessential hypertension. The structural requirements forhigh ACE activity within congeneric series of chemicalshave been derived from intensive SAR studies [4–6]. Thelatter, combined with crystallographic data from theanalogous enzyme, thermolysin, and its inhibitors, de-fined the minimal set of the active groups necessary for achemical to elcite ACE inhibition: a C-terminal carboxylgroup for ionic binding to the enzyme; a carbonyl oxygenwhich hydrogen bonds to some active site residue; someZn-binding functional groups such as a carboxylate,hydroxamate, phosphonate, or thiolate. This structural

information has been used to screen databases of struc-turally diverse classes of ACE-inhibitors to elicite thecommon three-dimensional geometry of the pharmacoph-oric sites consistent with their activity [7–9]. Still, therecognition of the 3-D spatial alocation of the pharma-cophore responsible for ACE inhibition potency is asubject of intensive QSAR studies employing a variety ofpattern recognition techniques.

A set of 28 inhibitors of angiotensin converting en-zyme, selected by Mayer et al. [8], were modelled by theactive analogue technique to elucidate the pharmacoph-ore, i.e., the essential 3-D arrangement of functionalgroups that a molecule must possess to be recognized bythe receptor under study. A strategy for systematic sam-pling of the conformational space has been used, thus,arriving at two plausible alternative active site hypo-theses [10, 11], demonstrating that the multiple searchesof the conformational space of flexible ligands is aneffective way to achieve increasing reliability of themodelling results.

The CoMFA methodology (comparative molecularfield analysis) is one of the most applied 3D-QSAR*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 687−699 687© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

approaches to analyse ACE activity of chemicals. De-Priest et al. [12] and Waller et al. [13] applied the CoMFAparadigm. They enlarged the initial set of 28 moleculeswith structures from additional classes of ACE inhibitorsto resolve some uncertainties associated with recenthypotheses for the pharmacophoric site [8]. The CoMFAmodels of DePriest et al. [7], derived from potentialfields, were found to be insufficient for accurately quan-tifying the enzyme-inhibitor interactions. The model andits predictive ability was improved by introducing a Zinkindicator variable explicitly describing the Zink-ligandinteraction. The derived 3D-QSAR models have beenfurther used for predicting ACE activity of chemicals notincluded in the initial training set [12].

Pharmacophore search methods and receptor-site map-ping, such as the active analogue technique and CoMFA,face significant challenges, which include the selection ofappropriate conformations and obtaining an alignment ofthese structures. There are a number of good techniquesfor superimposing molecules [14–17], but developing arobust alignment model is not trivial. Typically, hundredsof alignments are explored to reach an optimal outcome,which, if not carefully evaluated and explained in thecontext of a presumed mechanism of interaction with thereceptor may become susceptible to violation of thecriteria of Topliss and Edwards [18] for causality instructure-activity models. Alignment errors can also leadto models that are incorrect or are poorly predictive.Further, the use of the lowest-energy conformers inmethods such as CoMFA to assess similarity in pharmo-cophore search and receptor-mapping alogorithms seemsinappropriate because, in complex systems such as bio-logical tissues and fluids, chemicals are likely to exist ina variety of conformational states. In fact the lowest-energy, gas-phase, conformations might be the leastlikely to interact with the solvent or macromol-ecules [19], and solvation and binding interactions couldmore than compensate for energy differences among theconformers of a chemical [20–23].

In an attempt to address the issue of conformationalflexibility, Prendergast et al. [24] reported an augmentedactive analogue technique to identify specific conformersof ligands acting as antagonists to angiotensin-II. Allgeometrically reasonable conformers were assessed;however, conformational energies were not evaluated andan energy minimization was not performed during thesearch. In a sense, this methodology can be viewed as anaugmented version of the active analogue approachbecause it accounts for conformational flexibility and iteliminates the necessity of conformer alignment.

Recently, we developed a new pattern recognitionapproach, named COREPA (COmmon REactive PAttern)

employed as a pharamacophore mapping technique.COREPA is a generalization of the active analogueapproach and it provides an implicit exploration of areceptor’s stereoelectronic shape. As opposed to existingpharmacophore search methods, the COREPA is anattempt to circumvent the problems related to the selec-tion of appropriate conformations and obtaining appro-priate template alignments [25, 26]. For each stereoelec-tronic parameter identified to be associated with theendpoint studied, all energetically-reasonable conformerdistributions for the compounds from the training set aresuperimposed and the parameter ranges common for atleast one conformer from all of the compounds identified.The collection of common stereoelectronic parameterranges defines the common reactivity pattern. Due to theincreasing pharmacological interest in potent vasocon-strictor agents, the present work aims to employ theCOREPA method for the recognition of stereoelectronicrequirements for a high ACE inhibition activity.

2. Materials and methods

2.1. The COREPA algorithm

The methodology to elucidate chemical similarity isbased on the assumption that chemicals that elicit similarbiological behaviour through a common mechanism ofaction should also possess similarities in stereoelectronicdescriptors. Elucidation of this common reactivity patternwithin a set of biologically-similar chemicals requiresexamination of the conformational flexibility of thecompounds to evaluate molecular similarity in the con-text of the associated variability in specific stereoelec-tronic parameters.

The principal steps of the algorithm are presented anddiscussed in detail in Mekenyan et al. [25]. However, inorder to follow the forthcoming, they are summarized asfollows:

Step 1. Definition of the training set of chemicals. Adefined subset of chemicals in the reaction series underinvestigation are selected as the training set. The trainingset can include either the most or least active chemicals,as defined by a user-imposed threshold of biologicalactivity. This initial step establishes the extent of biologi-cal similarity among the chemicals from which stereo-electronic similarity will be discerned in the subsequentsteps of the algorithm.

Step 2. Evaluation of stereoelectronic parameters hy-pothesized to be associated with biologically similarcompounds. A restricted set of parameters, hypothesizedto be associated with biological activity, are evaluatedbased on the normalized sum of dynamic similarity

688

Figure 1. The training set of 28 AR ligands examined in this study (training set A).

689

indices [27] between each pair of molecules in thetraining set. The Tanimoto coefficient, TAN, and therelated Hodgkin-Richards similarity metric, 3D-TAN,cosine index, COSINE, and an index based on informa-tion theory, IF, were used as similarity indices, whereas,the Euclidean metric, DIST, was employed as a dissimi-larity measure [27]. The stereoelectronic parameters thatprovide the maximal measure of similarity among thechemicals in the training set are assumed to be mostclosely associated with the activity under considerationand are used in the subsequent step of the algorithm.

Step 3. Recognition of the common reactivity pattern.For each stereoelectronic parameter identified in step 2,the conformer distributions of the chemicals from thetraining set are superimposed and the parameter rangescommon for conformers from all of the chemicals iden-tified. The collection of common stereoelectronic param-eter ranges defines the common reactivity pattern.

2.2. ACE ligands and binding affınity

A training set of 28 AR ligands examined in this study(training set A) are those included in the learning set ofMayer et al. [8]. The structures are depicted infigure 1.The pIC50 values (the negative log of the inhibitionconcentration providing a standard biological response)

are listed above the structures infigure 1. In the followinganalyses, sets of known ACE inhibitors such as captopril,pivalopril, rentiapril, zofenopril, enalapril, ramapril,quinapril, perindopril, cilazapril, delapril, lisinopril, SQ29,852 and fosinopril [28–30] were used as an externalvalidation data set (test set B) to evaluate discriminationabilities of the derived reactivity pattern. Their structuresand associated activities (pIC50) are presented infigure 2.

2.3. ACE ligand conformations

A primary aspect of the COREPA approach is toevaluate the conformational space for the chemicalsunder study using a number of conformers that canreasonably be assumed to represent the diversity ofrelevant stereoelectronic character for the biological pro-cess of interest. Sampling of conformational space wasperformed, aiming to provide up to 50 structurally mostdistinct conformers for each of the structures included inthe training set. The conformational search was per-formed by making use of the 3DGEN algorithm thatinitiates from molecular topology and generates all con-formers consistent with steric constraints and expertrules [31]. Specific geometric constraints were imposedduring structure generation. Thus, torsion resolution (TR)around “saturated” (SP3-SP3) acyclic bonds was chosen

Figure 2. The test set of 13 captopril derivatives (test set B).

690

to be 120°, using an initial torsion angle of 60°, withrespect to the plane of the preceding three atoms. Up to 7rotational variables were chosen for the acyclic part ofeach chemical. For all chemicals, 1.5 Å was set as thedistance between non-bonded atoms, while 1.2–2.0 Åwas the range imposed for ring closure. After generationof the initial set of conformers, up to 50 of the structurallymost diverse conformers were screened for each chemicalbased on structural dissimilarity of conformers as as-sessed by the sum of the distances between their non-hydrogen atoms. Each of the generated conformers wassubmitted to force field optimization based on a simpleenergy-like function, where only the electrostatic termsare omitted. Subsequently, the conformational degen-eracy of the isomers was detected due to molecularsymmetry and geometry convergence. Subsequent geom-etry optimization of the conformers was employed withMOPAC 7 [32], using the AM1 Hamiltonian with the keywords “PRECISE” and “NOMM”. Finally, these con-formers were screened to eliminate those structures with∆Hf

o values that were 20 kcal/mol or higher than thatcalculated for the conformer associated with the “abso-lute” energy minimum (see Results and discussion for anexplanation of this threshold). The COREPA analyseswere performed on this resulting data set.

2.4. Molecular descriptors

Stereoelectronic parameters were calculated with MO-PAC 7 [32], augmented by a new computing module [33],that provides additional reactivity descriptors, using theAM1 all-valence electron semi-empirical Hamiltonian.The electronegativity (EN), dipol moment (µ), volumepolarizability (Vol.P), energy of frontier orbitals (EHOMO

and ELUMO), and electronic gap (EHOMO-LUMO) wereused as global electronic descriptors, whereas the atomiccharges (qi), frontier atomic charges (fi

HOMO and fiLUMO),

and acceptor superdelocalizability indices (SiE and Si

N),as well as atomic self-polarizabilities (πii ), were calcu-lated as local electronic indices (i denotes a specific atomin a molecule). As it was already mentioned, whensearching common patterns based upon local parameterdistributions, the atomic reactivity indices were notrestricted to specific rings in studied derivatives.

The delocalizabilities were calculated by using thefollowing equations:

SiE = (

j

occ

(a

i

Cja Cja /~ Eref − Ej ! (1)

SiN = (

j

vac

(a

i

Cja Cja /~ Eref − Ej ! (2)

where Ej are the molecular orbital (MO) energies; Cj arethe corresponding eigenvectors, and a pertains to theatomic orbitals of site i; Eref is a non-zero, fixed, energyreference level, that was set equal to the electronicmidgap level for benzene (–4.549 eV), i.e., Eref =0.5(EHOMO + ELUMO)benzene. Superdelocalizability indi-ces were used to assess the ability of a reactant to formbonds through charge transfer, i.e., donor superdelocaliz-ability indices are measures of the ability of molecules todonate electron density through orbital transfers, whileacceptor superdelocalizability indices are measures of theability of molecules to accept electron density.

Conformer structures were also assessed based on thesteric descriptors GW (sum of geometric distances [34]),Lmax (the greatest interatomic distance), dij (steric dis-tance between atoms i and j) and planarity (the normal-ized sum of torsion angles in a molecule [25]). Finally,Vol.Polar., defined as a sum of atomic self-polarizabilities, and thus, the averaged ability of a com-pound to change electron density at its atoms duringchemical interactions [35, 36], was selected as a physio-chemical descriptor. Lower values of Vol. Polar. (Vol.P>0) reflect higher charge localizations and more polariz-able (less lipophilic) molecules [36]. These descriptorswere selected because hydrophobicity, steric bulk andsize constraints have been reported as important criteriain predicting and interpreting ligand binding interactions.

2.5. Database screening approach

The common reactivity pattern, i.e., the pharmaco-phore, defined in terms of 2-D structural fragments andassociated 3-D stereochemical information (such as par-ity of stereocentres, distance between pharamacophoricsites, charge and delocalizability requirements) can befurther used to screen databases for the presence ofchemicals meeting these structural characteristics. A newchemical rule interpreter (CRI) has been developed forthis purpose allowing simultaneous search according allavailable 2-D and 3-D information [37]. With that aim,substructure search techniques were combined with therange requirements for numeric descriptors as introducedin the extended SMILES language [37]. These entities,combining structural fragments with range requirementfor numeric descriptors, are further called SD (structure-descriptor) screens. The CRI provides selective classifi-cation or screening of the chemicals from an OASISdatabase file (*.CMP) [33] according to a set of rules,given in terms of SD screens. A particular classificationscheme is developed and described in a text file (*.RULfile). A separate rule of the scheme can be used to selectfrom the *.CMP file those chemicals that satisfy it. The

691

application of the whole classification scheme results in acase-specific assignment of descriptors to the chemicalsof the database. The *.RUL file is comprised of threesections, namely, Defines, Rules and Apply sections.

In the Defines section, the pertinent SD screens aredescribed by means of extended SMILES language. It isvery often practical to combine multiple SD screens in asingle group that is used further as a single composite SDscreen. For this purpose, the program supports a dynamiclibrary of SD screens, partitioned in groups. Composite orsimple SD screens have user defined labels–screen iden-tifiers. Once defined, screen identifiers can be furtherused as parts of SMILES strings in order to define newSD screens. Screen identifiers are SD screens on theirown since they can serve for SD search or for definitionof new screen identifiers. When SD search is performed,each encountered screen identifier is recursively replacedby its actual contents, and the result is positive if at leastone match is found.

Explicit SD screens and predefined screen identifiersconstitute the basic logical variables of Boolean expres-

sions (BL rules). The BL rules in use are described in theRules section of the *.RUL file. An SD screen in thecontext of a BL rule takes logical values “true” or “false”if the SD screen is encountered in the current chemical ornot. A BL rule combines a Boolean expression SD screenrelated with the logical operators “and”, “or” and “not”.The default priority of Boolean operations can be re-defined by the use of brackets in unlimited nesting. As inthe case of SD screens, separate BL rules can be assignedto different rule identifiers. An explicit BL rule or a ruleidentifier can be selected, and applied to a *.CMP file toproduce in an output *.CMP file the subset of chemicalsthat match the rule.

In the Apply section, BL rules are employed inconditional if, then, elsestatements. The condition of astatement is a rule identifier or a Boolean expression ofrule identifiers.

Examples are given with some of the structural re-quirements for high ACE-activity included in Defines andRules sections of *.RUL file (table IV). Comments areincluded to provide explanation of the screens.

Defines:

RX: O, S, N, P Wild-card atom RX holds for any of O, S, N and P atoms.Z1: O_O{5.6< DISTANCE < 5.9} Two O-atoms in a distance range of 5.6–5.9 [Å].Z4: O{–0.39< Q < –0.25}_N{–0.39< Q < –0.25}{6.9 < DISTANCE < 7.4} O- and N-atoms having charges in the range of –0.39 to –0.25

[a.u.] to be in a distance range of 6.9–7.4 [Å].Z5: RX{–0.39 < Q < –0.25}_RX{–0.39< Q < –0.25}{8.7 < DISTANCE < 9.4} Two wild-card atoms RX (O, S, N and P) having charges in the range of –0.39 to

–0.25[a.u.] to be in a distance range of 8.7–9.4 [Å].Z7: O{0.22 < DONOR_DLC< 0.30}_O{0.22<DONOR_DLC< 0.30}{8.7 < DISTANCE < 9.4} Two O-atoms having donor delocalizabilties in the range of 0.22–0.30

[(a.u.)2/eV] to be in a distance range of 8.7–9.4 [Å].Z9: RX{0.22 < DONOR_DLC< 0.30}_RX{0.22 <DONOR_DLC< 0.30}{8.7 < DISTANCE < 9.4} Two wild-card atoms RX (O, S, N and P) having donor delocalizabilties in the

range of 0.22–0.30[(a.u.)2/eV] to be in a distance range of 8.7–9.4 [Å].Rules:r2: “Z7” and “Z8” Simultaneous fulfillment of the requirements “Z7” and “Z8”.r4: “Z3” or “Z7” Fulfillment of any one of requirements “Z3” or “Z7”.r5: (“Z3” or “Z4”) and (“Z7” or “Z8”) Simultaneous fulfillment of the structural combination in brackets. The first of

these combinations means fulfillment of any one of the requirements “Z3” or “Z4”whereas the second means fulfillment of any one of the requirements “Z7” or“Z8”.

r6: (“Z3” and “Z4”) or (“Z7” and “Z8”) Fulfillment of any one of the requirements in brackets. The first one meanssimultaneous fulfillment of the requirements “Z3” and “Z4”, whereas the secondone means simultaneous fulfillment of requirements “Z7” and “Z8”.

r15: (“Z5” or “Z6”) or (“Z9” or “Z10”)and not (“Z1” and “Z2”)

Fulfillment of any one of the first two requirements in brackets and at the sametime non fulfillment of the requirement defined in the third combination. The firstand third of the combinations means fulfillment of any one of the requirements“Z5” or “Z6”, and “Z9” or “Z10”, respectively, whereas the third one meanssimultaneous fulfillment of the requirements “Z1” and “Z2”.

692

3. Results and discussions

3.1. Conformational flexibility and electronic structure

Recently, the 20 kcal/mol threshold for∆∆Hfo was

assumed to result in an energetically reasonable set ofconformations, given the extent to which energy providedduring ligand binding could facilitate conformationaltransformations [22, 23]. In this respect, the range of∆∆Hf

o values for the conformers of the chemicals understudy was selected to be less than 20 kcal/mol(table I).

For a given compound, conformers within the specifiedrange of∆∆Hf

o, exhibited chemically-significant varia-tion in potentially relevant electronic descriptors, assummarized intable I. For example, the following pa-rameter ranges are produced by the conformers of com-pound7: 0.944 eV for ELUMO (from –0.236–0.708), 0.89eV for EHOMO, 0.702 eV for EHOMO-LUMO, 0.017(a.u.)2/eV for Vol.P and 6.155 D forµ. Significantparameter variations were also observed for the othercompounds. The observation that relatively small energydifferences between conformers can be associated withsignificant variations in electronic structure (i.e., molecu-

lar electronic descriptors) highlights the necessity ofincluding all energetically-reasonable conformers whendefining common reactivity patterns. Within the em-ployed approximation, these conformers were consideredas equally probable because it is difficult to relate their“gas-phase” energies with the conformer preference incomplex biological environments.

3.2. Application of the COREPA algorithm

According to the first step of the COREPA algorithm,two learning subsets were selected out of all 28 ACEinhibitors under investigation (training set A). The firstone includes the most active chemicals having pIC50 ≥8.9, whereas the second consists of the least potent(passive) ligands having pIC50 ≤ 7.42. Further, thedynamic similarity was calculated for each pair of chemi-cals belonging to the subsets of active and non-activeligands, by using molecular descriptors presented in thepreceding section. The normalized (over the pairs ofcompared chemicals) similarity indices, associated withmolecular descriptors are presented intable II.

Table I. ACE-inhibitors, observed binding affinities, number of conformer and parameter ranges for some significant stereoelectronicparameters.

Structure pIC50 Conformers Vol.Polar E (HOMO) E (LUMO) E(HOMO-LUMO)µ [D] ∆Hfo [kcal/mol]

# obs. [(a.u.)2/eV] [eV] [eV] [eV]

1 9.638 32 1.511 to 1.546 –9.713 to –8.564 –0.606 to 0.069 8.514 to 9.364 2.113 to 13.257 –213.709 to –194.3152 9.222 40 1.464 to 1.481 –9.224 to –8.812 –0.055 to 0.43 8.980 to 9.280 0.999 to 6.900 –190.112 to –170.2463 9.000 46 0.899 to 0.912 –10.943 to –9.827 –0.915 to 0.09 9.532 to 10.132 2.030 to 8.560 –162.691 to –144.8684 8.959 41 1.607 to 1.631 –8.853 to –8.474 –0.562 to –0.148 8.283 to 8.359 1.093 to 9.888 –177.810 to –158.0975 8.921 38 1.374 to 1.398 –9.817 to –9.229 0.041 to 0.588 9.506 to 9.907 1.360 to 9.822 –201.710 to –181.7166 8.921 37 1.648 to 1.664 –9.744 to –9.273 –0.166 to 0.513 9.488 to 9.895 1.296 to 10.996 –215.523 to –195.6977 8.854 45 0.799 to 0.816 –10.643 to –9.753 –0.236 to 0.708 10.040 to 10.742 1.395 to 7.550 –359.037 to –339.9138 8.796 33 1.910 to 1.923 –8.837 to –8.147 –0.253 to 0.468 8.566 to 8.633 1.441 to 8.794 –501.299 to –481.4099 8.585 49 1.242 to 1.259 –10.389 to –9.712 –0.278 to 0.869 9.850 to 10.870 1.375 to 8.565 –245.194 to –225.58310 8.553 41 1.589 to 1.610 –9.713 to –9.264 –0.587 to 0.045 8.892 to 9.345 2.394 to 9.494 –167.064 to –151.34111 8.523 44 1.240 to 1.263 –9.426 to –8.643 0.203 to 0.769 8.964 to 9.794 1.141 to 7.011 –184.961 to –164.97112 8.523 27 1.453 to 1.468 –9.743 to –9.293 –0.147 to 0.547 9.390 to 9.908 2.300 to 7.324 –206.092 to –188.25313 8.495 32 1.705 to 1.732 –9.723 to –8.994 –0.547 to 0.344 8.957 to 9.694 1.796 to 10.779 –152.300 to –132.99814 8.432 44 0.985 to 1.014 –9.378 to –8.718 –0.422 to 0.170 8.581 to 9.286 0.531 to 7.336 –92.215 to –73.73615 8.398 46 1.540 to 1.563 –9.295 to –8.762 –0.529 to 0.039 8.668 to 8.837 3.402 to 9.146 –247.923 to –229.78616 8.222 45 1.069 to 1.097 –9.338 to –8.629 –0.639 to 0.103 8.419 to 8.981 0.775 to 7.067 –128.738 to –109.57617 8.155 42 1.278 to 1.294 –9.833 to –9.155 0.069 to 0.512 9.291 to 9.927 2.037 to 8.545 –271.344 to –252.29118 8.046 50 0.774 to 0.801 –9.763 to –8.738 –0.246 to 0.321 8.818 to 9.798 0.881 to 6.955 –164.372 to –148.33019 8.000 44 0.936 to 0.948 –9.393 to –8.735 –1.278 to –0.518 8.026 to 8.216 1.850 to 9.097 –237.338 to –219.39820 7.921 35 1.812 to 1.835 –9.790 to –9.233 –0.516 to 0.125 8.737 to 9.753 1.415 to 7.814 –306.403 to –287.90321 7.700 42 1.328 to 1.352 –9.671 to –9.028 –0.684 to –0.032 8.856 to 9.263 2.230 to 8.691 –140.147 to –120.23422 7.638 47 0.778 to 0.799 –9.475 to –8.794 0.049 to 0.831 9.138 to 9.805 1.286 to 7.696 –135.315 to –118.16923 7.420 49 0.881 to 0.906 –9.629 to –9.054 –0.318 to 0.146 8.940 to 9.500 0.558 to 6.044 –125.260 to –106.60424 7.398 47 0.891 to 0.920 –9.627 to –8.713 –0.001 to 0.775 9.129 to 9.787 1.995 to 7.923 –139.030 to –119.15025 7.301 47 0.824 to 0.847 –9.538 to –8.712 –0.010 to 0.844 9.351 to 9.851 0.852 to 7.961 –113.348 to –95.44926 7.000 48 0.939 to 0.959 –9.526 to –8.771 –0.234 to 0.449 8.863 to 9.745 0.268 to 7.502 –112.093 to –97.39527 7.155 36 0.773 to 0.791 –9.792 to –9.270 –0.727 to –0.228 8.764 to 9.264 2.582 to 6.031 –59.732 to –47.80528 6.000 47 0.883 to 0.909 –9.755 to –9.022 –0.394 to 0.195 8.823 to 9.629 1.537 to 6.628 –113.843 to –93.921

693

The data summarised intable II suggest that the mostactive and inactive ACE-ligands are highly similaraccording to charges (qi), steric distances (dij), donordelocalizabilities (Si

E, SiN) frontier charges on

HOMO (fiHOMO) and LUMO (fi

LUMO) and atom polari-zabilities (πi). In this respect, the next search of thecommon reactivity pattern was based on conformerdistributions of chemicals across each of those localmolecular descriptors. As such, we have also studied theinteratomic distances dij between non-hydrogen atoms:N-X, O-X and X-X, where X stands for any non-hydrogen atom in the molecules.

To establish common reactivity patterns (step 3),conformer frequency distributions of compounds fromeach of the two training sets were subsequently examinedacross all local stereoelectronic descriptors, with resultsbased on qi and dij , illustrated infigure 3 and figure 4,respectively. To study the effect of parameter distributionpartitioning (i.e., size of parametric windows) on theobtained reactivity pattern, the number of parameterpartitions was set at 20 and 40 for each descriptor.

The intersections between conformer distributionswere identified separately for chemicals belonging toactive and non-active subsets. These intersections aredescribed in terms of parameter ranges occupied by atleast one conformer from each chemical belonging to the

learning subsets. The collection of those ranges repre-sents the common reactivity pattern necessary for elicit-ing similar (high or low) biological effect (ACE-inhibition). Stereoelectronic parameters producing themost distinct reactivity patterns for active and passivechemicals and the respective ranges are presented intable III.

Even though conformer distributions from the mostactive and least active sets of compounds overlap, thesubset of common partitions that contain conformersfrom each of the active compounds does not overlap withpartitions containing conformers from each of the leastactive compounds. Thus, the common atomic chargepattern associated with the most active ligands(figure 3a)deviates significantly from that of the least active chemi-cals(figure 3b). For the most active compounds, commonpartitions corresponding to oxygen and nitrogen atomshaving charges from –0.25 to –0.39 [a.u.](figure 3a)anddonor delocalizabilities from 0.22–0.30 [(a.u.)2/eV] wereobserved. In difference, the charge and donor delocaliz-ability patterns corresponding to O and N atoms fornon-active inhibitors were shifted to more positive chargevalues (–0.26 to –0.36 [a.u.];figure 3b) and lower delo-calizabilities. Significant differences have also beenfound in charge and donor delocalizability patterns, dueto the presence of S-atoms in non-active chemicals

Table II. Normalized similarity measurements between conformer distributions of most active and least active ACE-inhibitors.

Descriptors Most active ACE-inhibitors Least active ACE-inhibitors

TAN DIST IF 3D-TAN COSINE TAN DIST IF 3D-TAN COSINE

ACCEPT DLC 0.285 0.472 3.297 0.878 0.935 0.269 0.473 3.144 0.535 0.716ACCEPT MLK 0.272 0.473 3.282 0.865 0.928 0.279 0.472 3.155 0.408 0.649BOND ORDER 0.291 0.522 3.319 0.817 0.931 0.289 0.517 3.179 0.747 0.856DONOR DLC 0.277 0.472 3.289 0.655 0.792 0.277 0.472 3.289 0.655 0.792DONOR MLK 0.279 0.472 3.292 0.884 0.958 0.279 0.472 3.292 0.884 0.958POLAR 0.289 0.472 3.302 0.333 0.756 0.289 0.472 3.302 0.333 0.756POLAR MLK 0.292 0.472 3.305 0.600 0.866 0.292 0.472 3.305 0.600 0.866POP–HOMO 0.283 0.472 3.293 0.700 0.841 0.283 0.472 3.293 0.700 0.841POP–HOMO MLK 0.278 0.529 3.292 0.924 0.961 0.278 0.529 3.292 0.924 0.961POP–LUMO 0.284 0.514 3.296 0.579 0.937 0.284 0.514 3.296 0.579 0.937POP–LUMO MLK 0.283 0.492 3.297 0.652 0.894 0.283 0.492 3.297 0.652 0.894Q 0.286 0.478 3.296 0.750 0.949 0.286 0.478 3.296 0.750 0.949Q MLK 0.287 0.487 3.297 0.207 0.455 0.287 0.487 3.297 0.207 0.455SPECIAL DISTANCE 0.261 0.475 4.471 0.125 0.223 0.261 0.475 4.471 0.125 0.223E(HOMO) 0.061 0.969 0.505 0.154 0.272 0.061 0.969 0.505 0.154 0.272ELECTRONEGATIVITY 0.149 0.958 1.225 0.000 0.000 0.149 0.958 1.225 0.000 0.000∆E(HOMO–LUMO) 0.096 0.663 0.673 0.000 0.000 0.096 0.663 0.673 0.000 0.000GEOM. WIENER 0.012 0.750 0.244 0.000 0.000 0.012 0.750 0.244 0.000 0.000VOLUME POLARIZAB. 0.000 0.510 0.000 0.000 0.000 0.000 0.510 0.000 0.000 0.000CALC. HEAT FORM. 0.051 0.719 0.471 0.070 0.134 0.051 0.719 0.471 0.070 0.134DIPOLE MOMENT 0.257 0.472 1.665 0.056 0.113 0.257 0.472 1.665 0.056 0.113E(LUMO) 0.104 1.044 0.802 0.077 0.147 0.104 1.044 0.802 0.077 0.147

694

(figure 3b). The maximum common distance range be-tween oxygen atoms for active ligands was found to befrom 8.7–9.4 Å (light coloured bars infigure 4a), whereasbetween O and N atoms was from 6.9–7.4 Å (notillustrated). As shown infigure 4b, no such commonranges have been established between the electronegativesites in non-active chemicals. As can be seen intable III,less significant differences in reactivity patterns of activeand non-active chemicals (i.e., small differences in para-metric ranges for active and non-active ACE inhibitors)were established for donor delocalizabilities of carbonylcarbons and the remaining C-skeleton.

Figure 3. Common reactivity patterns for activea. and non-active b. ACE-inhibitors across the charges of hetero-atoms(oxygen, nitrogen and sulfur). Grey bars correspond to para-meter ranges populated by at least conformers of all chemicals(activities, in figure 3a, and nonactives infigure 3b), whereasthe black bars correspond to parameter ranges populated byconformers of some of the chemicals from active and non-active subsets.

Figure 4. Common reactivity patterns for activea. and nonac-tive b. ACE-inhibitors across the distance between oxygenatoms. Grey bars correspond to parameter ranges populated byat least conformers of all chemicals (actives infigure 4aandnon-actives infigure 4b), the black bars correspond to parame-ter ranges populated by conformers of some of the chemicalsfrom active and non-active subsets. The light grey bar infigure 4aare associated with the largest common distance rangefor active chemicals in the training set.

695

3.3. Validation of reactivity pattern

The complete lists of studied structural requirementsand the Boolean expressions based on the above definedcommon parameter ranges are presented intable IV.More or less conservative screens were imposed by usingflexibility of the Boolean expressions (i.e., the logicalcombination of the different structural rules by makinguse of the logical operators “and”, “or” and “not”).

The validation of the method was based on the appli-cations of stereoelectronic screens on chemicals of theinternal training set (A) (subset of which was used forderiving reactivity patterns) as well as to an externaldataset (B) with well known ACE-inhibitors, used aspharmaceutical agents. The results are presented inta-ble V.

The quality of screens is assessed by their ability to: (i)identify conformers of the chemicals from the learning

Table III. Common reactivity patterns for most active and least active ACE-inhibitors from training set A.

Most active ACE–inhibitors Least active ACE-inhibitorspIC50 ≥ 8.9 pIC50 ≤ 7.42

Descriptor interatomicdistances, [Å]

Atoms/fragm.

Descriptor ranges Descriptor ranges

Partitioning = 40 Partitioning = 20 Partitioning = 40 Partitioning = 20

O-O 8.94 to 9.17 8.74 to 9.39 5.75 to 5.88 5.63 to 5.88O-N 6.96 to 7.19 6.95 to 7.42 3.71 to 3.80 3.71 to 3.89

Charge [a.u.] X –0.39 to –0.26 (O,N); –0.39 to –0.25 (O,N); –0.36 to –0.26 (O,N); –0.36 to –0.26 (O,N);–0.013 to 0.010 (S); –0.024 to 0.021 (S);

Donor All 0.111 to 0.126 (C=(O)); 0.111 to 0.126 (C=(O)); 0.113 to 0.121 (C=(O)); 0.113 to 0.129 (C=(O));Delocalizability 0.149 to 0.212 (C); 0.142 to 0.22 (C); 0.146 to 0.163 (C(N)); 0.146 to 0.163 (C-(N));[(a.u.)2/eV] 0.227 to 0.290 (O,N) 0.22 to 0.30 (O,N) 0.171 to 0.188 (C(C)); 0.163 to 0.196 (C-(C));

0.230 to 0.255 (N,OH); 0.230 to 0.247 (N,OH);0.263 to 0.280 (O=(C)); 0.247 to 0.280 (O=(C));0.381 to 0.410 (S) 0.381 to 0.410 (S)

X - for non hydrogens and non-carbon atoms;ALL - the partitions of the local parameter were analysed for all atoms in the molecule.

Table IV. The structural requirements and their Boolean combinations used for selecting highly active ACE inhibitors from test series underinvestigation.

Expressions of the defined structural requirement used for selecting high ACEinhibition activity from test series under investigation.

Boolean expression of the structural requirements

RX:O, S, N, P r1: “Z3” and “Z4”Z1:O_O{5.6< DISTANCE < 5.9} r2: “Z7” and “Z8”Z2:O_N{3.7 < DISTANCE < 3.9} r3: “Z3” or “Z4”Z3:O{–0.39< Q < –0.25}_O{–0.39< Q < –0.25}{8.7 < DISTANCE < 9.4} r4: “Z3” or “Z7”Z4:O{–0.39< Q < –0.25}_N{–0.39< Q < –0.25}{6.9 < DISTANCE < 7.4} r5: (“Z3” or “Z4”) and (“Z7” or “Z8”)Z5:RX{–0.39 < Q < –0.25}_RX{–0.39< Q < –0.25}{8.7 < DISTANCE < 9.4} r6: (“Z3” and “Z4”) or (“Z7” and “Z8”)Z6:RX{–0.39 < Q < –0.25}_RX{–0.39< Q < –0.25}{6.9 < DISTANCE < 7.4} r7: (“Z3” or “Z4”) or (“Z7” or “Z8”)Z7:O{0.22 < DONOR_DLC< 0.30}_O{0.22< DONOR_DLC< 0.30}{8.7 < r8: (“Z5” or “Z9”)

DISTANCE<9.4} r9: (“Z5” and “Z9”)Z8:O{0.22 < DONOR_DLC< 0.30}_N{0.22 < DONOR_DLC< 0.30}{6.9 < r10: (“Z5” or “Z6” )

DISTANCE < 7.4} r11: (“Z5” and “Z6” )Z9:RX{0.22 < DONOR_DLC< 0.30}_RX{0.22 < DONOR_DLC< 0.30}{8.7 < r12: (“Z5” or “Z6”) or (“Z9” or “Z10”)

DISTANCE < 9.4} r13: (“Z5” and “Z6”) or (“Z9” and “Z10”)Z10:RX{0.22 < DONOR_DLC< 0.30}_RX{0.22 < DONOR_DLC< 0.30}{6.9 < r14: (“Z5” or “Z6”) and (“Z9” or “Z10”)

DISTANCE < 7.4} r15: (“Z5” or “Z6”) or (“Z9” or “Z10”) and not (“Z1” and “Z2”)r16: (“Z5” or “Z6”) or (“Z9” or “Z10”) and not (“Z1” or “Z2”)

696

subset; (ii) “catch” some chemicals with moderate activ-ity (not used for deriving the pattern); (iii) eliminate allconformers of the chemicals with low activity; and (iv)identify as active, chemicals from external databases,experimentally shown to elicit high activity (i.e., phar-maceutical agents). Beside selection of conformers ofmost active chemicals from training set A (having pIC50

≥ 8.9) used for deriving the common reactivity pattern,almost all of the structural rules screened also conformersof ACE ligands with lower activity, i.e., having pIC50 <8.9. Still, these are active and medium active inhibitorshaving 7.7≤ pIC50 ≥ 8.9 (up to chemical21, table I).Only few of the structural rules, classified as active, theinhibitors having pIC50 < 7.7. These are the “leastrestrictive” rules #10, 12, 14–16 with high “concentra-tion” of “or” logical operator, i.e., providing very lowscreen conservativeness. These rules, screened as active,chemicals having pIC50 ≥ 6.00. It is very difficult todefine precisely the pIC50 threshold according to which achemical could be classified as active. Usually, thesethresholds are user defined. All chemicals included intraining set A are ACE inhibitors and the differences intheir activity are only quantitative, not qualitative. In thisrespect, the obtained results are reasonable: the more

restrictive screens, selected as active, the inhibitors withhigher pIC50 values, whereas, the less restrictive screensselect as active all inhibitors under investigation.

The reactivity pattern derived by the most active ACEinhibitors from training set A was used for screening ofthe chemicals from the test set B. As it was alreadymentioned, these captopril derivatives are well knownACE inhibitors with different activity, most of them usedas pharmaceutical agents. It was hypothesized that thederived reactivity pattern should, screen as active, mostof the chemicals from this test set. As can be seen fromtable V, Rules # 3, 4, 7, 8, 11 and 13 selected 70% of theinhibitors from test set B, whereas Rules # 10, 12 and 16selected 85% of those chemicals. The fact that the lastfour rules were capable of screening a large percentage ofthe known ACE-inhibitors from the external dataset is anindication that they could be used for screening ofchemicals from databases as potential lead ACE inhibi-tors. A practical advise is to perform screenings with eachof the four individual rule files and then to look for theintersection of predicted subsets of chemicals. The latterare assumed to be the most potent ligands because theymeet simultaneously stereoelectronic requirements in-cluded in all rule files.

Table V. Screened ACE-inhibitors from test series according to the employed structural requirements, as described by the Booleanexpressions listed intable IV.

Rules Training set A Test set B

#a Numberof selectedchemicals

Numberof selectedconformers

Numberingof selectedchemicalsb

Numberof selectedchemicals

Numberof selectedconformers

Numberingof selectedchemicalsc

r1 10 47 1, 2, 4, 5, 6, 8, 10, 13, 20, 21 6 42 5–7, 9–11r2 12 75 1, 2, 4–8, 10, 12, 19–21 1 9 12r3 14 222 1–6, 8–10, 12, 13, 19–21 9 350 4–7, 9–13r4 16 186 1–10, 12, 13, 17, 19–21 9 248 4–7, 9–13r5 14 216 1–6, 8–10, 12, 13, 19–21 2 60 12, 13r6 12 65 1, 2, 4, 5, 7, 8, 10, 13, 19–21 7 51 5–7, 9–12r7 16 250 1–10, 12, 13, 17, 19–21 9 374 4–7, 9–13r8 16 222 1–11, 13, 17, 19–21 9 317 4–7, 9–13r9 14 188 1–6, 8–11, 13, 19–21 2 35 12, 13r10 19 428 1–6, 8–13, 16, 19–21, 23, 27, 2811 596 3–13r11 14 127 1–6, 8–11, 13, 19–21 9 134 4–7, 9–13r12 22 495 1–10, 12, 13, 15–17, 19–21, 23,

27, 2811 624 3–13

r13 15 156 1–11, 13, 19–21 9 162 4–7, 9–13r14 19 416 1–6, 8–11, 13, 16, 19–21, 23,

27, 282 79 12, 13

r15 22 484 1–13, 15–17, 19–21, 23, 27, 2810 619 3–12r16 22 350 1–13, 15–17, 19–21, 23, 27, 2811 265 3–13

bThe numbering of chemicals corresponds to those infigure 1cThe numbering of chemicals corresponds to those infigure 2

697

4. Summary and conclusions

The present search for the structural requirements forhigh ACE-inhibition is based on the COREPA algorithm.This last one is a generalization of the active analogueapproach circumventing the need to align conformers ofactive molecules, while it explicitly addresses variation inconformational flexibility in the context of varying bio-logical activity. The distributions of all energeticallyreasonable conformers of chemicals from the trainingsubsets across specific molecular descriptors found to berelevant to activity under study are analysed. Thus,instead of template alignment based on single conformerrepresentations of chemicals, their conformer distribu-tions are naturally aligned (ordered) across specific mo-lecular descriptors. The distribution intersections popu-lated by conformers of each of the biologically similarchemicals (actives and non-actives) form the commonreactivity pattern. Due to the population character ofconformer distributions in the COREPA approach it isdifficult to recognize the active conformers of chemicalsin classical chemical terms as it is done in other “dy-namic” techniques [20–22, 26, 38, 39].

The common reactivity pattern, in terms of charge anddistance ranges between electronegative atoms requiredfor high ACE-inhibition, was derived. It was based on thesubset of the most active chemicals from the training setA. Then, it was validated by screening all chemicals fromthe same training set, as well as an external set of activeanalogues known as pharmaceutical agents (training setB). The various logical combinations of the stereoelec-tronic rules provided the flexibility of the screeningprocess. The most restrictive screens selected as activethe inhibitors with pIC50 values higher or equal to theactivity threshold used for deriving reactivity patterns.The less restrictive screens in addition, selected as active,inhibitors with moderate activity. Six of the employedscreening rules selected as active 70% of the activeinhibitors from test set B, whereas four of the rulesselected 85% of those chemicals. The large percentage ofthe screened ACE-inhibitors from the test set by thesefour stereoelectronic screens is indicative that the asso-ciated reactivity pattern could be used for screening oflarge databases of 3-D structures for the search ofpotentially active ACE-inhibitors.

Individual screenings were performed by employingeach of the stereoelectronic screens on conformers of thechemicals from the test set. Chemicals meeting multiplestereoelectronic requirements, i.e., lying on the intersec-tion of the subsets predicted by single screens, areconsidered as the most likely candidates for active

ligands. Such studies are presently performed by ourlaboratory for the needs of the Chemicals and Pharma-ceutical Research Institute, Sofia, aiming to design origi-nal pharmaceutical agents. In large databases, however,chemicals are represented by single conformers becausethe conformational search for each compound is compu-tationally impractical. In these situations, a less restrictivescreening strategy is employed initially that assesses asingle conformer per chemical and which is designed tominimize the percentage of false negatives (i.e., com-pounds incorrectly predicted to be non-active). The“Tweak” technique is also used as a pre-screen alterna-tive, in which the rotatable bonds of the structures areadjusted to produce a conformation which matches asclosely as possible a given 3D requirement. Then, in asecond stage, more refined screens should be employedon a smaller set of conformationally multiplied chemi-cals, already passed the pre-screen.

Acknowledgements

This research was financially supported by the Chemi-cals and Pharmaceutical Research Institute, Sofia (grant #RD-09-173/20.12.1995). The authors also thanks DrKarabunarliev for developing the Chemical Rule Inter-preter software as well as Drs Matey Vitev, Neno Dimovand Kiril Ninov for valuable discussions.

References

[1] Bunning P., Kleemann S.G., Riordan J.F., Biochemistry 29 (1990)10488–10492.

[2] Chen Y.N.P., Riodan J.F., Biochemistry 29 (1990) 10493–10498.

[3] Soubrier F., Alhenc-Gelas F., Hubert C., Allegrini J., John M.,Tregear G., Corvol P., Proc. Natl. Acad. Sci. USA 85 (1988)9386–9390.

[4] Petrillo W.W., Trippodo N.C., De Forrest J.M., in: Robertson D.W.(Ed.), Annual Reports in Medicinal Chemistry Vol 25, AcademicPress, New York, 1989, pp. 51–60.

[5] Hangauer D.G., in: Perun T.J., Propst C.L. (Eds.), Computer-AidedDrug Design: Methods and Applications, Marcel Dekker, New York,1989, pp. 253–295.

[6] Wyvratt M.J., Patchett A.A., Med. Res. Rev. 5 (1985) 483–531.

[7] Depriest S.A., Shands R.F.B., Dammkeohler R.A., Marshall G.R.,in: Silipo C., Vittoria A. (Eds.), QSAR Rational Approaches to theDesign of Bioactive Compounds Vol 16, Elsevier Science, Amster-dam, 1991, pp. 405–414.

[8] Mayer D., Naylor C.B., Motoc I., Marshall J.R., J. Comput. -AidedMol. Des 1 (1987) 3–16.

[9] Andrews P.R., Carson J.M., Caselli A., Spark M.J., Woods R., J.Med. Chem. 28 (1985) 393–399.

[10] Dammkoehler R.A., Karasek S.F., Shands E.B.F., Marshall G.R., J.Comput. -Aided Mol. Des. 3 (1989) 3.

[11] Dammkoehler R.A., Karasek S.F., Shands E.B.F., Marshall G.R., J.Comput. -Aided Mol. Des. 9 (1995) 491–499.

698

[12] Depriest S.A., Mayer D., Naylor C.B., Marshall G.R., J. Am. Chem.Soc. 13 (1993) 115 5372–5384.

[13] Waller C.L., Shands E.B.F., Dammkoehler R.A., Marshall G.R., J.Comput. -Aided Mol. Des. 11 (1997) 121.

[14] Kearsely S.K., Smith G.M., Tetrahedron Comput. Methodol. 3(1990) 615–633.

[15] Blaney J.M., Dixon J.S., Perspect. Drug Discov. Des. 1 (1993)301–319.

[16] Perkins T.D., Deam P.M., J. Comput. -Aided Mol. Des. 7 (1993)155–172.

[17] Martain Y.C., Bures M.G., Danahar E.A., Delazzar J., Lico I., PavlikP.A., J. Comput. -Aided Mol. Des. 7 (1993) 83–102.

[18] Topliss J.G., Edwards R.P., J. Med. Chem. 22 (1979) 1238–1244.

[19] Eliel E.L., in: Warr W.A. (Eds.), Chemistry in Three Dimensions inChemical Structures, Vol. 1, Springer, Berlin, 1993, pp. 1–8.

[20] Mekenyan O.G., Ivanov J.M., Veith G.D., Bradbury S.P., Quant.Struct. -Act. Relat. 13 (1994) 302–307.

[21] Mekenyan O.G., Veith G.D., Call D.J., Ankley G.T., Environ. HealthPerspect. 104 (1996) 1302–1309.

[22] Bradbury S.P., Mekenyan O.G., Ankley G.T., Environ. Toxicol.Chem. 15 (1996) 1945–1954.

[23] Wiese T., Brooks S.C., J. Steroid. Biochem. Molec. Biol. 50 (1994)61–72.

[24] Prendergast K., Adams K., Greenlee W.J., Nachbar R.B., PatchettA.A., Underwood D.J., J. Comput. -Aided Mol. Des. 8 (1994)491–512.

[25] Mekenyan O.G., Ivanov J.M., Karabunarliev S.H., Bradbury S.P.,Ankley G.T., Karcher W., Environ. Sci. Technol. 31 (1997)3702–3711.

[26] Bradbury S.H., Mekenyan O.G., Ankley G.T., Environ. Toxicol.Chem. 17 (1998) 15–25.

[27] Mekenyan O.G., Ivanov J.M., Karabunarliev S.H., Hansen B.,Ankley G.T., Bradbury S.P., in: Chen F., Schüürmann G. (Eds.),Quantitative Structure Activity Relationships in Environmental Sci-ences VII, SETAC Press, Pensacola FL, USA, 1998, pp. 39.

[28] Cambell W.B., Brooks S.N., Pettinger W.A., Science 184 (1974)994–996.

[29] Ondetti M.A., Rubin B., Cushman D.W., Science 196 (1977)441–444.

[30] Natoff I.L., Redshaw S., Drugs Future 12 (1987) 475–483.

[31] Ivanov J.M., Karabunarliev S.H., Mekenyan O.G., J. Chem. Inf.Comput. Sci. 34 (1994) 234–243.

[32] Stewart J.J.P., MOPAC: A general molecular orbital packages;Version 7. 0. Software. Quantum Chemistry Program Exchange no.455, University of Indiana, Bloomington, IN, USA, (1995).

[33] Mekenyan O.G., Karabunarliev S.H., Ivanov J.M., Dimitrov D.N.,Comput. Chem. 18 (1994) 173–187.

[34] Mekenyan O.G., Peitchev D., Bonchev D., Trinajstic N., Bangov I.,Arzneim. Forsch. 36 (1986) 176.

[35] Schüürmann G., Quant. Struct. -Act. Relat. 59 (1990) 326–333.

[36] Lewis D.F.V., J. Comput. Chem. 10 (1989) 145–151.

[37] Karabunarliev S.H., Nikolov N., Mekenyan O.G., submitted.

[38] Mekenyan O.G., Schultz T.W., Veith G.D., Kamenska V.B., J. Appl.Toxicol. 16 (1996) 355.

[39] Veith G.D., Mekenyan O.G., Ankley G.T., Call D.J., Environ. Sci.Technol. 29 (1995) 126.

699

Original article

Synthesis of some new 1,4-benzothiazine and 1,5-benzothiazepine tricyclicderivatives with structural analogy with TIBO

and their screening for anti-HIV activity #

Giuliano Grandolini*, Luana Perioli, Valeria Ambrogi

Istituto di Chimica e Tecnologia del Farmaco, Università di Perugia, Via del Liceo 1, 06123 Perugia, Italy

(Received 10 November 1998; accepted 27 January 1999)

Abstract – Several new tricyclic derivatives with structural analogy to TIBO were prepared starting from properly substituted 1,4-benzothiazines and 1,5-benzothiazepine. All synthesized compounds were submitted to screenings for in vitro anti-HIV-1 activity. Only twocompounds showed moderate activity. © 1999 Éditions scientifiques et médicales Elsevier SAS

1,4-benzothiazine and 1,5-benzothiazepine derivatives / anti-HIV-1 activity / non-nucleoside reverse transcriptase inhibitor (NNRTI)

1. Introduction

Thedramatic increaseofspreadof theacquired immuno-deficiency syndrome (AIDS) has stimulated considerableefforts in the research in this field. As a result of thedifficulties encountered in the development of an effec-tive vaccine, research is aimed at the discovery of newchemotherapeutic agents.

Good results were obtained with the class of non-nucleoside reverse transcriptase inhibitors (NNRTIs) forboth their antiviral activity and their low toxicity.

These considerations prompted us to prepare some1,4-benzothiazine and 1,5-benzothiazepine derivativeswith structural analogy with TIBO [1–5], a non-nucleoside inhibitor of HIV-1 reverse transcriptase(fig-ure 1). In this paper we describe the synthesis and thepreliminary anti-HIV screening of some tricyclic deriva-tives (figure 1).

2. Chemistry

Starting materials for our work program werecompounds2 (a–f) in which the amino group is properlylocated to react with bifunctional reagents to give thedesired tricyclic compounds. The 5- or 6-aminoderivatives 2 (table I) were obtained by reductionof 7-methoxy-5-nitro-3,4-dihydro-2H-1,4-benzothiazin-3-ones or 8-methoxy-6-nitro-2,3,4,5-tetrahydro-1,5-benzothiazepin-4-one1 already synthesized by us [6](figure 2).

# A preliminary account of this work was presented at the 4thInternational Conference on Heteroatom Chemistry, Seoul,Korea, July 30–August 4, 1995*Correspondence and reprints

Figure 1. General structure of new tricyclic 1,4-benzothiazineand 1,5-benzothiazepine derivatives with structural analogywith TIBO.

Eur. J. Med. Chem. 34 (1999) 701−709 701© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Unfortunately, because of the problems encounteredduring the synthesis of1 [3], to date only compoundswith a methoxy substituent at the 7 or 8 position could beprepared.

Our first attempt to synthesize the imidazo derivatives4 (a–f) was performed by heating2 (a–f) with an excessof formic acid under reflux [7]. In this case theN-formylderivatives 3 (a–c, e and f) were obtained(table II). Tentative cyclizations were performed by treat-ing 3 with concentrated HCl but only for3a (R = H,n = 0) the desired tricyclic4a was obtained.

Treatment of3b (R = CH3, n = 0) with concentratedHCl gave rise to the replacement of the N-formylgroup with a chlorine atom with the formation ofthe 5-chloro-7-methoxy-2-methyl-3,4-dihydro-2H-1,4-benzothiazin-3-one5.

Another tentative synthesis of4 (a–f) was achieved byrefluxing compound2 (a–f) with triethyl orthoformate in

xylene (mixture of isomers) [7], in this way only com-pounds4e and4f were obtained(table III).

Condensation of compounds2 with triethyl ortho-acetate in xylene was successful only for compounds6a,b and e (table III). 5-Amino-2-ethyl-7-methoxy-3,4-dihydro-2H-1,4-benzothiazin-3-one2c and 5-amino-2-butyl-7-methoxy-3,4-dihydro-2H-1,4-benzothiazin-3-one2d afforded only a variety of decomposition pro-ducts which have not yet been identified. When the samereaction was performed with the benzothiazepine2f,the 2-ethoxy-9-methoxy-2-methyl-1H-5,6-dihydro-imi-dazo[3,4,5-e,f]-1,5-benzothiazepin-4-one7 (figure 3)wasobtained. Operating in the presence of pyridinium hydro-chloride, compound8 was formed(figure 3).

Compounds9a–c and e (table IV, figure 2) were ob-tained by refluxing the corresponding amines2a–candewith carbon disulfide in anhydrous pyridine. Reaction ofcarbon disulfide with2d and 2f either in anhydrous

Table I. Physical and chemical data of compounds2a–f.

Compound R n Yield M.p. Formula 1H-NMR, δ(%) (°C) (MW)

2a H 0 78 258–260 C9H10N2O2S(210.25)

3.35 (s, 2H, SCH2), 3.62 (s, 3H, OCH3), 5.25 (s, 2H,NH2), 6.10–6.20 (m, 2H, Ar), 9.50 (s, 1H, NH) (DMSO-d6)

2b CH3 0 89 178–179 C10H12N2O2S(224.28)

1.30 (d, J = 9.5 Hz, 3H, SCHCH3), 3.50 (q, 1H,SCHCH3), 3.65 (s, 3H, OCH3), 5.25 (s, 2H, NH2),6.10–6.20 (m, 2H, Ar), 9.48 (s, 1H, NH) (DMSO-d6)

2c CH2CH3 0 70 153–154 C11H14N2O2S(238.30)

1.05 (t, 3H, CH2CH3), 1.50–2.10 (m, 2H,CH2CH3),3.25 (q, 1H, SCH), 3.75 (s, 3H, OCH3), 4.05 (brs, 2H,NH2), 6.15 (d,J = 2.4 Hz, 1H, Ar), 6.35 (d,J = 2.4 Hz,1H, Ar), 9.40 (brs, 1H, NH) (CDCl3)

2d (CH2)3CH3 0 51 104–106 C13H18N2O2S(266.36)

0.90 (t, 3H, CH3), 1.22–1.70 (m, 4H, CH2CH2CH2CH3),1.85–2.03 (m, 2H,CH2CH2CH2CH3), 3.35 (q, 1H,SCH), 3.73 (s, 3H, OCH3), 3.88 (br s, 2H, NH2), 6.20 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.35 (d,Jmeta= 2.4 Hz, Ar), 8.75(br s, 1H, NH) (CDCl3)

2e C6H5 0 62 220–221 C15H14N2O2S(286.35)

3.61 (s, 3H, OCH3), 4.82 (s, 1H, SCHC6H5), 5.38 (s, 2H,NH2), 6.10 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.20 (d,Jmeta=2.40 Hz, 1H, Ar), 7.27 (s, 5H, Ar), 9.83 (s, 1H, NH)(DMSO-d6)

2f H 1 51 198–199 C10H12N2O2S(224.28)

2.33, 3.33 (A2B2 system, 4H, CH2CH2), 3.65 (s, 3H,OCH3), 5.12 (s, 2H, NH2), 6.30–6.40 (m, 2H, Ar), 8.67(s, 1H, NH) (DMSO-d6)

702

pyridine or in anhydrous xylene afforded only decompo-sition of the starting products. A further attempt wasrealized by lowering the reaction temperature, thus2dand2f were reacted with carbon disulfide in tetrahydro-furan in the presence of triethylamine. Only compound9f(table IV)was obtained, although in low yields. Unfortu-nately this attempt to synthesize the 2-butylderivativewas unsuccessful.

Diazotization reaction of the primary amino groupgave rise to the triazoloderivatives10a, b and d–f(table V). It is noteworthy that this reaction occurredimmediately for 2-unsubstituted-1,4-benzothiazine andneeded 2–40 h of stirring for 2-substituted benzothiaz-ines.

Annulation of a six-membered ring on the 1,4-benzothiazine or 1,5-benzothiazepine system was verydifficult. Many attempts were made with 1,2-dibromoethane, phenacylbromide, ethylchlorooxalate andethylpyruvate, but only reaction of2 with 1,2-dibromoethane was successful and gave rise to com-pounds11a–f (table VI), operating in the presence oftetrabutylammonium bromide and powdered potassium

hydroxide in tetrahydrofuran according to phase-transfercatalysis conditions (PTC) for11a–d and refluxing inxylene for11eand f.

It is noteworthy that the building of a third nucleus waseasier on the 1,5-benzothiazepine system than on the1,4-benzothiazine one and among the 1,4-benzothiazinesit was easier on the 2-unsubstituted than on the2-substituted derivatives, perhaps because of the hin-drance exerted by the substituent at the 2 position. Theannulation of a six-membered ring was more difficultthan that of a five membered one.

3. Biological investigation and results

All synthesized compounds were submitted to theNational Cancer Institute (NCI) of Bethesda (MD) andwere evaluated for in vitro anti-HIV-1 activity. Allcompounds were inactive with the exception of9c whichshowed moderate activity (CC50 = 0.155 mM, EC50 =0.0323 mM), the ratio between the two concentrations(therapeutic index = CC50/EC50) being 4.80.

Figure 2. Synthetic pathway for new tricyclic derivatives.

Figure 3. Tentative synthesis of 2-methyl-9-methoxy-4H-5,6-dihydroimidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one.

703

4. Experimental protocols

4.1. Chemistry

Melting points were taken on a Kofler hot-stageapparatus and are uncorrected.1H-NMR spectra wererecorded in the solvent indicated, using a Bruker AC-200(200 MHz) instrument. The chemical shift values arereported inδ (ppm) relative to tetramethylsilane as aninternal standard. Mass spectra were recorded on a VarianMAT 311A spectrometer. Elemental microanalyses wereperformed for C, H and N on a Carlo Erba ElementalAnalyser model 1106 and results were within± 0.4% ofthe theoretical values. The purity of the compounds waschecked by TLC (pre-coated silica-gel plates, MerckKieselgel 60 F254). Flash chromatographies were per-formed on columns packed with Merck silica gel,230–400 mesh.

4.1.1. General procedure for 5-amino-7-methoxy-3,4-dihydro-2H-1,4-benzothiazin-3-ones2a–e and 6-amino-8-methoxy-2,3,4,5-tetrahydro-1,5-benzothiazepin-4-ones2f (table I)

Nitrobenzothiazinederivative1a–e (0.03 mol) wasadded portionwise under stirring to a solution of stannouschloride dihydrate (6.77 g, 0.03 mol) in concentrated HCl(10 mL). For thiazepine compound1f (0.03 mol) a sus-pension of stannous chloride dihydrate (9.02 g, 0.04 mol)in tetrahydrofuran (30 mL) and concentrated HCl (2 mL)was used. The mixture was heated in a steam bath for10 min for compound1d, 30 min for1a, 1 h for1b, eandf, and 1.5 h for1c.

After cooling, in the case of1a, an abundant precipitatewas formed. It was collected by filtration, alkalinizedwith 5 N NaOH solution, filtered again, washed withwater and finally recrystallized from EtOH to give2a.

Table II. Physical and chemical data of compounds3a–c, eand f.

Compound R n Yield M.p. Formula 1H-NMR, δ(%) (°C) (MW)

3a H 0 50 208–210 C10H10N2O3S(238.26)

3.45 (s, 2H, SCH2), 3.75 (s, 3H, OCH3), 6.83 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.15 (d,Jmeta = 2.4 Hz, 1H, Ar),8.25 (s, 1H, NHCHO), 9.60, 9.77 (2s, 2H, 2NH)(DMSO-d6)

3b CH3 0 77 203–204 C11H12N2O3S(252.29)

1.29 (d, J = 7.8 Hz, 3H, CHCH3), 3.60 (q, 1H,CHCH3), 3.70 (s, 3H, OCH3), 6.80 (d,J = 2.4 Hz, 1H,Ar), 7.15 (d, J = 2.4 Hz, 1H, Ar), 8.25 (s, 1H,NHCHO), 9.61, 9.80 (2s, 2H, 2NH) (DMSO-d6)

3c CH2CH3 0 23 58–60 C12H14N2O3S(266.31)

0.75 (t, 3H, CH2CH3), 1.10–1.70 (m, 2H,CH2CH3),3.10–3.20 (m, 1H, SCH), 3.50 (s, 3H, OCH3), 6.70 (d,Jmeta = 2.4 Hz, 1H, Ar), 6.90 (d,Jmeta = 2.4 Hz, 1H,Ar), 8.00 (s, 1H, NHCHO), 8.10, 8.15 (2s, 2H, 2NH)(DMSO-d6)

3e C6H5 0 27 117–119 C16H14N2O3S(314.36)

3.72 (s, 3H, OCH3), 6.00 (s, 1H, SCHC6H5), 6.70 (brd, 1H, Ar), 6.88 (br d, 1H, Ar), 7.25–7.50 (m, 7H, C6H5+ NHCHO), 8.15 (s, 1H, NH) (DMSO-d6)

3f H 1 33 186–188 C11H12N2O3S(252.29)

2.55, 3.40 (A2B2 system, 4H, CH2CH2), 3.85 (s, 3H,OCH3), 7.00 (d,J = 2.4 Hz, 1H, Ar), 7.77 (d,J = 2.4Hz, 1H, Ar), 8.20 (s, 1H, NHCHO), 8.48, 8.56 (2s, 2H,2NH) (CDCl3)

704

In the case of1f, the mixture was evaporated in vacuo,the residue was dissolved in CHCl3 and extracted with 2N HCl. The aqueous solution was alkalinized with 50%NaOH solution, extracted with CHCl3, dried over anhy-drous Na2SO4, filtered and evaporated in vacuo. The puresolid residue was recrystallized from EtAc to give2f.

For other compounds, the reaction mixture was alka-linized with 30% ammonia solution. Compounds2c and2d crystallized and were collected by filtration, washedwith water and recrystallized from EtOH. In the othercases the alkalinized mixture was extracted with CHCl3,washed with water, dried over Na2SO4 and dried invacuo. The solid residue was recrystallized from EtOH.

4.1.2. General procedure for 5-formylamino-7-methoxy-1,4-benzothiazine derivatives3a–c and e and 6-formylamino-8-methoxy-2,3,4,5-tetrahydro-1,5-benzothia-zepin-4-one3f (table II)

A solution of 2 (0.01 mol) in 30 mL formic acid(excess) was refluxed under stirring for 40 min for2f, 2 hfor 2a–cand 9 h for2e. After cooling, the solution waspoured into ice-water and alkalinized with 30% ammonia

solution and the resulting precipitate was collected andrecrystallized from EtOH with the exception of3b whereMeOH was used.

4.1.3. 8-Methoxy-2H-4,5-dihydro-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-one4a (table III)

N-(formyl)derivative 3a (2.38 g, 0.01 mol) was re-fluxed in concentrated HCl (10 mL) for 1 h. After cool-ing, the solution was poured into ice-water and alkalin-ized with 30% ammonia solution. The resultingprecipitate was collected by filtration and recrystallizedfrom EtOH.

4.1.4. Tentative synthesis of 5-methyl-8-methoxy-4,5-dihydro-2H-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-one.Formation of 5-chloro-7-methoxy-2-methyl-3,4-dihydro-2H-1,4-benzothiazin-3-one5.

When the reaction described above (4.1.3) was per-formed using3b, the 5-chloro-7-methoxy-2-methyl-3,4-dihydro-2H-1,4-benzothiazin-3-one5 was obtained (38%yield), m.p. 143–144 °C, recrystallized from EtOH[(C10H10ClNO2S)]; 1H-NMR, δ (DMSO-d6): 1.5 (d,J =

Table III. Physical and chemical data of compounds4a, eand f and6a, b ande.

Compound R R1 n Yield M.p. Formula 1H-NMR, δ(%) (°C) (MW)

4a H H 0 71 210–212 C10H8N2O2S(220.25)

3.78 (s, 3H, OCH3), 3.97 (s, 2H, SCH2), 6.75 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.90 (d,Jmeta = 2.4 Hz, 1H, Ar),8.15 (s, 1H, N=CH)(DMSO-d6)

4e C6H5 H 0 44 158–159 C16H12N2O2S(296.34)

3.81 (s, 3H, OCH3), 5.90 (s, 1H, SCHC6H5), 7.04 (d,Jmeta = 2.4 Hz, 1H, Ar), 7.21 (d,Jmeta = 2.4 Hz, 1H,Ar), 7.39 (s, 5H, Ar), 8.87 (s, 1H, N=CH) (DMSO-d6)

4f H H 1 10 176–178 C11H10N2O2S(234.27)

3.17, 3.51 (A2B2 system, 4H, CH2CH2), 3.85 (s, 3H,OCH3), 6.87 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.10 (d,Jmeta= 2.4 Hz, 1H, Ar), 8.71 (s, 1H, N=CH) (CDCl3)

6a H CH3 0 54 151 C11H10N2O2S(234.27)

2.85 (s, 3H, CH3), 3.78 (s, 3H, OCH3), 4.10 (s, 2H,SCH2), 6.75 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.85 (d,Jmeta= 2.4 Hz, 1H, Ar) (DMSO-d6)

6b CH3 CH3 0 63 166–167 C12H12N2O2S(248.30)

1.65 (d,J = 6.3 Hz, 3H, SCHCH3), 2.85 (s, 3H, CH3),3.84 (s, 3H, OCH3), 4.06 (q, 1H, SCHCH3), 6.78 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.00 (d,Jmeta= 2.4 Hz, 1H, Ar)(CDCl3)

6e C6H5 CH3 0 50 140–141 C17H14N2O2S(310.37)

2.73 (s, 3H, CH3), 3.78 (s, 3H, OCH3), 5.76 (s, 1H,SCHC6H5), 6.94 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.05 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.36 (s, 5H, Ar) (DMSO-d6)

705

8.3 Hz, 3H, CH3), 3.55 (q,J = 8.3 Hz, 1H, CHCH3), 3.78(s, 3H, OCH3), 6.75 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.83 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.80 (br s, 1H, NH) ppm.

4.1.5. 8-Methoxy-2-phenyl-4,5-dihydro-2H-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-one4e and 9-methoxy-2,4,5,6-tetrahydro-imidazo[3,4,5-e,f]-1,5-benzothiazepin-5-one4f (table III)

A solution of triethyl orthoformate (1.50 g, 0.01 mol)in anhydrous xylene (5 mL) was added slowly understirring to a suspension of2e and f (0.005 mol) inanhydrous xylene (20 mL). The reaction mixture wasrefluxed with stirring for 4 h for2e and 7 h for2f. Aftercooling, the solvent was evaporated in vacuo and thesolid residue was recrystallized from EtOH.

4.1.6. General procedure for 2-methyl-8-methoxy-4,5-dihydro-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-ones6a,b ande (table III)

A solution of triethyl orthoacetate (3.24 g, 0.02 mol) inanhydrous xylene (5 mL) was added slowly and withstirring to a suspension of2a, b and e (0.01 mol) in

anhydrous xylene (100 mL). The mixture was refluxedfor 6–12 h, then filtered while hot. The filtrate wasconcentrated under reduced pressure until ca. 40 mL. Theresulting precipitate was isolated by filtration and recrys-tallized from EtOH.

4.1.7. Tentative synthesis of 2-methyl-9-methoxy-4H-5,6-dihydroimidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one. For-mation of 2-ethoxy-9-methoxy-2-methyl-1H-5,6-dihydro-imidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one7 and 6-{[(E)-1-ethoxyethylidene]amino}-8-methoxy-2,3,4,5-tetra-hydro-1,5-benzothiazepin-4-one8

4.1.7.1. Cyclocondensation reaction in anhydrous xyleneThe reaction mixture, prepared as described above

using 2f as starting material, was refluxed for 16 h andwas evaporated under reduced pressure. The residue waspurified by flash chromatography using CHCl3 as eluentand recrystallized from EtOH to give7 (19% yield), m.p.157–159 °C [(C14H18N2O3S)]; 1H-NMR, δ (CDCl3):1.33 (t, J = 6.9 Hz, 3H,CH2CH3), 1.82 (s, 3H, CH3),2.63, 3.44 (A2B2 system, 4H, CH2CH2), 3.78 (s, 3H,

Table IV. Physical and chemical data of compounds9a–c, eand f.

Compound R n Yield M.p. Formula 1H-NMR, δ(%) (°C) (MW)

9a H 0 48 242–243 C10H8N2O2S2(252.31)

3.78 (s, 3H, OCH3), 4.05 (s, 2H, SCH2), 6.50 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.80 (d,Jmeta = 2.4 Hz, 1H, Ar),13.18 (s, 1H, NH) (DMSO-d6)

9b CH3 0 89 246–248 C11H10N2O2S2(266.33)

1.48 (d,J = 6.6 Hz, 3H, CHCH3), 3.78 (s, 3H, OCH3),4.37 (q, 1H,CHCH3), 6.50 (d,Jmeta= 2.4 Hz, 1H, Ar),6.78 (d,Jmeta = 2.4 Hz, 1H, Ar), 13.20 (s, 1H, NH)(DMSO-d6)

9c CH2CH3 0 10 172–175 C12H12N2O2S2(280.36)

1.00 (t, 3H, CH2CH3), 1.75–2.05 (m, 2H,CH2CH3),3.52 (q, 1H, SCH), 3.80 (s, 3H, OCH3), 6.75 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.90 (d,Jmeta = 2.4 Hz, 1H, Ar),11.05 (br s, 1H, NH) (CDCl3)

9e C6H5 0 25 263–265 C16H12N2O2S2(328.41)

3.79 (s, 3H, OCH3), 5.65 (s, 1H, SCHC6H5), 6.52 (d,J = 2.4 Hz, 1H, Ar), 6.83 (d,J = 2.4 Hz, 1H, Ar), 7.35(m, 5H, Ar), 13.28 (s, 1H, NH) (DMSO-d6)

9f H 1 13 117–120 C11H10N2O2S2(266.34)

2.62, 3.15, 4.20 (AB2X system, 4H, CH2CH2), 3.73 (s,3H, OCH3), 6.75 (d,J = 2.4 Hz, 1H, Ar), 6.83 (d,J =2.4 Hz, 1H, Ar), 11.28 (s, 1H, NH) (CDCl3)

706

OCH3), 4.20 (q,J = 6.9 Hz, 2H,CH2CH3), 6.32 (d,Jmeta

= 2.4 Hz, 1H, Ar), 6.85 (d,Jmeta= 2.4 Hz, 1H, Ar) ppm.

4.1.7.2. Cyclocondensation reaction in anhydrous xylenein the presence of pyridinium hydrochloride

To the reaction mixture, prepared as above, was added1.16 g (0.01 mol) of pyridinium hydrochloride. The reac-tion mixture was refluxed for 14 h and finally wasevaporated under reduced pressure. The residue waspurified by flash chromatography using CHCl3 as eluentand was recrystallized from EtOH to give8 (72% yield),m.p. 234–235 °C [(C14H18N2O3S)]; 1H-NMR, δ(CDCl3): 1.25 (t,J = 6.9 Hz, 3H, CH2CH3), 2.60 (s, 3H,CCH3), 2.60, 3.20 (A2B2 system, 4H, CH2CH2), 3.80 (s,3H, OCH3), 4.16 (q,J = 6.9 Hz, 2H,CH2CH3), 6.90 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.98 (d,Jmeta= 2.4 Hz, 1H, Ar),8.90 (s, 1H, NH) ppm.

4.1.8. General procedure for 8-methoxy-4-oxo-1H,4,5-dihydro-imidazo[3,4,5-d,e]-1,4-benzothiazin-2-thiones9a–cande (table IV)

Carbon disulfide (20 mL) was added slowly to asolution of2a–cande (0.01 mol) in anhydrous pyridine(50 mL). The reaction mixture was refluxed for 8 h for2a, 14 h for2b andeand for 30 h for2c. After cooling the

solvent was evaporated in vacuo. The solid residue wasrecrystallized from EtOH to give9a, b ande. In the caseof 9c the oily residue was induced to crystallize by addinga few drops of EtOH. The compound was purified byflash chromatography using CHCl3 as eluent and finallywas recrystallized from EtOH.

4.1.9. 9-Methoxy-4-oxo-1,4,5,6-tetrahydroimidazo[3,4,5-e, f]-1,5-benzothiazepin-2-thione9f (table IV)

A mixture of 2f (2.24 g, 0.01 mol), carbon sulfide(20 mL) and triethylamine (1.52 g, 0.015 mol) in anhy-drous tetrahydrofuran (50 mL) was refluxed under stir-ring for 16 h. After cooling the reaction mixture wasevaporated in vacuo and the residue was purified by flashchromatography using CHCl3 as eluent. The purifiedcompound was crystallized from EtOH.

4.1.10. General procedure for 8-methoxy-4,5-dihydro-1,2,3-triazolo[3,4,5-d, e]-1,4-benzothiazin-4-ones10a, b,d and e and 9-methoxy-4H-5,6-dihydro-1,2,3-tria-zolo[3,4,5-e, f]-1,5-benzothiazepin-4-one10f (table V)

A solution of sodium nitrite (1.04 g, 15 mmol) in10 mL of water was added slowly to an ice-cooledsuspension of the aminoderivative2a, b, d–f (0.01 mol)in 2 N HCl (10 mL).

Table V. Physical and chemical data of compounds10a, bandd–f.

Compound R n Yield M.p. Formula 1H-NMR, δ(%) (°C) (MW)

10a H 0 69 250–251 C9H7N3O2S(221.23)

3.82 (s, 3H, OCH3), 4.08 (s, 2H, SCH2), 6.80 (d,J = 1.5Hz, 1H, Ar), 6.95 (d,J = 1.5 Hz, 1H, Ar) (DMSO-d6)

10b CH3 0 24 191–194 C10H9N3O2S(235.26)

1.50 (d,J = 7.5 Hz, 3H, CHCH3), 3.80 (s, 3H, OCH3),4.50 (q, 1H,CHCH3), 6.85–7.15 (m, 2H, Ar) (DMSO-d6)

10d (CH2)3CH3 0 35 192–195 C13H15N3O2S 0.85 (t, 3H, CH3), 1.20–1.50 (m, 4H, CH2CH2CH2CH3),1.70–1.95 (m, 2H,CH2CH2CH2CH3), 3.83 (s, 3H,OCH3), 4.25–4.42 (m, 1H, SCH), 6.95 (br s, 2H, Ar)(DMSO-d6)

10e C6H5 0 23 215–217 C15H11N3O2S 3.80 (s, 3H, OCH3), 5.85 (s, 1H,CHC6H5), 6.80–7.55(m, 7H, Ar) (DMSO-d6)(297.33)

10f H 1 10 85–86 C10H9N3O2S(235.26)

2.70, 3.40 (A2B2 system, 4H, CH2CH2), 3.88 (s, 3H,OCH3), 7.00 (d,Jmeta= 2.4 Hz, 1H, Ar), 7.05 (d,Jmeta= 2.4 Hz, 1H, Ar) (CDCl3)

707

After the addition was complete, compound10a wasimmediately obtained as an abundant precipitate whichwas collected by filtration and recrystallized from MeOH.In all the other cases the reaction mixture was stirred atroom temperature for 2 h for compound2d, for 5h forcompounds2b and f and for 40 h for2e. The resultingprecipitate was filtered to give compounds10b, d andewhich were recrystallized from EtOH (10d and e) orEtAc (10c andd).

In the case of the benzothiazepine derivative, as noprecipitate was formed, the solution was alkalinized withdilute NaHCO3, extracted with CHCl3, washed withwater, dried over anhydrous Na2SO4 and dried in vacuo.The oily residue was purified by flash chromatographyusing CHCl3 as eluent to give10f.

4.1.11. General procedure for 9-methoxy-2,3,6,7-tetrahydro-5H-[1,4]thiazino[4,3,2-d,e]quinoxalin-3-ones11a–d(table VI)

To a stirred solution of the aminoderivatives2a–d(0.01 mol), tetrabutylammonium bromide (0.32 g,0.001 mol) and 1,2-dibromoethane (1.88 g, 0.01 mol) intetrahydrofuran, finely powdered potassium hydroxide(0.56 g, 0.01 mol) was added. The reaction mixture waskept at room temperature for 20 h for2d, 48 h for2a, 3d for 2b and 6 d for2c,and then filtered. The filtrate wasevaporated in vacuo and the residue was taken up withchloroform, the chloroform extract washed with water,dried over anhydrous Na2SO4 and dried in vacuo. Theoily residue was purified by flash chromatography usingCHCl3 as eluent.

Table VI. Physical and chemical data of compounds11a–f.

Compound R n Yield M.p. Formula 1H-NMR, δ (CDCl3)(%) (°C) (MW)

11a H 0 18 136–138 C11H12N2O2S(236.29)

3.20–3.43 (superimposed d and t, 4H,CH2CH2 +SCH2), 3.67 (s, 3H, OCH3), 3.92 (t,J = 10.0 Hz, 2H,CH2CH2), 4.25 (s, 1H, NH), 6.00 (d,Jmeta= 2.5 Hz, 1H,Ar), 6.22 (d,Jmeta= 2.5 Hz, 1H, Ar)

11b CH3 0 27 118–120 C12H14N2O2S(250.32)

1.50 (d,J = 5.1 Hz, 3H, CHCH3), 3.40 (t, 2H,CH2CH2),3.55 (q, 1H,CHCH3), 3.73 (s, 3H, OCH3), 3.75–4.23(m, 3H, CH2CH2 + NH), 6.05 (d,Jmeta = 2.4 Hz, 1H,Ar), 6.25 (d,Jmeta= 2.4 Hz, 1H, Ar)

11c CH2CH3 0 5 oil C13H16N2O2S(264.34)

1.08 (t, 3H, CH2CH3), 1.50–2.10 (m, 2H,CH2CH3),3.25–3.45 (m, 2H, CH2), 3.60–3.80 (m, 2H, CH2), 3.72(s, 3H, OCH3), 4.08 (br s, 1H, NH), 4.20–4.35 (m, 1H,SCH), 6.05 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.25 (d,Jmeta=2.4 Hz, 1H, Ar)

11d (CH2)3CH3 0 4 oil C15H20N2O2S(292.40)

0.86 (t, 3H, CH3), 1.20–1.70 (m, 4H, CH2CH2CH2CH3),1.80–2.00 (m, 2H,CH2CH2CH2CH3), 3.30–3.46 (m,3H, NH + CH2CH2), 3.72 (s, 3H, OCH3), 4.15–4.28 (m,2H, CH2CH2), 6.05 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.25 (d,Jmeta= 2.4 Hz, 1H, Ar)

11e C6H5 0 24 160–162 C17H16N2O2S(312.39)

3.70 (s, 3H, OCH3), 3.82, 4.25 (A2B2 system, 4H,CH2CH2), 4.07 (s, 1H, NH), 4.68 (s, 1H, SCHC6H5),6.03 (d,Jmeta= 2.4 Hz, 1H, Ar), 6.25 (d,Jmeta= 2.4 Hz,1H, Ar), 7.20–7.40 (m, 5H, Ar)

11f H 1 45 oil C12H14N2O2S(250.32)

2.65–3.50 (m, 8H, SCH2CH2 + NCH2CH2), 3.75 (s, 3H,OCH3), 4.20 (s, 1H, NH), 6.15 (d,Jmeta = 2.4 Hz, 1H,Ar), 6.50 (d,Jmeta= 2.4 Hz, 1H, Ar)

708

4.1.12. Preparation of 9-methoxy-2-phenyl-2,3,6,7-tetra-hydro-5H-1,4-thiazino[4,3,2-d,e]quinoxalin-3-one11eand 10-methoxy-2,3,6,7-tetrahydro-1H,5H-1,4-thiazepino[4,3,2-d,e]quinoxalin-5-one11f (table VI)

A suspension of aminoderivative2e and f (0.01 mol),1,2-dibromoethane (1.88 g, 0.01 mol), NaHCO3 (2.52 g,0.03 mol) in anhydrous xylene was refluxed under stir-ring for 24 h. After cooling the reaction mixture wasfiltered, the filtrate was dried in vacuo and the resultingresidue was purified by flash chromatography usingCHCl3 as eluent.

5. Biological evaluation

The procedure [8] used in the NCI’s test for anti-HIV-1screening is designed to detect agents acting at any stageof the virus reproductive cycle.

Tested compounds were dissolved in dimethylsulfox-ide and then diluted 1:100 in cell culture medium andthen serial half-log10 dilutions were prepared. T4 lym-phocytes (CEM cell line) were added and after a briefinterval HIV-1 was added, resulting in a 1:200 finaldilution of the compound. Uninfected cells with thecompound were used as a toxicity control and uninfectedcells without the compound as basic controls. Cultureswere incubated at 37 °C in a 5% carbon dioxide atmo-sphere for 6 d.

The tetrazolium salt XTT was added to all wells andcultures were incubated to allow formazan colour develop-ment by viable cells. Individual wells were analysedspectrophotometrically to quantitate formazan productionand in addition were viewed microscopically for detec-tion of viable cells and confirmation of protective activity.

Drug-treated virus-infected cells were compared withdrug-treated noninfected cells and with other controls

(untreated-infected and noninfected cells, drug-containing wells without cells) on the same plate. Datawere reviewed in comparison with other tests done at thesame time and a determination about activity was made.

Anti-HIV-1 activity was expressed as 50% effectiveconcentration (EC50) against HIV-1 cytopathic effectsand drug cytotoxicity as 50% cytotoxic concentration(CC50).

Acknowledgements

The authors would like to express their gratitude andthanks to the staff of the anti-HIV screening division,National Cancer Institute, Bethesda, MD, USA for carry-ing out the in vitro anti-HIV-1 testing. Special thanks aredue to the Ministero dell’Università e della RicercaScientifica e Tecnologica (M.U.R.S.T.) and the ConsiglioNazionale delle Ricerche (C.N.R.), Rome, for financialsupport.

References

[1] Pauwels R., Andries K., Desmyter J., Schols D., Kukla M., Breslin H.et al., Nature 343 (1990) 470–474.

[2] Parker K.A., Coburn C.A., J. Org. Chem. 56 (1991) 4600–4601.

[3] De Clercq E., Clin. Microbiol. Rev. 10 (1997) 674–693.

[4] Pauwels R., in: Adams J., Merluzzi V.J., (Eds.), The Search forAntiviral Drugs, Birkhauser, Boston, 1993, pp. 71–104.

[5] De Clercq E., Int. J. Immunotherapy 10 (1994) 145–158.

[6] Grandolini G., Perioli L., Ambrogi V., Gazz. Chim. Ital. 127 (1997)411–413.

[7] Liu K.C., Shih B.J., Chern J.W., J. Heterocycl. Chem. 26 (1989)457–460.

[8] Weislow O.W., Kiser R., Fine D., Bader J., Shoemaker R.H., BoydM.R., J. Natl. Cancer Inst. 81 (1989) 577–586.

709

Original article

Synthesis of a new series of N-hydroxy, N-alkylamidesof aminoacids as ligands of NMDA glycine site

Eleonora Ghidinia*, Maurizio Delcanalea, Vittorino Servadioa, Claudio Pietraa,Marco Bergamaschia, Gino Villettia, Enrico Redentia, Paolo Venturaa, Lucio Merlinib

aR & D Department, Chiesi Farmaceutici S.p.A., Via Palermo 26/A, 43100 Parma, ItalybDISMA, University of Milano, Via Celoria 2, I-20133 Milan, Italy

(Received 16 November 1998; accepted 4 February 1999)

Abstract – A new series of N-hydroxy, N-alkylamides of aminoacids structurally related to the N-hydroxy-3-amino-2 pyrrolidone[(±)HA-966] was synthesised and evaluated for the ability to displace [3H]Glycine, [3H]CGS19755, [3H]AMPA and [3H]Kainate binding sites.The N-heptyl glycinamide5a was the most potent compound (IC50 = 4.5µM) in inhibiting [3H]Glycine binding. Compounds5b, 5d, 5m, 5p,5q and5r showed an activity similar to (±)HA-966, whereas5h, 5i, 5n and5sappeared less active. None of the compounds tested exhibiteda significant displacement of [3H]AMPA and [3H]Kainate binding sites. Compounds active in the [3H]Glycine binding inhibited, to a differentdegree, NMDA induced contractions in guinea-pig LMPP preparation. © 1999 E´ditions scientifiques et médicales Elsevier SAS

N-hydroxyamides of aminoacids / hydroxamic acids / aminoacids / glycine antagonists

1. Introduction

The NMDA receptor possesses a variety of potentialdrug binding sites, among which the strychnine-insensitive glycine regulatory site plays an important roleas up-regulator of the receptor function and as a co-transmitter site [1–7]. Therefore the identification of thissite has stimulated intensive effort to discover selectiveligands to be used as potential neuroprotective andanticonvulsant drugs. Examples of these selective com-pounds include the kynurenic acid analogueL-689,560 [8], the 3-substituted indole-2-carboxylate [9],the quinoxaline ACEA 1021 [10] and the 2-quinolonederivative l-701,324 [11]. Before the development ofthese selective glycine antagonists, manipulation of theN-hydroxy-3-amino-2-pyrrolidone [(±)HA-966] (fig-ure 1) moiety has been reported to generate compoundswith affinity and selectivity for the glycine site in both invitro and in vivo functional studies [12]. To this aim, inthis paper we describe the synthesis and the functional invitro activity with regard to the antagonism of the NMDAresponse of a novel series of N-hydroxy, N-alkylamides

of aminoacid derivatives5a–w obtained upon chemicalmanipulation of a structure corresponding to the openedring of HA-966. In most of these compounds, the featuresof HA-966, i.e. the hydroxamic moiety and the aminogroup, have been maintained, whereas the kind, lengthand size of the N-linked group has been variouslymodified. In a few cases, also theα-carbon and the NH2group of the starting aminoacid were alkylated.*Correspondence and reprints

Figure 1. Structure of N-hydroxy-3-amino-2-pyrrolidone[(±)HA-966].

Eur. J. Med. Chem. 34 (1999) 711−717 711© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

The synthetic pathways for the synthesis of the titlecompounds are illustrated infigure 2 (procedure A andB). The new derivatives synthesised are numbered intable I.

Usually, hydroxamic acids are prepared followingclassical methods of literature [13]. We have found thatN-alkylhydroxylamines undergo condensation with suc-cinic esters of aminoacids under very mild condi-tions [14]. The reaction proceeds through an intermediate3 (figure 2, route A), which is soon formed, derived fromthe attack of the hydroxylamine oxygen on the succinicester carboxylic group. In a few cases, the intermediatewas isolated and characterised by1H NMR. This inter-mediate slowly rearranges to the product4 [15].

The route B was followed when the rearrangement ofthe intermediate3, which is formed in route A, was tooslow to give the product. To avoid the attack of thehydroxylamine oxygen on the active ester, the former wasprotected by reaction with trimethylchlorosilane and the

silylated hydroxylamine was then reacted with the activeester in pyridine.

The appropriate monosubstituted N-hydroxylamineswere conveniently prepared by a well-described synthe-sis [16] from the corresponding oxime [17, 18], while theother reactions just involve classical reactivity of ami-noacids [19].

The structures of the products5a–wwere confirmed byIR and1H NMR.

Although hydroxamic derivatives may exist as a mix-ture of Z and E isomers, only the former was present inDMSO at room temperature and at the concentration usedfor measuring the NMR spectra, in accordance to what isreported in the literature [20]. The conformation in solu-tion was also confirmed byn.O.e.’s experiments on thecorresponding hydrochlorides in order to be able toobserve the OH signal. Indeed, when the spectra arerecorded on the corresponding free bases, the protons onheteroatoms are in rapid exchange with traces of waterfrom DMSO-d6.

Figure 2. Procedures A and B used in preparing4 and5a–w.

712

3. Pharmacology

Compounds5a–w were tested for their ability todisplace the [3H]Glycine binding accordingly to theprocedure described in the experimental protocol.Table Ishows the results in comparison with that obtained withthe parent compound (±)HA-966.5a was the most potentcompound in inhibiting [3H]Glycine binding, IC50 being4.5 µM. N-isopropylglycinamide5b and N-heptyl-(D)-alaninamide5d derivatives as well as compounds5m, 5p,5q and5r showed an activity similar to (±)HA-966. TheN-benzyl, N-cyclohexyl, N-cyclohexylmethyl, N-(4-Cl-benzyl)glycinamide derivatives, respectively5h, 5i, 5nand5s, appeared less active than (±)HA-966. The remain-ing compounds tested were inactive in displacing [3H]G-lycine binding up to a concentration of 1 mM.

As for the interaction with the competitive site ofNMDA receptors, only compounds5a, 5b, 5d, 5h and5i

showed a weak activity in the displacement of[3H]CGS19755 binding, their IC50s being 16, 57, 72, 95and 51µM, respectively.

No significant displacement of [3H]Kainate or[3H]AMPA binding sites was observed for all compounds(up to 1 mM) under examination. As expected, alsoHA-966 (up to 1 mM) was inactive on [3H]CGS19755,[3H]Kainate and [3H]AMPA binding. Compounds show-ing a significant displacement of [3H]Glycine or[3H]CGS19755 binding sites exhibited a different degreeof inhibition of NMDA induced contractions in theguinea-pig LMPP in vitro, when tested at the concentra-tion of 100 µM. Accordingly with the [3H]Glycinebinding data, compound5a was the most potent ininhibiting NMDA induced contractions. None of thecompounds under examination including (±) HA-966exhibited, per se, a contractile response.

Table I. Physical properties and observed biological activity of compounds5.

Compound R1 R2 R3 M.W. Salt Configuration Yield (%) IC50a (µM) % Inhibitionb

(±)HA-966 22 245a heptyl H H 188.27 64 4.5 1005b isopropyl H H 132.16 34 20 155c butyl H H 146.19 24 N.A. N.T.5d heptyl methyl H 202.3 R 22 49 105e heptyl methyl H 202.3 S 22 N.A. N.T.5f H H penthyl 160.22 43 N.A. N.T.5g nonyl H H 216.33 27 N.A. N.T.5h benzyl H H 180.21 51 91 195i cyclohexyl H H 172.23 59 51 225j ethoxycarbonylmethyl H H 176.17 45 N.A. N.T.5k heptyl H methyl 202.3 35 N.A. N.T.5l 4-methylbenzyl H H 194.24 HCl 25 N.A N.T.5m 2,3-dimethylpenthyl H H 188.27 HCl c 64 29 785n cyclohexylmethyl H H 186.26 74 62 455o heptyl butyl H 244.38 HCl RS 8 N.A. N.T.5p 4-methoxybenzyl H H 210.23 10 27 75q 4-carbomethoxybenzyl H H 238.25 16 28 205r 4-nitrobenzyl H H 225.21 HCl 16 26 105s 4-chlorobenzyl H H 214.65 HCl 71 42 135t heptyl phenyl H 264.37 11 N.A. N.T.5u 1-methylheptyl H H 202.3 HCl RS 16 N.A. N.T.5v heptyl hexyl H 272.43 RS 8 N.A N.T.5w heptyl butyl H 244.38 HCl S 26 N.A. N.T.

aPotencies of compounds in inhibiting [3H]Glycine binding in rat cortical homogenates. N.A.: not active up to 1 mM;bEffect on the inhibitionof NMDA (1mM)-induced contractions in guinea-pig LMPP. Each compound was tested at 100µM. N.T.: not tested;cAs diastereomericmixture.

713

The SAR in this class of compounds was explored,changing first the substituents on the nitrogen atom of thehydroxamic acids. The activity seems to depend rathercritically on the size and length of the alkyl group, thehighest activity being reached for the N-heptyl group(5a), whereas shortening (5c) or lengthening (5g) of thelinear chain was detrimental, as well asα-methylation(5u). Activity, although lower, was retained with bulkiergroups, such as isopropyl (5b), 2,3-dimethylpentyl (5m),cyclohexyl (5i), or cyclohexylmethyl (5n). Only a small(methyl) substituent seems to be tolerated on the ami-noacidα-carbon, (cf.5d vs. 5o, 5t, 5v, 5w) but this isdependant on the configuration (5d vs. 5e).

The only example of an unsubstituted hydroxamic acid5f, was also inactive.

Attempts to find a linear correlation with one or moreparameters representative of these substituents, such aslogP, WDW volume, molecular connectivity or lengthfailed (see experimental part). Similar negative resultswere obtained by applying a principal component analy-sis. Therefore, only a qualitative discussion of the resultsreported above seems possible.

All the compounds showed no, or scarce, bindingaffinity towards other excitatory aminoacid receptors,such as glutamate, kainate and AMPA receptors.

In conclusion, the N-alkylamides of aminoacids relatedto the structure of (±) HA-966 demonstrated a varyingdegree of affinity for the glycine binding site on theNMDA receptors,5a being the most potent and com-pounds5p, 5q, 5r and5smore selective versus the otherexcitatory aminoacids evaluated. The antagonist effect ofthese substances was also confirmed in the functionalstudies in vitro on guinea-pig LMPP. In this respect, thelevel of inhibition showed by (±) HA-966 is consistentwith that reported in other experiments [21]. Since (±)HA-966 is defined also as a glycine partial agonist, it islikely that these compounds behave in such a manner.However, this issue was not addressed in this study. Moreexperimental work at the electrophysiological level or atthe modulation on NMDA receptors in the presence orabsence of glycine need to be performed in order toclarify this point.

4. Experimental protocols

4.1. Chemistry

Melting points were determined on a Büchi capillarymelting point apparatus and are uncorrected. The IRspectra were determined on a Perkin Elmer 1310 spec-trophotometer.1H NMR spectra were recorded at 200.13MHz on a Bruker ACF 200 spectrometer; chemical shifts

are in δ (ppm), with tetramethylsilane as internal stan-dard. Mass spectra were measured with a Fisons-VG Trio2000 single quadrupole spectrometer equipped with adual EI/CI source. The [α]D were performed on aPerkin-Elmer 241 and 241-MC polarimeter. Sodium sul-phate was employed as a drying agent for ether extracts.The petroleum ether used throughout this work had aboiling point of 40–70 °C. TLC on silica gel plates(Merck, 60 F 254) was used to check product purity andthe spots detected with ninhydrin. The intermediate3 canbe monitored by TLC also, with potassium ferricyanide-ferric chloride giving a typical blue spot, whereas thefinal product4 gives a brown one with ferric chloride.Preparative chromatography was carried out on silica gelICN (32–63µm). The structures of all compounds wereconsistent with their analytical and spectroscopic data.

4.1.1. General synthesis of4 (route A)To a mixture of the appropriate N-monosubstituted

hydroxylamine 1 (85 mmol) in dichloromethane(360 mL) at 0 °C was added the succinic ester of thecorresponding N≠-protected aminoacid2 (82 mmol). Thesolution was stirred for 1 h at 10 °C and at roomtemperature for a few days. The composition of the crudeproduct was determined by TLC and infrared spectros-copy. The intermediate3 slowly rearranges to the product4. When the conversion was complete, the solution waspoured in water. The separated organic phase was washedwith a 5% sodium bicarbonate solution (2× 50 mL), withwater (2× 50 mL) and dried. The solvent was removed invacuo at 35 °C. The compound4 was finally deprotectedfollowing literature methods. The following compoundswere prepared according to this procedure from thecorresponding N-protected (Cbz or Boc) aminoacid.

4.1.1.1. N-Hydroxy, N-heptylglycinamide5aWhite solid; IR (Nujol) 3 355, 3 280, 1 630,

1 540 cm–1; 1H NMR (DMSO-d6) δ 0.9 (t, 3H), 1.0–1.4(s, 8H), 1.5 (m, 2H), 3.5 (t,J = 7.0 Hz, 2H), 3.9 (s, 2H),4.0–4.5 (br, 3H); EI+/MS (m/z) 188 (M+).

4.1.1.2. N-Hydroxy, N-heptylalaninamide5d and5ePale oil; IR (CHCl3) 3 060, 1 610, 1 580 cm–1; 1H

NMR (DMSO-d6) δ 0.8 (t, 3H), 1.0 (d,J = 6.8 Hz, 3H),1.1–1.4 (m, 8H), 1.5 (m, 2H), 3.3–3.6 (m, 2H), 3.8 (q,J= 6.8 Hz, 1H), 3.9–4.5 (br, 3H); EI+/MS (m/z) 203 (M+.);5d: [a]D = –10.5° (c = 0.6, EtOH);5e: [a]D = + 11.2° (c= 0.6, EtOH).

4.1.1.3. N-Hydroxy, N-benzylglycinamide5hWhite solid; m.p. 118–120 °C; IR (Nujol) 3 320, 3 290,

2 500, 1 800, 1 630, 1 550, 720, 685 cm–1; 1H NMR

714

(DMSO-d6) δ 3.4 (s, 2H), 4.7 (s, 2H), 3.5–4.5 (br, 3H),7.1–7.5 (m, 5H).

4.1.1.4. N-Hydroxy, N-cyclohexylglycinamide5iWhite solid; m.p. 128–129 °C; IR (CHCl3)

3 360–3 050, 1 620 cm–1; 1H NMR (DMSO-d6) δ0.8–1.6 (m, 8H), 1.7 (m, 2H), 3.3 (s, 2H), 4.1 (m, 1H),3.5–4.5 (br, 3H); EI+/MS (m/z) 173 (M+.).

4.1.1.5. N-Hydroxy, N-ethoxycarbonylmethylglycinamide5j

White solid; m.p. 159–162 °C; IR (Nujol)3 500–2 600, 1 720, 1 640, 1 600 cm–1; 1H NMR(DMSO-d6) δ 1.2 (t,J = 7 Hz, 3H), 3.8 (s, 2H), 4.1 (q,J= 7 Hz, 2H), 4.4 (s, 2H), 8.3 (br, 3H), 10.7 (br, 1H);EI+/MS (m/z) 176 (M+.).

4.1.1.6. N-Hydroxy, N-heptyl, N≠-methylglycinamide5kWhite solid; IR (Voltalef) 3 430, 3 300, 2 400, 1 850,

1 630 cm–1; 1H NMR (DMSO-d6) δ 0.9 (t, 3H), 1.1–1.5(m, 8H), 1.6 (m, 2H), 2.5 (s, 2H), 3.5 (s, 2H), 3.6 (t,J =7Hz, 2H), 3.6–4.4 (br, 3H); EI+/MS (m/z) 203 (MH+.).

4.1.1.7. N-Hydroxy, N-(4-methylbenzyl)glycinamide5lWhite solid; m.p. 197–199 °C; IR (Nujol)

3 200–3 000, 1 640, 880, 780 cm–1; 1H NMR (DMSO-d6)δ 2.3 (s, 3H), 3.8 (s, 2H), 4.7 (s, 2H), 7.0–7.2 (AB syst.,2H), 8.2 (br, 3H), 10.5 (br, 1H); EI+/MS (m/z) 194 (M+.).

4.1.1.8. N-Hydroxy, N-(2,3-dimethylpenthyl)glycinamide5m

White solid;1H NMR (DMSO-d6) δ 0.8 (d; d,J = 6.8Hz, 3H), 0.9 (d; t, 6H), 0.9–1.5 (m, 3H), 1.0 (m, 1H), 3.4(m, 2H), 3.8 (q,J = 7 Hz, 2H), 8.2 (br, 3H), 10.5 (br, 1H);EI+/MS (m/z) 188 (M+.).

4.1.1.9. N-Hydroxy, N-cyclohexylmethylglycinamide5nWhite solid; IR (Nujol) 3 500–2 300, 1 640 cm–1; 1H

NMR (DMSO-d6) δ 0.9 (m, 2H), 1.1–1.3 (m, 3H),1.5–1.9 (m, 6H), 3.3 (d,J = 6.8 Hz, 2H), 3.4 (s, 2H),3.2–4.2 (br, 3H); EI+/MS (m/z) 186 (M+.).

4.1.1.10. N-Hydroxy, N-heptyl, norleucinamide5oWhite solid; 1H NMR (DMSO-d6) δ 0.9 (t; t, 6H),

1.1–1.5 (m, 12H), 1.6 (m, 2H), 1.8 (m, 2H), 3.4 (m, 1H),3.7 (m, 1H), 4.2 (t,J = 7 Hz, 1H), 8.3 (br, 3H), 10.5 (br,1H); EI+/MS (m/z) 244 (M+.).

4.1.1.11. N-Hydroxy, N-(4-methoxybenzyl)glycinamide5p

White solid; IR (KBr) 3 320, 3 290, 1 640;1H NMR(DMSO-d6) δ 3.3 (s, 2H), 3.7 (s, 3H), 3.7–4.3 (br, 3H),4.6 (s, 2H), 6.9 (AÁ of AA≠XX≠ syst., 2H), 7.2 (XX≠ ofAA≠XX≠ syst., 2H); EI+/MS (m/z) 210 (M+.).

4.1.1.12. N-Hydroxy, N-(4-carbomethoxybenzyl)glyci-namide5q

White solid; IR (KBr) 3 330, 3 290, 1 650 cm–1; 1HNMR (DMSO-d6) δ 3.4 (s, 2H), 3.5–4.2 (br, 3H), 3.9 (s,3H), 4.8 (s, 2H), 7.4 (d,J = 8.2 Hz, 2H), 8.0 (d,J = 8.2Hz, 2H); EI+/MS (m/z) 239 (M+.).

4.1.1.13. N-Hydroxy, N-(4-nitrobenzyl)glycinamide5rWhite solid; m.p. 179–181 °C; IR (KBr) 3 320,

3 200–2 500, 1 680 cm–1; 1H NMR (DMSO-d6) δ 3.9 (s,2H), 4.9 (s, 2H), 7.6 (d,J = 8.2 Hz, 2H), 8.2 (d,J = 8.2Hz, 2H), 8.3 (br, 3H), 10.8 (br, 1H); EI+/MS (m/z) 225(M+.).

4.1.1.14. N-Hydroxy, N-(4-chlorobenzyl)glycinamide5sWhite solid; m.p. 183–185 °C; IR (KBr) 3 500–2 500,

1 660 cm–1; 1H NMR (DMSO-d6) δ 3.9 (s, 2H), 4.7 (s,2H), 7.2–7.5 (AB syst., 4H), 8.3 (br, 3H), 10.7 (br, 1H);EI+/MS (m/z) 214 (M+.).

4.1.1.15. N-Hydroxy, N-heptyl,2-phenylglycinamide5tWhite solid; IR (nujol) 3 310, 3 100–3 000,

1 630 cm–1; 1H NMR (DMSO-d6) δ 0.8 (t, 3H), 1.0–1.4(m, 8H), 1.5 (m, 2H), 3.3–3.7 (br, m, 5H), 5.0 (s, 1H);EI+/MS (m/z) 264 (M+.).

4.1.1.16. N-Hydroxy, N-(1-methylheptyl)glycinamide5uWhite solid; m.p.> 95 °C;1H NMR (DMSO-d6) δ 0.8

(t, 3H), 1.1 (d,J = 6.8 Hz, 3H), 1.1–1.5 (m, 9H), 1.6 (m,1H), 3.8 (s, 2H), 4.4 (m, 1H), 8.3 (br, 3H), 10.7 (br, 1H);EI+/MS (m/z) 202 (M+.).

4.1.1.17. N-Hydroxy, N-heptyl,2-hexylglycinamide5vWhite solid; IR (KBr) 3 320, 3 280, 3 000–2 800,

1 640, 1 620 cm–1; 1H NMR (DMSO-d6) δ 0.8 (t, 6H),1.1–1.5 (m, 16 H), 1.6 (m, 2H), 1.8 (m, 2H), 3.4 (m, 1H),3.5–4.2 (br, 3H), 3.7 (m, 1H), 4.2 (t, 1H); EI+/MS (m/z)272 (M+.).

4.1.1.18. N-Hydroxy, N-heptylnorleucinamide5wWhite solid; 1H NMR (DMSO-d6) δ 0.8 (t, 6H),

1.1–1.5 (m, 12H), 1.8 (m, 2H), 3.5 (m, 1H), 3.7 (m, 1H),8.3 (br, 3H), 10.7 (br, 1H); EI+/MS (m/z) 244 (M+.).

4.1.2.General synthesis of4 (route B)The appropriate hydroxylamine hydrochloride1

(80 mmol) was dissolved in dichloromethane (200 mL),added with potassium carbonate (80 mmol) and stirredfor 10 min at room temperature. The solution was filteredand the filtrate was cooled to –10 °C. Pyridine (350 mL)and trimethylchlorosilane (800 mmol) were added andthe solution was stirred at room temperature for 30 min.The solution was cooled again to –10 °C and the succinic

715

ester of the corresponding N≠-protected aminoacid2 wasadded (82 mmol). The reaction was stirred at roomtemperature for a further 2 d. The solution was treatedwith hydrogen chloride until pH = 1 and the product wasextracted with ethyl acetate (3× 70 mL). The organiclayer, that contained only one product, was dried and thesolvent evaporated in vacuo.

4.1.2.1. N-Hydroxy, N-isopropylglycinamide5bWhite solid; m.p. 58–59 °C; IR (CHCl3) 3 350,

1 640 cm–1; 1H NMR (DMSO-d6) δ 1.1 (d,J = 6.8 Hz,6H), 4.0 (s, 2H), 4.5 (m, 1H), 8.3 (br, 3H), 10.4 (br, 1H);EI+/MS (m/z) 132 (M+.).

4.1.2.2. N-Hydroxy, N-butylglycinamide5cWhite solid;1H NMR (DMSO-d6) δ 0.9 (t, J = 7 Hz,

3H), 1.3 (m, 2H), 1.6 (m, 2H), 3.6 (t,J = 7.0 Hz, 2H), 3.9(s, 2H), 4.1 (br, 3H); EI+/MS (m/z) 176 (M+.).

4.1.2.3. N-Hydroxy, N≠-penthylglycinamide5fPale oil; m.p. 130–132 °C; IR (oil) 3 300–2 300,

1 620 cm–1; 1H NMR (DMSO-d6) δ 0.8 (t, 3H), 1.1–1.5(m, 6H), 2.4 (t,J = 7.0 Hz, 2H), 3.0 (s, 2H), 3.1–3.6 (br,3H); EI+/MS (m/z) 161 (M+.).

4.1.2.4. N-Hydroxy, N-nonylglycinamide5gWhite solid; m.p. 118–120 °C; IR (Nujol) 3 320, 3 290,

2 500, 1 800, 1 630, 1 550, 720, 685 cm–1; 1H NMR(DMSO-d6) δ 3.4 (s, 2H), 4.7 (s, 2H), 3.5–4.5 (br, 3H),7.1–7.5 (m, 5H).

4.2. Pharmacology

Assays of [3H]Glycine binding were performed asdescribed by Snell et al. [22]. Rat brain cortices wereremoved from male Wistar rats in order to prepare amembrane fraction by standard techniques. A quantity of12.5 mg of membrane preparation was incubated with 20nM [3H]Glycine for 30 min at 4 °C. Non-specific bindingwas estimated in the presence of 10µM glycine. Mem-branes were filtered and washed 3 times and the filterswere counted to determine [3H]Glycine specificallybound.

Assays of [3H]CGS19755 binding were performed asdescribed by Murphy et al. [23]. 20 mg of membranepreparation was incubated with 10 nM [3H]CGS19755for 20 min at 4 °C. Non-specific binding was estimated inthe presence of 100µM L-glutamate. Membranes werefiltered and washed 3 times and the filters were counted todetermine [3H]CGS19755 bound.

The binding to kainic receptors was measured by using[3H]Kainate [24], a selective agonist that binds to the

kainate subtype of the ionotropic glutamate receptorspresent in rat brain. 15 mg of membrane preparation wasincubated with 5 nM [3H]Kainate for 1 h at 4 °C. Non-specific binding was estimated in the presence of 1 mMnative L-glutamate. Membranes were filtered and washed3 times to separate free from bound ligand and filterswere counted to determine [3H] Kainate bound.

The binding to AMPA receptors was measured byusing [3H]AMPA [25]. Briefly, 20 mg of membranepreparation was incubated with 10 nM [3H]AMPA for30 min at 4 °C. Membranes were filtered and washed 3times to separate free from bound ligand and filters werecounted to determine [3H]AMPA bound. All conditionswere tested in triplicate. For each compound, concentra-tion response changes in [3H]Glycine, [3H]Kainate or[3H]CGS19755 binding were fitted to a sigmoidal curveand IC50 values were estimated using a curve fittingprogram to the logistic equation described by De Lean etal. [26].

Functional studies in vitro were performed in theguinea-pig ileum muscle/myenteric plexus preparation(LMPP) [27]. Two cm lenghts of intestine were stretchedover a glass pipette. The LMPP was separated from theunderlying circular muscle by a cotton-tipped applicationand mounted in a 20 mL organ bath contained Mg++-freeKrebs’ buffer (composition in mM: NaCl 134, KCl 3.4,CaCl2 2.8, KH2PO4 1.3, NaHCO3 16 and glucose 7.7) at37 °C bubbled with 5% CO2 in oxygen. Contractionswere recorded isometrically (0.5 g tension) with trans-ducers connected to a poligraph. After 1 h of equilibrationperiod, tissues were contracted with NMDA 1mM. Com-pounds that showed a significant activity in the displace-ment of [3H]Glycine binding sites were evaluated in theirability to antagonise the NMDA response.

The molecular models of the compounds were built ona Silicon Graphics IRIS 35, using the program Insight II.The program Discover (Molecular Simulation Inc., SanDiego, CA) was used to generate low-energy conforma-tions. The values of van der Waals volumes (V) and thelength of R groups (L) were calculated on these refinedconformations via Insight II. Calculation of logP wascarried out using atomic parameters derived by Ghoseand Crippen [28, 29].

The structure-activity relationships were explored withthe program Systat 5.0 (Systat Inc.), used to find possiblesimple correlations between biological activity and eachof the above parameters logP, V or L and to perform aprincipal component analysis (PCA). In both cases nosignificant relationships were found.

716

Acknowledgements

The authors are grateful to Mr. Milco Lipreri fortechnical support.

References

[1] Johnson J.W., Ascher P., Nature 325 (1987) 529–531.

[2] Small D.L., Buchan A.M., Int. Rev. Neurobiol. 40 (1997) 137–171.

[3] Sucher N.J., Awobuluyi M., Choi Y.B., Lipton S.A., Trends Phar-macol. Sci. 17 (1996) 348–355.

[4] Leeson P.D., Iversen L.L., J. Med. Chem. 37 (1994) 4053–4067.

[5] Krogsgaardlarsen P., Ebert B., Lund T.M., Braunerosborne H., SlokF.A., Johansen T.N., Brehm L., Madsen U., Eur. J. Med. Chem. 31(1996) 515–537.

[6] Nankai M., Klarica M., Fage D., Carter C., Prog. Neuro-Psycopharmacol. Biol. Psychiatry 22 (1998) 35–64.

[7] Bigge C.F., Nikam S.S., Exp. Opin. Ther. Pat. 7 (1997) 1099–1114.

[8] Grimwood S., Moseley A.M., Carling R.W., Leeson P.D., FosterA.C., Mol. Pharmacol. 41 (1992) 923–930.

[9] Gaviraghi G., Pietra C., Ratti E., Trist D., Cerebrovascular Dis. 6(1996) P1220.

[10] Woodward R.M., Huettner J.E., Guastella J., Keana J.F.K., WeberE., Mol. Pharmacol. 47 (1995) 568–581.

[11] Kulagowski J.J., Baker R., Curtis N.R., Leeson P.D., Mawer I.M.,Moseley A.M.E.T. et al., J. Med. Chem. 37 (1994) 1402–1405.

[12] Fletcher E.J., Loger D., Eur. J. Pharmacol. 151 (1988) 161–162.

[13] Kehl H. (Ed.), Chemistry and Biology of Hydroxamic Acids, Karger,NewYork, 1982.

[14] Chiesi P., Ventura P., Servadio V., Italian Patent IT 01255240,(1992).

[15] Stroh R. (Ed.), Houben-Weyl Handbuch der Organischen ChemieVol 10/1, Georg Thieme Velag, Stuttgart, 1971.

[16] Borch R.F., Bernstein M., Durst D., J. Am. Chem. Soc. 93 (1971)2897–2904.

[17] Lachman A., Org. Synth. Coll. Vol. II (1943) 70–71.

[18] Kahr K., Berther C., Chem. Ber. 93 (1960) 132–134.

[19] Barrett G.C., Chemistry and Biology of Amino Acids, Chapman andHall, London, 1985.

[20] Brown D.A., Glass W.K., Mageswaran R., Mohammed S.A., J.Magn. Reson. 29 (1991) 40–45.

[21] Campbell B.G., Couceyro P., Keana J.F.W., Weber E., J. Pharmacol.Exp. Ther. 257 (1991) 754–766.

[22] Snell L.D., Morter R.S., Johnson K.M., Eur. J. Pharmacol. 156(1988) 105–110.

[23] Murphy D.E., Adler J., Br. J. Pharmacol. 95 (1988) 932–938.

[24] London E.D., Coyle J.T., Mol. Pharmacol. 15 (1979) 492–505.

[25] Olsen R.W., Szamraj O., Houser C.L., Brain. Res. 402 (1987)243–254.

[26] De Lean A., Munson P.J., Rodbard D., Am. J. Physiol. 235 (1978)E97–E105.

[27] Shannon H.E., Sawyer B.D., J. Pharmacol. Exp. Ther. 251 (1989)518–523.

[28] Ghose A.K., Pritchett A., Crippen G.M., J. Comp. Chem. 9 (1988)80–90.

[29] Viswanadhan V.N., Ghose A.K., Revankar G.R., Robins R.K., J.Chem. Inf. Comput. Sci. 29 (1989) 163–172.

717

Original article

Synthesis and binding affinitiesfor 5-HT 1A, 5-HT2A and 5-HT2C receptors of a series

of 1- and 2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole derivatives

Giuseppe Caliendoa*, Ferdinando Fiorinoa, Paolo Griecoa, Elisa Perissuttia,Vincenzo Santagadaa, Stefania Albriziob, Loredana Spadolab,

Giancarlo Brunic, Maria Rosaria Romeoc

aDipartimento di Chimica Farmaceutica e Tossicologica, Università di Napoli “Federico II”, Via D. Montesano, 49-80131 Napoli, ItalybDipartimento di Scienze Farmaceutiche, Università di Salerno, Piazza V. Emanuele 9, 84080, Penta di Fisciano, Salerno, Italy

cIstituto di Farmacologia, Università di Siena- Via delle Scotte, 6 - 53100 Siena, Italy

(Received 9 September 1998; accepted 4 February 1999)

Abstract – A number of 1- and 2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole derivatives (1–4) were prepared in order to obtaincompounds with a high affinity and selectivity for 5-HT1A receptors. 5-HT1A, 5-HT2A and 5-HT2C affinities were determined by radioligandbinding experiments and the most active compounds were also tested for binding affinities on dopaminergic D-1, D-2 and adrenergicα1, α2receptors. The modification of aromatic substituents, the length of the alkyl chain and its position on the 4-benzoyl-1,2,3-triazole ring wereexplored. Most of the considered compounds generally showed moderate to high affinity for the 5-HT1A receptor binding site. Threederivatives2c, 3c and3e bind to 5-HT1A receptors in the nanomolar range (IC50 values = 2, 7.2 and 2.6 nM respectively). The most activecompound,2c, presented a high degree of selectivity versus all considered receptors. It was found that the benzoyltriazole derivatives1h and4c are new selective ligands for 5-HT2A (IC50 = 89 nM) and 5-HT2C receptors (IC50 = 17 nM), respectively. © 1999 E´ditions scientifiqueset médicales Elsevier SAS

arylpiperazines / synthesis / 5-HT receptor / serotonin

1. Introduction

The neurotransmitter serotonin is involved in variousphysiological (e.g. sleep and thermoregulation) andpathological processes (e.g. migraine and depression). Aclassification of 5-HT receptor subtypes, their role invarious CNS activities and their respective ligands wererecently reviewed [1, 2]. One of the serotonin receptorsubtypes, 5-HT1A, plays an important role as the soma-todendritic autoreceptor (presynaptic) in the dorsal raphenucleus and as a postsynaptic receptor for 5-HT interminal areas [3]. Buspirone and ipsapirone, which areagonists and display high affinity to the 5-HT1A receptor(IC50 = 60 and 35 nM, respectively), are presently beingused as anti-anxiety agents [4, 5]. It seems useful to

develop ligands that are selective for the 5-HT1A subtypeto facilitate the study and characterization of this recep-tor, but also in view of their potential use as anxiolytics..

The most commonly used ligand, 8-OH-DPAT (8-hydroxy-2-(N-N-di-n-propylamino)tetralin), is a potent5-HT1A agonist, and in fact, the tritium-labelled com-pound is the ligand of choice for 5-HT1A receptor bindingstudies (Kd = 0.5 nM, rat hippocampal homogenates).

One class of compounds with affinity for the 5-HT1A

receptor is represented by the arylpiperazine deriva-tives [6] (general structureI ) where R is a heterocyclicnucleus.

*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 719−727 719© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

Affinity and specificity depend critically on the natureof the specific heterocyclic nucleus (R), on the (X)aromatic substitution and on the length (n) of the poly-methylene chain. In particular, it was observed thatdifferent terminal groups (R) play an important role in theinteraction with a corresponding hydrophobic region ofthe 5-HT1A receptor [7].

We recently reported a series of arylpiperazines ofgeneral structureI [8] having a benzotriazole group as aterminal (R) moiety with mixed 5-HT1A, 5-HT2A and5-HT2C receptor affinity.

Structure-affinity relationships in this series [8]showed that nanomolar affinity towards the 5-HT1A

receptor requires a methoxy group at theortho-positionof the phenyl ring. However, the most potent benzotria-zole ligands of the 5-HT1A receptor have poor selectivity,e.g., they possess affinities for the 5-HT2A receptor only10–50 times lower than those measured for 5-HT1A.

Considering the important role played by the (R)terminal group, in an attempt to increase both affinity andselectivity for the 5-HT1A receptor the benzotriazolenucleus of these arylpiperazine derivatives [8], was re-

placed by a 4-benzoyl-1,2,3-triazole group. These com-pounds (1–4) represent open chain analogues of theirbenzotriazole counterparts with bioisosteric properties.

Most of the arylpiperazine moieties used are thosepreviously reported [8] displaying the highest affinity forthe 5-HT1A receptor, typified by 1-(2- or 4-methoxy-phenyl)-, 1-phenyl-, 1-(2- or 3- or 4-chlorophenyl)pipe-razine. In this work we considered also a new electron-withdrawing substituent such as a fluoro groupintroduced into the 2- or 4- position of the phenylpipera-zine moiety. The piperazine is connected to the terminalbenzoyltriazole system via bridges of two (series1 and2)or three (series3 and4) methylene groups.

Herein we report the synthesis of a series of 1- and2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole de-rivatives 1–4, (table I) and the affinities for 5-HT1A,5-HT2A and 5-HT2C receptors obtained by radioligandbinding studies. The most active compounds were alsotested for their affinity at dopaminergic D-1, D-2 andadrenergicα1 and α2 receptors to substantiate theirpharmacological profile in view of potential therapeuticuses.

Table I. Physicochemical properties of 4-benzoyl-1,2,3-triazole derivatives.

1-Substituted 4-benzoyl-1,2,3-triazoles 2-Substituted 4-benzoyl-1,2,3-triazoles

X n Formulaa M.W. Compoundb M.p. (°C) Yieldc % Compoundb M.p. (°C) Yieldc %

H 2 C21H23N5O·HCl 397.91 1a 201 – 202 25 2a 206 – 207 35o-Cl 2 C21H22ClN5O·HCl 432.34 1b 208 – 210 23 2b 212 – 214 32m-Cl 2 C21H22ClN5O·HCl 432.34 1c 204 – 206 24 2c 180 – 182 35p-Cl 2 C21H22ClN5O·HCl 432.34 1d 236 – 238 25 2d 191 – 193 37o-OCH3 2 C22H25N5O2·HCl 428.01 1e 210 – 212 28 2e 206 – 208 40p-OCH3 2 C22H25N5O2·HCl 428.01 1f 218 – 220 20 2f 200 – 201 38o-F 2 C21H22N5OF·HCl 415.89 1g 142 – 143 28 2g 164 – 165 42p-F 2 C21H22N5OF·HCl 415.89 1h 216 – 217 24 2h 196 – 197 38

H 3 C22H25N5O·HCl 411.93 3a 183 – 184 27 4a 251 – 252 42o-Cl 3 C22H24ClN5O·HCl 446.35 3b 334 – 335 27 4b 177 – 179 45m-Cl 3 C22H24ClN5O·HCl 446.35 3c 231 – 232 22 4c 178 – 180 38p-Cl 3 C22H24ClN5O·HCl 446.35 3d 217 – 219 27 4d 248 – 249 40o-OCH3 3 C23H27N5O2·HCl 441.95 3e 207 – 209 28 4e 153 – 155 41p-OCH3 3 C23H27N5O2·HCl 441.95 3f 220 – 221 22 4f 208 – 210 34o-F 3 C22H24N5OF·HCl 429.92 3g 188 – 189 28 4g 179 – 180 42p-F 3 C22H24N5OF·HCl 429.92 3h 187 – 188 25 4h 208 – 209 39

aSatisfactory microanalyses obtained: C, H, N, Cl, F values are within± 0.4% of the theoretical ones;bAll compounds were crystallized bymethyl alcohol and diethyl ether.cYield refers to the single structural isomer after separation by chromatography as free bases.

720

2. Chemistry

4-benzoyl-1,2,3-triazole derivatives1a–h, 2a–h, 3a–hand 4a–h (table I) were synthesized as described infigure 1. The parent compound 4-benzoyl-1,2,3-triazole(7) was obtained in 95% yield by a modified methoddescribed in the literature [9, 10]. Oxidation of thephenylethynylcarbinol5 with chromium trioxide andconcentrated sulphuric acid produced phenyl ethynylketone6, which was successively treated with NaN3 inanhydrous dimethylacetamide providing the desired com-pound. Alkylation of the aromatically substituted 1-(2-chloroethyl)- or 1-(3-chloropropyl)-4-phenylpiperazines,prepared as described in the literature [8], with4-benzoyl-1,2,3-triazole in butan-2-one in the presence ofpotassium carbonate afforded a mixture of the expected1-, 2- and 3- triazole isomers with an overall yield in therange of 45–80%. The three isomers were separated bychromatography on a silica gel column usingn-hexane/diethyl ether 95:5 as eluent. The faster moving2-substituted isomers (2 and 4) were collected with ahigher yield with respect to the slower moving1-substituted isomers (1 and3). The yield of intermediatemoving 3-substituted isomers (8) was so low that theywere not further considered. The free bases were con-verted into their corresponding hydrochlorides by usualmethods. All the final products were further purified bycrystallization from a mixture of diethyl ether and ethylalcohol. Synthesized compounds listed intable I werecharcterized by1H-NMR spectroscopy.1H-NMR differ-entiated clearly between 1-, 2- and 3- substituted

4-benzoyl-1,2,3-triazole derivatives. In fact, it should bepointed out that there is a difference in the chemical shiftvalues among the protons in the 5- position of thebenzoyltriazole ring in the series of 1-, 2- and the protonin 4-position of the benzoyltriazole ring of the3-substituted compounds. The triazole proton of the 1-isomer appears always as a singlet at lower field withrespect to the position of the same proton of analogues 2-substituted, whereas, that of the 3-isomer resonated as asinglet at higher field with respect to the corresponding 1-and 2-isomers. This evidence is in accordance with toliterature data [11, 12]. All compounds gave satisfactoryanalyses (C, H, N, Cl, F).

3. Biological activity

The compounds reported intable I (1–4) were testedfor in vitro affinity on serotonin 5-HT1A, 5-HT2A and5-HT2C receptors by radioligand binding assays. Themore active compounds on serotonin receptors have beenselected and evaluated for their affinity on dopaminergic(D-1 and D-2) and adrenergic (α1 andα2) receptors. Allthe compounds were used as hydrochloride salts and werewater-soluble. The following specific radioligands andtissue sources were used: (a) serotonin 5-HT1A receptors,[3H]-8-OH-DPAT, rat brain cortex membranes; (b) sero-tonin 5-HT2A receptors, [3H] ketanserin, rat brain cortexmembranes; (c) serotonin 5-HT2C receptors, [3H] me-sulergine, rat brain cortex membranes; (d) dopamine D-1receptors [3H]SCH-23390, rat strial membranes; (e)

Figure 1. Reagents: (a) CrO3/H2SO4; (b) NaN3; (c) N-arylpiperazine/K2CO3; (d) HCl (anhydrous).

721

dopamine D-2 receptors [3H]spiroperidol, rat strial mem-branes; (f)α1 adrenergic receptors [3H]prazosin, rat braincortex membranes; (g)α2 adrenergic receptors [3H]yo-himbine, rat brain cortex membranes. Concentrations

required to inhibit 50% of radioligand specific binding(IC50) were determined through 2–4 independent experi-ments with samples in triplicate using 7–9 differentconcentrations of the title compounds. Specific binding,

Table II. Binding affinities and selectivities

IC50 nM (± SEM) Selectivity vs. 5-HT1A receptorCompound X n IC50 ratio

5-HT1A 5-HT2A 5-HT2C 5-HT2A 5-HT2C

[3H] 8-OH-DPAT [3H] ketanserin [3H] mesulergine

1-substituted1a H 2 3 200± 300 24 000± 1 900 2 900± 350 7.5 0.91b o-Cl 2 24± 2 > 105 37 ± 6 > 4 103 1.51c m-Cl 2 240± 20 30 000± 2 850 960± 86 125 4.01d p-Cl 2 180± 16 9 600± 810 480± 34 53 2.71e o-OCH3 2 480± 35 46 000± 4 400 180± 20 96 0.41f p-OCH3 2 > 105 > 105 > 105 1 11g o-F 2 10 000± 1 200 1 100± 170 610± 55 0.1 0.061h p-F 2 > 105 89 ± 6 > 105 > 9 10-4 12-substituted2a H 2 73± 5.8 > 105 1 100± 170 > 103 152b o-Cl 2 36± 3 > 105 12 000± 1 100 > 3 103 3332c m-Cl 2 2 ± 0.2 > 105 1 900± 290 > 104 9502d p-Cl 2 16 000± 1 400 > 105 250± 23 > 6.2 0.022e o-OCH3 2 62± 7 > 105 3 000± 280 > 103 48.42f p-OCH3 2 > 105 > 105 > 105 1 12g o-F 2 4 800± 390 240± 20 8 800± 780 0.08 1.82h p-F 2 13 000± 1 200 1 000± 120 580± 57 0.05 0.041-substituted3a H 3 20± 1.8 1 300± 140 17± 2.0 65 0.83b o-Cl 3 660± 60 2 000± 360 120± 19 3.0 0.23c m-Cl 3 7.2± 0.8 770± 67 230± 30 107 323d p-Cl 3 200± 15 3 000± 250 310± 30 15 1.53e o-OCH3 3 2.6± 0.2 8 100± 700 980± 72 3115 3773f p-OCH3 3 > 105 > 105 > 105 1 13g o-F 3 5 200± 480 7 800± 670 > 105 1.5 > 193h p-F 3 11 000± 1 000 6 800± 620 29 000± 2 750 0.6 32-substituted4a H 3 91± 6.8 22 000± 2 100 9 500± 880 242 1044b o-Cl 3 180± 20 34 000± 3 100 930± 72 189 5.24c m-Cl 3 710± 65 8 200± 740 17± 2 11.5 0.024d p-Cl 3 10 000± 1 150 12 000± 1 100 1 100± 120 1.2 0.14e o-OCH3 3 93± 6.8 > 105 2 200± 150 > 103 244f p-OCH3 3 > 105 > 105 > 105 1 14g o-F 3 95± 7.2 11± 1.2 460± 40 0.1 54h p-F 3 1 500± 160 800± 74 4 400± 380 0.5 38-OH-DPAT 2.1± 0.1 – – – –ketanserin – 1.7± 0.3 – – –cinanserin – 1.6± 1.0 – – –mesulergine – – 1.2± 0.2 – –

722

defined as described in the experimental section, repre-sented more than 75% of total binding in all three assays.The obtained IC50 values are listed intables II and III .

4. Results and discussion

In this work we examined affinity and selectivity of 4series of 4-benzoyl-1,2,3-triazole derivatives (1–4) to-wards 5-HT1A, 5-HT2A and 5-HT2C receptors by apreliminary binding screen. As mentioned, these com-pounds can be regarded as open chain analogues ofpreviously reported benzotriazole derivativesI [8].

The IC50 nM values listed intable II reveal that severalcompounds exhibit nanomolar affinity for the 5-HT1A

receptor, whereas generally only a few bind to 5-HT2A

and 5-HT2C receptors with comparable affinity.As far as 5-HT1A receptor affinity is concerned, the

results reported intable II indicate that the structuresendowed with highest 5-HT1A affinity are 1b, 2a–c, 2e,3a, 3cand3e. With respect to the length of the alkyl chainbetween the arylpiperazine moiety and the terminal4-benzoyl-1,2,3-triazole nucleus, the spacer effect is dif-ferent for 1- or 2-substituted compounds. For the ana-logues having an ethyl [(CH2)2] connecting group be-tween the triazole and piperazine rings, several2-substituted compounds have lower IC50 values than thecorresponding 1-substituted derivatives. For example,compare2a vs. 1a (IC50 nM = 73 and 3 200, respec-tively); 2cvs.1c (IC50 nM = 2 and 240, respectively) and2evs.1e (IC50 nM = 62 and 480, respectively). The onlypartial exceptions are compounds2b and 2d whoseaffinities are in fact lower than those of1b and 1drespectively.

By contrast, for the analogues having a propyl con-necting group [(CH2)3] between the triazole and thepiperazine rings, several 1-substituted compounds havelower IC50 values than the corresponding 2-substituted

derivatives. For example, compare3a vs. 4a (IC50 nM =20 and 91, respectively);3c vs. 4c (IC50 nM = 7.2 and710, respectively);3e vs. 4e (IC50 nM = 2.6 and 93,respectively). Onlyortho chloro (3b) andortho andparafluoro substituted compounds (3g and3h) exhibited IC50

values higher than 2-substituted analogues. It seems thatin these series the alkyl chain is not interacting directlywith the receptor on the basis of its hydrophobicity but ismore likely acting as a spacer.

The nature and the position of the substituent on thephenyl ring also has a significant influence on affinity. Forall series,p-substitution (by Cl, F, or OCH3) of thepiperazinyl-aromatic ring led to compounds with supranM affinity for the 5-HT1A receptor (but note, compound1a with no substituent did not bind with high affinity).For compounds of structure2 and 3, m-Cl substituentsafforded analogues with high 5-HT1A receptor affinity,whereas, in the benzotriazole series it had little effect onaffinity for the same receptor [8]. Ano-OCH3 substituentalso appeared to support 5-HT1A receptor binding forthese compounds.

Lòpez-Rodrìguez et al. [13] reported that theo-methoxy group on the phenyl ring in varioushydantoin-phenylpiperazine homologues exerted effectson 5-HT1A affinity strongly related to the overall ligandstructure. Taken together, these findings suggest that thereare different possible orientations within the 5-HT1A

binding site accessible to the X substituent on the phenylring. Such orientations are dictated by the global confor-mational arrangement of the bound ligand. As a result,contribution of X to affinity depends on the properties ofX as well as on the nature of the receptor subsite intowhich such a substituent is docked.

Among compounds evaluated intable II, 2c (m-Cl) and3e (o-OCH3) had the highest binding affinity and showthe most favourable affinity-selectivity profiles for the5-HT1A receptor (5-HT2A/5-HT1A IC50 ratios are about

Table III. Binding affinities and selectivities for D-1, D-2,α1 andα2 receptors in compounds1h, 2c, 3c, 3e and4c.

IC50 nM (± SEM)Compound X n

D-1 D-2 α1 α2

[3H] SCH-23390 [3H] spiroperidol [3H] prazosin [3H] yohimbine

1h p-F 2 > 105 > 105 7 200± 610 24 000± 1 8002c m-Cl 2 > 105 > 105 > 105 24 000± 1 7003c m-Cl 3 11 000± 1 100 10 000± 1 500 > 105 220± 203e o-OCH3 3 7 300± 620 1 500± 210 1 700± 250 43± 74c m-Cl 3 1 900± 15 1 800± 270 120± 15 260± 28spiroperidol 3.6± 0.5 4.6± 0.7prazosin 1.3± 0.7yohimbine 22 ± 2

723

10 000 and 3 000, respectively, while 5-HT2C/5-HT1A

IC50 ratios are about 1 000 and 300, respectively). Whilefluorine substitution generally led to compounds with lowbinding affinity to all 5-HT receptors, paradoxically,compound4g (o-fluoro) had IC50 values of 95 nM and 11nM in the 5-HT1A and 5-HT2A binding assays respec-tively.

The other two highly potent ligands of the 5-HT1A

receptor,1b and 3c, are moderately selective as theyretain appreciable affinities for 5-HT2C receptors. Acomparison between affinity profiles of benzoyltriazolederivatives2c and 3e versus those of the above men-tioned benzotriazole counterparts [8] (IC50 ratios alwaysless than 50) clearly show that the 5-HT1A/5-HT2A and5-HT1A/5-HT2C selectivity ratios of the former com-pounds are significantly superior. This result might bepartly ascribed to poor steric and/or electronic comple-mentarity between the benzoyltriazole residue and the5-HT2A and 5-HT2C binding sites compared to thebenzotriazole moiety in compoundsI .

The results of this study indicate that the affinity for the5-HT2A receptor is usually lower than the affinity for5-HT1A receptors. The introduction of the electron-withdrawing substituents in theortho andpara position,such as the fluoro group, increased the binding affinity forthe 5-HT2A receptor in all series except3. Only 4gexhibits nanomolar affinity (IC50 = 11 nM), but thiscompound binds also to the 5-HT1A receptors withappreciable potency. A remarkable selectivity (1 000-fold) for 5-HT2A versus 5-HT1A and 5-HT2C receptorswas achieved with1h, although the affinity of thiscompound for the 5-HT2A receptor fell within the sub-micromolar range (IC50 = 89 nM).

As far as the 5-HT2C receptor is concerned, nanomolarpotency was observed for compounds1b, 3a and 4c(IC50s< 50 nM). Binding of the former two ligands to the5-HT1A receptor was, however, slightly stronger withrespect to the 5-HT2C receptor. The compound4cwas theonly one showing appreciable selectivity towards the5-HT2C receptor (5-HT1A/5-HT2C and 5-HT2A/5-HT2C

IC50 ratios were about 40 and 500, respectively).As mentioned in the introduction section, in order to

evaluate potential therapeutic uses, it is crucial to checkwhether high potency is flanked by undesirable affinityfor other receptors, e.g., adrenergic receptors [14, 15].Thus the most active compounds on 5-HT1A (2c, 3c, and3e),5-HT2A (1h) and 5-HT2C (4c) were further evaluatedfor their affinity at dopaminergic and adrenergic recep-tors. Results are summarized intable III. As far as thedopaminergic system was concerned, the D-1 and D-2receptor affinity consistently showed IC50 values above10–6 M. As regards the affinity forα1 adrenergic recep-

tors, the IC50 values were high for all selected com-pounds, with the exception of4c (IC50 = 120 nM). Theaffinity for α2 receptors were in some cases quite consid-erable (3c, 3e and4c). In particular,o-OCH3-Ph deriva-tive 3eshowed affinity for theα2 receptor which was onlyone order of magnitude lower than the affinity toward the5-HT1A receptor.

However, compound2c, the most potent 5-HT1A

ligand showed the best selectivity profile versus allconsidered receptors.

In conclusion, some of the newly investigated benzoyl-triazole derivatives were endowed with high affinity andselectivity for the 5-HT1A receptor. Particularly, the IC50

values of2c and3e measured on this receptor are 2 and2.6 nM, and in terms of selectivity, these two ligandsrepresent a substantial improvement over previouslydescribed benzotriazole analogues [8].

These observations suggest that both the arylpiperazinepharmacophore and the terminal benzoyltriazole systemcontribute to the 5-HT1A interaction of the compounds.

Further synthesis and biological in vivo evaluation ofderivatives with new substituents are currently inprogress, and the results will be reported in due course.

5. Experimental protocols

5.1. Chemistry

Melting points were determined using a Kofler hot-stage apparatus and are uncorrected. Kieselgel 60 wasused for column chromatography and kieselgel 60 F254

plates from Merck were used for TLC. Where analysesare indicated only by the symbols of the elements, resultsobtained are within± 0.4% of the theoretical values. Thepurity of compounds were carefully assessed using ana-lytical TLC and the structure verified spectroscopicallyby proton NMR spectra recorded on a Bruker AMX-500instrument. Chemical shifts in ppm are referenced to theDMSO signal at 2.50 ppm. Elemental analyses (C, H, N,Cl, F) were determined within 0.4% of the theoreticalvalues.

The following compounds were synthesized by pub-lished procedures: aromatically substituted 1-(2-chloro-ethyl)-4-phenylpiperazine and aromatically substituted1-(3-chloropropyl)-4-phenylpiperazine derivatives [8].

5.1.1. Phenyl ethynyl ketone(6)A solution of chromium trioxide (0.10 mol) in water

(30 mL) and concentrated sulphuric acid (8.5 mL) wasslowly added to a stirred and cooled solution of propyl-ethynylcarbinol5 (0.15 mol) in acetone (50 mL). Thereaction mixture was carried out at 4 °C in an atmosphere

724

of nitrogen. After stirring for 7 h, water was added todissolve the precipitated chromium salts and the productwas extracted with chloroform. Evaporation of the or-ganic solution gave a yellow solid which was recrystal-lized from aqueous methanol to give 16.6 g (85%) of6 aspale yellow needles. The physical data are in agreementwith those given in reference [9].

5.1.2. 4-benzoyl-1H-1,2,3-triazole(7)To a stirred and heated solution of NaN3 (0.10 mol) in

anhydrous dimethylacetamide (80 mL) was slowly addedphenyl ethynyl ketone6 (0.10 mol) dissolved in dimethyl-acetamide, anhydrous (80 mL). The reaction mixture waskept at 100 °C for 2 h. After stirring for a further 12 h atroom temperature, evaporation of the solvent underreduced pressure gave a liquid residue which was dilutedwith water. The aqueous layer was acidified to pH = 5with 10% HCl and extracted with ether (3× 200 mL). Thecombined organic layers were dried over anhydrousNa2SO4. After evaporation of the solvent, the solidresidue was purified by crystallization from ethyl alcoholto give 16.4 g (95%) of7. The physical data are inagreement with those given in reference [10].

5.1.3. General procedure for the preparation of 1- and2-[2-[4-(X-Phenyl)piperazinyl]ethyl](4-benzoyl)-1,2,3-tri-azole(1a–hand2a–h)

To a stirred solution of 4-benzoyl-1,2,3-triazole(0.1 mol) and appropriate aromatically substituted 1-(2-chloroethyl)-4-phenylpiperazines (0.15 mol) in 80 mL ofbutan-2-one was added anhydrous K2CO3 (0.3 mol). Themixture was refluxed for 20 h and the solvent wasremoved under reduced pressure. The residue was dilutedwith H2O and extracted with CHCl3 (3 × 200 mL). Thecombined organic layers were washed with water anddried over Na2SO4. Evaporation of the solvent in vacuoprovided the crude residue consisting of the expectedtriazole isomers. The mixture of isomers was separatedand purified using a silica-gel column chromatography(diethyl ether/n-hexane 95:5 as eluent) to give 1-isomers1 and 2-isomers2 in the relative yields provided intable I. Spectral data of the title compounds refer to thehydrochloride salts.

5.1.3.1. 1-[2-[4-(Phenyl)piperazinyl]ethyl](4-benzoyl)-1,2,3-triazole⋅HCL (1a)

M.p. 201–202 °C;1H NMR (DMSO-d6) δ 3.34 (m,4H, 2CH2-pip), 3.71 (t,J = 7.5 Hz, 2H, CH2N-pip), 3.94(m, 4H, CH2-pip), 5.22 (t,J = 7.5 Hz, 2H, CH2N-triaz),6.97–8.35 (mm, 10H, ArH), 9.17 (s, 1H, H-triaz).

5.1.3.2. 2-[2-[4-(Phenyl)piperazinyl]ethyl](4-benzoyl)-1,2,3-triazole⋅HCL (2a)

M.p. 206–207 °C;1H NMR (DMSO-d6) δ 3.34 (m,4H, 2CH2-pip), 3.71 (t,J = 7.5 Hz, 2H, CH2N-pip), 3.96(m, 4H, 2-CH2-pip), 5.32 (t,J = 7.5 Hz, 2H, CH2N-triaz),6.97–8.30 (mm, 10H, ArH), 8.60 (s, 1H, H-triaz).

5.1.4. General procedure for the preparation of 1- and2-[3-[4-(X-Phenyl)piperazinyl]propyl](4-benzoyl)-1,2,3-triazole (3a–hand4a–h)

These compounds were prepared in a manner similar tothat mentioned above, starting from 4-benzoyl-1,2,3-triazole (0.1 mol) and aromatically substituted 1-(3-chloropropyl)-4-phenyl-piperazines (0.1 mol). The crudemixture of isomers was chromatographed on a silica-gelcolumn (diethyl ether/n-hexane 95:5 as eluent) to yieldpure 3 and 4 in the relative yields shown intable I.Spectral data of title compounds refer to the hydrochlo-ride salts.

5.1.4.1. 1-[3-[4-(Phenyl)piperazinyl]propyl](4-benzoyl)-1,2,3-triazole⋅HCL (3a)

M.p. 183–184 °C;1H NMR (DMSO-d6) δ 2.58 (m,2H, CH2CH2CH2), 3.27 (m, 4H, 2CH2-pip), 3.32 (t,J =7.5 Hz, 2H, CH2N-pip), 3.67 (m, 4H, 2CH2-pip), 4.76 (t,J = 7.5 Hz, 2H, CH2N-triaz), 6.95–8.37 (mm, 10H, ArH),9.31 (s, 1H, H-triaz).

5.1.4.2. 2-[3-[4-(Phenyl)piperazinyl]propyl](4-benzoyl)-1,2,3-triazole⋅HCL (4a)

M.p. 251–252 °C;1H NMR (DMSO-d6) δ 2.60 (m,2H, CH2CH2CH2), 3.26 (m, 4H, 2CH2-pip), 3.45 (t,J =7.5 Hz, 2H, CH2N-pip), 3.67–3.88 (m, 4H, 2CH2-pip),4.81 (t, J = 7.5 Hz, 2H, CH2N-triaz), 6.84–8.28 (mm,10H, ArH), 8.56 (s, 1H, H-triaz).

5.1.5. Hydrochloride salts: general procedureThe hydrochloride salts were prepared by adding an

HCl ethereal solution to an ethanolic solution of freebases. Recrystallization solvent, formulae and meltingpoints are reported intable I. They were obtained aswhite crystals.

5.2. Biological methods

5.2.1. 5-HT1A binding assayRadioligand binding assays were performed following

a published procedure [16]. Cerebral cortex from maleSprague-Dawley rats (180–220 g) was homogenized in20 volumes of ice-cold Tris-HCl buffer (50 mM, pH 7.7at 22 °C) with a Brinkmann Polytron (setting 5 for 15 s),and the homogenate was centrifuged at 50 000g for10 min. The resulting pellet was then resuspended in the

725

same buffer, incubated for 10 min at 37 °C, and centri-fuged at 50 000g for 10 min. The final pellet wasresuspended in 80 volumes of the Tris-HCl buffer con-taining 10µM pargyline, 4 mM CaCl2, and 0.1% ascor-bate. To each assay tube was added the following: 0.1 mLof the drug dilution (0.1 mL of distilled water if nocompeting drug was added), 0.1 mL of [3H]-8-hydroxy-2-(di-n-propylamino)tetralin ([3H]-8-OH-DPAT) in buffer(containing Tris, CaCl2, pargyline, and ascorbate) toachieve a final assay concentration of 0.1 nM, and 0.8 mLof resuspended membranes. The tubes were incubated for30 min at 37 °C, and the incubations were terminated byvacuum filtration through Whatman GF/B filters. Thefilters were washed twice with 5 mL of ice-cold Tris-HClbuffer, and the radioactivity bound to the filters wasmeasured by liquid scintillation spectrometry. Specific[3H]-8-OH-DPAT binding was defined as the differencebetween binding in the absence and presence of 5-HT(10 µM).

5.2.2. 5-HT2A and 5-HT2C binding assaysRadioligand binding assays were performed as previ-

ously reported by Herndon et al. [17]. Briefly, frontalcortical regions of male Sprague-Dawley rats(200–250 g, Charles River) were dissected on ice andhomogenized (1:10 w/v) in ice-cold buffer solution(50 mM Tris HCl, 0.5 mM EDTA, and 10 mM MgCl2 atpH 7.4) and centrifuged at 3 000g for 15 min. The pelletwas resuspended in buffer (1:30 w/v), incubated at 37 °Cfor 15 min and then centrifuged twice more at 3 000g for10 min (with resuspension between centrifugations). Thefinal pellet was resuspended in buffer that also contained0.1% ascorbate and 10–5 M pargyline.

Assays were performed in triplicate in a 2.0 mLvolume containing 5 mg wet weight of tissue and 0.4 nM[3H] ketanserin (hyphen) (76 Ci/mmol; New EnglandNuclear) for 5-HT2A receptor assays, and 10 mg wetweight of tissue and 1 nM [3H]mesulergine (75.8Ci/mmol; Amersham) for 5-HT2C receptor assays. Cin-anserin (1.0µM) was used to define nonspecific bindingin the 5-HT2A assay. In the 5-HT2C assays, mianserin (1.0µM) was used to define nonspecific binding, and 100 nMspiperone was added to all tubes to block binding to5-HT2A receptors. Tubes were incubated for 15 min at37 °C, filtered on Schliecher and Schuell (Keene, NH)glass fibre filters presoaked in poly(ethylene imine), andwashed with 10 mL of ice-cold buffer. Filters werecounted at an efficiency of 50%.

5.2.3. D-1 dopaminergic binding assayThe binding assay for D-1 dopaminergic receptors was

that described by Billard et al. [18]. Corpora striata were

homogenized in 30 vol. (w/v) ice cold 50 mM Tris-HClbuffer (pH 7.7 at 25 °C) using a Polytron PT10 (setting 5for 20 s). Homogenates were centrifuged twice for10 min at 50 000g with resuspension of the pellet in freshbuffer. The final pellet was resuspended in 50 mM icecold Tris-HCl containing 120 mM NaCl, 5 mM KCl,2 mM CaCl2, 1 mM MgCl2, 0.1% ascorbic acid and 10µM pargyline (pH 7.1 at 37 °C). Each assay tube con-tained 50µL [3H]SCH-23390 to achieve a final concen-tration of 0.4 nM, and 900µL resuspended membranes(3 mg fresh tissue). The tubes were incubated for 15 minat 37 °C and the incubation was terminated by rapidfiltration under vacuum through Whatman GF/B glassfibre filters. The filters were washed three times with5 mL ice-cold 50 mM Tris-HCl buffer (pH 7.7 at 25 °C).The radioactivity bound to the filters was measured by aliquid scintillation counter. Specific [3H]SCH-23390binding was defined as the difference between binding inthe absence or in the presence of 0.1µM piflutixol.

5.2.4. D-2 dopaminergic binding assayThe procedure used in the radioligand binding assay

was reported in detail by Creese et al. [19]. Corporastriata were homogenized in 30 vol. (w/v) ice cold50 mM Tris-HCl buffer (pH 7.7 at 25 °C) using a Poly-tron PT10 (setting 5 for 20 s). Homogenates were centri-fuged twice for 10 min at 50 000g with resuspension ofthe pellet in fresh buffer. The final pellet was resuspendedin 50 mM ice cold Tris-HCl containing 120 mM NaCl,5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 0.1% ascorbicacid and 10µM pargyline (pH 7.1 at 37 °C). Each assaytube contained 50µL [3H]spiroperidol to achieve a finalconcentration of 0.4 nM, and 900µL resuspended mem-branes (3 mg fresh tissue). The tubes were incubated for15 min at 37 °C and the incubation was terminated byrapid filtration under vacuum through Whatman GF/Bglass fibre filters. The filters were washed three timeswith 5 mL ice-cold 50 mM Tris-HCl buffer (pH 7.7 at25 °C). The radioactivity bound to the filters was mea-sured by a liquid scintillation counter. Specific [3H]spiro-peridol binding was defined as the difference betweenbinding in the absence or in the presence of 1µM(+)-butaclamol.

5.2.5.α1 adrenergic binding assayThe procedure used in the radioligand binding assay

has been reported in detail by Greengrass and Brem-ner [20]. Brain cortex was homogenized in 30 vol. (w/v)ice-cold 50 mM Tris-HCl buffer, (pH 7.2 at 25 °C) usinga Polytron PT10 (setting 5 for 20 s). Homogenates werecentrifuged twice for 10 min at 50 000g with resuspen-sion of the pellet in fresh buffer. The final pellet was

726

resuspended in 50mM ice-cold Tris-HCl, (pH 7.4 at25 °C). Each assay tube contained 50µL drug solution,50 µL [3H]prazosin to achieve a final concentration of 0.4nM, and 900µL resuspended membranes (10 mg freshtissue). The tubes were incubated for 30 min at 25 °C andthe incubation was terminated by rapid filtration undervacuum through Whatman GF/B glass fibre filters. Thefilters were washed three times with 5 mL ice-cold50 mM Tris-HCl, buffer (pH 7.2 at 25 °C). The radioac-tivity bound to the filters was measured by a liquidscintillation counter. Specific [3H]prazosin binding wasdefined as the difference between binding in the absenceor in the presence of 10µM phentolamine.

5.2.6.α2 adrenergic binding assayThe procedure used in the radioligand binding assay

was reported in detail by Perry and U’Prichard [21].Brain cortex was homogenized in 30 vol. (w/v) ice-cold5mM tris-HCl, 5mM EDTA buffer (pH 7.3 at 25 °C)using a polytron PT10 (setting 5 for 20 s). Homogenateswere centrifuged three times for 10 min at 50 000g withresuspension of the pellet in fresh buffer. The final pelletwas resuspended in 50 mM ice-cold Tris-HCl, 0.5 mMEDTA (pH 7.5 at 25 °C). Each assay tube contained50 µL drug solution, 50µL [3H]yohimbine to achieve afinal concentration of 1 nM, and 900µL resuspendedmembranes (10 mg fresh tissue). The tubes were incu-bated for 30 min at 25 °C and the incubation was termi-nated by rapid filtration under vacuum through WhatmanGF/B glass fibre filters. The filters were washed threetimes with 5 mL ice-cold 50 mM Tris-HCl, 0.5 mMEDTA buffer (pH 7.5 at 25 °C). The radioactivity boundto the filters was measured by a liquid scintillationcounter. Specific [3H]yohimbine binding was defined asthe difference between binding in the absence or in thepresence of 10µM phentolamine.

Acknowledgements

This work was supported by a grant from RegioneCampania ai sensi della L.R. 31 December 1994. TheNMR spectral data were provided by Centro di Ricerca

Interdipartimentale di Analisi Strumentale, Universitàdegli Studi di Napoli “Federico II”. The assistance of thestaff is gratefully acknowledged.

References

[1] Baumgarten H.G., Gothert M. (Eds.), Serotoninergic Neuron and5-HT Receptors in the CNS, Springer-Verlag, Berlin, 1997.

[2] van Wijngaarden I., Soudin W., Serotonin Receptors and TheirLigands, Elsevier, Amsterdam, 1997.

[3] Hamon M., Gozlan H., El Mestikawy S., Emerit M.B., Boslanos F.,Schecther L., Am. N.Y. Acad. Sci. 600 (1990) 114–131.

[4] Eison A.S., Eison M.S., Prog. Neuropsychopharmacol. Biol. Psy-chiatry 18 (1994) 47–62.

[5] Eison M.S., Psychopathology 1 (1989) 13–20.

[6] Glennon R.A., Drug Dev. Res. 26 (1992) 251–274.

[7] van Steen B.J., van Wijngaarden I., Tulp M.T.M., Soudijn W., J.Med. Chem. 36 (1993) 2751–2760.

[8] Caliendo G., Greco G., Grieco P., Novellino E., Perissutti E.,Santagada V., Barbarulo D., De Blasi A., Eur. J. Med. Chem. 31(1996) 207–213.

[9] Kenneth B., Heilbron I.M., Jones E.R.H., Weedon B.C.L., J. Chem.Soc. Part I (1946) 39–45.

[10] Nesmeyanov A.N., Rybinskaya M.I., Dokl. Akad. Nauk. SSSR. 158(1964) 408–410.

[11] Livi O., Biagi G., Ferretti M., Lucacchini A., Barili P.L., Eur. J. Med.Chem. -Chim. Ter. 18 (1983) 471–475.

[12] Biagi G., Ferretti M., Livi O., Scartoni V., Lucacchini A., MazzoniM., Il Farmaco Ed. Sc. 41 (1986) 388–400.

[13] Lòpez-Rodrìguez M.L., Rosado M.L., Benhamù B., Morcillo M.J.,Feràndez E., Schaper K.J., J. Med. Chem. 40 (1997) 1648–1656.

[14] Lòpez-Rodrìguez M.L., Rosado M.L., Benhamù B., Morcillo M.J.,Sanz A.M., Orensanz L., Beneitez M.E., Fuentes J.A., ManzanaresJ., J. Med. Chem. 39 (1996) 4439–4450.

[15] Perrone R., Berardi F., Leopoldo M., Tortorella V., J. Med. Chem. 39(1996) 3195–3202.

[16] Schlegel J.R., Peroutka S.J., Biochem. Pharmacol. 35 (1986)1943–1949.

[17] Herndon J.L., Ismaiel A., Ingher S.P., Teitler M., Glennon R.A., J.Med. Chem. 35 (1992) 4903–4910.

[18] Billard W., Ruperto V., Grosby G., Iorio L.C., Barnett A., Life Sci.35 (1985) 1885–1893.

[19] Creese I., Schneider R., Snyder S.H., Eur. J. Pharmacol. 46 (1977)377–381.

[20] Greengrass P., Brenner R., Eur. J. Pharmacol. 55 (1979) 323–326.

[21] Perry B.D., U’Prichard D.C., Eur. J. Pharmacol. 76 (1981) 461–464.

727

Original article

Synthesis and QSAR of substituted 3-hydroxyanthranilic acid derivativesas inhibitors of 3-hydroxyanthranilic acid dioxygenase (3-HAO)

Mats Linderberga, Sven Hellberga, Susanna Björka, Birgitta Gotthammara, Thomas Högberga,Kerstin Perssona, Robert Schwarczc, Johan Luthmanb, Rolf Johanssona*

aMedicinal Chemistry, Preclinical R& D, Astra Arcus, S-15185 Södertälje, SwedenbCell Biology, Preclinical R& D, Astra Arcus, S-15185 Södertälje, Sweden

cMaryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, MD 21228, USA

(Received 21 January 1999; revised 12 March 1999; accepted 18 March 1999)

Abstract – Novel 4,5-, 4,6-disubstituted and 4,5,6-trisubstituted 3-hydroxyanthranilic acid derivatives were synthesized and their ability toreduce the production of the excitotoxin quinolinic acid (QUIN) by inhibition of brain 3-hydroxyanthranilic acid dioxygenase (3-HAO) wassubsequently investigated. The potency of the compounds to inhibit 3-HAO was assayed in rat brain homogenate, while chemical stability ofcertain compounds was studied by HPLC. The data were used to generate quantitative structure-activity relationship (QSAR) models forpotency of 3-HAO inhibition and compound stability. Compounds with longer half-lives were obtained when the difference between theHOMO and LUMO was increased, while electron withdrawing groups in the 4- and 5-positions increased the potency of 3-HAO inhibition.Selected compounds that showed high potency in vitro were also found to be efficacious inhibitors in vivo after cerebral administration in rats.© 1999 Editions scientifiques et médicales Elevier SAS

3-HAO / QSAR / hydroxyanthranilic acid / enzyme inhibitors / neuroprotection

1. Introduction

The cytosolic, non-haem ferrous (Fe2+) enzyme3-hydroxyanthranilic acid dioxygenase (3-HAO; EC1.13.11.6) plays an important role in the metabolictransformation of L-tryptophan to nicotinamide in thekynurenine pathway (figure 1). The enzyme oxidizes andring opens 3-hydroxyanthranilic acid (3-HANA) to pro-duce α-amino-â-carboxymuconic acidω-semialdehyde,which subsequently, spontaneously cyclizes and formsquinolinic acid (QUIN) [1]. 3-HAO seems mainly to belocalized to hepatic tissue, but it has also been demon-strated that the enzyme is expressed both in the brain [2,3] and in inflammatory cells [2, 4] where the productionof QUIN has been shown to be stimulated by certaincytokines [5–9]. Biochemical and immunological analy-sis in the rat suggest that the brain and liver 3-HAO areidentical proteins [10]. Moreover, there appears to be a

high degree of homology (94% similarity) between therat and the human 3-HAO amino acid sequence [11].

QUIN is an NMDA receptor agonist and an excito-toxin [12] that has been reported to cause convul-sions [13] or neurodegeneration [14] after intracerebraladministration. It has been shown that intrastriatal injec-tions of QUIN in rats induce biochemical and morpho-logical alterations similar to those observed in Hunting-ton’s disease [15, 16]. Furthermore, increases in neuronalHuntingtin immunoreactivity [17] have recently beenreported to occur after administration of QUIN into thestriatum of mice, providing further support for a role ofQUIN in the pathogenesis of Huntington’s disease. It isalso possible that QUIN is involved in epilepsy. Forexample, the levels of QUIN are higher in epilepsy-pronemice in a transgenic model of epilepsy than in thecorresponding wild type mice [18]. Moreover, increasedQUIN levels, or enhanced kynurenine pathway activity,have been implicated in inflammatory diseases of thecentral nervous system. For instance, QUIN has been*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 729−744 729© 1999 Editions scientifiques et médicales Elevier SAS. All rights reserved

associated with neurological dysfunction occurring fol-lowing various viral or bacterial infections [19–23].

To elucidate the possible role of 3-HAO-mediatedformation of QUIN in different neurological conditions,inhibitors of the enzyme are valuable tools, and may alsofind therapeutic utility. In this report, the synthesis ofseveral analogues of 3-HANA and their structure-activityrelationship as inhibitors of 3-HAO are presented. Themain objective was to develop inhibitors with improvedchemical and pharmacological properties compared topreviously described 4-halogenated 3-HANA ana-logues [24]. A particular problem associated with3-HANA analogues is the propensity of 3-HANA forauto-oxidation leading to its degradation and the forma-tion of oxidized cinnabarinic acid [25, 26]. Hence, thestability of some of the synthesized compounds wasstudied by HPLC. The potency of the compounds toinhibit 3-HAO was assessed in rat brain tissue homoge-nates, while the ability of selected compounds to inhibit3-HAO in vivo was also studied following intracerebraladministration.

2. Chemistry

A series of 4,5-, 4,6-disubstituted and 4,5,6-trisubstituted 3-hydroxyanthranilic acids were prepared.The 4,5-dihalo substituted compounds were synthesizedfrom the known 4-chloro- and 4-bromo-3-hydroxy-anthranilic acids. The strong directing effect of the aminosubstituent gives access to derivatives11 and 12 afterreaction with bromine and chlorine, respectively. When

the 5-position is substituted, the 4-substituted halogenscan be obtained by a regio-selective reaction at the4-position. Accordingly, compounds8 and 10 were ob-tained by direct bromination of the 5-substituted-2-hydroxyanthranilic acids or esters (figures 2and 3). Toconfirm the identity of the structures, 5-bromo-3-hydroxyanthranilic acid was chlorinated and shown togive an identical product as formed by bromination of4-chloro-3-hydroxyanthranilic acid. The identity of the4,5-dichloro and 4,5-dibromo derivatives was confirmedby comparison with the corresponding 4,6-disubstitutedcompounds and by synthesis either from the 4- or5-halogenated 3-HANA leading to identical products(figure 2).

Compound 39 was synthesized by bromination of5-hydroxy-2-methoxybenzoic acid, which after nitrationand reduction of the nitro group gave the desired product

Figure 1. The biotransformation pathway for tryptophan yielding the excitotoxin QUIN.

Figure 2. i) Br2/HBr/HOAc, ii) Br2/HBr/HOAc orCl2/HCl/HOAc.

730

(figure 4). Compounds6, 34and44 (figures 5and6) wereprepared by nitration of suitably substituted mono ordihalogenated benzoic acid. The benzoic acids32 and42were obtained by a palladium catalysed carbonylation ofthe corresponding triflates. The directing effect of thephenol gave the desired regio-selectivity for compounds6 and 44. The more easily oxidized compounds17, 25and36 (figure 7and8) were obtained after halogenationof an intermediate without the strongly directing andactivating amino-function. To make the tri-substitutedcompounds21and29and the 4,6-disubstituted50and54the carboxylic acid was introduced via ring closure of thecorresponding isonitrosoacetanilid followed by oxidativering-opening of the isatin-derivative (figure 9).

Figure 7. i) Cl2/CHCl3, ii) BnBr/K2CO3, iii) Cu(I)Cl/KBH 4.

Figure 8. i) Br2/HOAc/NaOAc, ii) H2/PtS2.

Figure 9. i) CCl3CH(OH)2/NH2OH/DMF/H2O, ii) H2SO4/Dio-xane or polyphosphoric acid, iii) NaOH/H2O2, iv) H2/Pd orBBr3.

Figure 3. i) Br2/HOAc, ii) KOH/EtOH.

Figure 4. i) Br2/HOAc, ii) NaNO3/La(NO3)3/HCl(aq.)/diethylether, iii) H2/PtS2/EtOH.

Figure 5. i) Pd(OAc)2/CO/MeOH/dppp, ii) KOH/MeOH, iii)BBr3 or HBr, iv) HNO3/CH3NO2, v) Pd/H2.

Figure 6. i) EtSNa/DMF, ii) HCl/MeOH, iii) HNO3/CH3NO2/CH2Cl2, iv) KOH/EtOH, v) SnCl2xH2O/EtOH.

731

3. QSAR

A quantitative structure-activity relationship was de-veloped for the 16 compounds presented intable I. Theσof the R1, R2 and R3 substituents, theπ and MR for theR4, R5 and R6 substituents were obtained from theliterature [27]. The pKa values for the CO2H, NH2 andOH substituents and logP for the molecules were calcu-lated using the Pallas software [28]. Semi-empirical cal-culations using AM1 (Spartan) [29] provided values forHOMO and LUMO energies, e-neg, hardness, heat offormation, polarizability, surface area, molecular volumeand ovality [30] which were subsequently used as de-scriptors in the QSAR calculations.

4. Pharmacology

The compounds were tested for their ability to inhibit3-HAO in homogenates of rat brain tissue according tothe method of Foster et al. [18]. The production ofradioactive QUIN was measured after addition of[1-14C]3-HANA to determine inhibition of the enzyme.

Test compound concentrations resulting in a 50% inhibi-tion of the enzyme (IC50) are reported.

Also, the ability of selected compounds to inhibitcerebral 3-HAO in vivo was studied following intracere-broventricular (i.c.v.) administration of the compounds inrats. The de novo formation of QUIN in hippocampaltissue after concomitant i.c.v. administration of the testcompounds and QUIN was measured by GC/MS.

5. Results and discussion

A PLS analysis# (Codex, AP Scientific Service) ofpotency resulted in a four component model, accountingfor 97% of the variance in pIC50 (cross-validated Q2 =0.85) (figure 10). All the components in the model werestatistically significant according to cross-validation. Thetwo most important components accounted for 86% ofthe variance. A plot of the PLS-weights for these com-ponents indicated that the most important factors influ-

# Descriptors used in the PLS analysis are available on theInternet on request

Table I. Inhibition of 3-HAO and chemical stability of 3-HANA derivatives.

Compound R4 R5 R6 IC50a nM Stabilityb

55 [24, 32] Cl H H 6 3856 [24] Br H H 257 [24] F H H 246 Br OEt H 568 Br Br H 0.3 2110 Br Me H 2.311 Cl Br H 0.2612 Cl Cl H 0.336 Br H Br 5.8 9039 Br H MeO 120 044 Cl H Cl 10.1 8950 Cl H Ph 11 000 7921 Cl (CH2)2 (CH2)2 20029 Cl Me Me 8.2 0c

34 Cl Me Cl 4.4 6254 Cl H Me 7.8 53

aIn vitro inhibition of 3-HAO in rat cortex.bPercent remaining after 24 h in a PBS buffer at pH 7.5 and 37 °C.cPercent remaining after 24 hin a PBS buffer at pH 7.5 and 50 °C.

732

encing the IC50 values are the size of the R6 substituent,the lipophilicity of the R5 substituent and the pKa of thephenol. This implies that small, electron-withdrawinggroups at R6 and lipophilic, electron-withdrawing groupsat R5 would increase the potency of the compounds.

Nine of the compounds were used to develop a QSARmodel for chemical stability. The PLS analysis resulted ina two-component model, accounting for 94% of thevariance in stability. The single most important factor forthe model was the difference between the HOMO andLUMO energies (hardness) of the molecules (figure 11).The very labile compound39, having the stronglyelectron-donating 6-methoxy substituent could not beaccounted for in the model. This may indicate a differentmechanism for the degradation of this compound. Theincreased instability observed for compounds with asmall difference in HOMO and LUMO-orbitals is inagreement with the observation that dimerization hasbeen reported to be involved in the degradation of4-bromo-3-hydroxyanthranilic acid [25, 26].

Several compounds that were found to be potent invitro inhibitors of 3-HAO were also efficacious in vivoafter intracerebral administration. There was a generalagreement between the in vitro and in vivo ability of the3-HANA analogues to inhibit 3-HAO (tables I and II ).Hence, compounds which showed low nM IC50 valueswere able to almost totally inhibit de novo production ofQUIN in vivo when given at doses that were equivalent toor higher than those of 3-HANA, while21, which showeda high nM IC50 value in vitro, displayed no in vivoactivity. On the other hand, certain differences in potencyobserved in vitro were not reflected in vivo. This may, atleast partially, depend on the inherent differences inchemical stability of the compounds.

For example, the low stability of the very potentcompound8 may have reduced its in vivo actions, butdistribution and local metabolism in vivo could alsocontribute to the discrepancies seen.

Importantly, the finding that potency and stabilitydepend on separate structural features present possibili-ties to further improve the characteristics of 3-HANAanalogues as useful inhibitors of 3-HAO.

Figure 10. Predicted vs. observed IC50 values for 16 3-HAOinhibitors.

Figure 11. Plot of stability versus hardness.

Table II. In vivo effects of 3-HAO inhibitors on 3-HANA-inducedde novo production of QUIN after i.c.v. administration in rat.

Compound Ratio dosea % Inhibition3-HANA:compound

55 1:1 96.9± 0.18 1:1 87.4± 2.036 1:0.3 67.2± 6.6

1:1 81.9± 1.41:30 95.2± 1.4

44 1:30 91.9± 1.921 1:30 0± 1329 1:30 95.1± 1.234 1:0.3 69.6± 7.9

1:30 93.9± 0.6

a3-HANA was given in a dose of 10 nmol i.c.v., while the differentcompounds were given in doses of either 3, 10 or 300 nmol i.c.v.at the same time. The cerebral levels of QUIN at 2 h afteradministration were determined in hippocampal tissue by GC/MS.Data are the mean± SEM of 5–7 determinations.

733

6. Experimental protocols

6.1. Chemical methods

Melting points were determined on a Büchi SMP-20apparatus.1H and 13C NMR spectra were recorded atambient temperature on a Varian Unity 400 or VarianGemini 300 instrument. Chemical shifts are given in ppmfrom internal standards. For1H NMR and 13C NMRspectra the internal references were tetramethylsilane (0.0ppm), CDCl3 (δ 7.26 orδ 77.0 ppm), CD3OD (δ 3.38 orδ 49.3 ppm) or DMSO-d6 (δ 2.49 or δ 39.5 ppm),respectively. Coupling constants are given in Hertz, andthe splitting patterns are designated as follows: s, singlet;d, doublet; dd, doublet of doublets; t, triplet; q, quartetand br, broad. Mass spectra were obtained on an LKB2091 (ELI, 70 eV or CI/CH4) or on a Finnigan-MAT TSQ70 (thermospray) spectrometer. Elemental analyses wereperformed by MIKRO KEMI AB, Uppsala, Sweden. Thevalues were within± 0.4% of theoretical if not otherwiseindicated.

6.1.1. Ethyl 4-bromo-3,5-diethoxybenzoate14-Bromo-3,5-dihydroxybenzoic acid [31] (4.68 g,

20 mmol), potassium carbonate (8.28 g) and diethyl sul-fate (7.84 mL) in dry acetone (50 mL) under nitrogenatmosphere were stirred at room temperature for 20 minand then refluxed for 8 h. The salts were filtered off andthe solvent was evaporated, and the remainder wasdissolved in diethyl ether. The organic layer was washedwith water, sodium hydroxide (aq., 5%), ammoniumhydroxide (conc.), hydrochloric acid, water and finallydried (Na2SO4). The organic phase was then evaporatedto give the title compound (2.63 g, 51%).1H NMR(CDCl3) δ 7.22 (s, 2H), 4.39 (q,J = 7.1 Hz, 2H) 4.18 (q,J = 7.0 Hz, 4H), 1.5 (t,J = 7.0 Hz, 6H), 1.41 (t,J = 7.1Hz, 3H),13C NMR (CDCl3) δ 166.8, 157.0, 130.8, 107.9,106.8, 65.4, 61.6, 14.7, 14.4.

6.1.2. 4-Bromo-5-ethoxy-3-hydroxybenzoic acid2A solution of 1 (1.02 g, 3.2 mmol) inN,N-dimethyl-

formamide (5 mL) was added to sodium ethanethiolate(4.53 g, 51.2 mmol) inN,N-dimethylformamide (25 mL)under nitrogen atmosphere. The mixture was stirred at100 °C for 8 h. After cooling, the mixture was poured intowater. The aqueous phase was acidified by the addition ofhydrochloric acid (1 M) and extracted with diethyl ether.The organic phase was dried (Na2SO4) and evaporatedThe remainder was purified by flash chromatography(SiO2, toluene-acetic acid 5:1) to give the title compound(115 mg, 14%).1H NMR (DMSO-d6) δ 7.13 (d,J = 1.8Hz, 1H), 6.95 (d,J = 1.8 Hz, 1H) 4.08 (q,J = 7.0 Hz,2H), 3.4 (br, 2H), 1.30 (t,J = 7.0Hz, 3H), 13C NMR

(DMSO-d6) δ 167.5, 156.5, 155.8, 131.3, 109.7, 104.5,104.3, 64.6, 14.6, MS (TSP) m/z: 263/261 (M + 1).

6.1.3. Methyl 4-bromo-5-ethoxy-3-hydroxybenzoate3A solution of 2 (115 mg, 0.44 mmol) in dry methanol

(10 mL) was saturated with hydrogen chloride at 0 °Cand the mixture was stirred overnight at room tempera-ture. The solution was evaporated and the remainder wasdissolved in chloroform. The organic phase was washedwith sodium bicarbonate (sat.), water, brine and thendried (Na2SO4) and evaporated to give the title com-pound (94 mg, 79%).1H NMR (CDCl3) δ 7.35 (d,J = 1.8Hz, 1H), 7.14 (d,J = 1.8 Hz, 1H) 5.85 (br, 1H), 4.17 (q,J = 7.0 Hz, 2H), 3.91 (s, 3H), 1.49 (t,J = 7.0 Hz, 3H),13C NMR (CDCl3) δ 167.2, 156.5, 154.4, 130.8, 110.1,106.0, 105.5, 65.4, 52.7, 14.7, MS (TSP) m/z: 277/275(M + 1).

6.1.4. Methyl 4-bromo-5-ethoxy-3-hydroxy-2-nitrobenzo-ate4

A solution of3 (89.7 mg, 0.326 mmol) in nitromethane(1 mL) and dichloromethane (1 mL) was stirred withnitric acid (90%, 15µL) at 40 °C for 10 min. Ice wasadded, and the layers were separated. The aqueous layerwas extracted with chloroform, and the combined organiclayer was washed with water and evaporated. Purificationof the remainder by flash chromatography (SiO2, toluene-ethyl acetate 10:1) gave the title compound (37 mg,35%).1H NMR (CDCl3) δ 6.59 (s, 1H), 4.21 (q,J = 7.0Hz, 2H) 3.90 (s, 3H), 1.47 (q,J = 7.0 Hz, 3H),13C NMR(CDCl3) δ 167.1, 162.2, 154.4, 132.2, 126.9, 105.4,103.3, 66.6, 53.9, 14.6, MS (TSP) m/z: 339/337 (M +NH4).

6.1.5. 4-Bromo-5-ethoxy-3-hydroxy-2-nitrobenzoic acid5Methyl 4-bromo-5-ethoxy-3-hydroxy-2-nitrobenzoate

4 (30.2 mg, 0.094 mmol) was dissolved in ethanol(1.2 mL) and potassium hydroxide (62 mg) in water(0.54 mL) was added. The mixture was stirred at 40 °Cfor 16 h and then cooled and acidified to pH 2 by theaddition of hydrochloric acid. The remainder was tritu-rated with water, filtered and dried to give the titlecompound (24.6 mg, 86%).1H NMR (CD3OD) δ 6.82 (s,1H), 4.74 (br, 2H) 4.00 (q,J = 7.0 Hz, 2H), 1.26 (t,J =7.0 Hz, 3H),13C NMR (CD3OD) δ 167.3, 159.7, 151.0,135.3, 128.2, 106.9, 106.3, 67.4, 15.4, MS (EI, 70 eV)m/z: 307/305 (M +, 99/100).

6.1.6. 4-Bromo-5-ethoxy-3-hydroxyanthranilic acid6A mixture of5 (19.8 mg, 0.065 mmol) and SnCl2xH2O

(73 mg) in ethanol (2.0 mL) was heated at 70 °C for 5.5 hunder nitrogen atmosphere. After cooling, water (2.0 mL)was added and the pH was adjusted to 7 by adding

734

sodium bicarbonate (aq., 5%). The mixture was extractedwith ethyl acetate, and the organic phase was washedwith brine and dried (Na2SO4). The organic phase wasevaporated to give the title compound (8.4 mg, 47%).1HNMR (CD3OD) δ 6.93 (s, 1H), 4.80 (br, 4H) 3.89 (q,J =7.0 Hz, 2H), 1.30 (t,J = 7.0 Hz, 3H),13C NMR (CD3OD)δ 171.7, 147.7, 145.1, 137.9, 109.5, 108.7, 107.2, 67.0,15.7, MS (TSP) m/z: 278/276 (M + 1).

6.1.7. 5-Bromo-3-hydroxyanthranilic acid7To 3-hydroxyanthranilic acid (5 g, 32.6 mmol) in ace-

tic acid (450 mL) was bromine (10.4 g, 65.2 mmol) inacetic acid (50 mL) added drop-wise. The thick slurrywas stirred for 1 h at room temperature and was thenevaporated. The remainder was dissolved in methanol(60 mL) and water was added. The precipitate wasfiltered and dried to give the title compound (7.38 g,98%), m.p.: 210–212 °C.1H NMR (DMSO-d6) δ 10.20(br, 1H), 7.30 (d,J = 2.4 Hz, 1H) 6.88 (d,J = 2.4 Hz,1H), 13C NMR (DMSO-d6) δ 169.2, 146.4, 141.2, 123.2,118.9, 110.9, 104.2, MS (EI, 70 eV) m/z: 233/231 (M +,14/13).

6.1.8. 4,5-Dibromo-3-hydroxyanthranilic acid85-Bromo-3-hydroxyanthranilic acid 7 (7.38 g,

31.8 mmol) was dissolved in methanol (120 mL) anddichloromethane (250 mL) saturated with hydrogen bro-mide was added. More dichloromethane (300 mL) wasadded and the mixture was stirred for 48 h whereafter thesolvent was evaporated. The residue (8.92 g) was dis-solved in acetic acid (400 mL) and hydrobromic acid(48%, 100 mL) and bromine (5.5 g, 34.4 mmol) in aceticacid (50 mL) were added drop-wise for 20 min. Themixture was heated to 70 °C and additional bromine(5.50 g, 34.4 mmol) in acetic acid (50 mL) was added.After 14 h at 90 °C the solvent and excess reagent wereevaporated. The remainder was crystallized from ethanol-water to give the title compound (8.45 g, 85%), m.p.:223–223.5 °C.1H NMR (DMSO- d6) δ 9.3 (br, 1H), 7.56(s, 1H), 6.0 (br, 1H), 5.3 (br, 1H).13C NMR (DMSO-d6)δ 168.9, 143.5, 142.5, 125.9, 117.9, 110.3, 107.2. MS (EI,70 eV) m/z: 313/311/309 (M +, 51/100/50). Anal. for thehydrochloride C7H5Br2NO3xHCl (C, H, N).

6.1.9. Methyl 4-bromo-3-hydroxy-5-methylanthranilate9To a solution of methyl 3-hydroxy-5-methylanthra-

nilate hydrochloride (50 mg, 0.23 mmol) dissolved inacetic acid (2.5 mL) under argon atmosphere was bro-mine (36 mg, 0.23 mmol) in acetic acid (2.5 mL) addeddrop-wise. After 3 h at room temperature, additionalbromine (18 mg, 0.11 mmol) in acetic acid (1 mL) wasadded. The solvent and excess reagent were evaporated,and the remainder was purified by repeated preparative

thin layer chromatography (SiO2, chloroform-methanol-ammonium hydroxide 300:10:1, chloroform-methanol1 000:1) to give the title compound (9 mg, 15%).1HNMR (DMSO- d6) δ 7.22 (s, 1H), 6.25 (br, 3H), 3.74 (s,3H), 2.17 (s, 3H), MS (EI, 70 eV) m/z: 261/259 (M +,85/86).

6.1.10. 4-Bromo-3-hydroxy-5-methylanthranilic acid10Methyl 4-bromo-3-hydroxy-5-methylanthranilate9

(14 mg, 0.05 mmol) was dissolved in ethanol (0.7 mL)and flushed with argon. Potassium hydroxide (aq,0.35 mL, 0.05 mmol) was added and the reaction wasstirred at 40 °C for 6 h. The mixture was acidified to pH2 by the addition of hydrochloric acid (2 M). The solventwas removed and the remainder was triturated withice-water. The crude product was purified by preparativeHPLC (Lichrosorb C18) using methanol-ammonium hy-droxide (0.05 M) (40:60) as the eluent to give the titlecompound (2 mg, 16%).1H NMR (DMSO-d6) δ 7.23 (s,1H), 3.35 (br, 2H), 2.13 (s, 2H), 2.05 (s, 3H), MS (EI, 70eV) m/z: 247/245 (M +, 82/86). High resolution MSC8H8BrNO3.

6.1.11. 5-Bromo-4-chloro-3-hydroxyanthranilic acid11To 4-chloro-3-hydroxyanthranilic acid [23] (50 mg,

0.27 mmol) dissolved in acetic acid (2.0 mL) were hy-drobromic acid (47%, 0.02 mL, 0.54 mmol) and thenbromine (86 mg, 0.54 mmol) in acetic acid (3 mL) added,and the mixture was stirred at room temperature andunder nitrogen atmosphere for 2 d. The solvent wasevaporated and the remainder was crystallized fromethanol (40%), to give the title compound (37 mg, 51%),m.p.: 227–228 °C.1H NMR (CD3OD) δ 7.72 (s, 1H), 5.0(br, 4H), 13C NMR (CD3OD) δ 170.3, 143.5, 143.0,127.2, 125.4, 111.0, 106.7. MS (TSP) m/z: 270/268/266(M +, 28/100/80). Anal. C7H5BrClNO3 (C, H, N, O).

6.1.12. 4,5-Dichloro-3-hydroxyanthranilic acid12To a solution of 4-chloro-3-hydroxyanthranilic

acid [32] (150 mg, 0.80 mmol) in acetic acid (12 mL) andunder argon, were hydrochloric acid (12 M, 336µL,4.8 mmol) and then chlorine (1.76 mmol) in acetic acid(1.78 mL) added drop-wise. The reaction was stirred for20 h and then the precipitated material was filtered. Thecrude material was recrystallized from ethanol (50%)giving the title compound (14 mg, 8%).1H NMR(CD3OD) δ 7.57 (s, 1H), 4.98 (br, 4H),13C NMR(CD3OD) δ 170.7, 143.6, 143.4, 124.0, 118.7, 110.6,109.5. MS (TSP) m/z: 224/222 (M +, 53/100). Anal.C7H5Cl2NO3 (C, H, N).

735

6.1.13. 6-Methoxy-7-nitrotetralin13To a solution of 6-methoxytetralin (6.80 g, 41.9 mmol)

in dichloromethane (135 mL) under argon atmosphereand at 0 °C was nitric acid (90%, 3.92 mL, 83.8 mmol)added drop-wise. The mixture was stirred for 25 h wheniodomethane (7.85 mL, 126 mmol), tetra-n-butyl-ammonium hydrogensulfate (28.4 g, 83.8 mmol) and so-dium hydroxide (2 M, 8 mL, 16 mmol) were added andthe mixture was heated at reflux for 4 h. The organic layerwas collected, washed with brine, sodium hydroxide (2M), brine, dried (MgSO4) and evaporated. The remainderwas purified by flash chromatography (SiO2, chloroform-hexane 2:1) to give the title compound (2.62 g, 30%).1HNMR (CDCl3) δ 7.60 (s, 1H), 6.74 (s, 1H), 3.90 (s, 3H),2.79 (m, 2H) 2.71 (m, 2H) 1.81–1.77 (m, 4H),13C NMR(CDCl3) δ 151.5, 145.5, 137.6, 130.0, 126.8, 114.2, 57.0,30.5, 28.8, 23.3 23.1, MS (EI, 70 eV) m/z: 207 (M +, 59).

6.1.14. 6-Hydroxy-7-nitrotetralin14To 13 (2.92 g, 14.1 mmol) dissolved in dichlo-

romethane (450 mL) at –65 °C was boron tribromide(1 M, 28.2 mmol) in dichloromethane (28.2 mL) addedfor 15 min under an atmosphere of argon. The solutionwas stirred for 2 h when the temperature was graduallyincreased to 0 °C. The organic phase was collected,washed with sodium bicarbonate (sat., 200 mL), brine,dried (MgSO4) and evaporated. The remainder was puri-fied on a short column (SiO2, chloroform-hexane 2:1) togive the title compound (2.10 g, 77%).1H NMR (DMSO-d6) δ 10.47 (br, 1H), 7.62 (s, 1H), 6.80 (s, 1H), 2.7–2.6(m, 4H), 1.65–1.71 (m, 4H),13C NMR (DMSO-d6) δ150.0, 146.1, 134.0, 128.3, 124.7, 118.5, 28.9, 27.5, 22.3,22.0, MS (EI, 70 eV) m/z: 193 (M +, 100).

6.1.15. 5-Chloro-6-hydroxy-7-nitrotetralin15To a solution of14 (2.44 g, 12.6 mmol) in chloroform

(290 mL) under argon atmosphere was chlorine (0.99 M,25.6 mmol) in chloroform (25.6 mL) added. The mixturewas stirred at room temperature for 6 h when the solventwas evaporated. The remainder was purified by flashchromatography (SiO2, chloroform-hexane 1:1) to givethe title compound (2.29 g, 80%).1H NMR (DMSO-d6) δ10.62 (br, 1H), 7.71 (s, 1H), 2.75–2.69 (d, 4H), 1.77–1.64(d, 4H), 13C NMR (DMSO-d6) δ 146.3, 143.6, 134.6,129.7, 123.4, 122.8, 28.3, 27.9, 21.7, 21.6, MS (EI, 70eV) m/z: 229/227 (M +, 32/100).

6.1.16. 6-Benzyloxy-5-chloro-7-nitrotetralin16To a solution of15 (2.28 g, 10.0 mmol) in dryN,

N-dimethylformamide (40 mL) under argon atmosphere,were benzyl chloride (11.5 mL, 100 mmol), tetra-n-butylammonium hydrogensulfate (95 mg, 0.25 mmol)and potassium carbonate (41.5 g, 30.0 mmol) added. The

mixture was stirred at room temperature for 24 h, the saltswere then filtered and the solvent was evaporated. Theremainder was dissolved in ethyl acetate, and the organicphase was washed with brine, dried (MgSO4) and evapo-rated. The crude material was purified by flash chroma-tography (SiO2, chloroform-hexane 1:1) to give the titlecompound (2.20 g, 69%), m.p.:80–82 °C.1H NMR(DMSO-d6) δ 7.73 (s, 1H), 7.48–7.37 (m, 5H), 5.06 (s,2H), 2.79–2.75 (m, 4H), 1.79–1.69 (m, 4H).13C NMR(DMSO-d6) δ 144.8, 142.1, 135.8, 135.6, 129.6, 128.4,123.4, 106.2, 105.5, 28.6, 27.6, 21.6, 21.4. MS (TSP)m/z: 337/335 (M + NH4, 30/100). Anal. C17H16ClNO3

(C, H, N).

6.1.17. 7-Amino-6-benzyloxy-5-chlorotetralin17To a suspension of16 (2.56 g, 8.33 mmol) in methanol

at 1 °C was copper(I) chloride (4.95 g, 25.0 mmol) andpotassium borohydride (3.25 g, 60.1 mmol) addedportion-wise for 8 h. The mixture was filtered and evapo-rated. The remainder was dissolved in ethyl acetate andthe organic phase was washed with water, brine, dried(Na2SO4) and evaporated. The crude material was puri-fied by flash chromatography (SiO2, chloroform) to givethe title compound (1.82 g, 76%).1H NMR (DMSO-d6) δ7.55 (dd,J = 1.6 Hz,J = 8.1 Hz, 2H), 7.41–7.34 (m, 3H),6.42 (s, 1H), 4.84 (s, 2H), 4.79 (s, 2H), 2.58–2.52 (m,4H), 1.71–1.60 (m, 4H),13C NMR (DMSO-d6) δ 140.1,139.3, 137.3, 134.0, 128.3, 128.2, 128.1, 127.9, 127.0,121.8, 114.0, 72.8, 29.0, 26.2, 22.7, 22.3, MS (TSP) m/z:290/288 (M + 1, 27/100). Anal. C17H18ClNO (C, H, N).

6.1.18. 6-Benzyloxy-5-chloro-7-isonitrosoacetamino-tetralin 18

To a solution of 17 (1.84 g, 6.41 mmol) in N,N-dimethylformamide (80 mL) and water (8 mL) underargon atmosphere, were hydrochloric acid (12 M,0.53 mL) and chloral hydrate (1.17 g, 7.05 mmol) added.The mixture was heated to 110 °C when hydroxylaminehydrochloride (1.78 g, 25.6 mmol) in water (8 mL) wasadded. The mixture was heated to 100 °C for 1 h and thesolvent was evaporated. The remainder was dissolved inethyl acetate, and the organic phase was washed withwater, brine, dried (Na2SO4) and evaporated. The crudematerial was purified by flash chromatography (SiO2,ethyl acetate-chloroform 1:5) to give the title compound(1.35 g, 59%).1H NMR (DMSO-d6) δ 12.33 and 9.72(E/Z) (2s, 1H), 9.20 and 8.28 (E/Z) (2s, 1H), 7.84 and7.72 (E/Z) (2s, 1H), 7.60 (s, 1H), 7.57–7.37 (m, 5H), 4.90and 4.88 (E/Z) (2s, 2H), 2.71–2.65 (m, 4H), 1.75–1.68(m, 4H), 13C NMR (DMSO-d6) δ 160.3, 160.0, 143.4,143.0, 136.6, 136.1, 134.4, 134.3, 131.2, 130.6, 129.8,129.5, 128.8, 128.3, 128.2, 128.1, 127.0, 120.7, 120.3,

736

74.8, 74.3, 29.1, 26.7, 22.2, 22.0, MS (TSP) m/z: 361/359(M + 1, 28/100).

6.1.19. 9-Benzyloxy-8-chloro-1H-4,5,6,7-tetrahydro[e]benzindole-2,3-dione19

To sulfuric acid (conc., 5 mL) at 60 °C,18 (500 mg,1.39 mmol) was added portion-wise for 1 h. The mixturewas poured on ice (50 mL) and extracted with ethylacetate (200 mL). The organic phase was dried (Na2SO4)and evaporated. The remainder was dissolved inN,N-dimethylformamide (3 mL) under argon atmosphere,and benzyl bromide (0.16 mL, 1.39 mmol) and potassiumcarbonate (192 mg, 1.39 mmol) were added. The mixturewas stirred at room temperature for 18 h. The salt wasfiltered off and the solvent was evaporated. The remain-der was purified by repeated flash chromatography (SiO2,ethyl acetate-methanol 20:1, chloroform-methanol 50:1)to give the title compound (56 mg, 12%).1H NMR(DMSO-d6) δ 11.40 (s, 1H), 7.58 (d,J = 7.3 Hz, 2H),7.42–7.36 (m, 3H), 4.90 (s, 2H), 2.92 (t,J = 7.3 Hz, 2H),2.63 (t,J = 6.0 Hz, 2H), 1.74–1.66 (m, 4H),13C NMR(DMSO-d6) δ 183.8, 159.4, 142.3, 137.3, 137.1, 136.2,136.2, 129.9, 128.8, 128.2, 128.1, 114.7, 74.6, 26.7, 25.4,21.7, 20.8, MS (EI, 70 eV) m/z: 343/341 (M +, 5/15).

6.1.20. 6-Amino-7-benzyloxy-5-carboxy-8-chlorotetralin20

To a suspension of19 (51 mg, 0.15 mmol) in sodiumhydroxide (0.68 M, 0.60 mmol) and water (0.46 mL) at10 °C, was hydrogen peroxide (30%, 86µL, 0.85 mmol)in sodium hydroxide (0.68 M, 0.90 mmol) added. Themixture was stirred for 3 h and then filtered. Water andacetic acid were added and the mixture was extractedwith ethyl acetate (40 mL). The organic phase waswashed with brine, dried (Na2SO4) and evaporated togive the title compound (35 mg, 70%).1H NMR (DMSO-d6) δ 7.56 (d,J = 7.0 Hz, 2H), 7.43–7.36 (m, 3H), 4.84(s, 2H), 3.3 (br, 3H), 2.73 (m, 2H), 2.59 (m, 2H)1.70–1.61 (m, 4H),13C NMR (DMSO-d6) δ 169.3, 139.9,139.5, 136.8, 133.1, 129.5, 128.3, 128.1, 122.3, 116.5,73.1, 28.1, 26.8, 22.2, 22.0, MS (EI, 70 eV) m/z: 333/331(M +, 7/18).

6.1.21. 6-Amino-5-carboxy-8-chloro-7-hydroxytetralin21

6-Amino-7-benzyloxy-5-carboxy-8-chlorotetralin20(33 mg, 0.10 mmol) in ethanol (3 mL) was hydrogenatedat room temperature and at atmospheric pressure for 2 hwith Pd/C (5%, 4 mg) as the catalyst. The mixture wasfiltered, evaporated and dried to give the title compound(21 mg, 87%), m.p.: 147 °C (dec.).1H NMR (DMSO-d6)δ 7.9 (br, 4H), 2.66 (t,J = 7 Hz, 2H), 2.54 (t,J = 7 Hz,2H), 1.70–1.64 (m, 2H), 1.63–1.57 (m, 2H),13C NMR

(DMSO-d6) δ 169.7, 138.1, 136.4, 128.1, 123.3, 121.4,115.2, 27.9, 26.9, 22.4, 22.2, MS (EI, 70 eV) m/z:333/331 (M +, 21/65).

6.1.22. 4,5-Dimethyl-2-nitrophenol22To a solution of 3,4-dimethylphenol (20.0 g,

164 mmol) in dichloromethane (400 mL) under argonatmosphere and at 2 °C was nitric acid (90%, 7.7 mL,165 mmol) added drop-wise. The mixture was stirred for90 min while the temperature was slowly increased toambient temperature. To the mixture was dichlo-romethane (500 mL) added and the organic phase waswashed with brine, water, dried (MgSO4) and evaporated.The remainder was purified by flash chromatography(SiO2, ethyl acetate-hexane 1:1) to give the title com-pound (10 g, 36%), m.p.: 82–83 °C.1H NMR (DMSO-d6) δ 11.5 (br, 1H), 7.71 (s, 1H), 6.92 (s, 1H), 2.22 (s,3H), 2.17 (s, 3H),13C NMR (DMSO-d6) δ 150.8, 146.3,133.4, 128.0, 124.9, 119.7, 19.7, 18.1, MS (EI, 70 eV)m/z: 167 (M +, 100).

6.1.23. 2-Chloro-3,4-dimethyl-6-nitrophenol23To a solution of22 (7.0 g, 41.9 mmol) dissolved in

chloroform (300 mL) under argon atmosphere was chlo-rine (0.99 M, 83.7 mmol) in chloroform (84.8 mL) added.The mixture was stirred for 26 h and then the solvent andexcess reagents were evaporated. The remainder wasdissolved in dichloromethane (400 mL) and the organicphase was washed with brine, dried (MgSO4) and evapo-rated. The crude material was purified by flash chroma-tography (SiO2, chloroform-hexane 1:1) to give the titlecompound (6.47 g, 77%), m.p.: 62–63 °C.1H NMR(DMSO-d6) δ 11.5 (br, 1H), 7.80 (s, 1H), 2.34 (s, 3H),2.27 (s, 3H). 13C NMR (DMSO-d6) δ 146.9, 143.8,134.1, 128.8, 123.7, 122.9, 19.4, 17.4. MS (EI, 70 eV)m/z: 203/201 (M +, 37/100).

6.1.24. 2-Benzyloxy-3-chloro-4,5-dimethylnitrobenzene24

To a solution of23 (6.44 g, 31.9 mmol) in dryN,N-dimethylformamide (105 mL) under argon atmo-sphere, were benzyl bromide (4.17 mL, 35.1 mmol) andpotassium carbonate (13.2 g, 95.7 mmol) added. Themixture was stirred at room temperature for 8 h and wasthen filtered. The solvent was evaporated, and the remain-der was purified by flash chromatography (SiO2,chloroform-hexane 1:1) affording the title compound(8.44 g, 91%), m.p.: 74–75 °C.1H NMR (DMSO-d6) δ7.80 (s, 1H), 7.48–7.37 (m, 5H), 5.06 (s, 2H), 2.37 (s,3H), 2.34 (s, 3H),13C NMR (DMSO-d6) δ 145.2, 142.2,142.1, 135.7, 134.7, 129.6, 128.4, 128.4, 123.6, 75.7,19.7, 17.1. MS (EI, 70 eV) m/z: 293/291 (M +, 8/37).

737

6.1.25. 2-Benzyloxy-3-chloro-4,5-dimethylaniline25To a solution of24 (3.00 g, 10.3 mmol) in methanol

(420 mL) at 2 °C, were copper(I) chloride (6.11 g,30.8 mmol) and potassium borohydride (4.4 g,81.6 mmol) added in portions. The mixture was filteredand evaporated. The remainder was dissolved in ethylacetate, and the organic phase was washed with water,dried (MgSO4) and evaporated. The crude material waspurified by flash chromatography (SiO2, chloroform) togive the title compound (2.14 g, 79%), m.p.: 57–58 °C.1H NMR (DMSO-d6) δ 7.54 (d, J = 6.5 Hz, 2H),7.42–7.34 (m, 3H), 6.52 (s, 1H), 4.83 (s, 2H), 4.79 (s,2H), 2.12 (s, 6H),13C NMR (DMSO-d6) δ 140.0, 138.9,137.3, 133.0, 128.3, 128.2, 127.9, 127.3, 121.2, 115.4,72.7, 20.2, 15.4, MS (EI, 70 eV) m/z: 286/284 (M + 23,42/100).

6.1.26. 2-Benzyloxy-3-chloro-4,5-dimethylisonitroso-acetanilide26

To a solution of 25 (2.14 g, 8.19 mmol) in N,N-dimethylformamide (60 mL) and water (2 mL) werehydrochloric acid (conc., 0.68 mL, 8.19 mmol) and chlo-ral hydrate (1.49 g, 9.00 mmol) added. The mixture washeated to 105 °C and hydroxylamine hydrochloride(2.28 g, 32.8 mmol) dissolved in water (4 mL) wasadded. The mixture was stirred for 1 h and the solventwas evaporated. The remainder was dissolved in ethylacetate, and the organic layer was washed with water,dried (MgSO4) and evaporated. The crude material waspurified by flash chromatography (SiO2, chloroform-ethyl acetate 10:1) affording the title compound (1.28 g,47%).1H NMR (DMSO-d6) δ 12.33 and 9.71 (E/Z) (2s,1H), 9.20 and 8.27 (E/Z) (2s, 1H), 7.91 and 7.83 (2s, 1H),7.60 (s, 1H), 7.65–7.36 (m, 5H), 4.90 (s, 1H), 4.87 (s,1H), 2.26–2.23 (m, 6H),13C NMR (DMSO-d6) δ 160.2,160.0, 143.4, 143.1, 136.6, 136.1, 133.4, 133.3, 131.0,130.4, 129.7, 129.4, 128.7, 128.4, 128.4, 128.3, 128.2,128.1, 128.0, 127.4, 127.2, 121.5, 121.2, 74.7, 74.2, 20.3,16.1, MS (EI, 70 eV) m/z: 334/332 (M +, 4/12).

6.1.27. 7-Benzyloxy-6-chloro-4,5-dimethylisatin27To sulfuric acid (conc., 6 mL) at 80 °C was26

(700 mg, 2.10 mmol) added. The mixture was stirred at80 °C for 10 min and then poured into ice-water(200 mL). The mixture was extracted with ethyl acetate(200 mL) and the organic phase was dried (MgSO4) andevaporated. The remainder was dissolved inN,N-dimethylformamide and benzyl bromide (0.28 mL,2.30 mmol) and potassium carbonate (318 mg,2.30 mmol) were added. The mixture was stirred atambient temperature for 30 h and filtered. Acetic acid(1.5 mL) was added, and the solvents were evaporated.

The remainder was purified by repeated flash chromatog-raphy (SiO2, dichloromethane-methanol 50:1) to give thetitle compound (13 mg, 2%).1H NMR (DMSO-d6) δ11.42 (s, 1H), 7.57 (d,J = 6.2 Hz, 2H), 7.42–7.35 (m,3H), 4.92 (s, 2H), 2.44 (s, 3H), 2.22 (s, 3H),13C NMR(DMSO-d6) δ 184.3, 159.3, 141.9, 137.0, 137.0, 136.3,135.4, 130.2, 128.8, 128.3, 128.2, 115.6, 74.6, 15.3, 14.2,MS (EI, 70 eV) m/z: 317/315 (M +, 3/7).

6.1.28. 3-Benzyloxy-4-chloro-5,6-dimethylanthranilicacid 28

To a suspension of27 (13 mg, 0.04 mmol) in dioxane(0.5 mL) and sodium hydroxide (0.68 M, 0.14 mmol) at10 °C, was hydrogen peroxide (30%, 7µL, 0.22 mmol)dissolved in sodium hydroxide (0.41 mL) added in por-tions. The mixture was concentrated in a stream ofnitrogen, when acetic acid (38µL) was added. Themixture was dissolved in ethyl acetate (3 mL) and theorganic phase was washed with brine, dried (MgSO4) andevaporated to give the title compound (11 mg, 90%).1HNMR (DMSO-d6) δ 7.55 (d,J = 7.0 Hz, 2H), 7.43–7.35(m, 3H), 4.82 (s, 2H), 2.21 (s, 3H), 2.19 (s, 3H), MS (EI,70 eV) m/z: 307/305 (M +, 7/19).

6.1.29. 4-Chloro-5,6-dimethyl-3-hydroxyanthranilic acid29

3-Benzyloxy-4-chloro-5,6-dimethylanthranilic acid28(10 mg, 0.03 mmol) in ethanol (1.5 mL) was hydrogen-ated at ambient temperature and atmospheric pressure for5 h with Pd/C (10%, 2 mg) as the catalyst. The mixturewas filtered and the solvent was evaporated. The remain-der was purified by preparative HPLC (Lichrosorb-C18,methanol-phosphate buffer, pH 3, 1:1). The fractionscontaining the product were pooled and evaporated. Tothe residue was ethyl acetate added and the organic phasewas washed with sodium carbonate (sat), dried andevaporated to give the title compound (3 mg, 46%).1HNMR (DMSO-d6) δ 3.35 (br, 1H), 3.16 (s, 2H), 2.16 (s,3H), 2.15 (s, 3H),13C NMR (DMSO-d6) δ 169.9, 137.9,135.2, 126.4, 123.5, 121.4, 117.3, 17.7, 16.0, MS (EI, 70eV): m/z 217/215 (M +, 21/63).

6.1.30. 2,4-Dichloro-5-methoxy-3-methylphenyl trifluoro-methanesulfonate30

To a solution of 2,4-dichloro-5-methoxy-3-methyl-phenol [33] (7.73 g, 37.3 mmol) in dichloromethane(180 mL) were triethylamine (10.4 mL, 74.7 mmol) andN,N-dimethylaminopyridine (10 mg, 0.08 mmol) added.The solution was cooled to –78 °C and trifluoromethane-sulfonic anhydride (9.4 mL, 74.7 mmol) was added drop-wise. The mixture was then stirred for 20 min while thetemperature slowly reached 0 °C. To the solution wasadded dichloromethane (200 mL) and the organic phase

738

was washed with water (150 mL), brine, dried (MgSO4)and evaporated. The remainder was purified by flashchromatography (SiO2, ethyl acetate-hexane 1:3) afford-ing the title compound (12.3 g, 97%), m.p.: 58.5–59.0 °C.1H NMR (DMSO-d6) δ 7.30 (s, 1H), 3.92 (s, 3H), 2.49 (s,3H), 13C NMR (DMSO-d6) δ 154.2, 143.9, 137.0, 122.8,118.1, 118.0 (q,J = 321Hz), 105.3, 57.2, 18.0, MS (EI,70 eV) m/z: 340/338 (M +, 47/64).

6.1.31. Methyl 2,4-dichloro-5-methoxy-3-methylbenzoate31

To a solution of30 (7.60 g, 22.4 mmol) in dioxane(75 mL) were 1,3-bis(diphenylphosphino)propane(0.37 g, 0.90 mmol) and palladium acetate (0.20 g,0.90 mmol) added. The mixture was flushed with carbonmonoxide, then triethylamine (6.90 mL, 49.4 mmol) andmethanol (23 mL) were added. The mixture was stirred at70 °C and atmospheric pressure for 25 h and was thenfiltered and evaporated. The remainder was dissolved indiethyl ether and the organic phase was washed withammonium hydroxide (2 M, 150 mL), brine (150 mL),dried (MgSO4) and evaporated. The crude material waspurified by flash chromatography (SiO2, ethyl acetate-hexane 1:3) to give the title compound (4.18 g, 75%),m.p.: 73.5–74 °C.1H NMR (DMSO-d6) δ 7.33 (s, 1H),3.88 (s, 3H), 3.86 (s, 3H), 2.45 (s, 3H),13C NMR(DMSO-d6) δ 165.7, 153.3, 136.1, 130.5, 125.2, 123.0,110.8, 56.6, 52.7, 17.9, MS (EI, 70 eV) m/z: 250/248(M +, 53/80).

6.1.32. 2,4-Dichloro-5-hydroxy-3-methylbenzoic acid32

To a solution of31 (0.24 g, 0.96 mmol) in methanol(30 mL) under argon atmosphere, was potassium hydrox-ide (0.31 g, 4.8 mmol) added and the mixture was stirredat 50 °C for 19 h. The solvent was evaporated, andhydrobromic acid (48%, 30 mL) was added and themixture was stirred at 110 °C for 3 d. The acid wasevaporated, and the remainder was dissolved in ethylacetate (40 mL). The organic phase was extracted withammonium hydroxide (dil., 11 mL) and the pH of theaqueous phase was lowered to 1. The aqueous phase wasextracted with ethyl acetate (40 mL), the organic phasewas washed with brine, dried (MgSO4) and evaporated togive, after crystallization from diethyl ether, the titlecompound (0.20 g, 94%), m.p.: 203.5–204.5 °C.1HNMR (DMSO-d6) δ 10.68 (br, 1H), 7.15 (s, 1H), 2.42 (s,3H), 13C NMR (DMSO-d6) δ 166.7, 151.8, 135.9, 131.2,123.7, 121.2, 114.4, 17.9, MS (EI, 70 eV) m/z: 222/220(M +, 56/100).

6.1.33. 4,6-Dichloro-3-hydroxy-5-methyl-2-nitrobenzoicacid 33

To a solution of 32 (90 mg, 0.41 mmol) in ni-tromethane (9 mL) at 40 °C, was nitric acid (90%, 20 mL,0.43 mmol) added. The mixture was stirred for 4 h, whenthe solvent was removed by evaporation. The remainderwas purified by flash chromatography (SiO2, ethylacetate-acetic acid 30:1) to give the title compound(79 mg, 72%), m.p.: 197–199 °C.1H NMR (DMSO-d6) δ2.45 (s, 3H). 13C NMR (DMSO-d6) δ 164.3, 147.1,139.6, 136.3, 128.4, 126.5, 118.9, 18.6. MS (EI, 70 eV)m/z: 267/265 (M +, 66/100).

6.1.34. 4,6-Dichloro-3-hydroxy-5-methylanthranilic acid34

4,6-Dichloro-3-hydroxy-5-methyl-2-nitrobenzoic acid33 (69 mg, 0.26 mmol) in acetic acid (10 mL) and hydro-chloric acid (conc., 33 mL) was hydrogenated at atmo-spheric pressure and room temperature for 2 h with Pd/C(10%, 10 mg) as the catalyst. Methanol was added andthe mixture was filtered. The solvent was evaporated andthe remainder was purified by flash chromatography(SiO2, ethyl acetate-acetic acid 45:1) affording the titlecompound (55 mg, 90%), m.p.: 192 °C (dec.).1H NMR(DMSO-d6) δ 2.27 (s, 3H),13C NMR (DMSO-d6) δ167.4, 139.2, 136.0, 122.9, 121.3, 120.4, 116.9, 17.0, MS(EI, 70 eV) m/z: 237/235 (M +, 45/79).

6.1.35. 4,6-Dibromo-3-hydroxy-2-nitrobenzoic acid35To a cooled solution of 3-hydroxy-2-nitrobenzoic acid

(10.5 g, 0.057 mol) and sodium acetate (9.85 g, 0.57 mol)in acetic acid (100 mL) was bromine (6.15 mL, 0.12 mol)added drop-wise. The mixture was stirred at 60 °C for68 h and then cooled and filtered. The solution wasevaporated, and the remainder was dissolved in ethylacetate. The organic phase was washed with dilutehydrochloric acid, dried (MgSO4) and evaporated. Thecrude product was purified by flash chromatography(SiO2, toluene-acetic acid 5:1). The pure compound wascrystallized from methanol to give the title compound(15.4 g, 79%), m.p.: 201–202 °C (dec.).1H NMR(DMSO-d6) δ 8.20 (s, 1H),13C NMR (DMSO-d6) δ164.4, 146.9, 139.1, 130.5, 116.2, 108.1, MS (EI, 70 eV)m/z: 343/341/339 (M +, 46/98/49). Anal. C7H3Br2NO5

(C, H, N).

6.1.36. 4,6-Dibromo-3-hydroxyanthranilic acid364,6-Dibromo-3-hydroxy-2-nitrobenzoic acid 35

(4.09 g, 12 mmol) in ethanol (150 mL) was hydrogenatedat atmospheric pressure and room temperature for 45 h,with PtS2 (0.16 g, 0.62 mmol) as the catalyst. The mix-ture was filtered and the solvent was removed by evapo-ration. The remainder was purified by flash chromatog-

739

raphy (SiO2, toluene-acetic acid 5:1) affording, aftercrystallization from methanol-water, the title compound(2.51 g, 63%), m.p.: 162–164.5 °C.1H NMR (DMSO-d6)δ 6.96 (s, 1H),13C NMR (DMSO-d6) δ 167.7, 140.6,140.0, 121.8, 117.2, 112.5, 110.4. MS (EI, 70 eV) m/z:313/311/309 (M +, 36/72/34). Anal. C7H5Br2NO3 (C, H,N).

6.1.37. 4-Bromo-5-hydroxy-2-methoxybenzoic acid37To a solution of 5-hydroxy-2-methoxybenzoic

acid [34] (1.24 g, 73.7 mmol) in acetic acid (100 mL)was bromine (0.38 mL, 7.4 mmol) added drop-wise. Themixture was stirred for 3 h and the solvent was thenevaporated. The remainder was purified by flash chroma-tography (SiO2, toluene-acetic acid 10:1) to give the titlecompound (1.5 g, 83%).1H NMR (CD3OD) δ 7.38 (s,1H), 7.26 (s, 1H), 3.84 (s, 3H),13C NMR (CD3OD) δ169.0, 154.1, 149.6, 121.3, 119.6, 119.0, 116.6, 57.7.

6.1.38. 4-Bromo-3-hydroxy-6-methoxy-2-nitrobenzoicacid 38

To a solution of sodium nitrate (361 mg, 4.25 mmol),lanthanum nitrate hexahydrate (18 mg, 0.042 mmol) andhydrochloric acid (12 M, 4 mL) in water (4 mL) at 0 °C,was 37 (1.05 g, 4.25 mmol) in diethyl ether (20 mL)added in portions. The mixture was stirred for 7 h, whilethe solution reached ambient temperature. Dichlo-romethane (90 mL) and water (20 mL) were added, andthe organic phase was collected, dried (MgSO4), filteredand evaporated. The remainder was purified by flashchromatography (SiO2, toluene-ethyl acetate-acetic acid8:2:1) affording the title compound (600 mg, 48%).1HNMR (CD3OD) δ 7.57 (s, 1H), 3.85 (s, 3H), MS (EI, 70eV) m/z 293/291 (M +, 21/19).

6.1.39. 4-Bromo-3-hydroxy-6-methoxyanthranilic acid394-Bromo-3-hydroxy-6-methoxy-2-nitrobenzoic acid

38 (52 mg, 0.18 mmol) in ethanol (7 mL) was hydrogen-ated at atmospheric pressure and room temperature for18 h with PtS2 as the catalyst. The solution was filteredand the solvent was evaporated. The remainder waspurified by flash chromatography (SiO2, toluene-ethylacetate-acetic acid 8:2:1) to give the title compound(30 mg, 64%).1H NMR (CD3OD) δ 6.40 (s, 1H), 3.88 (s,3H), 13C NMR (CD3OD) δ 170.8, 154.8, 145.6, 137.8,115.5, 102.1, 101.6, 57.6, MS (EI, 70 eV): m/z 263/261(M +, 79/80).

6.1.40. 2,4-Dichloro-5-methoxyphenyl trifluoromethane-sulfonate40

To a solution of 2,4-dichloro-5-methoxyphenol [35](4.43 g, 22.9 mmol), triethylamine (6.40 mL, 45.9 mmol)and N,N-dimethylaminopyridine (5 mg, 0.04 mmol) in

dichloromethane (10 mL) at –70 °C, was trifluro-methanesulfonic anhydride (5.79 mL, 34.4 mmol) added.The mixture was stirred for 20 min while the temperaturewas increased to ambient temperature. Dichloromethane(150 mL) was added and the organic phase was washedwith water (100 mL), brine (100 mL), dried (MgSO4) andevaporated. The remainder was purified by flash chroma-tography (SiO2, dichloromethane-hexane 1:2) affordingthe title compound (6.37 g, 86%).1H NMR (DMSO-d6) δ7.99 (s, 1H), 7.42 (s, 1H), 3.92 (s, 3H),13C NMR(DMSO-d6) δ 154.6, 144.0, 130.9, 122.3, 118.0 (q,J =321 Hz), 117.1, 108.2, 57.3, MS (EI, 70 eV) m/z:328/326/324 (M +, 6/37/55).

6.1.41. Methyl 2,4-dichloro-5-methoxybenzoate41To a solution of 40 (6.35 g, 19.5 mmol) in N,

N-dimethylformamide (65 mL) flushed with carbon mon-oxide were 1,3-bis(diphenylphosphino)propane (314 mg,0.76 mmol), palladium acetate (171 mg, 0.76 mmol), tri-ethylamine (6.0 mL) and methanol (14.5 mL) added. Thereaction was stirred at 70 °C and atmospheric pressure for5 h, the solvent was then evaporated. The remainder wasdissolved in diethyl ether (600 mL) and the organic phasewas washed with ammonium hydroxide (2 M, 300 mL),brine (200 mL), dried (MgSO4) and evaporated. Thecrude product was purified by flash chromatography(SiO2, dichloromethane-hexane 2:3) affording the titlecompound (1.65 g, 36%), m.p.: 88–89 °C.1H NMR(DMSO-d6) δ 7.73 (s, 1H), 7.49 (s, 1H), 3.90 (s, 3H),3.86 (s, 3H), 13C NMR (DMSO-d6) δ 164.8, 153.4,131.3, 129.6, 125.3, 123.2, 114.2, 56.7, 52.7, MS (EI, 70eV) m/z: 238/236/234 (M +, 8/55/83).

6.1.42. 2,4-Dichloro-5-hydroxybenzoic acid42To a solution of41 (600 mg, 2.55 mmol) in dichlo-

romethane (10 mL) at –70 °C and under argon atmo-sphere was boron tribromide (0.72 mL, 7.66 mmol)added. The mixture was stirred for 4 h while the tempera-ture slowly reached ambient temperature. Dichloro-methane (20 mL) and sodium hydroxide (2 M, 15 mL)were added and stirring was continued for 30 min. Theaqueous phase was washed with dichloromethane(10 mL) and was then acidified to pH 1 by the addition ofhydrochloric acid (12 M). The aqueous phase was ex-tracted with ethyl acetate (40 mL), and the organic phasewashed with brine, dried (MgSO4) and evaporated to givethe title compound (350 mg, 66%), m.p.: 200.5–201.5 °C.1H NMR (DMSO-d6) δ 10.8 (br, 1H) 7.56 (s, 1H), 7.38(s, 1H), 13C NMR (DMSO-d6) δ 165.8, 152.0, 131.3,130.3, 123.6, 121.6, 118.1, MS (EI, 70 eV) m/z:210/208/206 (M +, 11/64/100).

740

6.1.43. 4,6-Dichloro-3-hydroxy-2-nitrobenzoic acid43To a solution of42 (280 mg, 1.35 mmol) in nitro-

methane (35 mL) at 45 °C was nitric acid (90%, 63µL,1.35 mmol) added. The mixture was stirred for 4 h andthen the solvent was evaporated. The remainder wasdissolved in ethyl acetate (100 mL), and the organicphase was washed with water (5 mL), brine (5 mL), dried(MgSO4) and evaporated. The crude product was purifiedby flash chromatography (SiO2, ethyl acetate-acetic acid50:1) affording the title compound (323 mg, 95%), m.p.:186 °C (dec.).1H NMR (DMSO-d6) δ 7.88 (s, 1H),13CNMR (DMSO-d6) δ 163.8, 147.2, 139.0, 133.1, 128.1,126.3, 118.2, MS (EI, 70 eV) m/z: 253/251 (M +, 34/49).

6.1.44. 4,6-Dichloro-3-hydroxyanthranilic acid hydro-chloride44

4,6-Dichloro-3-hydroxy-2-nitrobenzoic acid 43(223 mg, 0.88 mmol) dissolved in ethanol (50 mL) washydrogenated at atmospheric pressure and room tempera-ture for 1 h with Pd/C (10%, 30 mg) as the catalyst. Themixture was filtered and the solvent was evaporated. Theremainder was purified by flash chromatography (SiO2,ethyl acetate-acetic acid 50:1) affording the free base. Thebase was dissolved in tetrahydrofuran (0.6 mL) andethereal hydrogen chloride (3 M, 1 mL) was added togive the title compound (32 mg, 14%), m.p.: 231 °C(dec.).1H NMR (DMSO-d6) δ 6.68 (s, 1H),13C NMR(DMSO-d6) δ 167.2, 140.3, 139.1, 122.5, 121.9, 116.2,114.0, MS (EI, 70 eV) m/z: 223/221 (M +, 42/65). Anal.C7H5Cl2NO3xHCl (C, H, N).

6.1.45. 3-Chloro-2-methoxy-5-phenylnitrobenzene45A solution of 2-chloro-6-nitro-4-phenylphenol [36]

(21.2 g, 0.08 mol), potassium carbonate (17.6 g,0.13 mol) and iodomethane (13.5 mL, 0.22 mol) in dryN,N-dimethylformamide (150 mL) was stirred at roomtemperature under nitrogen atmosphere for 19 h, whenthe solvent was evaporated. The remainder was purifiedby flash chromatography (SiO2, hexane→ ethyl acetate-hexane 10:1) affording, after crystallization from ethylacetate-hexane, the title compound (22.2 g, 99%), m.p.:76–77.5 °C.1H NMR (CDCl3) δ 7.88 (d,J = 2.3 Hz, 1H),7.81 (d,J = 2.3 Hz, 1H), 7.53–7.40 (m, 5H), 4.05 (s, 3H),13C NMR (CDCl3) δ 148.8, 145.5, 138.1, 137.1, 132.8,130.8, 129.4, 129.2, 128.7, 126.8, 121.8, 62.6, MS (EI,70 eV) m/z: 265/263 (M +, 30/98). Anal. C13H10ClNO3

(C, H, N).

6.1.46. 3-Chloro-2-methoxy-5-phenylaniline463-Chloro-2-methoxy-5-phenylnitrobenzene45 (9.1 g,

34.5 mmol) was hydrogenated in tetrahydrofuran(100 mL), methanol (100 mL) and hydrochloric acid(2 M, 34.5 mL) at atmospheric pressure and room tem-

perature for 3 h with Pd/C (10%, 180 mg) as the catalyst.The mixture was filtered and the solvent was evaporated.The remainder was dissolved in ethyl acetate and theorganic phase was washed with ammonium hydroxide (2M), dried (MgSO4) and evaporated. The crude materialwas crystallized from ethyl acetate-hexane affording thetitle compound (5.05 g, 62%), m.p.: 65–66.5 °C.1HNMR (CDCl3) δ 7.49–7.31 (m, 5H), 6.96 (d,J = 2.1 Hz,1H), 6.82 (d,J = 2.1 Hz, 1H), 3.97 (s, 2H), 3.86 (s, 3H),13C NMR (CDCl3) δ 142.5, 141.4, 139.9, 138.4, 128.7,127.9, 127.4, 126.8, 118.2, 112.9, 59.8, MS (EI, 70 eV)m/z: 235/233 (M +, 24/48). Anal. C13H12ClNO (C, H, N).

6.1.47. 3-Chloro-2-methoxy-5-phenylisonitrosoacetani-lide 47

To a stirred solution of chloralhydrate (1.08 g,6.53 mmol) and sodium sulfate (5.0 g, 35.2 mmol) inwater (20 mL) was46 (1.01 g, 4.33 mmol) dissolved in amixture of water (5 mL), N,N-dimethylformamide(10 mL) and hydrochloric acid (conc., 0.43 mL) added.The mixture was stirred at 95 °C for 90 min, whenhydroxylamine hydrochloride (1.36 g, 19.5 mmol) wasadded, and the stirring was continued for another 22 h.The mixture was cooled and evaporated and then ethylacetate was added. The organic phase was washed withwater, dried (MgSO4) and evaporated. The remainder waspurified by flash chromatography (SiO2, dichloromethane→ ethyl acetate) to give the title compound (1.04 g,79%).1H NMR (DMSO-d6) δ 9.51 (s, 1H), 8.39 (s, 1H),7.80 (s, 1H), 7.60 (d, 2H), 7.51–7.34 (m, 5H), 3.83 (s,3H), 13C NMR (DMSO-d6) δ 160.4, 145.3, 143.6, 138.4,137.3, 132.8, 129.0, 128.8, 127.9, 127.0, 126.6, 126.4,123.3, 118.6, 60.8, MS (EI, 70 eV) m/z: 306/304 (M +,6/20). Anal. C15H13ClN2O3 (C, H, N).

6.1.48. 6-Chloro-7-methoxy-4-phenylisatin483-Chloro-2-methoxy-5-phenylisonitrosoacetanilide47

(6.5 g, 21.3 mmol) and polyphosphoric acid (66 g) werestirred at 80 °C for 3 h when ice, water and ethyl acetatewere added. The organic phase was washed with water,dried (Na2SO4) and evaporated. The remainder waspurified by flash chromatography (SiO2, dichloromethane→ ethyl acetate-hexane 1:1) to give the title compound(5.06 g, 83%).1H NMR (DMSO-d6) δ 11.54 (br s, 1H),7.54–7.42 (m, 5H), 7.10 (s, 1H), 3.82 (s, 3H),13C NMR(DMSO-d6) δ 181.8, 159.3, 145.4, 139.8, 138.0, 135.6,134.9, 128.9, 128.8, 128.1, 124.4, 114.5, 61.2, MS (EI,70 eV) m/z: 289/287 (M +, 27/100). Anal. C15H10ClNO3,(C, H, N).

6.1.49. 4-Chloro-3-methoxy-6-phenylanthranilic acid49To a solution of48 (2.15 g, 7.47 mmol) in dioxane

(20 mL) and sodium hydroxide (0.68 M, 60 mL) at 0 °C

741

and under nitrogen atmosphere, was hydrogen peroxide(30%, 4.0 mL, 39.2 mmol) in sodium hydroxide (0.68 M,50 mL) added drop-wise. The solution was stirred atambient temperature for 6 h when acetic acid was addedto pH 5. The mixture was extracted with ethyl acetate, theorganic phase was dried (Na2SO4) and evaporated. Theremainder was purified by flash chromatography (SiO2,ethyl acetate-hexane 1:2→ toluene-ethyl acetate-aceticacid 4:1:1) to give the title compound (1.37 g, 66%).1HNMR (DMSO-d6) δ 7.38–7.25 (m, 5H), 6.52 (s, 1H),3.75 (s, 3H), 13C NMR (DMSO-d6) δ 169.3, 142.9,141.6, 141.1, 139.0, 128.0, 127.9, 127.1, 117.5, 114.6,59.4, MS (EI, 70 eV) m/z: 280/278 (M +, 15/47). Anal.C14H12ClNO3 (C, H, N).

6.1.50. 4-Chloro-3-hydroxy-6-phenylanthranilic acid50To 49 (110 mg, 0.40 mmol) dissolved in diethyl ether

was ethereal hydrogen chloride added. The solvent wasevaporated and dichloromethane (3 mL) was added. Thesolution was cooled to –70 °C under nitrogen atmosphereand boron tribromide (0.08 mL, 0.84 mmol) was added.The solution was stirred for 40 min while the temperaturereached ambient temperature. The mixture was left overnight, and then sodium bicarbonate (sat, 15 mL) wasadded and the aqueous phase was stirred with dichlo-romethane for 1 h. Then the pH was lowered to 4 by theaddition of hydrochloric acid (2 M), and the aqueousphase was extracted with dichloromethane, the organicphase was dried (MgSO4) and evaporated. The crudematerial was purified by flash chromatography (SiO2,toluene-ethyl acetate-acetic acid 4:1:1) affording the titlecompound (23 mg, 22%).1H NMR (DMSO-d6) δ7.33–7.22 (m, 5H), 6.52 (s, 1H),13C NMR (DMSO-d6) δ172.5, 143.6, 141.2, 141.1, 137.2, 129.6, 129.2, 128.0,123.0, 119.8, 115.2, MS (EI, 70 eV) m/z: 265/263 (M +,23/79).

6.1.51. 3-Chloro-2-methoxy-5-methylisonitrosoaceta-nilide 51

To a stirred solution of chloral hydrate (0.99 g,6.0 mmol) and sodium sulfate (3.4 g, 24 mmol) in water(15 mL) was 3-chloro-2-methoxy-5-methylaniline [37](0.51 g, 3.0 mmol) inN,N-dimethylformamide (7 mL),water (9 mL) and hydrochloric acid (conc., 0.37 mL)added. The mixture was stirred at 90 °C for 25 min, whenhydroxylamine hydrochloride (1.24 g, 18 mmol) wasadded and stirring was continued for another 18 h. Themixture was cooled and ethyl acetate (100 mL) wasadded. The organic phase was washed with water(100 mL), dried (Na2SO4) and evaporated. The remainderwas purified by flash chromatography (SiO2, ethylacetate-hexane 1:1) affording the title compound (0.62 g,

85%).1H NMR (DMSO-d6) δ 9.36 (s, 1H), 7.90 (d,J =1.6 Hz, 1H), 7.74 (s, 1H), 7.07 (d,J = 2.1 Hz, 1H), 3.75(s, 3H), 2.25 (s, 3H),13C NMR (DMSO-d6) δ 160.2,143.6, 134.8, 132.1, 125.9, 125.5, 120.8, 60.7, 20.5,MS (TSP) m/z: 262/260 (M + 18, 30/100). Anal.C10H11ClN2O3 (C, H, N).

6.1.52. 6-Chloro-7-methoxy-4-methylisatin523-Chloro-2-methoxy-5-methylisonitrosoacetanilide51

(0.49 g, 2.0 mmol) and polyphosphoric acid (6.3 g) werestirred at 80 °C for 3 h when ice was added. Ethyl acetate(150 mL) was added and the organic phase was washedwith water (50 mL), dried (Na2SO4) and evaporated. Theremainder was purified by flash chromatography (SiO2,ethyl acetate-hexane 1:2→ ethyl acetate) to give the titlecompound (0.37 g, 82%).1H NMR (DMSO-d6) δ 11.41(br, 1H), 7.00 (s, 1H), 3.74 (s, 3H), 2.38 (s, 3H),13CNMR (DMSO-d6) δ 183.8, 159.4, 144.5, 138.6, 136.0,135.8, 125.1, 116.1, 61.1, 16.6, MS (TSP) m/z: 245/243(M + 18, 37/100). Anal. C10H8ClNO3 (C, H, N).

6.1.53. 4-Chloro-3-methoxy-6-methylanthranilic acid53To a solution of52 (320 mg, 1.4 mmol) in dioxane

(5 mL) at 5 °C and under nitrogen atmosphere washydrogen peroxide (30%, 0.73 mL, 7.2 mmol) in water(10 mL) and sodium hydroxide (0.67 M, 26 mL) addeddrop-wise. The mixture was stirred for 3 h while thetemperature slowly reached ambient temperature. Water(25 mL) was added and the pH was adjusted to 5 by theaddition of acetic acid. The mixture was extracted withethyl acetate (120 mL) and the organic phase was dried(Na2SO4) and evaporated. The remainder was purified byflash chromatography (SiO2, ethyl acetate→ ethylacetate-acetic acid 1 000:1) affording the title compound(210 mg, 70%).1H NMR (CD3OD) δ 6.45 (s, 1H), 3.76(s, 3H), 2.37 (s, 3H),13C NMR (CD3OD) δ 171.9, 146.4,142.9, 138.3, 131.0, 120.3, 114.5, 60.2, 23.2, MS (TSP)m/z: 218/216 (M + 1, 34/100). Anal. C9H10ClNO3 (C, H,N).

6.1.54. 4-Chloro-3-hydroxy-6-methylanthranilic acid54To 53 (49 mg, 0.23 mmol) in diethyl ether was ethereal

hydrogen chloride (3 M) added. The solvent was evapo-rated and dichloromethane (2 mL) was added. The solu-tion was cooled to –65 °C under nitrogen atmospherewhen boron tribromide (0.1 mL, 1.12 mmol) was added.The reaction was stirred for 43 h while the temperaturereached ambient temperature. Sodium bicarbonate (sat.,15 mL) was added, and the pH of the aqueous phase wasadjusted to 2 by the addition of hydrochloric acid (2 M).The aqueous phase was extracted with ethyl acetate, andthe organic phase was dried (Na2SO4) and evaporated.The remainder was purified by flash chromatography

742

(SiO2, toluene-acetic acid 10:1→ 5:1) affording the titlecompound (22 mg, 47%).1H NMR (CD3OD) δ 6.43 (s,1H), 2.35 (s, 3H),13C NMR (CD3OD) δ 172.2, 142.9,139.9, 133.4, 123.8, 119.8, 113.8, 22.9, MS (TSP) m/z:204/202 (M + 1, 28/100). Anal. C8H8ClNO3 (C, H, N).

6.2. Stability assay

The compounds were dissolved in PBS buffer at pH7.5 and the solution was left in the dark for 24 h at 37 °C.Aliquots were analysed by HPLC (Bondapak C18, 150×3.9, phosphate buffer pH 3-acetonitrile-acetic acid75:25:1) with UV detection (254 nm).

6.3. Pharmacological methods

6.3.1. In vitro measurement of 3-HAO inhibitionMale Sprague-Dawley rats (150–200 g) were anaesthe-

tized and perfused before decapitation. The cortex wasrapidly dissected out on ice and stored at –70 °C. Toprepare a homogenate, the cortex was thawed, mincedand sonicated in 9 volumes (w/v) of ice-cold distilledwater by 3 times 5 s bursts of sonication using a celldisrupter (Branson sonifier). Aliquots of the crude homo-genate were stored in vials (400µL/vial) at –70 °C priorto assay. To prepare cell-free homogenate, the vials werethawed in an ice-bath, and to each vial, 600µL 150 mM(N-morpholino)-2-ethanesulfonic acid (MES)/NaOHbuffer (pH 6.5), was added. The vials were then vortexedand centrifuged at 15 000g for 10 min at 4 °C. 100µL ofthe supernatant was incubated at 37 °C under gentlyshaking in a water bath for 1 h in a solution containingFeSO4 (0.3 mM), ascorbic acid (aq, 0.1%) and [14C]3-HANA (specific activity of 5.5 mCi/mmol, 5µM, AstraArcus AB, final volume 200µL). Test compounds wereadded in a volume of 10µL prior to the substrate. Theincubation was terminated by addition of HClO4 (aq.,6%, 50µL), the tubes cooled on ice and the precipitateremoved by centrifugation at 10 000g for 4 min at 4 °C.An aliquot (230µL) of the supernatant was applied to aDowex 50W (200–400 mesh) cation-exchange column(0.5 × 2 cm), which was thereafter washed with distilledwater (1 mL) to elute the [14C]-QUIN. Scintillation fluid(10 mL, Beckman, Ultima Gold) was added to the eluate,and the radioactivity was determined by liquid scintilla-tion spectrophotometry.

6.3.2. In vivo measurement of 3-HAO inhibitionMale Sprague-Dawley rats (150–200 g) were acclima-

tized to the animal quarters for at least 7 d beforeinitiation of the experiments. The rats were housed fiveanimals per cage under controlled conditions of tempera-ture (21 °C), relative humidity (55–65%) and light-dark

cycle (12:12 h, lights on 6 a.m.). Food and tap water wereavailable ad libitum. The rats were anaesthetized withenflurane (Efrane; Abbott) in a flow mixture of O2 andN2O. They were thereafter placed in a stereotaxic frame.The anaesthesia was maintained by free breathing into amask fitted over the nose of the rat of 3.5–5.0% ofenflurane maintained at a flow of 3 L/min of O2 and7 L/min of N2O. Body temperature was kept at 37 °Cusing a heating pad controlled via a rectal thermometer(CMA/12). The skull was oriented with the horizontalplane passing through bregma and lambda and throughthe intraural line. The skin over the skull was opened anda unilateral hole on the right side of the skull bone wasmade by drilling with a 1 mm burr. A Hamilton microlitresyringe (gauge 22S, 25µL) was lowered 3.9 mm verti-cally from the surface of the brain at AP 0 mm (bregma)and L 1.0 mm. An intracerebroventricular (i.c.v.) injec-tion of 3, 10, or 300 nmol of test compound together with10 nmol 3-HANA dissolved in 10µl 10 mM phosphatebuffer, containing 105 mM NaCl, 2.5 mM KCl, 1.18 mMMgCl2, 1.26 mM CaCl2 and 0.1% ascorbic acid wasperformed for 1 min. The syringe was then removed andthe skin was closed by wound clips. At various time-points after the i.c.v. injection, the rats were decapitated,and their brains rapidly removed and placed on anice-chilled petri-dish. The right hippocampus was dis-sected out and stored at –70 °C until QUIN determina-tions.

For QUIN determination, the samples were homog-enized in 10 volumes (w/w) 0.3 M formic acid with[18O]4-quinolinic acid added as internal standard. Afterfiltration, the samples were applied to a pretreated anion-exchanger column (0.5 mL Bio-Rad AG 1× 8, 200–400mesh, formate form). The column was washed with water(4 mL) before use. QUIN and internal standard wereeluted with 3 mL 6.0 M formic acid. After evaporation,the samples were esterified at 70 °C with hexafluoroiso-propanol and trifluoroacetylimidazole for 1 h. The esterswere partitioned between water (0.4 mL) andn-heptane(3 mL) by whirlimixing for 1 min followed by centrifu-gation at 1 000g for 5 min. The organic phase wastransferred to a new tube and washed with 0.4 mL ofwater by whirlimixing. After a second centrifugation, thewater was discarded and the samples evaporated in astream of nitrogen at room temperature to about 5 mL.Two µL of the final solution was assayed using gaschromatography coupled to a mass spectrometer(GC/MS). The GC (Carlo Erba MFC 500) was operatedin splitless injection mode, and a capillary column(NB-225, 25 m, i.d. 0.32 mm) was used. Helium wasemployed as carrier gas. The oven was run from70–135 °C, with a ramp of 20 °C/min. The MS (VG,

743

TS-250) was adjusted to record m/e 467 (QUIN) and m/e471 (internal standard) by negative chemical ionization.Methane was used as reagent gas. The peak height ratio(QUIN/internal standard) from standards with knownamounts of QUIN was calculated. The calibration curvewas plotted as peak height ratio against concentration ofQUIN in the standards. QUIN amount in the samples wasthen calculated from the standard curve.

Acknowledgements

The authors are grateful for the analysis of QUIN byMr Göran Fredriksson, for surgical assistance by Ms EvaVänerman and to Mr Göran Stening for skillful technicalassistance. 4-Halogenated 3-HANA analogues were gen-erously provided by Dr Barry Carpenter, Cornell Univer-sity.

References

[1] Bokman A.H., Schweigert B.S., Arch. Biochem. Biophys. 33 (1951)270–276.

[2] Roberts R.C., McCarthy K.E., Du F., Ottersen P., Okuno E.,Schwarcz R., J. Neurosci. 15 (1995) 1150–1161.

[3] Heyes M.P., Chen C.Y., Major E.O., Saito K., Biochem. J. 326(1997) 351–356.

[4] Alberati-Giani D., Ricciardi-Castagnoli P., Köhler C., Cesura A.M.,J. Neurochem. 66 (1996) 996–1004.

[5] Saito K., Markey S.P., Heyes M.P., Brain Res. 546 (1991) 151–154.

[6] Saito K., Chen C.Y., Masana M., Crowley J.S., Markey S.P., HeyesM.P., Biochem. J. 291 (1993) 11–14.

[7] Saito K., Markey S.P., Heyes M.P., Neurosci. Lett. 178 (1994)211–215.

[8] Heyes M.P., Saito K., Major E.O., Milstien S., Markey S.P., VickersJ.H., Brain 116 (1993) 1425–1450.

[9] Blight R.T., Cohen T.I., Saito K., Heyes M.P., Brain 118 (1995)735–752.

[10] Okuno E., Köhler C., Schwarcz R., J. Neurochem. 49 (1987)771–780.

[11] Malherbe P., Köhler C., Da Prada M., Lang G., Kiefer V., SchwarczR., Lahm H.W., Cesura A.M., J. Biol. Chem. 269 (1994)13792–13797.

[12] Stone T.W., Pharmacol. Rev. 45 (1993) 309–379.

[13] Lapin I.P., J. Neuronal Trans. 42 (1978) 37–43.

[14] Schwarcz R., Whetsell W.O. Jr., Mangano R.M., Science 219 (1983)316–318.

[15] Beal M.F., Ferrante R.J., Swartz K.J., Kowall N.W., J. Neurosci. 11(1991) 1649–1659.

[16] Huang Q., Zhou D., Sapp E., Aizawa H., Ge P., Bird E.D., VonsattelJ.P., Di Figlia M., Neuroscience 65 (1995) 397–407.

[17] Tatter S.B., Galpern W.R., Hoogeveen A.T., Isacson O., Neuroreport6 (1995) 1125–1129.

[18] Foster A.C., White B.J., Schwarcz R., J. Neurochem. 47 (1986)22–30.

[19] Heyes M.P., Jordan E.K., Lee K., Saito K., Frank J.A., Snoy P.J.,Markey S.P., Gravell M., Brain Res. 570 (1992) 237–250.

[20] Halperin J.J., Heyes M.P., Neurology 42 (1992) 43–50.

[21] Eastman C.L., Urbánska E., Löve A., Kristensson K., Schwarcz R.,Exper. Neurol. 125 (1994) 119–124.

[22] Heyes M., Saito K., Major E.O., Milstien S., Markey S.P., VickersJ.H., Brain 116 (1993) 1425–1450.

[23] Heyes M.P., Saito K., Milstien S., Schiff S.J., J. Neurol. Sciences133 (1995) 112–118.

[24] Todd W.P., Carpenter B.K., Schwarcz R., Prep. Biochem. 19 (1989)155–165.

[25] Manthey M.K., Pyne S.G., Truscott R.J.W., J. Org. Chem. 53 (1988)1486–1488.

[26] Manthey M.K., Pyne S.G., Truscott R.J.W., Biochim. Biophys. Acta1034 (1990) 207–212.

[27] Hansch C., Leo A., Hoekman D., Exploring QSAR: Hydrophobic,Electronic and Steric Constants, ACS, Washington DC, 1995.

[28] Pallas – CompDrug Chemistry Ltd.

[29] Dewar M.J.S., Zoenisch E.G., Healy E.F., Stewart J.J.P., J. Am.Chem. Soc. 107 (1985) 3902–3909.

[30] Ovality is defined as ovality = A/4p(3V/4p) 2/3, where A is themolecular surface area and V is the molecular volume.

[31] Cannon J.R., Cresp T.M., Metcalf B.W., Sargent M.V., VinciguerraG., J. Chem. Soc. C (1971) 3495–3504.

[32] Melikian A., Boigegrain R., Kan J.P., Soubrie P., Eur. J. Med.Chem./Chim. Ther. 25 (1990) 267–270.

[33] Calam C.T., Oxford A.E., J. Chem. Soc. (1939) 280–284.

[34] Schmidt U., Boekens H., Lieberknecht A., Griesser H., TetrahedronLett. 22 (1981) 4949–4952.

[35] Castelfranchi G., Perrotti E., Ann. Chim. 47 (1957) 1201–1214.

[36] Colbert J.C., Meigs W., Stuerke A.H., J. Amer. Chem. Soc. 56(1934) 2128–2130.

[37] Watanabe Y., Nakajima K., Seki T., Ozawa H., Chem. Pharm. Bull.18 (1970) 2208–2215.

744

Short communication

Phenylsulfonylnitromethanes as potent irreversible inhibitors of aldose reductase

Nada H. Saaba, Isaac O. Donkora, Libaniel Rodriguezb, Peter F. Kadorb, Duane D. Millera*aDepartment of Pharmaceutical Sciences, The University of Tennessee, Memphis, TN 38163, USA

bLaboratory of Ocular Therapeutics, National Eye Institute, National Institutes of Health, Bethesda, MD 20892, USA

(Received 27 October 1998; revised 4 March 1999; accepted 4 March 1999)

Abstract – Aldose reductase (AR) inhibition provides a viable pharmacologically direct mode for the treatment of diabetic complications. Wehave synthesized a series of N-4 substituted analogues (15–21) of the known aldose reductase inhibitor phenyl-sulfonylnitromethane. Thecompounds are potent inhibitors of AR with IC50s between 0.01 and 0.19µM. Some of the compounds are also potent affinity labels for AR.Compound19exhibits the highest and almost complete irreversible inhibition of AR known to date. © 1999 E´ditions scientifiques et médicalesElsevier SAS

aldose reductase / diabetic complication / phenylsulphonyl-nitromethane / irreversible inhibitors / aldose reductase inhibitors

1. Introduction

Over 50% of diabetics develop tissue-damaging com-plications [1]. These complications develop in tissuescapable of insulin-independent glucose uptake and resultin retinopathy, nephropathy, cataract, keratopathy, neur-opathy and angiopathy. Results of the recent DiabetesControl and Complications Trial suggest that tight controlof blood sugar levels can reduce the incidence andseverity of diabetic complications [1]; however, on apractical basis, tight control is difficult to maintain. Thishas spurred efforts toward the development of alternativetreatments with agents acting by mechanisms indepen-dent of the control of blood glucose. Aldose reductaseinhibitors (ARIs) have provided therapeutically usefulagents [2], which in long-term animal studies demon-strate beneficial prevention or delay in the onset andprogression of diabetic complications with no significantadverse effects [3]. The search for clinically useful ARIshas been an on-going process since the late 1960s. Thiseffort has led to the discovery of a number of structurallydiverse compounds as ARIs(figure 1).

Kinetic studies suggest that despite their structuralvariations, ARIs exhibit either uncompetitive or noncom-

petitive inhibition and bind to a site on the enzyme whichis independent of the substrate and NADPH cofactorbinding sites [4–6]. Using affinity-labelling studies, Ka-dor et al. [7] demonstrated that there are three distinctbinding sites on the AR enzyme, namely, the substratesite, cofactor site and inhibitor site. This observation is,however, at odds with recent X-ray crystallographicstudies involving the ternary complex between AR,NADPH and zopolrestat in which the inhibitor was foundcompletely sequestered into the substrate site [8]. In aneffort to study the nature of the inhibitor binding site ofAR, we previously reported a series of affinity labels(7–10) (table I) based on the reversible inhibitor alresta-tin [9]. Based on the remarkable irreversible inhibitoryactivity of 5-iodoacetamidoalrestatin (10) in our previousstudies [7, 9] and the possibility of using affinity labels tolocate the binding site(s) of ARIs on AR, we synthesizedaffinity labels (15–17) and (19–21) along with theirknown acetylated derivatives (14 and 18) (table I), de-rived from a new class (sulfonylnitromethanes) of potentAR inhibitors, and examined the compounds for ARinhibitory activity.

2. Chemistry

Sulfonylnitromethanes have been synthesized by threeprincipal synthetic routes. These include the sulfonation*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 745−751 745© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

of α-halo nitro compounds [10], the oxidation ofα-nitrosulfide [11] and the alkaline nitration of sulfones [12].Alkaline nitration, which is based on a free radical chainprocess rather than nucleophilic displacement of iodideby the nitro anion, was the most fruitful method for oursynthesis. The synthesis of affinity labels15–17 and19–21 was achieved as depicted infigure 2. The aminomoiety of dimethylaniline (22) was protected by treat-ment with acetic anhydride to give23,which was reactedwith chlorosulfonic acid to yield the corresponding sul-fonyl chloride25. Compound25 as well as the commer-cially available N-acetylsulfonyl chloride (24) were re-duced using sodium sulfite in the presence of sodiumbicarbonate buffer to give the corresponding sulfinicacids26 and27 [13]. The acids were converted to theirsodium salts and then reacted with two-fold excess ofnitromethane in the presence of iodine and sodiummethoxide at 0 °C to give the corresponding N-(acetyl-phenyl)sulfonyl nitromethane derivatives14 and18 [13,14]. Basic hydrolysis of the acetyl group afforded12 and13. The latter compounds were treated with thiophosgenein acetone to give the corresponding isothiocyanateanalogues15 and19, respectively. Treatment of12 or 13with chloro- or iodoacetic anhydride afforded the corre-sponding chloro- or iodoacetamido derivatives16 and20or 17 and21, respectively.

3. Results and discussion

The sulfonylnitromethane analogues were found to bepotent inhibitors of recombinant rat lens AR with IC50sranging between 0.01 to 0.19µM (table I). Subsequentgel filtration studies for irreversible binding also indi-cated that compounds15–17and19–21are potent affinitylabels for AR (table I). Comparison of the activities ofcompounds15–17 with that of compounds19–21 indi-cates thatorthodimethyl substitution potentiates irrevers-ible inhibition of AR in this class of compounds. Incontrast to the affinity labels (7–10, table I) derived fromalrestatin, the irreversible inhibitory activity for thepresent compounds is not directly linked to the chemicalreactivity of their electrophilic groups. Irreversible in-hibitory activity for both sets of compounds decreased inorder: isothiocyanato analogues> iodoacetamido ana-logues> chloroacetamido analogues. No correlation be-tween reversible and irreversible inhibitory activities wasobserved.

Results from mass spectrometry and molecular mod-elling studies [7], carboxymethylation studies [15], sitedirected mutagenesis studies [16], as well as kineticdata [4–6] suggest the possibility of multiple inhibitorbinding sites on AR which are distinct from the substratebinding site. Structurally diverse, selective affinity labels

Figure 1. Aldose reductase inhibitors.

746

can be useful tools for investigating the possibility ofmultiple inhibitor sites on AR. The present results indi-cate that compounds19 and21 are more potent affinitylabels than 5-iodoacetamidoalrestatin (10) and that appar-ently complete irreversible inhibition of AR can beachieved with compound19. Future determination of theexact site(s) of interaction with the enzyme by thesecompounds and identification of the reactive nucleophilicresidue(s) that bind to these compounds will provideinsight into the possibility of multiple inhibitor bindingsites on AR.

4. Experimental protocols

4.1. Chemistry

Melting points were determined on a Thomas-Hoovercapillary melting point apparatus and are uncorrected.Infrared (IR) spectra were recorded on a Perkin ElmerSystem 2000 FT-IR spectrophotometer. Proton nuclearmagnetic resonance (1H NMR) spectra were recorded ona Bruker AX 300 spectrometer. Chemical shift values are

reported in parts per million (δ) relative to tetramethyl-silane (TMS) as an internal standard. Spectral data areconsistent with assigned structures. Elemental analyseswere performed by Atlantic Microlab Inc., Norcross GA,and experimentally determined values are within± 0.4%of the theoretical values. Routine thin-layer chromatog-raphy (TLC) was performed on silica gel GHIF plates(Analtech Inc., Newark DE). Flash chromatography wasperformed on silica gel (Merck, grade 60, 230–400 mesh,60 Å). Acetonitrile (MeCN) was dried by distillation,dimethylformamide (DMF) was dried by distillation fromP2O5. All solvents (except anhydrous MeCN and DMF)were stored over 3 or 4 Å molecular sieves.

4.1.1. N-Acetyl-3,5-dimethylaniline23Acetic anhydride (60 mL) was added slowly to 3,5-

dimethylaniline (40 mL, 0.32 mol). The hot reactionmixture was allowed to cool to room temperature to give42 g (80%) of 23 as a white solid: m.p. 137–139 °C(Lit. [13] 138 °C). IR (KBr, cm–1) 1 663 (C=O).1H NMR(CDCl3) δ 7.12 (s, 2H, aromatic), 6.75 (s, 1H, aromatic),2.28 (s, 6H, 2CH3), 2.15 (s, 3H, COCH3), 1.63 (br, 1H,NHCO).

Table I. Reversible (IC50) and irreversible (%) inhibition of recombinant rat lens AR.

# R1 R2 Ra

7 NCS NDb 56.08 NHCOCH2Cl 0.55± 0.04 0.09 NHCOCH2Br 0.60± 0.05 46.010 NHCOCH2I 0.40± 0.02 89.011 H CH3 NDb NDb

12 NH2 H 0.19± 0.66c 0.013 NH2 CH3 0.13± 0.04c 0.014 NHCOCH3 H 0.09± 0.09 0.015 NCS H 0.05± 0.05 80.716 NHCOCH2Cl H 0.06± 0.07 58.017 NHCOCH2l H 0.09± 0.03 73.818 NHCOCH3 CH3 0.09± 0.09 0.019 NCS CH3 0.13± 0.03 99.520 NHCOCH2Cl CH3 0.01± 0.04 86.621 NHCOCH2l CH3 0.03± 0.001 96.8

a The alrestatin derivatives were previously reported [9].b ND, not determined.c Reported values from [13, 14].

747

Figure 2. Synthetic route to phenylsulfonylnitromethanes.

748

4.1.2. (4-Acetamido-2,6-dimethylphenyl)sulfonyl chlo-ride 25

In a three-necked round bottom flask fitted with amechanical stirrer was placed chlorosulfonic acid (100 g,0.85 mol). The flask was cooled (12–15 °C) by means ofan ice-bath and N-acetyl-3,5-dimethylaniline (28.07 g,0.172 mol) was added slowly over a 15 min period. Afterthe addition was complete, the mixture was allowed towarm to room temperature and then was heated to 60 °Cwith stirring. After 2.5 h, the mixture was added slowly toice-water mixture (330 mL) with stirring. The off-whiteprecipitate that formed was filtered and dried in adesiccator under vacuum overnight to give 22 g of crudeproduct. An analytical sample was recrystallized frombenzene: m.p. 157–159 °C. IR (KBr, cm–1) 3 322 (NH),1 683 (C=O), 1 360 (SO2) and 1 172 (SO2).

1H NMR(CDCl3)δ 7.65 (br, 1H, NH), 7.42 (s, 2H, aromatic), 2.72(s, 6H, 2CH3), 2.22 (s, 3H, COCH3).

4.1.3. (4-Acetamidophenyl)sulfinic acid26N-Acetylsulfanilyl chloride (11.9 g, 0.05 mol) was

added to a vigorously stirred solution of sodium bicar-bonate (10 g, 0.119 mol) and sodium sulfite (14.29 g,0.113 mol) in water (60 mL) at 70–80 °C. The mixturewas heated for 1 h and allowed to cool to room tempera-ture. The white precipitate that separated out was redis-solved in water and acidified with 60% sulfuric acid untila white precipitate appeared. The suspension was left in arefrigerator overnight. The solid was recovered by filtra-tion and recrystallized from water to give 7.23 g (72%) of26 as white needle-like crystals: m.p. 149–151 °C. IR(KBr, cm–1) 3 321 (NH), 1 668 (C=O), 1 320 (SO2),1 084 (SO2).

1HNMR (DMSO-d6) δ 10.20 (s, 1H,NHCO), 7.73 (d, 2H, aromatic), 7.58 (d, 2H, aromatic),2.06 (s, 3H, COCH3).

4.1.4. (4-Acetamido-2,6-dimethylphenyl)sulfinic acid27Sulfonylchloride25 (22 g, 0.084 mol) was transformed

to sulfinic acid27 as described above for the synthesis of26. The crude product was recrystallized from water togive 5.5 g (29%) of 27 as white crystals: m.p.158–159 °C. IR (KBr, cm–1) 3 321 (NH), 1 668 (C=O),1 320 (SO2), 1 084 (SO2).

1H NMR (CDCl3) δ 7.31 (s,2H, aromatic), 2.59 (s, 6H, 2CH3), 2.10 (s, 3H, COCH3).

4.1.5. (4-Acetamidophenyl)sulfonylnitromethane14Sodium metal (0.62 g, 0.027 mol) was dissolved in dry

MeOH and 26 (5.5 g, 0.027 mol) was added and themixture was stirred overnight. The solvent was removedunder reduced pressure and the residue was kept in adesiccator. Another sample of sodium metal (1.09 g,0.047 mol) was dissolved in dry MeOH, the solvent wasstripped off under reduced pressure, and the residue was

dissolved in dry DMF (15 mL) and cooled to 0 °C.Nitromethane (3.22 g, 0.052 mol) was dissolved in dryDMF (15 mL) and added slowly to the cooled NaOMesolution. The reaction mixture was stirred at 0 °C for15 min during which period a bright yellow precipitateseparated out. The previously prepared sodium salt of26was dissolved in dry DMF, cooled to 0 °C, and added tothe bright yellow mixture followed immediately by theaddition of iodine (6.08 g, 0.024 mol). The reaction wasstirred at 0 °C for 4 h and then allowed to warm to roomtemperature over 30 min. The dark solution was pouredinto ice-water and decolourized with sodium sulfite. Itwas acidified slowly to pH 1.5 with 2 N HCl and theresulting suspension was stored in a refrigerator over-night followed by filtration to recover the product whichwas dried and recrystallized from MeOH to give a 63%yield of 14: m.p. 220–222 °C (dec.) (Lit. [14]228–229 °C). IR (KBr, cm–1) 3 259 (NH), 1 674 (C=O),1 590 (NO2), 1 553 (NO2).

1H NMR (DMSO-d6) δ 10.54(s, 1H, NHCO), 7.90 (s, 4H, aromatic), 6.57 (s, 2H,CH2NO2), 2.13 (s, 3H, COCH3).

4.1.6. (4-Acetamido-2,6-dimethylphenyl)sulfonylnitro-methane18

Compound27 (2 g, 8.8 mmol) was transformed into18as described above for the synthesis of14. The off-whitesolid obtained was purified by column chromatographyover silica gel with hexane/ethyl acetate (1:2) as theeluant followed by recrystallization from ethanol to yieldwhite crystals: m.p. 177–179 °C (Lit. [13] 179–180 °C).IR (KBr, cm–1) 3 308 (NH), 1 677 (C=O), 1 534 (NO2).1H NMR (acetone-d6) δ 9.50 (s, 1H, NHCO), 7.59 (s, 2H,aromatic), 6.15 (s, 2H, CH2NO2), 2.60 (s, 6H, 2CH3),2.05 (s, 3H, COCH3).

4.1.7. (4-Aminophenyl)sulfonylnitromethane12A solution of 14 (0.2 g, 0.77 mmol) in 2 N NaOH

(2 mL) was heated for 1 h at 80 °C. The mixture was thenpoured into ice-water mixture (10 mL) containing aceticacid (0.3 mL) to yield a solid which was extracted withEtOAc, dried (MgSO4), and evaporated under reducedpressure. The residue was purified by column chromatog-raphy over silica gel with hexane/EtOAc (1:1) as theeluant followed by recrystallization from EtOH to give a78% yield of12: m.p. 128–130 °C. IR (KBr, cm–1) 3 396(NH2), 1 554 (NO2).

1H NMR (DMSO-d6) δ 7.51 (s, 2H,aromatic), 6.67 (s, 2H, aromatic), 6.44 (s, 2H, CH2NO2),6.30 (s, 2H, Ar-NH2).

4.1.8. (4-Amino-2,6-dimethylphenyl)sulfonylnitromethane13Compound18 (0.146 g, 0.51 mmol) was transformed

into 13 in 81% yield as described above for the synthesisof 12: m.p. 131–133 °C (Lit. [13] 132–133 °C). IR (KBr,

749

cm–1) 3 474 (NH2), 3 376 (NH2), 1 551 (NO2).1H NMR

(DMSO-d6) δ 7.54 (s, 2H, aromatic), 6.40 (s, 2H,CH2NO2), 2.62 (s, 6H, 2CH3), 6.32 (s, 2H, Ar-NH2).

4.1.9. (4-Isothiocyanatophenyl)sulfonylnitromethane15Thiophosgene (12 drops) was added to a solution of12

(0.06 g, 0.28 mmol) in dry acetone (4.5 mL) and themixture was stirred at room temperature for 2 h. Thesolvent was removed under reduced pressure and thecrude product was purified by column chromatographyover silica gel with hexane/EtOAc (1:1) as the eluant toyield 0.06 g (87%) of15 as a bright yellow solid: m.p.159–160 °C. IR (KBr, cm–1) 2 131 (NCS). 1H NMR(acetone-d6) δ 8.09 (d, 2H, aromatic), 7.72 (d, 2H,aromatic), 6.35 (s, 2H, CH2NO2). MS (EI, m/z) 258+.Anal. C8H6N2S2O4 (C, H, N, O, S).

4.1.10. (4-Isothiocyanato-2,6-dimethylphenyl)sulfonyl-nitromethane19

Thiophosgene (8 drops) was added to a solution of13(0.06 g, 2.4 mmol) in dry acetone (3 mL) and the mixturewas stirred at room temperature for 2 h. The solvent wasremoved under reduced pressure and the crude productwas purified by column chromatography over silica gelwith hexane/EtOAc (1:1) as the eluant to yield 0.065 g(93%) of19 as a white solid: m.p. 110–112 °C. IR (KBr,cm–1) 2 009 (NCS).1H NMR (acetone-d6) δ 7.35 (s, 2H,aromatic), 6.27 (s, 2H, CH2NO2), 2.68 (s, 6H, 2CH3). MS(EI, m/z) 286+. Anal. C10H10N2S2O4 (C, H, N, O, S).

4.1.11. (4-Chloroacetamidophenyl)sulfonylnitro-methane16

Chloroacetic anhydride (73 mg, 0.43 mmol) was addedto a solution of12 (0.06 g, 0.27 mmol) in dry CH3CN(4 mL) and the mixture was stirred at room temperaturefor 18 h. The solvent was concentrated to precipitate pure16 in 0.044 g (55%) yield as a white solid: m.p. 212 °C.IR (KBr, cm–1) 3 279 (NH), 1 684 (C=O).1H NMR(DMSO-d6) δ 10.86 (s, 1H, NHCO), 7.92 (d, 4H,aromatic), 6.60 (s, 2H, CH2NO2), 4.34 (s, 2H, CH2Cl).MS (EI, m/z) 292+. Anal. C9H9N2SO5Cl (C, H, N, O, S,Cl).

4.1.12. (4-Chloroacetamido-2,6-dimethylphenyl)sulfonyl-nitromethane20

Compound13 (0.05 g, 2.0 mmol) was treated as de-scribed for the synthesis of16 to yield 0.06 g (87%) of20as a bright yellow solid: m.p. 160–161 °C. IR (KBr,cm–1) 3 366 (NH), 1 723 (C=O).1H NMR (DMSO-d6) δ10.66 (s, 1H, NHCO), 7.53 (s, 2H, aromatic), 6.50 (s, 2H,CH2NO2), 4.31 (s, 2H, CH2Cl), 2.56 (s, 6H, 2CH3). MS(EI, m/z) 320+. Anal. C11H13N2SO5Cl (C, H, N, O, S, Cl).

4.1.13. (4-Iodoacetamidophenyl)sulfonylnitromethane17Iodoacetic anhydride (0.106 g, 0.29 mmol) was added

to a solution of12 (0.05 g, 0.23 mmol) in dry CH3CN(4 mL) and the mixture was stirred at room temperaturefor 18 h. The mixture was concentrated and CHCl3 wasadded and kept in a refrigerator overnight to precipitatepure17as pale yellow crystals in 95% yield: m.p. 165 °C.IR (KBr, cm–1) 3 356 (NH), 1 690 (C=O).1H NMR(DMSO-d6) δ 10.88 (s, 1H, NHCO), 7.50 (d, 4H,aromatic), 6.61 (s, 2H, CH2NO2), 3.87 (s, 2H, CH2I). MS(EI, m/z) 384+. Anal. C9H9N2SO5I (C, H, N, O, S, I).

4.1.14. (4-Iodoacetamido-2,6-dimethylphenyl)sulfonyl-nitromethane21

This compound was synthesized as described for thesynthesis of17by reacting iodoacetic anhydride (0.094 g,0.26 mmol) with13 (0.05 g, 2.0 mmol). Compound21was obtained in 95% yield as white crystals. M.p. 200 °C.IR (KBr, cm–1) 3 366 (NH), 1 723 (C=O).1H NMR(DMSO-d6) δ 10.68 (s, 1H, NHCO), 7.50 (s, 2H, aro-matic), 6.49 (s, 2H, CH2NO2), 3.84 (s, 2H, CH2I), 2.51 (s,6H, 2CH3). Anal. C11H13N2SO5I (C, H, N, O, S, I).

4.2. Biological studies

4.2.1. Enzyme purificationRecombinant rat lens AR was purified by a series of

chromatographic procedures as previously de-scribed [17]. Briefly, AR was released fromE. coli bysonication and the mixture was centrifuged at 10 000gfor 10 min. The supernatant was then subjected to gelfiltration on a Sephadex G-75 column (2.5× 90 cm),equilibrated with 10 mM imidazole-HCl buffer, pH 7.5containing 7 mM 2-mercaptoethanol and eluted with thesame imidazole buffer. The eluent was collected in220-drop aliquots (ca. 10 mL) and fractions containingAR activity were applied to a Matrex Gel Orange Aaffinity column (2.5× 15 cm). The affinity column waswashed with the imidazole buffer (ca. 500 mL) and theenzyme was eluted with the same imidazole buffercontaining 0.1 mM NADPH. Fractions eluted withNADPH were chromatofocused on a Mono P (HR 5/20)column developed at a flow rate of 1 mL/min withPolybuffer 74 (diluted 10-fold and containing 7 mM2-mercaptoethanol). The protein concentration of theeluent was monitored at 280 nm and peaks containing ARactivity were collected and concentrated on Centricon 10filters.

4.2.2. Enzyme assayReductase activity was spectrophotometrically assayed

on a Shimadzu UV 2100U spectrophotometer by follow-ing the decrease in the absorption of NADPH at 340 nm

750

over a 4-min period with DL-glyceraldehyde as sub-strate [9]. Each 1 mL cuvette contained equal units ofenzyme, 0.10 M Na, K phosphate buffer, pH 6.2, 0.3 mMNADPH with/without 10 mM substrate and inhibitor.Appropriate controls were employed to negate potentialchanges in the absorption of nucleotide and/or proteinmodification reagents or ARIs at 340 nm in the absence ofsubstrate.

4.2.3. Affınity bindingRat lens AR was passed through NAP-5 desalting

columns equilibrated with 0.1 M sodium phosphatebuffer, pH 7.6, to remove potentially reactive 2-mercapto-ethanol which stabilizes AR [9]. Equal aliquots of en-zyme were then combined with either reversible orirreversible inhibitor dissolved in 0.1 M sodium phos-phate buffer, pH 7.6, and the reaction was allowed toproceed at room temperature for 15 min. Reversible andunreacted inhibitor were then removed from the proteinby gel filtration with a NAP-5 desalting column with0.1 M phosphate buffer, pH 7.0, containing 10 mM2-mercaptoethanol. Residual AR activity was spectropho-tometrically measured. All experiments were conductedat least in triplicate.

Acknowledgements

This work was supported in part by NIH grant 1 R15EY0936-02 (I.O.D). We thank Mr John Miller for run-ning the Mass Spectra of the compounds.

References

[1] Diabetes Control and Complications Trial Research Group, NewEngl. J. Med. 329 (1993) 977–986.

[2] Tomlinson D.R., Stevens E.J., Diemel L.T., Trends Pharm. Sci.(England) 15 (1994) 293–297.

[3] Kador P.F., Robinson W.G. Jr., Kinoshita J.H., Ann. Rev. Pharmacol.Toxicol. 25 (1985) 691–714.

[4] Kador P.F., Sharpless N.E., Mol. Pharmacol. 24 (1983) 521–531.

[5] Okuda J., Miwa I., Inagaki K., Horie T., Nakayama M., Biochem.Pharmacol. 31 (1982) 3807–3822.

[6] Kador P.F., Nakayama T., Sato S., Smar M., Miller D.D., Enzymol.Mol. Bio. Carbonyl Metabolism 2 (1989) 237–250.

[7] Kador P.F., Lee Y.S., Rodriquez L., Sato S., Bartoszko-Malik A.,Abdel-Ghany Y.S., Miller D.D., Bioorg. Med. Chem. 3 (1995)1313–1324.

[8] Wilson K.D., Tarle I., Petrash M.J., Quiocho F.A., Proc. Natl. Acad.Sci. USA 90 (1993) 9847–9851.

[9] Smar M.W., Ares J.J., Nakayama T., Itabe H., Kador P.F., MillerD.D., J. Med. Chem. 35 (1992) 1117–1120.

[10] Troger J., Notle E., J. Prakt. Chem. 101 (1920) 136.

[11] Kharasch N., Cameron J.L., J. Am. Chem. Soc. 75 (1953)1077–1081.

[12] Truce W.E., Klingler T.C., Paar J.E., Feuer H., Wu D.K., J. Org.Chem. 34 (1969) 3104–3107.

[13] Brittain D.R., Brown S.P., Cooper A.L., Longridge J.L., Morris J.J.,Preston J., Slater L., UK Patent Application (GB 2227745), 1990.

[14] Kelley J.L., McLean E.W., Williard K.F., J. Hetero. Chem. 14 (1977)1415–1416.

[15] Liu S.Q., Bhatnagar A., Srivastavs S.K., Biochem. Biophy. Acta.1120 (1992) 329–336.

[16] Bohren K.M., Page J.L., Shankar R., Henry S.P., Gabby K.H., J.Biol. Chem. 266 (1991) 24031–24037.

[17] Old S.E., Sato S., Kador P.F., Carper D.A., Proc. Natl. Acad. Sci.USA 87 (1990) 4942–4945.

751

Short communication

QSAR study on antibacterial and antifungal activitiesof some 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones)using physicochemical, quantumchemical and structural parameters

Ömer Gebana, Hamide Ertepinara, Mine Yurtseverb, Seçkin Özdenc*, Fatma Gümüsd

aMiddle East Technical University, Faculty of Education, Department of Science Education, 06531 Ankara, TurkeybIstanbul Technical University, Faculty of Arts and Sciences, Department of Chemistry, Istanbul, Turkey

cAnkara University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 06100 Tandogˇan- Ankara, TurkeydGazi University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 06100 Ankara, Turkey

(Received 19 October 1998; revised 3 March 1999; accepted 11 March 1999)

Abstract – This work demonstrated the quantitative structure-activity relationships of 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) using quantum chemical parameter R(I), hydrophobicity descriptor and structural parameters. Semiempirical molecular orbitalcalculations were used to determine the quantum chemical parameter R(I), which is the electron density of HOMO at the sulfur and oxygenin position 1 of the compounds investigated, divided by the orbital energy of HOMO. It was shown that the electron density of HOMO at thesulfur and oxygen of the molecules was strongly related to the biological activities of these molecules. The results obtained from the QSARapplication indicated that there was a parabolic dependence between the biological activities and the R(I) index. The structural factor IY whichshows the presence of a sulfur atom in position 1 was the dominant predictor for the antibacterial and antifungal activities. On the other hand,the other structural variable IX which shows the presence of a sulfur atom double bonded to the C atom in position 5 caused a decrease, butthe hydrophobicity of the whole molecule (Σ) caused an increase in activity. © 1999 E´ditions scientifiques et médicales Elsevier SAS

3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) / QSAR / antibacterial activity / antifungal activity / quantum chemicalcalculations

1. Introduction

A series of compounds having the 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5-(4H)-ones(thiones) structure(figure 1) was found to show antibacterial and antifungalactivities [1]. These compounds were tested in vitro fortheir ability to inhibit the growth of representative gram-positive and gram-negative bacteria and various fungalspecies.

In order to identify the substituent effects to thechemical reactivity and the biological activity, we haveexamined the QSAR of a series of these compounds usingthe quantum chemical index R(I), physicochemical andstructural parameters.

Recently, much research which concerns the quantumchemical parameters in QSAR studies have been re-ported [2–6]. Theoretical calculations have been used for

the description of the mechanism of interactions at themolecular level. The electrophilic and nucleophilic super-delocalizability of nitrogen [6], sulfur and oxygenatom [5] and electron densities on the sulfur and oxygenatoms [4] in the investigated structures have been used as*Correspondence and reprints

Figure 1. Structures of compounds1–14.

Eur. J. Med. Chem. 34 (1999) 753−758 753© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

the independent parameters in QSAR analysis. The re-sults obtained from these researches showed that the maintype of interaction at the binding site should be viainteraction of the lone pairs of heteroatoms in thecompounds which were investigated with any bindingsite(s).

2. Chemistry

2.1. Chemicals and biological activity

The compounds investigated in this work were synthe-sized previously [7–9]. The antibacterial activity againstStaphylococcus aureus, Bacillus substilisandPseudomo-nas aeruginosaand antifungal activity againstCandidaalbicans, Candida pseudotropicalis, Candida parapsilo-sis and Crytococcus neoformanswere determined [1]using the tube dilution method, and given as minimuminhibitory concentration (MIC, g/mL). The potency wasdefined as log1/C in the QSAR analysis, where C is themolar MIC value of the compound and is used as thedependent variable in the QSAR study. The log1/C valuesare given intable I.

2.2. Molecular orbital calculations

The AM1 [10] molecular orbital calculations werecarried out using the MOPAC program [11] running on aSilicon Graphics Work station (IRIS 4D, INDIGO,POWER INDIGO 2). Initially the geometries of themolecules were fully optimized. After optimization, ISFcalculations with VECTORS option were carried out toprint all eigen values, eigen vectors, and HOMO energies(table II).

2.3. Quantitative structure-activity relationships (QSAR)

QSAR of 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4H)-ones(thiones) was investigated using the quantum-chemical parameter R(I), physicochemical parameter (π)and structural parameters.

The electron density of HOMO at active sites of themolecules (position 1) are strongly related to the antibac-terial and antifungal activities of the molecules. Theseactive sites are important in such a way that they bind tosome specific SH enzymes to achieve the role of inhibi-tion. Previously Nakayama et al. [4] introduced an indexR(I) which shows the electron density of HOMO at atomi which is one of the active sites (in the molecule) playinga role in the possible chemical reactions. It was defined as

R~ I ! = ~ fr~ i !/ − EHOMO ! × 102

where fr(i) is the frontier electron density of HOMO atatom i and –EHOMO is the energy of HOMO in eVmeasured from the zero level, f(i) is the measure of thedelocalization of electron density of HOMO on atom iand –EHOMO is the energy of the highest occupied (mostenergetic) molecular orbital. These two parameters are ameasure of the relative reactivity of HOMO at atom iwithin a single molecule. Then, the ratio of fr(i) and–EHOMO at atom i can be related to the biologicalreactivity of a series of molecules concerned.

Oxygen and sulfur atoms embedded in the five mem-bered ring of the concerned structure were thought to bethe effective part of these compounds. Then, to investi-gate the contribution of this index to the biologicalactivities of the examined molecules, the R(I) values atthe S and O atom in position 1 were calculated using theMOPAC program and the values are listed intable II.

Table I. Biological activities of the compounds1–14.

Compound R1 R2 S.a.a B.s.b P.a.c C.a.d C.p.e C.p.f C.n.g

1 Ph Me 3.58 3.58 3.58 3.89 3.89 3.89 4.192 2-Pyr Me 3.89 3.89 3.89 4.19 4.19 4.19 4.193 2-Pyr Ph 4.01 4.01 4.01 4.31 4.31 4.31 4.314 2-Pyr Et 3.92 4.22 4.22 3.92 4.22 4.22 4.225 4-Pyr p-tolyl 3.73 3.73 3.73 4.03 4.03 4.03 4.036 Ph H 3.89 3.89 3.89 3.89 4.19 3.89 4.197 Ph p-tolyl 4.06 4.06 4.06 4.06 4.36 4.06 4.368 2-Pyr Et 2.72 2.72 3.02 3.02 3.02 3.02 3.329 2-Pyr Me 2.99 2.99 2.99 3.29 3.29 3.29 3.2910 2-Pyr n-pr 3.05 3.05 3.05 3.35 3.35 3.35 3.6511 Me Ph 2.98 3.28 2.98 3.28 3.28 3.28 3.2812 4-Pyr Et 3.32 3.32 3.32 3.02 3.02 3.02 3.0213 4-Pyr Me 2.99 2.99 2.69 2.99 3.29 3.29 3.2914 2-Pyr Ph 3.11 3.11 3.11 3.41 3.41 3.41 3.41

aS. aureus,bB. subtilis,cP. aeruginosa,dC. albicans,eC. pseudotropicalis,fC. parapsilosis,gC. neoformans.

754

Concerning physicochemical parameters, the hydro-phobicity of the compounds (π) were used in the QSARanalysis.π values of the substituents were taken from thetables given by Hansch and Leo [12] andΣ was used forthe hydrophobicity of the whole molecule. The structuralparameter (IY) expresses the presence of S and O atomsin position 1. IY is defined as 1 in the case of the presenceof S in the position 1 and 0 for the O in the same position.The other structural variable (IX) represents the atomwhich is double bonded to the C atom in position 5.Similarly IX has a value of 1 for the S atom and 0 for theO atom.Table II shows the parameters used in this study.

3. Results and discussion

3.1. R(I) index and the biological activities of 3-4-disubsti-tuted 1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones)

Antibacterial and fungicidal activities of these com-pounds are given intable I. From table I it can be seenthat molecules that contain O in position 1 (8–14) havelower activities than the molecules which have S inposition 1 (1–7). Similarly the R(I) index is much lowerfor the O containing molecules than the others. Accord-ingly there is a correlation between R(I) and biologicalactivities measured. Namely, sulfur containing moleculesin position 1 have more electron density on the sulfuratom under consideration coming from HOMO, whichmakes that region more susceptible to chemical reactions,especially to the ones in which binding to another atom orelectron transfer reactions are considered.

3.2. QSAR

Correlation and regression analyses of the QSAR studywere done on a microcomputer using the SPSS/PC [13].

In the equations, the figures in parantheses are thestandard errors of the regression coefficients,n is thenumber of compounds, r is the multiple correlationcoefficient, s is the standard error of estimate, and F is theratio of regression and residual. The level of significanceaccepted in all of the analyses was 0.05.

The best equation(s) was selected among other equa-tions by considering the various statistical crite-ria [14–17]. All possible combinations of parameterswere considered, except that square terms were onlyallowed in equations containing the corresponding linearterms. This method provided a total of 22 possibleequations for each activity. The results showed that therewere two best equations including the same terms foreach activity. The first best fitted equations included R(I),R(I)2, Σ, and IX for the activities againstS. aureus, B.subtilis, P. aeruginosa, C. albicans, C. pseudotropicoli,C. parapsilosis,and C. neoformans.The second bestfitted equation for each activity included only the struc-tural parameter IY. Table III shows the best equations forthe antibacterial and antifungal activities of the examinedstructures. In these equations,n represents the number ofcompounds analysed, r, the multiple correlation coeffi-cient, s, the standard deviation, and F, the ratio ofregression and residual. The figure in parantheses is the95% confidence interval.

The equations including R(I), R(I)2, Σ, and IX togetherfor the activities againstS. aureus, B. subtilis, P. aerugi-nosa, C. albicans, C. pseudotropicoli, C. parapsilosis,and C. neoformanswere all significant because theoverall F statistics for these equations were statisticallysignificant, and they have high r values (about 0.90).Also, the individual F statistics for the coefficients ofR(I), R(I)2, Σ, and IX in these equations were significantat P < 0.05 (table IV).

Table II. QSAR parameters examined in this study for the compounds1–14.

X Y EHOMO(eV) R(I) ^π IX IY

1 O S –9.105 5.37 2.52 0 12 O S –9.559 4.32 1.06 0 13 O S –9.310 1.03 2.46 0 14 O S –9.560 4.37 1.52 0 15 O S –9.410 0.83 3.01 0 16 S S –8.961 3.55 1.96 1 17 S S –8.824 3.55 4.65 1 18 S O –9.067 1.04 1.52 1 09 S O –9.008 1.03 1.06 1 010 S O –9.059 1.03 2.05 1 011 S O –9.017 0.81 2.52 1 012 S O –9.274 1.06 1.34 1 013 S O –9.285 1.05 0.88 1 014 S O –8.984 0.93 2.46 1 0

755

The predictor variables R(I), R(I)2, Σ, and IX accountedfor a significant portion of the variance in the activities.Also, each predictor contributed a significant proportionof the additional variance in the presence of the othervariables for each activity. The structural parameter IX

induced some influences on the activity. But it showedthat S, which is double bounded to the C atom in position5, caused a decrease in the activity. R(I) and its squareproduced an additive contribution to the activity. Therewas a parabolic dependence between the biological ac-tivities and the R(I) index. The parabolic relationship ofthe activity againstS. aureuson the reactivity of HOMOon the sulfur and oxygen of 3,4 disubstituted 1,2,4-oxa(thia)-diazole-5(4 H)-one(thione) derivatives suggeststhat log1/C initially increases as the reactivity of HOMOincreases up to an ideal R(I) value, and then it decreasesafter this point.Σ is the summation of the hydrophobicparameter of the substituents R1 and R2 in the observed

structure and its positive coefficient suggested that higherhydrophobicity was favourable for the activity.

An examination of the intervariable correlations in-volving all predictor variables intable Vdemonstrates therelatively large correlation between R(I) and IY (r = 0.71)(colinearity).

When high colinearity exists, the regression analysesusing the given set of independent variables can not be

Table III. The best equations.

Number Equation n r s F p

The best equations againstS. aureus1 log1/C = 0.85 (± 0.20) IY + 3.02 14 0.94 0.17 84.61 0.00002 log1/C = 1.05 (± 0.46) R(I) – 0.18(± 0.085) R(I)2 + 14 0.95 0.17 20.19 0.0002

0.14 (± 0.11) π – 0.66 (± 0.28) IX + 2.59

The best equations againstB. subtilis3 log1/C = 0.85 (± 0.24) I Y + 3.06 14 0.91 0.20 58.39 0.00004 log1/C = 1.05 (± 0.51) R(I) – 0.17 (± 0.09) R(I)2 + 14 0.94 0.19 17.49 0.0003

0.14 (± 0.12) π – 0.68 (± 0.31) IX + 2.65

The best equations againstP. aeruginosa5 log1/C = 0.89 (± 0.23) I Y + 3.02 14 0.92 0.20 68.94 0.00006 log1/C = 1.13 (± 0.44) R(I) – 0.19 (± 0.079) R(I)2+ 14 0.96 0.17 26.65 0.0001

0.14 (± 0.11) π – 0.72 (± 0.27) IX + 2.57

The best equations againstC. albicans7 log1/C = 0.85 (± 0.19) I Y + 3.19 14 0.94 0.17 87.35 0.00008 log1/C = 0.75 (± 0.45) R(I) – 0.13 (± 0.083) R(I)2 + 14 0.95 0.17 21.16 0.0001

0.17 (± 0.11) π – 0.76 (± 0.28) IX + 3.06

The best equations againstC. pseudotropicalis9 log1/C = 0.93 (± 0.18) I Y + 3.24 14 0.95 0.16 121.29 0.000010 log1/C = 1.07 (± 0.44) R(I) – 0.18 (± 0.08) R(I)2 + 14 0.96 0.17 27.45 0.0000

0.18 (± 0.11) π – 0.71 (± 0.27) IX + 2.79

The best equations againstC. parapsilosis11 log1/C = 0.85 (± 0.19) I Y + 3.24 14 0.94 0.16 99.18 0.000012 log1/C = 0.82 (± 0.38) R(I) – 0.14 (± 0.069) R(I)2 + 14 0.97 0.15 30.99 0.0000

0.13 (± 0.09) π – 0.79 (± 0.23) IX + 3.14

The best equations againstC. neoformans13 log1/C = 0.89 (± 0.18) I Y + 3.32 14 0.95 0.15 120.33 0.000014 log1/C = 0.79 (± 0.46) R(I) – 0.12 (± 0.085) R(I)2 + 14 0.95 0.17 21.67 0.0001

0.18 (± 0.11) π – 0.61 (± 0.28) IX + 2.99

Table V. Correlation matrix of all predictor variables used in theequations.

R(I) ^π IY IX

R(I) 1.000^π 0.121 1.000IY 0.712 0.399 1.000IX –0.483 –0.033 –0.745 1.000

756

performed effectively [14, 18]. Also, the correlationsbetween the biological activities and predictor variablesdemonstrates the highest correlation between the biologi-cal activities and structural parameter IY (r values about0.90). For these reasons, the variable IY was entered intothe correlation equation separately, and the second bestequations were obtained(table III). F-test values for eachactivity were significant. It showed the presence of thesulfur atom in position 1 was a dominant predictor for theactivities againstS. aureus, B. subtilis, P. aeruginosa,C. albicans, C. pseudotropicoli, C. parapsilosis,andC. neoformans.

In order to judge the validity of the predictive power ofthe QSAR, a cross-validation method was applied to theoriginal data set for equations 1–14 and the resultingPRESS (predicted residual sum of squares) for eachequation was calculated according the following equa-tion [14, 19].

PRESS= (i

@ ~ yi − yi !2 /~ 1 − hii !

2#

here yi and yi are the response (activity) values ofobservation i (i = 1,2,..., n), observed and calculated bythe best equation, respectively. The diagonal elements ofthat matrix are denoted byhii in the equation, andcalculated by the SPSS computer program. The calcu-lated overall cross-validated r2 and standard deviationvalues for each equation are given intable VI.

The cross validation technique was used to check thevalidation of these equations, because this techniqueseems to substantially reduce the probability of chancecorrelation relative to multiple regression. In the presentstudy, cross validation did not confirm some of thefour-term equations. The cross-validated r2 values had amaximum at one-term equations. Cross validation resultsusing IY as the only independent variable were muchbetter than those including four parameters R (I), R(I)2, Σ,and IX. One parameter equations were very stable,leading to cross-validated values in the range between0.768–0.877, whereas the four-parameter equations gavecross-validated values between 0.383–0.806. Thisshowed that one-term equations indicated higher predic-tive ability, as shown by cross validation. On the otherhand, the greater the number of variables tested, thegreater role chance will play in the observed correla-tion [20]. With an increasing number of observations, theprobable degree of chance correlation is steadily reduced.

Table IV. Thet statistics (P < 0.05) for the coefficients of variablesin the best equations 2, 4, 6, 8, 10, 12, 14.

t P

Equation 2R(I) 5.128 0.0006R(I)2 4.688 0.0011^π 2.850 0.0191IX 5.264 0.0005

Equation 4R(I) 4.652 0.0012R(I)2 4.228 0.0022^π 2.539 0.0317IX 4.921 0.0008

Equation 6R(I) 5.863 0.0002R(I)2 5.333 0.0005^π 2.974 0.0156IX 6.085 0.0002

Equation 8R(I) 3.720 0.0048R(I)2 3.454 0.0072^π 3.523 0.0065IX 6.223 0.0002

Equation 10R(I) 5.556 0.0004R(I)2 5.043 0.0007^π 3.695 0.0050IX 6.062 0.0002

Equation 12R(I) 4.925 0.0008R(I)2 4.547 0.0014^π 3.139 0.0119IX 7.757 0.0000

Equation 14R(I) 3.863 0.0038R(I)2 3.300 0.0092^π 3.530 0.0064IX 4.880 0.0009

Table VI. The calculated r2 PRESSand sPRESS.

r2 PRESS s PRESS

Equation 1 0.831 0.201Equation 2 0.472 0.409Equation 3 0.768 0.242Equation 4 0.756 0.286Equation 5 0.798 0.234Equation 6 0.751 0.299Equation 7 0.836 0.198Equation 8 0.578 0.366Equation 9 0.877 0.185Equation 10 0.637 0.367Equation 11 0.853 0.186Equation 12 0.806 0.246Equation 13 0.877 0.177Equation 14 0.383 0.457

757

The number of variables correlating by chance increasesas the number of observations decreases. In this study, forfour parameter equations to be tested, the requirednumber of observations to avoid undue risk of chancecorrelations can be insufficient. For these reasons, oneparameter equations were sound models, and explainedthe data better.

The other one-term, two-term, three-term, four-term,and five-term equations were significant, but most ofthem have lower r values and higher s values than theequations given intable III. However, these one-term,two-term, three-term, four-term, and five-term equationscontain at least one coefficient with a poor F statistic(P > 0.05).

4. Conclusion

The activity contributions of 3,4 disubstituted-1,2,4-oxa(thio)-diazole-5(4H)-one(thione) derivatives indi-cated that S containing compounds in position 1 havehigher biological activities against some bacteria andfungal species. QSAR studies which concerned the quan-tum chemical, physical and structural parameters con-firmed this result.

The highest correlation between the biological activi-ties and structural parameter IY revealed that the maininteractions between the binding sites should be on thelone pair electrons of oxygen and sulfur atoms in position1 of the five membered ring of some 3,4 disubstituted-1,2,4 oxa(thio)-diazole-5(4 H)-one(thione) derivatives.

References

[1] Gümüs F., Dürüst Y., Dürüst N., Abbasoglu U., Pharmazie 48 (1993)867–868.

[2] Greco G., Novellino E., Silipo C., Vittoria A., Quant. Struct.-Act.Relat. 11 (1992) 461–477.

[3] Mekenyan O.G., Veith G.D., Bradbury S.P., Russom C.L., Quan.Struct.-Act. Relat. 12 (1993) 132–136.

[4] Nakayama A., Hagiwara K., Hashimoto S., Shimoda S., Quant.Struct.-Act. Relat. 12 (1993) 251–255.

[5] Sener E., Turgut H., Yalçin Y., Ören Y., Türker L., Çelebi N., AkinA., Int. J. Pharm. 110 (1994) 109–115.

[6] Veith G.D., Mekenyan O.G., Quant. Struct.-Act. Relat. 12 (1993)349–356.

[7] Agirbas H., Dürüst Y., Sümengen D., Sulfur and Silicon 66 (1992)321–324.

[8] Dürüst Y., Agirbas H., Sümengen D., Phosphorus Sulfur and Silicon62 (1991) 47–51.

[9] Sümengen D., Agirbas H., Dürüst Y., Dogan N., Chim. Acta Turc. 20(1992) 17–23.

[10] Dewar M.J.S., Stewart J.J.P., Zocbisch E.G., Healy E.F., J. Am.Chem. Soc. 107 (1985) 3902–3909.

[11] Stewart J.J.P., MOPAC Manual, Frank J. Seiler Research Labora-tory, United States Air Force Academy, USA, (1990).

[12] Hansch C., Leo A., Substituent Constants for Correlation Analysis inChemistry and Biology, John Wiley and Sons, New York, 1979.

[13] Norusis M.J., The SPSS Guide to Data Analysis for SPSS/PC+ (2nded.), SPSS Inc., Chicago, 1991, pp. 849–872.

[14] Rawlings J.O., Applied Regression Analysis: A Research Tool,Wadsworth Inc., California, 1988, pp. 327–336.

[15] Anderson T.W., Stanley L.S., An Introduction to the StatisticalAnalysis of Data, Houghton Mifflin Company, Boston, 1978, pp.619–637.

[16] Darper N.R., Smith H., Applied Regression Analysis, John Wileyand Sons Inc., New York, 1981, pp. 294–342.

[17] Howell D.C., Statistical Methods for Psychology, PWS-Kent Pub-lishing Company, Boston, 1987, pp. 463–507.

[18] Norman H.I., Hull C.H., Jenkins J.G., Steinbrenner K., Bent D.H.,SPSS Statistical Package for the Social Sciences, McGraw Hill Inc,New York, 1975, pp. 340–341.

[19] Wold S., Quant Struct.-Act. Relat. 10 (1991) 191–193.

[20] Topliss J.G., Costello, R.J., J. Med. Chem. 15 (1972) 1066–1068.

758

Laboratory note

Hydroxypyridone iron chelators, L1 and analogues. An17O NMR study

Giovanni Cerioni*a, Antonio Plumitallob, Maria G. Batzellac, Francesca Loccid

aDipartimento di Scienze Chimiche, Università di Cagliari, Cittadella Universitaria di Monserrato,SS 554 Bivio per Sestu, I-09024 Monserrato (Ca), Italy

bDipartimento Farmaco Chimico Tecnologico, Università di Cagliari, Via Ospedale 72, I-09124 Cagliari, ItalycCentro Trasfusionale, Ospedale di San Gavino Monreale, Via Roma 236, I-09037 San Gavino Monreale, Italy

dDipartimento di Citomorfologia, Università di Cagliari, Cittadella Universitaria di Monserrato,SS 554 Bivio per Sestu, I-09024 Monserrato (Ca), Italy

(Received 29 October 1998; revised 3 March 1999; accepted 3 March 1999)

Abstract – Some iron chelators of the class of hydroxypyridones, including the experimental drug 1,2-dimethyl-3-hydroxy-4-pyridone (L1),have been studied by17O NMR spectroscopy. The reciprocal influence of the keto and hydroxy groups has been examined in order to gaina better comprehension of their bonds. A strong intramolecular hydrogen bond between 3-OH and 4-CO groups has been observed for L1 andits analogues. A comparison has also been made on the data of three different solvents. © 1999 E´ditions scientifiques et médicales ElsevierSAS

iron chelators / L1 / hydrogen bond / 17O NMR

1. Introduction

Humans have a very limited capacity for iron excre-tion [1], even when iron stores are markedly increased.Thus, congenital or acquired iron-loading anaemias cancause severe tissue damage and eventual death, unless theiron overload is removed. One such disease,â-thalassaemia, still has a very large diffusion in ourregion of Sardinia. Desferrioxamine (DFO) is the onlyiron-chelating agent currently in widespread clinical use.Its administration, however, requires cumbersome andexpensive methods of slow subcutaneous or intravenousinfusion. This fact has expedited the research for oral ironchelators [2]. One of the most promising among them,1,2-dimethyl-3-hydroxy-4-pyridone (L1) [2, 3], is cur-rently in clinical tests by our group.

Dioxobidentate ligands are generally iron (III)ligands [4], and the majority of them incorporate thecarbonyl and the hydroxy groups. It would be interestingto directly investigate these groups, which actually bondto the metal atom. Moreover, one of the most usefulconcepts [5] to develop new chelating agents is the theory

of hard and soft bases [6]. Its application is helped by aknowledge of the electronic (and charge) distribution, e.g.in the bases [5].

A very valuable method to study oxygenated groups is17O NMR spectroscopy [7], and both carbonyl and hy-droxy groups have been extensively studied by thistechnique [8–11]. We thus decided to apply such amethod to the hydroxypyridone iron chelators, L1 andanalogous molecules, as a starting point for a systematicstudy of actual and possible new iron chelating drugs.

2. Chemistry

Structures of studied compounds are shown infigure 1and their17O NMR spectroscopic data are shown intableI. The OH structure has also been used for the groups inpositions 2 or 4, without any sign of a preference betweenthe hydroxyl or the carbonyl structure.

17O NMR shifts of the 3-OH group in H2O/D2O arequite similar for all the derivatives bearing this groupexcept for5. The observed shifts are in the range of thephenols [12–15], albeit1–3 are among the most shieldedones.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 759−763 759© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

The comparatively strong deshielding noted for5 isonly partly due to the lack of an intramolecular hydrogenbond with a contiguous carbonyl group and/or to adifferent extent of intermolecular hydrogen bonding withthe solvent. The intramolecular hydrogen bond betweenthe 3-hydroxy and 4-keto groups of1–3 should scarcelyinfluence the chemical shift of the 3-hydroxy group,within 10–15 ppm, as discussed e.g. in two studies onsalicylaldehydes and salicylanilides by Boykin and co-workers [12, 13].

Phenol itself in H2O has aδ = 77 ppm shift [14]. Itsshift ranges, in some solvents, from 73.5 (MeCN) [12] to83.6 ppm (DMSO-d6) [15]. All these values are thus moresimilar among themselves, and to those obtained for5 inpyridine-d5 and in CDCl3, than to the shift of5 inH2O/D2O. These data on phenol also suggest that inter-molecular hydrogen bonding is not a sufficient cause forthe observed deshielding.

Without re-opening a somewhat old and controversialproblem [16–20] on the real structure of5, we agree onan equilibrium, in H2O solution, between5a and5b withan estimated ca. 1:1 ratio [20](figure 2).

According to the Karplus-Pople equation [21], struc-ture5b with an increased electronic density at the oxygen

Figure 1. Structures of studied compounds.

Table I. 17O NMR chemical shifts (ppm) of studied compounds in H2O/D2O, C5D5N and CDCl3.a

Compound 2 OH 2 OH 2 OH 3 OH 3 OH 3 OH 4 OH 4 OH 4 OH(H2O/D2O) (C5D5N) (CDCl3) (H2O/D2O) (C5D5N) (CDCl3) (H2O/D2O) (C5D5N) (CDCl3)

1 54 53.5 50.6 224 306.2 280.32 60 51.0 47.5 229.6 305.0 277.43 59 53.6 52.1 227.5 303.0 280.34 224.7 268.6 245.75 105.2 77.7 78.06 266.3 251.9 ca. 2907 ca. 215 252.2 b ca. 65 69.8 b8 218 272 b 104 110 b

a) Linewidths range 200–600 Hz, but for5 (880 Hz) and6 (ca. 3 000Hz) in CDCl3 and for7 and8 in H2O/D2O (not measurable, very faintsignals).b) Insoluble, not measured.

Figure 2. Structures of compound5.

760

atom should cause an upfield shift. The observeddeshielding can be due to a considerable weight ofresonance for structures5c–e, which are alike to thecommonly accepted resonance structures of the phenoland phenate anion. Actually, sodium phenate in H2Osolution, in a trial experiment, has given a strongdeshielding (146 ppm) compared to phenol (ca. 77 ppm).An important C–O double bond character has thus to betaken into account for the observed deshielding of5.

Compounds 1–3 and 7, in H2O/D2O, for their3-hydroxy groups, show a narrow shift range, 54–65ppm, and are thus more shielded than most other phe-nols [12–15]. Several causes could co-operate to give thisresult. As discussed, the influence of intramolecularhydrogen bonding can partly account for this shielding,even if a possible inductive effect of the ring nitrogenshould operate in an opposite sense. Whilst to the best ofour knowledge data for meta-substituted phenols arelacking in the literature, an indication can be obtainedfrom the reported data on anisoles [23]. In their case, ifwe assume ametanitro group as a comparable substituteof the ring nitrogen of our compounds, a 15 ppmdeshielding has been observed. Moreover,para-substitued phenols are similarly deshielded by electron-withdrawing groups [12]. The two contributions to thechemical shifts seem opposite and comparable in magni-tude. The influence of the adjacent “carbonyl” group isthus apparent, irrespectively of its 2– or 4– position. Itspresence causes the resonance structures with a C–Odouble bond in the 3– position to become negligible,allowing an understanding of the shielding observedcompared to other phenols [12–15]. This point is alsoimportant for Fe(III) co-ordination, since the hardness asa donor of the oxygen atom is increased.

Derivatives 4 and 6 are known to exist in solutionmainly as 2– and 4–pyridones [20]. The17O NMRspectrum of4 in MeCN has been reported [24] and itsshift (269.5 ppm) is in good agreement with our value inpyridine-d5 (table I). The observed shifts of the4-carbonyl group of derivatives1–3 are thus in substan-tial agreement with the usually accepted pyridone struc-ture of these compounds. We observe that in H2O/D2Othe C=O groups of1–3 are much more upfield than in theother two solvents, particularly in pyridine-d5 (table I).On the contrary, 3–OH groups do not show any signifi-cant or systematic variation with the solvent. All theseobservations are coherent with a strong intermolecularhydrogen bond formation, at the carbonyl level, inH2O/D2O but without a rupture of the intramolecularhydrogen bond. An alternative and/or complementary

cause of the observed shielding could be an increasedimportance of the resonance formulas type1b for 1–3(figure 3).

Again, structures type1b should increase the hardnessof the oxygen atom.

The importance of the intermolecular hydrogen bond isevident by4, which shows the same trend of variation ofthe shifts with all the solvents. Its∆d values(table II)however are smaller than those of the other derivatives.Data for 6 are difficult to interpret. On going fromH2O/D2O to less polar solvents, a retreat towards thehydroxy tautomer would be predicted [25]. On the otherhand, a strong intermolecular hydrogen bonding betweenthe CO group and the solvent is present in H2O/D2O andits effect is shielding [26]. The two effects are thuscounteracting and it is not possible to know a priori whichone will prevail. Moreover, in CDCl3 there is surely anincreased importance of self-association, as shown by thequite large value of the line-width of the signal.

Spectra of8 clearly show that only one of the substitu-ents is in the keto form. In accordance with litera-ture [27], we can assign a 2-pyridone structure at thisderivative.

Figure 3. Resonance formulas of compound1.

Table II. Chemical shifts differences (∆δ)a of CO groups ofselected compounds.

Compound ∆δP-W (ppm) ∆δC-W (ppm) ∆δP-C (ppm)

1 +82.2 +56.3 +25.92 +75.4 +47.8 +27.63 +75.5 +52.8 +22.74 +43.9 +21.0 +22.96 –14.4 +23.7 –38.1

a) Subscripts: P–W, chemical shifts difference between C5D5N andH2O/D2O solutions, C–W, chemical shifts difference betweenCDCl3 and H2O/D2O solutions, P–C, chemical shifts differencebetween C5D5N and CDCl3 solutions.

761

3. Conclusion

In this report we have shown the potential of17O NMRspectroscopy, at natural isotopic abundance, for studyingthe bonds and solvent interactions in small oxygen-containing molecules, like the drug L1. It has beenpossible to determine that the intramolecular hydrogenbond between the 3–OH and the 4–CO groups in thismolecule is strong enough not to be broken even inaqueous solution. The mutual interactions between hy-droxyl and carbonyl groups have been examined, obtain-ing indications of the relative importance of their reso-nance structures and their influence on their ability to actas Fe(III) complexing agents. These data can thus facili-tate both the understanding of the activity of L1 as well asthe development of new and more efficient iron chelatingdrugs. The importance of these points is evident fromvery recent literature data [28, 29]. A possible hepatotox-icity of L1 has in fact been reported [28] from one of themain groups involved in clinical experimental use of thisdrug and this point has been challenged in an editorialpaper [29], stressing the importance of ascertaining thesafety and efficacy of this agent.

4. Experimental

All studied derivatives are known compounds.1–3have been prepared according to Kontoghiorghes proce-dure [30].4–8 were purchased from Aldrichy and usedwithout further purification.

17O NMR spectra were recorded, in the Fourier trans-form mode, on a Varian VXR 300 spectrometer equippedwith a Sun 3/60 computer and with a 10 mm broad bandprobe at 65 °C (probe temperature = 338 K) for H2O/D2Oand pyridine-d5 solutions and at 45 °C (probe tempera-ture = 318 K) for CDCl3 solutions and at natural isotopicabundance. Saturated solutions were used in all experi-ments. The instrumental settings were: 40.662 MHzfrequency, spectral width 36 KHz, acquisition time10 ms, pre-acquisition delay 100µs, pulse angle 90°(pulse width 28µs). Number of scans varied largely (4×104–3 × 106) as a function of solvent and solubility. Thespectra were recorded with sample spinning and withoutlock. The signal to noise ratio was in most cases im-proved by applying a 30 Hz exponential broadeningfactor (l.b.) to the FID prior to Fourier transformation. Insome experiments (H2O/D2O solutions of1–3 and7 andCDCl3 solution of6) an l.b. up to 120 Hz was necessary.The data point resolution was improved by zero filling to16 K data points. Chemical shifts are expressed in ppmand referred to external tap water by the substitutionmethod. The reproducibility of the chemical shift data is

estimated to be± 1 ppm (± 3 ppm when l.b. = 120 Hz hasbeen used).

Acknowledgements

This work was supported by the Regione Autonomadella Sardegna, Assessorato dell’Igiene e Sanità edell’Assistenza Sociale, Grant No. 3250.

References

[1] Green R., Charlton R., Seftel H., Bothwell T., Mayet F., Adams B.,Finch C., Layrisse M., Am. J. Med. 45 (1968) 336–353.

[2] Bergeron R.J., Brittenham G.M., The Development of Iron Chela-tors for Clinical Use, CRC Press, Boca Raton, 1994.

[3] Kontoghiorghes G.J., Ann. NY Acad. Sci. 612 (1990) 339–350.

[4] Hider R.C., Hall A.D., Prog. Med. Chem. 28 (1991) 41–173.

[5] Martell A.E., Motekaitis R.J., Sun Y., Clarke E.T., in: Bergeron R.J.,Brittenham G.M. (Eds.), The Development of Iron Chelators forClinical Use, CRC Press, Boca Raton, 1994, pp. 329–351.

[6] Pearson R.G., J. Am. Chem. Soc. 85 (1963) 3533–3539.

[7] Boykin D.W., CRC Press, Boca Raton,17O NMR Spectroscopy inOrganic Chemistry 1991.

[8] Chandrasekaran S., in: Boykin D.W. (Ed.),17O NMR Spectroscopyin Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 141–204.

[9] Boykin D.W., Baumstark A.L., in: Boykin D.W. (Ed.),17O NMRSpectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991,pp. 39–68.

[10] Boykin D.W., Baumstark A.L., Boykin D.W. (Ed.),17O NMRSpectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991,pp. 69–94.

[11] Boykin D.W., Baumstark A.L., in: Boykin D. W. (Ed.),17O NMRSpectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991,pp. 205–232.

[12] Boykin D.W., Chandrasekaran S., Baumstark A.L., Magn. Reson.Chem. 31 (1993) 489–494.

[13] Nowak-Widra B., Allison L.W., Kumar A., Boykin D.W., J. Chem.Res. (S) (1994) 490–491.

[14] St Amour T.E., Burgar M.I., Valentine B., Fiat D., J. Am. Chem. Soc.103 (1981) 1128–1136.

[15] Frey J., Eventova I., Rappoport Z., Müller T., Takai Y., Sawada M.,J. Chem. Soc. Perkin Trans. II (1995) 621–637.

[16] Paoloni L., Tosato M.L., Cignitti M., Theoret. Chim. Acta 14 (1969)221–231.

[17] Berthier G., Lévy B., Paoloni L., Theoret. Chim. Acta 16 (1970)316–318.

[18] Cignitti M., Paoloni L., Theoret. Chim. Acta 25 (1972) 277–288.

[19] Vögeli U., von Philipsborn W., Org. Magn. Reson. 5 (1973)551–559.

[20] Johnson C.D., in: Boulton A.J., McKillop A. (Eds.), ComprehensiveHeterocyclic Chemistry, The Structure, Reactions, Synthesis andUses of Heterocyclic Compounds, 2., Part 2A, Pergamon Press,Oxford, 1984, pp. 99–164.

[21] Karplus M., Pople J.A., J. Chem. Phys. 38 (1963) 2803–2807.

[22] Wheland G.W., in: Resonance in Organic Chemistry, Wiley, NewYork, 1955, pp. 341–342.

[23] Katoh M., Sugawara T., Kawada Y., Iwamura H., Bull. Chem. Soc.Jpn. 52 (1979) 3475–3476.

762

[24] Boykin D.W., Sullins D.W., Pourahmady N., Eisenbraun E.J.,Heterocycles 29 (1989) 307–312.

[25] Boulton A.J., McKillop A., in: Boulton A.J., McKillop A., (Eds.),Comprehensive Heterocyclic Chemistry, The Structure, Reactions,Synthesis and Uses of Heterocyclic Compounds, 2., Part 2A,Pergamon Press, Oxford, 1984, p. 26.

[26] Baumstark A.L., Boykin D.W., in: Boykin D.W., (Ed.),17O NMRSpectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991,pp. 111–112.

[27] De Kowalewski D.G., Contreras R.H., De Los Santos C., J. Mol.Struct. 213 (1989) 201–212.

[28] Olivieri N.F., Brittenham G.M., McLaren C.E., Templeton D.M.,Cameron R.G., McClelland R.A., Burt A.D., Fleming K.A., N. Engl.J. Med. 339 (1998) 417–423.

[29] Kowdley K.V., Kaplan M.M., N. Engl. J. Med. 339 (1998) 468–469.

[30] Kontoghiorghes G.J., Sheppard L., Inorg. Chim. Acta 136 (1987)L11–L12.

763

Laboratory note

Antimicrobial activities of new analogues of benzalkonium chloride

J. Pernaka*, I. Mirskab, R. Kmiecika

aPoznan University of Technology, Sklodowskiej-Curie 2, 60-965 Poznan, PolandbK. Marcinkowski University of Medical Sciences, Sieroca 10, 61-771 Poznan, Poland

(Received 29 December 1998; revised 15 March 1999; accepted 18 March 1999)

Abstract – (Alkoxymethyl)dimethyl{2-hydroxy-5-[(4-X-phenyl)azo]benzyl}ammonium chlorides were prepared in high yield. All thesechlorides, new analogues of benzalkonium chloride, showed antimicrobial activity. Activity depends on the length and kind of substituent atthe quaternary nitrogen atom. © 1999 E´ditions scientifiques et médicales Elsevier SAS

analogue of benzalkonium chloride / 4-hydroxyazobenzenes / Mannich bases / chloromethylalkyl ether / antimicrobial activity

1. Introduction

Benzalkonium chloride (BAC) is the product of anucleophilic substitution reaction of alkyldimethylaminewith benzyl chloride [1]. Chemically, it is monoalkyldi-methylammonium chloride with one long-chain alkylgroup representing a mixture of the alkyls from C8H17 toC18H37. Following Domagk’s publication in 1935 [2], alarge number of application areas were developed forBAC. It is used as a pharmaceutical aid (preservative),cationic surface active agent, germicide, antiseptic (topi-cal), antiseptic for skin preoperatively or for wounds,burns, etc. BAC is often present as a preservative orstabilising agent in nebulizer solutions used to treatasthma and chronic obstructive pulmonary disease [3].Also it is widely used as an antimicrobial agent in thetreatment of common infections of the mouth and throat.

We now report the synthesis and antimicrobial activi-ties of new quaternary ammonium compounds, the ana-logues of benzalkonium chloride. We plan to find com-pounds with antimicrobial activity which are diluted inwater giving a coloured solution. Commercial productswith these compounds will not have to contain any dye.

2. Chemistry

Mannich reaction (or aminomethylation) of variouslysubstituted phenols is a well known process and compre-hensive reviews have been published [4–7]. A few Man-nich bases of phenolic azobenzenes have demonstratedcytotoxicity towards murine and human cancers [8].Mannich bases methiodides have a promising cytotoxicactivity in a wide variety of tumours [9].

The new analogues of benzalkonium chloride (3–10)were prepared by the reaction of 2-[(dimethylamino)-methyl]-4-[(4-chlorophenyl)azo]phenol (2a) or 2-[(dime-thylamino)methyl]-4-[(4-methylphenyl)azo]phenol (2b)with chloromethylalkyl or chloromethylcycloalkyl ethersgiving yields between 80–98%. In this case, Mannichbase is a compound which is easily transformed into aquaternary ammonium salt.

Chloromethylalkyl and chloromethylcycloalkyl etherswere synthesised from the corresponding alcohols. Man-nich bases (2a–b) were prepared by treatment of the4-hydroxyazobenzenes with equimolar quantities offormaldehyde and dimethylamine in 75% yield.

The 4-hydroxyazobenzenes (azo dyes) were preparedby several authors in the 19th century [10]. At the presenttime, only one compound, 4-hydroxyazobenzene (Sol-vent Yellow 7), is commercially available.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 765−771 765© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

3. Antimicrobial activity

All synthesised quaternary ammonium chlorides(3–10) were tested for antimicrobial activity againstcocci, rods and fungi.

4. Results and discussion

4-Hydroxyazobenzenes react readily with formalde-hyde and secondary amines to give Mannich bases. Thestructures of two prepared mono Mannich bases (2a–b)were characterized by their microanalysis CHN and bytheir 1H and13C NMR spectroscopy.

Quaternization of2a and2b with chloromethylalkyl orchloromethylcycloalkyl ethers produced red crystallinequaternary ammonium chlorides3–10(table I) diluted inwater. The water solution of these chlorides is stable andorange in colour.1H and 13C NMR-spectral analysis ofprepared chlorides allowed easy elucidation of theirstructure.

The minimal inhibitory concentration (MIC) and theminimal bactericidal concentration (MBC) values deter-mined for all forty chlorides are given intables IIandIII .The chlorides studied were divided into four groups withrespect to the kind of substituent: group 1, chlorides withan alkoxymethyl substituent with an even number ofcarbon atoms (3 and7); group 2, chlorides with the samesubstituent but with an odd number of carbon atoms (4and 8); group 3, chlorides with a cycloalkoxymethyl

substituent (5 and 8); and group 4, chlorides withCH2O(CH2)nC6H11 substituent (6 and10). The calculatedaverage MIC values for cocci, rods and fungi are shownin figures 1and2. As shown by the results in these tablesand figures, all the chlorides studied are very activeagainst cocci and active against rods and fungi. Themicrobial activity depends on the length and kind ofsubstituent at the quaternary nitrogen atom.Figures 1and2 reveal a decrease in the MIC value to the optimumvalue and these values increase for chlorides from groups1, 2 and 3. The same correlation is observed for the MBCvalues. Generally the MBC values are slightly higherthan MIC values. To the most active compounds in group1 belong chlorides which have butoxymethyl, hexyl-oxymethyl and octyloxymethyl substituent, in group 2chloride with heptyloxymethyl chain, and in group 3chlorides which have cyclopentyloxymethyl, cyclohexyl-oxymethyl and cycloheptyloxymethyl substituent. To theworst compounds belong chlorides with dodecyloxy-methyl and cyclododecyloxymethyl chain. Chloridesfrom group 4 have comparable values of MIC. In thisgroup the microbial activity does not depend on thesubstituent – CH2O(CH2)nC6H11 wheren = 1 or 2.

The most active chlorides against microorganismswere: {5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl-(cyclohexyloxymethyl)ammonium (5b), {5-[(4-chloro-phenyl)azo]-2-hydroxybenzyl}dimethyl(cyclohexyl-methyloxymethyl)ammonium (6a), {2-hydroxy-5-[(4-methylphenyl)azo]benzyl}dimethyl(hexyloxymethyl)-

766

ammonium (7c), {2-hydroxy-5-[(4-methylphenyl)azo]ben-zyl}dimethyl(cyclopentyloxymethyl)ammonium (9a).

The results presented demonstrate that the new BACanalogues are very active against cocci. Their activitiesare similar to the activity of BAC. The antimicrobialactivities of BAC (Aldrich product, in which R representsa mixture of alkyls from C8H17 to C18H37) against cocci,Micrococcus luteus, Staphylococcus epidermidisandSta-phylococcus aureusas measured in the same MIC test are1.5, 3.0 and 1.5 mmol/L, respectively. Tomlinson andcoworkers [11] reported the antibacterial activities ofhomologues series (C8–C18) of alkylbenzyldimethyl-ammonium chlorides againstPseudomonas aeruginosa.BAC resistance is a potential problem for application, forexample in the food processing industry [12]. We foundcompounds with large molecular weights, crystalline,diluted in water and the orange water solution is stable.

5. Experimental protocols

5.1. Chemistry

NMR spectra were recorded on a Varian Model XL 300spectrometer at 300 MHz for1H and 75 MHz for13C at20 °C with tetramethylsilane as internal reference. Satis-factory elemental analyses were obtained: C± 0.32,H ± 0.29 and N± 0.24.

Chloromethylalkyl ethers and chloromethylcycloalkylethers were prepared via the procedures which werereported earlier [13]. The percentage of ether in a crudeproduct was determined by an alkalimetric method [14].

2-[(Dimethylamino)methyl]-4-(phenylazo)phenols (2a–b);general procedure:

Pathway A. To a solution of dimethylamine (30 mmol)in 95% EtOH (10 mL) paraformaldehyde powder (0.9 g,30 mmol) was added. The mixture was stirred and heatedwhen the paraformaldehyde had dissolved, then thecorresponding 4-hydroxyazobenzene (30 mmol) in100 mL EtOH was added. The reaction mixture wasstirred for 1 h at 60 °C.

Pathway B. To 2-[(dimethylamino)methyl]phenol(20 mmol) in 50 mL MeOH was added diazonium saltprepared from dimethylamine (20 mmol). The mixturewas stirred at room temperature for 2 h. The solidsubstrate was filtered and then recrystallized from EtOH.

2-[(Dimethylamino)methyl]-4-[(4-chlorophenyl)azo]-phenol (2a): m.p. 120–122 °C,1H NMR (CDCl3) δppm = 11.6 (s, OH), 7.83 (dd,J = 6 Hz, 1H), 7.82 (d,J = 9 Hz, 2H), 7.61 (d,J = 2 Hz, 1H), 7.45 (d,J = 8 Hz,2H), 6.95 (d,J = 8 Hz, 1H), 3.72 (s, 2H), 2.36 (s, 6H);13C NMR δ ppm = 161.8, 151.0, 145.5, 135.7, 129.1,125.2, 123.6, 122.6, 122.1, 116.6, 62.6, 44.4.

2-[(Dimethylamino)methyl]-4-[(4-methylphenyl)azo]-phenol (2b): m.p. 99–100 °C, lit. 103 °C [10].

Table I. (Alkoxymethyl)dimethyl{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}ammonium chlorides (3–6) and (alkoxymethyl)dimethyl{2-hydroxy-5-[(4-methylphenyl)azo]benzyl}ammonium (7–10) chlorides.

Chloride R1 R2 Yield (%) Chloride R1 R2 Yield (%)

3a Cl C2H5 80 7a CH3 C2H5 803b Cl C4H9

a 96 7b CH3 C4H9a 86

3c Cl C6H13a 98 7c CH3 C6H13

a 903d Cl C8H17

a 90 7d CH3 C8H17a 80

3e Cl C10H21a 85 7e CH3 C10H21

a 803f Cl C12H25

a 86 7f CH3 C12H25a 96

4a Cl C3H7a 90 8a CH3 C3H7

a 804b Cl C5H11

a 80 8b CH3 C5H11a 87

4c Cl C7H15a 80 8c CH3 C7H15

a 804d Cl C9H19

a 90 8d CH3 C9H19a 80

4e Cl C11H23a 96 8e CH3 C11H23

a 905a Cl C5H9

b 80 9a CH3 C5H9b 88

5b Cl C6H11b 80 9b CH3 C6H11

b 845c Cl C7H13

b 90 9c CH3 C7H13b 92

5d Cl C8H15b 80 9d CH3 C8H15

b 805e Cl C12H23

b 80 9e CH3 C12H23b 86

5f Cl C6H10CH3b 80 9f CH3 C6H10CH3

b 846a Cl CH2C6H11

b 90 10a CH3 CH2C6H11b 80

6b Cl CH2CH2C6H11b 86 10b CH3 CH2CH2C6H11

b 806c Cl CH2CH2CH2C6H11

b 85 10c CH3 CH2CH2CH2C6H11b 80

a linear alkyl, b alicyclic.

767

The quaternary ammonium chlorides3–10 were pre-pared by dissolving 2-[(dimethylamino)methyl]-4-(X-phenylazo)phenol in CH2Cl2 and adding an equimolaramount of the appropriate chloromethylalkyl or chloro-methylcycloalkyl ether. The mixture was stirred at roomtemperature for 24 h. The solvent was evaporated and thecrude product was extracted three times with hexane.Finally, the products were crystallized fromCH3COOC2H5/MeOH and dried in vacuum oven.

{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl-(dodecyloxymethyl)ammonium chloride (3f): m.p.124–126 °C,1H NMR (DMSO-d6) δ ppm = 11.9 (s, OH),8.06 (d,J = 2 Hz, 1H), 7.96 (dd,J = 7 Hz, 1H), 7.86 (d,J = 9 Hz, 2H), 7.66 (d,J = 9 Hz, 2H), 7.46 (d,J = 9 Hz,1H), 4.82 (s, 2H, CH2N), 4.58 (s, 2H, NCH2O), 3.86 (t,J = 7 Hz, 2H), 3.02 (s, 6H), 2.00 (m, 2H), 1.63 (m, 18H),0.86 (t,J = 7 Hz, 3H); 13C NMR δ ppm = 161.4, 150.5,144.6, 135.2, 130.5, 129.5, 125.9, 123.8, 117.1, 114.9,90.0, 73.0 (NCH2O), 57.9 (CH2N), 46.3 [N(CH3)2], 33.7,31.3, 29.2, 29.0, 28.8, 28.7, 25.3, 22.1, 13.9.

{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl-(cyclohexyloxymethyl)ammonium chloride (5b): m.p.120–123 °C,1H NMR (DMSO-d6) δ ppm = 11.5 (s, OH),8.07 (d,J = 2 Hz, 1H), 7.96 (dd,J = 7 Hz, 1H), 7.85 (d,J = 9 Hz, 2H), 7.58 (d,J = 9 Hz, 2H), 7.38 (d,J = 9 Hz,1H), 4.86 (s, 2H, CH2N), 4.60 (s, 2H, NCH2O), 3.84 (m,1H), 3.05 (s, 6H), 1.92 (m, 2H), 1.74 (m, 2H), 1.49 (m,6H); 13C NMR δ ppm = 159.3, 148.7, 143.0, 133.6,128.3, 127.4, 124.4, 121.8, 115.2, 113.2, 82.5, 78.0(NCH2O), 56.1 (CH2N), 44.4 [N(CH3)2], 33.4, 27.7,21.9.

(Cycloheptyloxymethyl)dimethyl{2-hydroxy-5-[(4-methyl-phenyl)azo]benzyl}ammonium chloride (9c): m.p.152–154 °C,1H NMR (DMSO-d6/CDCl3) δ ppm = 11.3(OH), 8.03 (d,J = 2 Hz, 1H), 7.92 (dd,J = 6 Hz, 1H),7.76 (d,J = 9 Hz, 2H), 7.34 (d,J = 9 Hz, 3H), 4.82 (s,2H, CH2N), 4.59 (s, 2H, NCH2O), 4.02 (m, 1H), 3.05 (s,6H), 2.42 (s, 3H), 1.99 (m, 2H), 1.81 (m, 4H), 1.58 (m,4H), 1.54 (m, 2H);13C NMR δ ppm = 158.7, 148.4,

Table II. The MICa and MBCa values of examined chlorides(3–6).

Strainsb Chlorides3a 3b 3c 3d 3e 3f 4a 4b 4c 4d 4e 5a 5b 5c 5d 5e 5f 6a 6b 6c

CocciM. luteus MIC 21 10 18 17 32 59 78 37 9 33 61 19 36 18 34 479 9 9 8 16

MBC 21 10 18 34 32 118 78 37 18 518 61 19 71 18 34 479 9 9 8 16

S. epidermidis MIC 21 10 9 8 8 30 156 73 9 16 31 19 18 9 8 8 9 18 8 16MBC 21 18 18 17 16 30 156 146 18 128 31 19 18 9 17 118 9 18 8 64

S. aureus MIC 260 38 217 267 252 238 156 146 9 128 245 9 9 9 214 8 18 18 8 16MBC 260 76 217 534 252 477 1256 219 36 518 490 19 18 9 214 8 36 36 8 129

RodsP. aeuruginosa MIC 260 304 284 267 504 1908 2512 1750 275 518 980 73 71 553 429 1915 137 137 133 520

MBC 520 304 568 534 504 1908 2515 4695 550 2074 980 146 142 553 429 1915 256 553 268 520

P. vulgaris MIC 81 76 18 132 252 477 628 219 9 259 245 19 18 9 268 16 9 18 8 16MBC 161 152 36 267 504 954 2512 438 36 518 490 73 71 137 536 1915 68 137 66 129

K. pneumoniae MIC 260 304 141 132 504 954 1256 875 9 64 490 38 36 36 133 239 36 36 66 64MBC 520 304 284 267 504 954 2512 1750 36 1037 980 146 71 36 133 958 68 36 133 129

E. coli MIC 81 38 141 132 125 477 628 438 136 259 490 38 71 18 66 958 36 18 34 32MBC 326 76 141 267 252 954 628 875 275 1037 980 73 142 18 133 1915 36 36 66 64

S. marescens MIC 1302 608 568 534 1008 3816 5025 1750 275 1037 1960 73 142 256 1073 3831 1106 68 133 1041MBC 1302 1216 568 534 4032 3816 5025 4695 550 2074 1960 146 285 553 2146 3831 256 137 268 1041

FungiC. albicans MIC 325 304 70 267 504 3816 1256 875 136 259 490 146 71 68 268 3831 256 68 34 129

MBC 651 608 141 534 1008 3816 2512 1750 1101 4149 980 146 142 256 536 3831 1106 137 66 260

T. mentagrophytes MIC 161 76 18 66 125 477 1256 438 36 64 122 73 36 36 268 1915 137 68 34 64MBC 161 152 18 132 252 477 1256 875 36 2074 245 73 36 36 286 1915 256 68 34 129

Rh. rubra MIC 161 152 36 132 252 3816 628 219 36 259 490 38 36 36 268 1915 256 36 17 64MBC 161 304 70 267 504 3816 1256 219 68 2074 490 73 71 36 268 3831 256 68 34 64

a in mmol/L, b the number of microorganisms in mL ranged from 104–105.

768

143.3, 139.0, 127.9, 127.8, 124.4, 120.5, 115.2, 112.5,86.6, 81.1, 56.4, 44.5, 31.9, 26.1, 20.3, 19.4 (CH3).

5.2. Antimicrobial activity

Microorganisms used: eleven standard strains repre-sentative of cocci;Micrococcus luteusATCC 9341,Staphylococcus epidermidisATCC 12228,Staphylococ-cus aureusATCC 6538, rods;Pseudomonas aeruginosaATCC 15442,Proteus vulgarisNCTC 4635,KlebsiellapneumoniaeATCC 4352,Escherichia coliNCTC 8196,Serratia marcescensATCC 8100, yeast-like fungi;Can-dida albicansATCC 10231,Rhodotorula rubraPhB, anddermatophytesTrichophyton mentagrophytesvar. gyp-seumATCC 9533.

Standard strains were supplied by National Collectionof Type Cultures (NCTC), London and American TypeCulture Collection (ATCC).Rhodotorula rubra(PHB)strain was taken from the Department of Pharmaceutical

Microbiology, K. Marcinkowski University of MedicalSciences, Poznan.

Antimicrobial activity was determined by the tubedilution method. The method shows, the lowest concen-tration of a chloride inhibiting cell multiplication (MIC)or killing them (MBC). Two-fold dilutions of the chlo-rides were prepared in the Mueller-Hinton broth medium(bacteria) or in the Sabouraud broth medium (fungi). Asuspension of the standard microorganisms preparedfrom 24 h cultures of bacteria in the Mueller-Hinton brothmedium and from 5 and 10 day cultures in the Sabouraudagar medium for fungi at a concentration of 105 cfu/mLwere added to each dilution in a 1:1 ratio. Growth (or itslack) of the microorganisms was determined visuallyafter incubation for 24 h at 37 °C (bacteria) or 5–10 daysat 28–30 °C (fungi). The lowest concentration at whichthere was no visible growth (turbidity) was taken as theMIC.

Table III. The MICa and MBCa values of examined chlorides(7–10).

Strainsb Chlorides7a 7b 7c 7d 7e 7f 8a 8b 8c 8d 8e 9a 9b 9c 9d 9e 9f 10a 10b 10c BAC

CocciM. luteus MIC 22 20 9 9 65 61 42 20 18 9 63 20 10 9 9 249 9 9 9 17 1.5

MBC 22 40 9 9 65 61 82 20 18 34 126 40 19 37 9 249 9 9 9 17 3

S. epidermidis MIC 22 10 9 9 33 248 42 10 9 34 8 20 19 9 9 8 9 9 9 17 3MBC 22 79 9 18 33 248 42 10 9 67 8 40 38 18 18 16 18 9 9 17 3

S. aureus MIC 11 20 19 9 16 16 42 10 9 134 126 77 19 9 9 249 9 37 9 17 1.5MBC 170 40 19 9 262 16 82 20 36 270 126 154 38 18 9 997 9 72 9 17 6

RodsP. aeuruginosa MIC 344 158 147 138 2103 3972 331 76 143 1083 2043 1239 148 143 139 1994 143 143 280 1088 96

MBC 688 319 298 279 4206 3972 662 308 288 4333 4086 1239 299 289 1122 1994 289 143 561 1088 192

P. vulgaris MIC 85 79 19 69 262 496 662 308 9 270 255 20 19 9 280 16 9 18 9 17 12MBC 688 319 298 69 262 993 662 154 143 1083 1021 619 299 18 280 249 289 143 139 561 12

K. pneumoniae MIC 344 79 19 138 262 993 331 39 18 67 63 40 38 18 36 997 18 72 18 67 12MBC 344 158 38 138 529 1986 331 76 36 1083 255 154 74 71 139 997 37 72 139 134 24

E. coli MIC 85 40 19 69 529 1986 82 20 36 270 255 77 19 18 36 498 9 37 36 134 3MBC 85 79 19 138 1051 1986 164 76 143 1083 510 154 38 37 70 997 18 37 69 134 6

S. marescens MIC 688 319 147 279 4206 3972 1324 152 143 2166 4086 1239 148 289 1122 1994 289 143 561 2176 12MBC 1376 638 596 279 4206 3972 1324 308 143 2166 2043 2478 299 289 1122 3988 289 143 561 2176 24

FungiC. albicans MIC 688 319 38 36 1051 1986 331 152 288 541 1021 309 148 37 36 1994 72 143 139 272 3

MBC 1376 319 38 69 2103 1986 662 308 576 2166 1021 619 299 143 70 1994 143 289 280 272 3

T. mentagrophytes MIC 170 40 38 9 262 248 164 39 288 1083 510 154 38 18 9 489 37 72 69 272 3MBC 170 79 74 18 529 248 331 152 288 1083 255 308 148 71 9 489 72 143 69 282 6

Rh. rubra MIC 344 158 19 9 1051 1986 164 20 143 270 255 154 74 37 18 997 72 72 69 272 12MBC 344 158 38 36 1051 1986 331 39 143 1083 255 154 74 37 36 1994 72 72 131 272 12

a in mmol/L, b the number of microorganisms in mL ranged from 104–105.

769

Then from each tube, one loopful was cultured on anagar medium with inactivates [14] (0.3% lecithin, 3%polysorbate 80 and 0.1% cysteine L) and incubated for48 h at 37 °C (bacteria) or for 5–10 days at 28–30 °C(fungi). The lowest concentration of the chloride support-ing no colony formation was defined as the MBC.

Acknowledgements

This investigation received financial support from thePolish Committee of Scientific Research, Grant KBN 3T09B 010 15.

Figure 1. The MIC mean values for (alkoxymethyl)dimethyl-{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}ammonium chlori-des (3–5).

Figure 2. The MIC mean values for (alkoxymethyl)dimethyl-{2-hydroxy-5-[(4-methylphenyl)azo]benzyl}ammonium chlo-rides (7–9).

770

References

[1] Dunn C.C., Proc. Soc. Exp. Biol. Med. 35 (1936) 427–429.

[2] Domagk G., Dtsch. Med. Wochenschr. 61 (1935) 829–832.

[3] Beasley R., Fishwick D., Miles J.F., Hendeles L., Pharmacotherapy18 (1998) 130–139.

[4] Tramontini M., Synthesis (1973) 703–775.

[5] Tramontini M., Angiolini L., Tetrahedron 46 (1990) 1791–1837.

[6] Tramontini M., Angiolini L., Mannich-Bases, Chemistry and Uses,CRC, Boca Raton USA, 1994.

[7] Arend M., Westermann B., Risch N., Angew. Chem. Int. Ed. 37(1998) 1045–1070.

[8] Dimmock J.R., Erciyas E., Kumar P. et al., Eur. J. Med. Chem. 32(1997) 583–594.

[9] Dimmock J.R., Pandeya S.N., Allen T.M., Koa G.Y., Pharmazie 53(1998) 201–202.

[10] a. Kimich C., Ber. 8 (1875) 1026–1032; b. Nolting E., Kohn O., Ber.17 (1884) 351–369; c. Heumann K., Oeconomides L., Ber. 20(1887) 904–909; d. Krause M., Ber. 32 (1899) 124–127; e. FarmerR., Hantzsch A., Ber. 32 (1899) 3089–3101.

[11] Tomlinson E., Brown M.R.W., Davis S.S., J. Med. Chem. 20 (1977)1277–1282.

[12] Heir E., Sundheim G., Holck A.L., J. Appl. Bacteriol. 79 (1995)149–156.

[13] Bedford C.D., Harris R.N., Howd R.A. et al., J. Med. Chem. 32(1989) 493–516.

[14] Skrzypczak A., Brycki B., Mirska I., Pernak J., Eur. J. Med. Chem.32 (1997) 661–668.

771

New products

New 2-sulfonamidothiazoles substituted at C-4: synthesis of polyoxygenated arylderivatives and in vitro evaluation of antifungal activity

Pierre Beucheta, Martine Varache-Lembègea, Arlette Neveua, Jean-Michel Légerb,Joseph Vercauterena, Stéphane Larrouturea, Gérard Deffieuxa, Alain Nuhricha*

aGroupe d’Étude des Substances Naturelles d’Intérêt Thérapeutique (GESNIT), Faculté de Pharmacie,Université Victor-Segalen (Bordeaux 2), 146, rue Léo-Saignat, 33076 - Bordeaux cedex, France

bLaboratoire de Chimie Analytique, Faculté de Pharmacie, Université Victor-Segalen (Bordeaux 2),3, Place de la Victoire. 33076 -Bordeaux cedex, France

(Received 29 September 1998; revised 4 February 1999; accepted 10 February 1999)

Abstract – Polymethoxylated and polyhydroxylated derivatives of 2-amino-4-arylthiazoles bearing a halogenobenzenesulfonamide moiety atposition 2 were synthesized as azole antifungal analogues. X-ray crystallography studies revealed the predominance of the 2-imino-2,3-dihydrothiazole form in the amino/imino tautomerism. In vitro assays against various pathogenic fungal strains (CandidaandTrichophytonspecies) showed no activity in comparison to econazole as reference. These results are discussed on the basis of the estimated globallipophilicity of the molecules (Rekker’s method) and theπ-electron distribution (Mulliken population analysis, AM1 method) within thefive-membered heterocycle. © 1999 E´ditions scientifiques et médicales Elsevier SAS

2,4-disubstituted thiazoles / sulfonamides / antifungal testing

1. Introduction

The most commonly used imidazole antifungal agents,such as miconazole1, econazole2 and ketoconazole3(figure 1), exhibit side effects including interference withsteroidogenesis or hepatic toxicity [1]. Over the past fewyears, the discovery of triazole compounds includingitraconazole4 and fluconazole5 with better tolerancegave rise to the synthesis of numerous related ana-logues [2].

Since the sulfonamide moiety seems to be a possiblepharmacophore in fungicidal agents [3], we decided tointroduce this functionality in position 2 of a thiazolenucleus. On the other hand, the lack of hydrosolubilitywhich limits systemic distribution of azoles led us toinclude hydrophilic fragments and thus to synthesize newthiazole derivatives with a polyoxygenated phenyl com-ponent (6–9), as represented infigure 1.

*Correspondence and reprintsFigure 1. Structure of azole antifungals (1–5) and target com-pounds (6–9).

Eur. J. Med. Chem. 34 (1999) 773−779 773© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

The target compounds (table I) were synthesized fol-lowing figure 2. The key step in the preparation ofderivatives6 and8 resulted in the reaction of bromom-ethylketones13 with the appropriate arylsulfonylthio-ureas 12 in refluxing THF, according to a modifiedHantzsch condensation [4]. The thioureas12 were ob-tained by reaction ofN-cyanoarylsulfonamide sodiumsalts 11 with thiosulfuric acid (generated in situ upontreatment of sodium thiosulfate with aqueous-H2SO4 [5]).

3. Structural data

As an example,1H and 13C NMR data for 6a areshown in table II. The chemical shifts for the carbonatoms at positions7, 8 and9 are consistent with data ofsubstituted thiazoles [6] and therefore confirm the hetero-cyclisation reaction. The main information is provided bythe long-range1H -13C correlations (HMBC spectrum).The quaternary carbon7 exhibits correlation with protonH6, belonging to the oxygenated ring The7,9-

functionalization on the heterocyclic ring was thereforeestablished.

X-ray diffraction analysis of6b (figure 3)confirmed astructure close to that of a thiazoline derivative [7].Noteworthy, the exocyclic C(9)-N(12) bond shows sig-nificant shortening from the value of the endocyclicN(8)-C(9) (table III). Consequently, this phenomenon isindicative of an exocyclic C=N double bond and confirmsthat the imino form is the predominant tautomer, at leastin the crystalline state. This conclusion agrees withprevious reports on 2-substituted thiazoles [8, 9].

4. Antifungal evaluation and discussion

The antifungal activity of compounds6–9 was evalu-ated in vitro, against yeasts (Candida albicansATCC 10231 andCandida kruseiCBS 573) and filamen-tous fungi (Trichophyton mentagrophytesIP 1468-83 andT. rubrum IP 2073-92), according to described meth-ods [11, 12]. By comparison with econazole, used asreference substance, all the tested thiazoles were foundinactive (data not shown). These results confirm thatlipophilicity plays a fundamental role for antifungal

Table I. Chemical and physical properties of compounds6–9.

Compound (RO)n Hal M.p.(°C) Yield (%) Formula (MW)a log Pcalc c

6a

2,5-OMe

4≠-Cl 130 75 C17H15ClN2O4S2 (410.90) 2.446b 4≠-F 136 36 C17H15FN2O4S2 (394.43) 1.926c 2≠,4≠-Cl 148–150 51 C17H14Cl2N2O4S2 (445.35) 3.186d 2≠,4≠-F 140 31 C17H14F2N2O4S2 (412.44) 2.13

7a

2,5-OH

4≠-Cl 260 45 C15H11ClN2O4S2 (382.84) 1.257b 4≠-F 218 47 C15H11FN2O4S2 (366.39) 0.737c 2≠,4≠-Cl 240 56 C15H10Cl2N2O4S2 (417.29) 1.997d 2≠,4≠-F 232 78 C15H10F2N2O4S2 (384.38) 0.94

8a

3,4,5-OMe

4≠-Cl b 58 C18H17ClN2O5S2 (440.92) 2.518b 4≠-F b 67 C18H17FN2O5S2 (424.47) 1.998c 2≠,4≠-Cl 165–166 56 C18H16Cl2N2O5S2 (475.37) 3.258d 2≠,4≠-F 178–180 20 C18H16F2N2O5S2 (442.46) 2.20

9a

3,4,5-OH

4≠-Cl 220 67 C15H11ClN2O5S2 (398.85) 0.739b 4≠-F 218–220 53 C15H11FN2O5S2 (382.39) 0.209c 2≠,4≠-Cl b 65 C15H10Cl2N2O5S2 (433.29) 1.479d 2≠,4≠-F 219 84 C15H10F2N2O5S2 (400.38) 0.42

aAll compounds were analysed for C, H, N, S and when present Cl or F; analytical results are within± 0.4% of the theoretical values.bDecomposition over 80 °C.cAccording to Rekker fragmental system [10].

774

properties [13]: the estimated logP values for com-pounds6–9, ranging from 0.2–3.25 (table I) are signifi-cantly lower than that calculated for econazole (logP =5.17).

We have also investigated the electron density distri-bution in econazole and compound6b, respectively. TheAM1 method [14] with Mulliken population analysis wasselected for its usefulness in predicting the aromaticity of5-membered heterocycles [15]. From the data presentedin figure 4, it is shown that the imidazole ring ofeconazole displays a high level of aromaticity (π overlappopulations ranging from 23–40%), while in the thiazolemoiety, the π-electrons are unequally spread over theintracyclic bonds (furthest values: 11–44%), thus indica-tive of a weak aromatic character for6b.

In conclusion, it appears that the association of apolyoxygenated phenyl ring and an arylsulfonamido moi-ety on a thiazolic framework does not lead to antifungalactivity. The electronic properties of the five-memberedring, which differ strongly from those of the imidazole

antifungals, might be a determining factor in the biologi-cal response.

5. Experimental protocols

Melting points were determined with a Köfler appara-tus and are uncorrected. NMR spectra were obtained inCDCl3, at 29 °C, using a Bruker AMX 500 spectrometer(1H: 500 MHz,13C: 125 MHz). Chemical shifts are givenin ppm from TMS as an internal standard (couplingconstantsJ, in Hz). Multiplicities in 13C NMR spectrawere derived from JMOD experiments. When necessary,HMQC, HMBC and 1H-1H COSY experiments wereused for structural assignments.

5.1. Chemistry

5.1.1. General procedure for N-aminothioxomethyl-arylsulfonamides12a–d

To a stirred suspension of sodium cyanamide [16](0.08 mol) in anhydrous Et2O (20 mL) was added appro-priate arylsulfonyl chloride10a–d (0.04 mol). Stirringwas continued for 6 h under reflux, and the insoluble Nasalt of N-cyanoarylsulfonamide11a–dwas washed (dryether, then acetone) and dried (yields: 65–75%).

A cooled solution (0 °C) of the convenient sodium salt11 (20 mmol) in H2O (10 mL) was neutralized with 2 NH2SO4 (10 mL). Solid Na2S2O3, 5H2O (60 mmol) wasadded in one portion under stirring and the mixture wastreated with more 2 N H2SO4 (15 mL). After furtherstirring overnight at 20 °C, the resulting precipitate waswashed with cold H2O (2 × 10 mL), and recrystallizedfrom methanol/H2O (50:50) to afford the expected prod-uct 12 as white needles (45–61%).

5.1.1.1. N-Aminothioxomethyl-4-chlorobenzenesulfon-amide12a

M.p. = 149–151 °C.1H-NMR δ 7.61 (d,J = 8.8, 2H);7.94 (d,J = 8.8, 2H);13C- NMR δ 130.0 (C2, C6); 130.6(C3, C5); 139.3 (C4); 141.3 (C1); 182.3 (C7).

5.1.1.2. N-Aminothioxomethyl-4-fluorobenzenesulfon-amide12b

M.p. = 120 °C.1H- NMR δ 7.33 (dd,J = 8.8, 8.6, 2H);8.02 (dd,J = 9.0, 5.0, 2H);13C- NMR δ 117.5 (C3, C5);131.4 (C2, C6); 136.8 (C1); 167.1 (C4); 182.4 (C7).

5.1.1.3. N-Aminothioxomethyl-2,4-dichlorobenzenesulfon-amide12c

M.p. = 205 °C.1H- NMR δ 7.81 (dd,J = 2.0, 8.6, 1H);7.94 (d,J = 2.0, 1H); 8.34 (d,J = 8.6, 1H);13C- NMR δ128.5 (C2); 132.5 (C4); 133.8 (C6); 134.3 (C1); 136.6(C5); 140.9 (C3); 181.1 (C7).

Figure 2. General synthetic pathway for 2,4-functionalizedthiazoles6–9.

775

5.1.1.4. N-Aminothioxomethyl-2,4-difluorobenzenesulfon-amide12d

M.p. = 102 °C.1H- NMR δ 7.03–7.15 (m, 2H); 7.95(ddd, J = 8.0, 6.0, 6.0, 1H);13C- NMR δ 106.4 (C3);112.4 (C5); 122.7 (C1); 131.8 (C6); 160.1 (C2); 166.9(C4); 179.6 (C7).

Table III. Selected bond lengths (Å) for6ba

C(1)–C(7) 1.477 (4) S(10)–C(11) 1.740 (3)

C(7)–C(11) 1.322 (4) N(12)–S(13) 1.588 (2)C(7)–N(8) 1.395 (4) S(13)–O(15) 1.441 (2)N(8)–C(9) 1.346 (3) S(13)–O(14) 1.444 (2)C(9)–N(12) 1.325 (4) S(13)–C(16) 1.762 (3)C(9)–S(10) 1.737 (3)

aEstimated standard deviations are given in brackets

Table II. NMR dataa and main HMQC, HMBC, and1H-1H COSY correlations for6a (CDCl3).

Positions 1H 13C HMQC HMBC

1 – 117.0 C8, H8 C1, H82 – 149.6 – C2, H6; C2, H43 6.94 (d, 8.8)b 116.0 C3, H3 –4 6.92 (dd, 9.1, 2.6)b 113.0 C4, H4 C4, H65 – 154.1 – C5, H6; C5, H46 7.02 (d, 2.6)b 113.2 C6, H6 C6, H47 – 134.4 – C7, H6; C7, H88 6.62 (s) 102.2 C8, H8 –9 – 168.4 – C9, H810 – 140.8 – C10, H12; C10, H1411 7.91 (d, 8.6)b 127.8 C11,H11 C11, H1512 7.42 (d, 8.6) 128.9 C12, H12 C12, H1413 – 138.4 – C13, H11; C13, H1514 7.42 (d, 8.6) 128.9 C14, H14 C14, H1215 7.91 (d, 8.6)b 127.8 C15, H15 C15, H11

aChemical shifts, ppm (multiplicity,J in Hz). b1H-1H COSY correlations.cThe methoxy protons appeared at 3.79 and 3.91 ppm, as singlets.dThe NH-proton was not obviously identifiable.

Figure 3. Molecular conformation of compound6b with spe-cial atom-numbering used in the crystallographic analysis.(displacement ellipsoids are shown for non-hydrogen atoms at50% probability level).

Figure 4. π-Interatomic overlap populations (%) calculatedusing the Mulliken population analysis for the heterocycles ofeconazole2 and compound6b (AM1 results from X-raystructure data). Atomic charges are indicated in italic numbers.

776

5.1.2. General procedure for 1-aryl-2-bromoethanones13These products were obtained by bromination of com-

mercial acetophenones in CCl4 [17], and chromato-graphed on silica gel column (eluent: hexane/CH2Cl2, 70:30) prior to use.

5.1.2.1. 1-(2,5-Dimethoxyphenyl)-2-bromoethanone13aM.p. = 80 °C.1H- NMR δ 3.80 (s, 3H); 3.93 (s, 3H);

4.65 (s, 2H); 7.05 (m, 2H); 7.40 (d,J = 8.0, 1H).

5.1.2.2. 1-(3,4,5-Trimethoxyphenyl)-2-bromoethanone13bM.p. = 71 °C.1H- NMR δ 4.00 (s, 9H); 4.46 (s, 2H);

7.30 (s, 2H).

5.1.3. General procedure for N-(4-aryl-2,3-dihydrothiazol-2-ylidene) arylsulfonamides6a–d and8a–d

A suspension ofN-aminothioxomethylbenzenesulfon-amide12a–d(4 mmol) and 1-aryl-2-bromoethanone13aor 13b (4 mmol) in anhydrous THF (30 mL) was heatedat reflux for 5 h under constant stirring. The mixture wasthen cooled to 20 °C and after removal of the volatilematerials under vacuo, the crude residue was chromato-graphed on a silica-gel column (eluent: CH2Cl2/EtOAc,80:20), then recrystallized from heptane (yields:40–51%).

5.1.3.1. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thia-zol-2-ylidene]-4-chlorobenzenesulfonamide6a (seetable II for spectral data)

5.1.3.2. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thia-zol-2-ylidene]-4-fluorobenzenesulfonamide6b

1H- NMR δ 3.80 (s, 3H); 3.92 (s, 3H); 6.62 (s, 1H);6.92 (dd,J = 9.0, 2.6, 1H); 6.95 (d,J = 9.0, 1H); 7.03 (d,J = 2.6, 1H); 7.13 (dd,J = 8.8, 8.6, 2H); 7.99 (dd,J = 8.8,5.2, 2H). 13C- NMR δ 55.8; 56.4; 102.2; 113.0; 113.3;115.6; 115.8; 116.0; 117.0; 128.9; 129.0; 134.4; 138.5;149.6; 154.1; 164.8 (d,1JC, F = 250); 168.4.

5.1.3.3. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thia-zol-2-ylidene]-2,4-dichlorobenzenesulfonamide6c

1H-NMR δ 3.80 (s, 3H); 3.92 (s, 3H); 6.64 (s, 1H);6.92 (dd,J = 9.1, 2.6, 1H); 6.95 (d,J = 8.8, 1H); 7.04 (d,J = 2.4, 1H); 7.34 (dd,J = 8.0, 2.0, 1H); 7.47 (d,J = 2.0,1 H); 8.16 (d,J = 8.5, 1H).13C-NMR δ 55.8; 56.3; 102.2;113.0; 113.3; 116.0; 117.0; 126.8; 131.1; 131.2; 133.5;134.5; 138.4; 149.7; 154.1; 169.0.

5.1.3.4. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thia-zol-2-ylidene]-2,4-difluorobenzenesulfonamide6d

1H-NMR δ 3.80 (s, 3H); 3.92 (s, 3H); 6.65 (s, 1H);6.89 (m, 2H); 6.95 (m, 1H); 7.05 (d,J = 2.7, 1H); 8.03

(m, 1H); 13C-NMR δ 55.8; 56.3; 102.3; 105.2 (dd,2JC, F

= 25, 2JC, F = 25); 111.1; 113.0; 113.3; 116.0; 117.0;126.7; 131.1; 134.5; 149.7; 154.1; 160.1 (dd,1JC, F= 250,3JC, F = 10); 165.1 (dd,1JC, F = 250,3JC, F = 10); 169.0.

5.1.3.5. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thia-zol-2-ylidene]-4-chlorobenzenesulfonamide8a

1H-NMR δ 3.88 (s, 9H); 6.46 (s, 1H); 6.71 (s, 2H);7.38 (d,J = 8.6, 2H); 7.89 (d,J = 8.6, 2H).13C-NMR δ56.3; 60.8; 101.2; 103.2; 123.9; 127.9; 128.9; 137.5;138.6; 139.5; 140.2; 153.7; 168.7.

5.1.3.6. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thia-zol-2-ylidene]-4-fluorobenzenesulfonamide8b

1H-NMR δ 3.85 (s, 9H); 6.46 (s, 1H); 6.73 (s, 2H);7.07 (dd,J = 8.6, 8.6, 2H); 7.97 (dd,J = 8.8, 5.1, 2H).13C-NMR δ 56.3; 60.8; 101.3; 103.2; 115.7; 115.9; 124.5;129.1; 137.6; 137.8; 139.2; 153.7; 164.8 (d,1JC, F = 250);168.7.

5.1.3.7. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thia-zol-2-ylidene]-2,4-dichlorobenzenesulfonamide8c

1H-NMR δ 3.88 (s, 9H); 6.49 (s, 1H); 6.71 (s, 2H);7.32 (d,J = 4.2, 1H); 7.45 (s, 1H); 8.10 (d,J = 4.2, 1H).13C-NMR δ 56.3; 60.9; 101.4; 103.2; 123.9; 126.8;131.1; 131.2; 133.4; 137.1; 137.8; 138.6; 139.5; 153.8;169.3.

5.1.3.8. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thia-zol-2-ylidene]-2,4-difluorobenzenesulfonamide8d

1H-NMR δ 3.85 (s, 9H); 6.50 (s, 1H); 6.67 (s, 2H);6.82 (m, 1H); 6.91 (m, 1H); 7.98 (m, 1H);13C-NMR δ56.3; 60.8; 101.3; 103.2; 105.4 (dd,2JC, F = 25, 2JC, F =23); 111.2 (d,2JC, F = 22); 123.9; 126.7 (d,2JC, F = 23);131.2 (d,3JC, F = 10); 137.3; 139.3; 153.7; 159.9 (dd,1JC,

F = 250,3JC, F = 10); 165.3 (dd,1JC, F = 250,3JC, F = 10);169.4.

5.1.4. General procedure for phenolic compounds7a–dand9a–d

To a cooled (0 °C) solution of6a–dor 8a–d (1 mmol)in anhydrous CH2Cl2, boron tribromide was added drop-wise (2.2 mmol for dimethoxy compounds6a–d, or3.3 mmol for trimethoxy compounds8a–d) under nitro-gen. The mixture was stirred at 0 °C for 1 h, thenquenched cautiously with H2O (50 mL). Following15 min hydrolysis, the resulting precipitate was collected,washed with cold H2O and recrystallized from 50%ethanol, affording the expected polyhydroxylated com-pound in 53–78% yield.

777

5.1.4.1. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thia-zol-2-ylidene]-4-chlorobenzenesulfonamide7a

1H-NMR δ 4.77 (s, 2H); 6.69 (dd,J = 8.7, 2.8, 1H);6.74 (d,J = 8.7, 1H); 6.88 (d,J = 2.8, 1H); 6.93 (s, 1H);7.50 (d,J = 8.6, 2H); 7.88 (d,J = 8.6, 2H).13C-NMR δ105.3; 114.7; 117.1; 118.3; 118.7; 129.1; 130.0; 137.1;139.4; 142.2; 148.8; 151.6; 170.

5.1.4.2. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thia-zol-2-ylidene]-4-fluorobenzenesulfonamide7b

1H-NMR δ 4.76 (s, 2H); 6.69 (dd,J = 8.7, 2.8, 1H);6.74 (d,J = 8.7, 1H); 6.89 (d,J = 2.8, 1H); 6.92 (s, 1H);7.22 (dd,J = 8.9, 8.8, 2H); 7.99 (dd,J = 8.9, 5.1, 2H).13C-NMR δ 105.2; 114.7; 116.7; 116.9; 117.1; 118.3;118.7; 130.2; 130.3; 139.7; 148.7; 151.6; 166.2 (d,1JC, F

= 250); 170.0.

5.1.4.3. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thia-zol-2-ylidene]-2,4-dichlorobenzenesulfonamide7c

1H-NMR δ 6.93 (dd,J = 8.7, 2.9, 1H); 7.03 (d,J = 8.7,1H); 7.24 (s, 1H); 7.25 (d,J = 2.9, 1H); 7.69 (dd,J = 8.5,2.0, 1H); 7.77 (d,J = 2, 1H); 8.3 (d,J = 8.5, 1H).13C-NMR δ 104.8; 114.4; 116.8; 118.5; 118.6; 127.9;131.9; 132.2; 134.0; 138.7; 140.1; 147.9; 151.7; 171.4.

5.1.4.4. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thia-zol-2-ylidene]-2,4-difluorobenzenesulfonamide7d

1H-NMR δ 6.69 (dd,J = 8.8, 2.8, 1H); 6.75 (d,J = 8.8,1H; 6.90 (d,J = 2.8, 1H); 6.95 (s, 1H); 7.05–7.13 (m,2H); 6.69 (dd,J = 8.4, 6.3, 1H);13C-NMR δ 105.5; 106.5(dd, 2JC, F = 25; 2JC, F = 23); 112.2; 114.7; 117.0; 118.3;118.8; 127.7; 132.3; 137; 148.8; 151.7; 161.2 (dd,1JC, F

= 250,3JC, F = 10); 166.9 (dd,1JC, F = 250,3JC, F = 10);170.4.

5.1.4.5. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thia-zol-2-ylidene]-4-chlorobenzenesulfonamide9a

1H-NMR δ 6.55 (s, 1H); 6.57 (s, 2H); 7.49 (d,J = 8.8,2H); 7.86 (d,J = 8.8, 2H). 13C-NMR δ 101.7; 106.4;121.4; 129.0; 130.0; 136.1; 139.4; 139.7; 142.3; 147.4;171.1.

5.1.4.6. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thia-zol-2-ylidene]-4-fluorobenzenesulfonamide9b

1H-NMR δ 6.55 (s, 1H); 6.57 (s, 2H); 7.21 (dd,J = 8.8,8.7, 2H); 7.93 (dd,J = 8.8, 5.2, 2H);13C-NMR δ 101.6;106.4; 116.7; 116.8; 121.4; 130.1; 130.2; 136.0; 139.7;139.8; 147.4; 166.2 (d,1JC, F = 250); 171.0.

5.1.4.7. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thia-zol-2-ylidene]-2,4-dichlorobenzenesulfonamide9c

1H-NMR δ 6.58 (s, 2H); 6.59 (s, 1H); 7.45 (dd,J = 8.5;2.0, 1H); 7.59 (d,J = 2.0, 1H); 8.10 (d,J = 8.5, 1H);

13C-NMR δ 101.9; 106.4; 121.3; 129.0; 132.4; 134.5;136.1; 139.5; 139.6; 139.7; 147.5; 171.4.

5.1.4.8. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thia-zol-2-ylidene]-2,4-difluorobenzenesulfonamide9d

1H-NMR δ 6.57 (s, 2H); 6.60 (s, 1H); 7.06–7.11 (m,2H); 7.98 (ddd,J = 8.0, 6.0, 6.0, 1H);13C-NMR δ 102.0;106.3 (dd,2JC, F = 25, 2JC, F = 22); 106.4; 112.2 (d,2JC, F = 22); 121.3; 126.9; 132.3; 136.0; 139.7; 147.5;161.2 (dd,1JC, F = 250,3JC, F = 10); 166.8 (dd,1JC, F =250,3JC, F = 10); 171.4.

5.2. X-ray structure determinationA suitable crystal of6b (size 0.30× 0.15 × 0.01 mm)

was obtained from a CHCl3 solution. C17H15FN2O4S2, M= 394.43, triclinic, space group P-1,a = 7.026(4),b =9.068,c = 14.442(7) Å,α = 88.74(4),â = 77.70(5),γ =80.63(4)°,V = 886.9(8) Å3, Z = 2,D = 1.477 Mg m–3. Thedata collection was performed on a Nonius CAD4 dif-fractometer using graphite monochromated Cu Kα radia-tion (λ = 1.54178 Å). 2 977 independent reflexions weremeasured, of which 2 461 were used in the refinements.

Refinement was carried out by the full-matrix least-squares method based onF2 with the SHELXL-93program [18]. (Final R indices: R1 = 0.0466, WR2 =0.1546). Full experimental details are available as supple-mentary material.

5.3. Computational studiesCalculations were performed using the MOPAC [19]

program implemented in Chem3D, V 3.5 [20]. Theatomic coordinates of the heavy atoms considered in thecalculations were those obtained by crystallographic data(compound6b: see above; econazole: see reference [21]).The atomic charges were analysed using the AM1 poten-tial function with the “Mulliken charges” option and the“PI” keyword.

Acknowledgements

The authors thank E. Bezançon for his able technicalassistance in providing high resolution NMR spectra. Thesecretarial help of A. Carrère is gratefully acknowledged.

References

[1] Hoeprich P.D., Prog. Drug Res. 44 (1995) 87–127.

[2] Lartey P.A., Moehle C.M., Ann. Rep. Med. Chem. 32 (1997)151–160.

[3] Habib O.M.O., Kandel E.M., Askou S.A., Moawad E.B., Hung. J.Ind. Chem. 14 (1986) 477–483.

778

[4] Das Gupta P.K., Gupta P., J. Indian Chem. Soc. 23 (1946)13–15.

[5] Földi Z., Földi T., Földi A., Acta Chim. Acad. Sci. Hung. 13 (1957)111–115.

[6] Faure R., Galy J.P., Vincent E.J., Elguero J., Can. J. Chem. 56 (1978)46–55.

[7] Kalcheva V., Tosheva M., Hadjieva P., Liebigs Ann. Chem. (1993)1319–1322.

[8] Form G.R., Paper E.S., Downie T.C., Acta Cryst. B 30 (1974)342–348.

[9] Argay G., Kalman A., Lazar D., Ribar B., Toth G., Acta Cryst. B 33(1977) 99–105.

[10] Rekker R.F., Mannhold R., Calculation of Drug Lipophilicity. TheHydrophobic Fragmental Approach, VCH, Weinheim, 1992.

[11] Scheven M., Senf L., Mycoses 37 (1994) 205–207.

[12] Drouhet E., Barale T., Bastide J.E. et al., Bull. Soc. Fr. Mycol. Med.10 (1981) 131–134.

[13] Wahbi Y., Tournaire C., Caujolle R., Payard M., Linas M.D., SeguelaJ.P., Eur. J. Med. Chem. 29 (1994) 701–706.

[14] Dewar M.J.S., Zoebisch E.G., Healy E.F., Stewart J.P., J. Am. Chem.Soc. 107 (1985) 3902–3909.

[15] Bean G.P., J. Org. Chem. 58 (1993) 7336–7340.

[16] Drechsel E., J. Prakt. Chem. 11 (1875) 284–353.

[17] Menassé R., Klein G., Erlenmeyer H., Helv. Chim. Acta 38 (1955)1289–1291.

[18] Sheldrick G.M., SHELXL-93, University of Göttingen, (1993).

[19] Stewart J.J.P., QCPE Bull. 9 (1989) 10–15.

[20] Chem3D: Cambridge Soft Corporation, Cambridge, (1996).

[21] Freer A.A., Pearson A., Salole E.G., Acta Cryst. C42 (1986)1350–1352.

779

Original article

Demonstration of the strength of focused combinatorial libraries in SARoptimisation of growth hormone secretagogues

Michael Ankersena*, Birgit Sehested Hansenb, Thomas Kruse Hansena, Jesper Laua,Bernd Peschkea, Kjeld Madsena, Nils Langeland Johansena

aMedChem Research, Health Care Discovery, Novo Nordisk A/S, Novo Nordisk Park, DK-2760 Måløv, DenmarkbAssay and Cell Technology, Health Care Discovery, Novo Nordisk A/S, Novo Nordisk Park, DK-2760 Måløv, Denmark

(Received 9 December 1998; accepted 10 February 1999)

Abstract – A series of 96 growth hormone secretagogues, derived from ipamorelin are described. The compounds are prepared as a 6× 4 ×4 member library on solid support using a PAL resin. The compounds are all acylated dipeptides, based on two aromatic amino acids and afree amino N-terminal. All compounds are characterised by HPLC, LC-MS and their ability to release GH in a pituitary cell based assay. Themost potent compounds show EC50 values at 1 nM and are full agonists. We demonstrate the strength of focused combinatorial libraries andconfirm the pitfall in broad SAR exploration by giving examples where selected fragments obviously show poor receptor interaction exceptin very defined structural arrangements. © 1999 Éditions scientifiques et médicales Elsevier SAS

GHRP-2 / GHRP-6 / MK677 / ipamorelin / growth hormone secretagogue / combinatorial libraries

1. Introduction

Since Bowers [1] in 1977 disclosed the discovery of anew class of compounds with the ability to release growthhormone (GH) from the pituitary in a manner distinctfrom GH releasing hormone (GHRH), an increasinginterest in this type of compound has emerged. Thesecompounds, called GH releasing peptides (GHRP) or GHsecretagogues (GHS) [2], may offer a new potentialtreatment for a number of conditions due to their abilityto maintain the physiological pulsatile pattern of GHsecretion. Such conditions could be GH deficiency, os-teoporosis or obesity [3–5]. The most prominent mem-bers of this class of GHSs are GHRP-2, GHRP-6 andMK677 [6]. Recently we discovered a new series ofacylated dipeptides derived from ipamorelin [7] via apeptidomimetic strategy [8–10]. Representatives of thisseries are the acylated dipeptides1, 2 and3, which are allbased on a free aminogroup at the N-terminal, twoN-methylated aromatic D-amino acids and methylamidein the C-terminal(figure 1).

To optimize this series, a compound library of singleisolated compounds was constructed in a 6× 4 × 4 matrix,based on a total combination of six N-terminals and twosets of N-methylated aromatic D-amino acids (4 buildingblocks in each), to give 96 compounds with the genericstructure as shown infigure 2.

Of the 96 compounds in the library, 10 compoundswere prepared in solution and have been describedpreviously [8–10], and 86 compounds were prepared onsolid support. Herein we report the synthesis of the 86compounds prepared on solid support and discuss thestructure-activity relationship of the entire library (ie. 96compounds).

2. Chemistry

The 86 compounds, prepared on solid phase, wereprepared in parallel on N-Me-PAL [11, 12](figure 3)assolid support using an Fmoc/Boc peptide synthesisstrategy with traditional amide coupling conditions[O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluroniumhexafluorophosphate, 1-hydroxy-7-azabenzotriazole, di-isopropylethylamine in dimethylformamide, deprotection*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 783−790 783© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

with piperidine in dimethylformamide for the Fmoc-protected building blocks B (1–4), i.e. B1, B2, B3 and

B4, and C (1–4) and deprotection/cleavage with TFA forthe Boc-protected A (1–6) (figures 4and 5)]. Six Boc-protected N-terminals (A (1–6)), two sets of four Fmoc-protected aromatic D-amino acids [B (1–4) and C (1–4)(figure 1)] and N-methylamine as a fixed C-terminal wereused for the 3 step synthesis.

The compounds were prepared in parallel as singlecompounds on an automated shaker with overnight runsfor each coupling. Each well was furnished with0.045 mmol of N-methyl-PAL-Resin (75 mg, load: 0.6)and each coupling was carried out with 0.09 mmol of therespective building block, 0.09 mmol of O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexa-fluorophosphate, 0.09 mmol of 1-hydroxy-7-azabenzo-

Figure 1. Peptidomimetic GH secretagogues.

Figure 2. General scaffold holding three building blocks (N-terminal and two aromatic groups) which constitute the 96member GH secretagogue library.

784

triazole and 0.18 mmol of diisopropylethylamine in atotal of 6 mL of dimethylformamide. After deprotection,the resin was treated with TFA at 0 °C and the obtainedcrude product was applied to a C-18 Sep-Pak ClassicCartridge (Waters) and a gradient on water, acetonitrileand TFA. The yield was measured by weight(table I)after lyophilisation and all compounds were characterisedby two HPLC systems(table II). The molecular weight ofall compounds sent for testing was in accordance with themass obtained by LC-MS.Figure 3. The N-Me PAL resin used in the preparation of the

96 member GH secretagogue library.

Figure 4. Building blocks A (1–6), B (1–4) and C (1–4). The building blocks A (1–6) were all Boc-protected and the building blocksB (1–4) and C (1–4) were Fmoc protected for synthesis.

785

3. Results and discussion

A total of 86 compounds were prepared on solidsupport. The synthesis of 5 of the 86 compounds (notedSF in table I) failed completely. The remaining 81 com-pounds were purified by a SepPac Classic Cartridge(Waters) and isolated in amounts ranging from 1–25 mgand purities from 70–100% as shown intable I. Althoughno characterisation of intermediates was carried out inthis study, it is our experience that the preparation of the

sequence B (1–4)-C (1–4)-NHCH3 usually runs withoutproblems, while the coupling of the last amino acid A(1–6) and in particular the final deprotection/cleavagewith TFA causes several problems and should be carriedout with great care (temp.≈ –10–0 °C). Thus, it isinteresting, that three of 12 compounds containing A6failed, and that only two of 12 compounds gave yieldsover 5 mg. In contrast 13 of 14 compounds containing A4gave yields over 5 mg, and if we consider thoseN-terminals which contain ether linkages, i.e. A3, A4 andA5, and compare them with those containing conjugation,i.e. A1, A2 and A6, we clearly see a better final yield inthe ether linkage group, indicating that the final yielddepended on the N-terminal.

All compounds were characterised by two HPLCsystems and LC-MS. The retention-times for the preparedcompounds are given intable II in two different systems.It should be noted that although the building block A1 isa racemic mixture, only one peak was observed in bothHPLC systems. The molecular weight of all compoundswas determined by LC-MS (electrospray) and were inaccordance with the obtained mass, except in those caseswhere the synthesis is denoted as failed. In such cases nofurther characterisation was carried out. HPLC data forthe compounds prepared in solution have been describedelsewhere, and are not included intable II.

The GH releasing properties of the prepared com-pounds, and those in the series previously prepared(marked with a reference) in solution are presented intable III by EC50 values as is the in vitro potency in a ratpituitary cell assay [7, 13].Figures 6–8 are graphic

Figure 5. Reaction scheme for preparation of the 96 memberGH secretagogue library.

Table I. Isolated amount (mg) and purity (%) after SepPac purification of the 86 compounds prepared on solid support.

Entry A1 A2 A3 A4 A5 A6

B1–C1 4 mg, 100 % 5 mg, 100 % 10 mg, 88 % 9 mg, 89 % 11 mg, 90 % NSB2–C1 3 mg, 100 % 2 mg, 100 % SF 8 mg, 100 % 6 mg, 100 % 2 mg, 100 %B3–C1 10 mg, 87 % 8 mg, 70 % 10 mg, 100 % 11 mg, 93 % 12 mg, 87 % 2 mg, 100 %B4–C1 2 mg, 100 % 12 mg, 85 % 14 mg, 90 % 11 mg, 85 % 13 mg, 71 % 13 mg, 70 %B1–C2 5 mg, 100 % 3 mg, 100 % 10 mg, 90 % 12 mg, 93 % 12 mg, 95 % NSB2–C2 4 mg, 81 % 3 mg, 95 % 11 mg, 100 % 7 mg, 90 % 8 mg, 100 % SFB3–C2 5 mg, 100 % 5 mg, 100 % 24 mg, 93 % 26 mg, 100 % 20 mg, 79 % SFB4–C2 10 mg, 81 % 7 mg, 85 % 23 mg, 78 % 21 mg, 71 % 12 mg, 89 % 5 mg, 95 %B1–C3 NS NS NS NS NS NSB2–C3 4 mg, 100 % 3 mg, 100 % 1 mg, 100 % 3 mg, 100 % 3 mg, 100 % 1 mg, 100 %B3–C3 16 mg, 100 % 13 mg, 90 % 15 mg, 94 % 18 mg, 100 % 17 mg, 100 % NSB4–C3 8 mg, 92 % 10 mg, 90 % 13 mg, 95 % 21 mg, 90 % NS 25 mg, 89 %B1–C4 8 mg, 100 % 4 mg, 90 % 18 mg, 86 % 24 mg, 91 % 16 mg, 88 % 3 mg, 75 %B2–C4 3 mg, 100 % 6 mg, 95 % 8 mg, 100 % 11 mg, 90 % 10 mg, 100 % 4 mg, 94 %B3–C4 6 mg, 100 % 7 mg, 100 % 2 mg, 100 % 19 mg, 100 % 21 mg, 84 % 4 mg, 100 %B4–C4 8 mg, 88 % 7 mg, 92 % 6 mg, 100 % 10 mg, 100 % SF SF

NS = not synthesised; SF = synthesis failed; purities based on HPLC system A(table II) at 215 nm.

786

illustrations of all 96 compounds in three different sky-line views to clarify the significance of each group ofbuilding blocks. Each bar depicts the potency of thecompound expressed as how many times the compound ismore potent than 10 nM, i.e., a bar height of 5 equals apotency of 2 nM. This highlights the most potent com-pounds and neglects the less potent compounds.

As illustrated infigures 6–8andtable III, a number ofcombinations are preferable. The combination of thebiphenyl derivative B4 with C1, C2 or C3 gives com-pounds with potencies in the range of 1–55 nM indepen-dent of the six building blocks A (1–6), while thecombination of the naphthyl-derivative B1 only with C2or C3 gives compounds with potencies in the range of2–75 nM. In the same mannerfigures 6–8illustrate anumber of unfavourable combinations. The combination

Table II. HPLC retention-time (system A, system B; min) of the 86 compounds prepared on solid support.

Entry A1 A2 A3 A4 A5 A6

B1–C1 31.05, 33.00 30.32, 32.22 30.13, 31.98 30,07, 31.88 29.77, 31.62 NSB2–C1 30.62, 32.57 29.82, 31.73 SF 29.42, 31.23 29.22, 31.00 29.62, 31.50B3–C1 28.85, 30.67 30.45, 30.43 28.23, 29.65 27.77, 29.40 28.50, 29.23 28.03, 29.77B4–C1 34.58, 36.67 33.92, 35.20 33.57, 34.47 33.30, 35.24 27.68, 30.22 28.48, 30.03B1–C2 32.00, 33.98 31.30, 33.23 31.25, 33.02 31.90, 32.73 30.83, 32.63 NSB2–C2 31.60, 33.57 30.82, 32.77 30.62, 32.45 30.47, 32.30 30.33, 32.10 SFB3–C2 31.07, 32.97 30.25, 32.10 29.32, 30.97 29.45, 30.87 30.17, 30.77 SFB4–C2 35.40, 37.58 34.70, 36.08 34.13, 36.08 27.65, 30.18 34.32, 36.13 27.68, 30.20B1–C3 NS NS NS NS NS NSB2–C3 31.15, 33.10 30.45, 32.28 30.48, 32.16 30.47, 31.97 30.18, 31.68 30.57, 32.18B3–C3 30.87, 32.50 30.02, 31.65 29.38, 30.85 29.00, 30.45 29.15, 30.57 NSB4–C3 35.25, 36.93 34.63, 36.38 34.30, 36.10 34.47, 36.17 NS 34.05, 35.78B1–C4 31.32, 32.92 30.62, 32.15 30.78, 32.27 30.85, 32.40 30.37, 31.90 30.42, 32.10B2–C4 30.90, 32.52 30.15, 31.67 30.18, 31.72 30.20, 31.70 29.87, 31.35 30.42, 31.62B3–C4 30.82, 31.82 36.23, N/A 36.15, 38.22 28.65, 30.02 28.80, 30.17 28.37, N/AB4–C4 34.23, N/A 34.21, 35.90 33.63, 35.65 33.73, 35.71 SF SF

NS = not synthesised; SF = synthesis failed; N/A = data not available.

Table III. Potency (EC50, nM) of the growth hormone secretago-gue library. Mean of at least two separate experiments.

Entry A1 A2 A3 A4 A5 A6

B1–C1 14 155 20 10 220 2a

B2–C1 2 80 SF 25 435 70B3–C1 600 inactive 40 1 000 inactive 405B4–C1 2 18 10 2 40 10B1–C2 11 2 5 13 15 75*B2–C2 75 145 70 155 28 SFB3–C2 40 22 inactive 1 370 500 SFB4–C2 7 9 10 1 2 18B1–C3 13b 40b 22b 16b 13b 18a

B2–C3 125 105 195 725 620 75B3–C3 600 975 1 850 150 650 5a

B4–C3 9 40 1 4 55* 3B1–C4 285 775 525 150 590 19B2–C4 700 inactive 840 inactive 2 850 120B3–C4 360 inactive inactive inactive inactive 650B4–C4 8 30 31 80 SF SF

NS = not synthesised; SF = synthesis failed; N/A = data notavailable;aprepared in solution and described in ref. [9];bpreparedin solution and described in ref. [10];*prepared in solution as inrefs. [9, 10] but not previously described; inactive defined as EC50> 2 000 nM.

Figure 6. 3-Dimensional graphic illustration of the potenciesof the growth hormone secretagogue library. The y-axis expres-ses how many times the compound is more potent than 10 nM.

787

of the benzyloxy-derivative B3 with C1 or C4, and thebenzothiophene-derivative B2 with C3 or C4, with any ofthe six building blocks A (1–6) gave compounds whichonly showed potencies in the range of 40 nM–inactive. Itis noteworthy that some building blocks generally giveinactive or low potent compounds, except for a fewparticular combinations, where an active compound canbe obtained. This is, for instance, illustrated by thep-methoxyphenyl-derivative C4, which in combinationwith any of the four building block B (1–4) and any of the

six building block A (1–6) gave compounds with poorpotencies, except when combined as A6-B1-C4 (19 nM)or A1-B4-C4 (8 nM). A similar pattern is observed, whenthe phenyl derivative C3 is combined with B2 or B3,where the combinations with all six building blocks A(1–6) show compounds with poor potencies, except whencombined as A6-B3-C3 (5 nM). The building blocks B2and B3 give, in general, compounds with poor potencies,except when combined as A1-B2-C1 (2 nM) or A6-B3-C3 (5 nM) (figures 6and7).

By observing each column A1–A6 intable III andfigure 6, it is interesting to see that the most potentcompound in each column never contains the samecombination of building blocks, i.e. A1 gives the mostpotent compound combined with B4-C1, A2 gives themost potent compound combined with B1-C2, and A3gives the most potent compound combined with B4-C3etc. Likewise, it is interesting to note that the twobuilding blocks A3 and A4, which give the two mostpotent compounds in combination with B4-C3 (1 nM)and B4-C2 (1 nM), respectively, give inactive compoundswhen combined with B3-C2 or B3-C4 (for A3) andB2-C4 or B3-C4 (for A4). Altogether, these observationsindicate the danger in conserving one part of the moleculewhile modifying the other part, without simultaneouslymodifying the remaining part of the molecule. Forinstance, it seems obvious that the p-methoxyphenyl-derivative C4 as a building block has poor receptorinteraction since in 23 combinations the compounds showpoor potencies, but nevertheless the building block incombination with A1-B4 shows good potency. Also, thecombination of B4-C3 with A3 gives a very potentcompound (1 nM), while combined with A2 the com-pound shows poor potency, indicating that A2 as anN-terminal is a bad choice. But when A2 is combinedwith B1-C2 it gives a compound more potent than whenB1-C2 is combined with A3. Other discrepancies may behighlighted studyingfigures 6–8and table III.

The use of combinatorial chemistry and high-throughput screening has increased tremendously overthe last few years. Combinatorial chemistry began withthe concept of huge libraries of mixtures and the decon-volution of biologically active mixtures to detect newleads. Nowadays, the automated parallel synthesis ofspecially designed and focused small libraries, made upfrom single compounds, is at the forefront of re-search [14]. Most combinatorial libraries are prepared forlead-finding purposes, and the diversity of the library is ofutmost importance since the chance of finding activecompounds in a library increases with the diversity anddissimilarity of the compounds in the library. However, ifa known lead structure is to be optimised, the library may

Figure 7. 3-Dimensional graphic illustration of the potenciesof the growth hormone secretagogue library. The y-axis expres-ses how many times the compound is more potent than 10 nM.

Figure 8. 3-Dimensional graphic illustration of the potenciesof the growth hormone secretagogue library. The y-axis expres-ses how many times the compound is more potent than 10 nM.

788

have comparatively little diversity [15]. Such small fo-cused combinatorial libraries for optimisation of lead-compounds have been described in the literature earlier( [15] and references therein), but the strength of such astrategy has never been manifested so clearly as wedescribe here.

4. Conclusion

We have described a fast and reliable method tosynthesize GH secretagogues derived from ipamorelin onsolid phase. All prepared compounds were tested for theirability to release GH from rat pituitary cells. The com-pounds show in vitro potencies with EC50 values from 1nM–inactive. Most striking is the unpredictability in theSAR, demonstrated by the fact that some structuralelements (i.e. building blocks) in one combination showhigh potency but totally lack potency in combinationswhich previously had seemed favourable. This fact em-phasizes that even small structural modifications maycause dramatic change in the ligand-receptor interaction,and it demonstrates the necessity for a very definedstructural arrangement to get optimal potency. Illustratedby compounds with potencies at 1 nM and by a numberof discrepancies in the SAR, we have clearly demon-strated the strength of focused combinatorial libraries inSAR optimisation.

5. Experimental protocols

All compounds were prepared in analogy to thesynthesis of 3-(1-aminoethyl)-N-methyl-N-((1R)-1-(N-methyl-N-((1R)-1-(methylcarbamoyl)-2-(2-thienyl)ethyl)-carbamoyl)-2-(2-naphthyl)ethyl)benzamide described be-low using the following buildingblocks: Boc-A1, 3-(1-(tert-butyloxycarbonylamino)ethyl)benzoic acid (racemicmixture); Boc-A2, 3-(tert-butyloxycarbonylaminomethyl)benzoic acid; Boc-A3, (2-(2S)-(tert-butoxycarbonyl-amino)butoxy)acetic acid; Boc-A4, (2S)-2-(((carboxy)methoxy)methyl)pyrrolidin-1-carboxylic acid tert-butylester; Boc-A5, (2-tert-butoxycarbonylamino-2-methyl-propoxy) acetic acid; Boc-A6, (2E)-5-(tert-butyl-oxycarbonylamino)-5-methylhex-2-enoic acid; Fmoc-B1,2-[(9H-fluoren-9-ylmethoxycarbonyl)-methyl-amino]-3-naphthalen-2-yl-propionic acid; Fmoc-B2, (2R)-3-(benzo[b]thiophen-2-yl)-2-(N-(9H-fluoren-9-ylmethoxy-carbonyl)-N-methylamino)propionic acid; Fmoc-B3,(2R)-2-(N-((9H-fluoren-9-ylmethoxy)carbonyl)-N-methyl-amino)-3-benzyloxypropionic acid; Fmoc-B4, (2R)-3-(biphenyl-4-yl)-2-(N-((9H-fluoren-9-yl)methoxycarbonyl)-N-methylamino)propionic acid; Fmoc-C1, (2R)-2-(N-

(9H-fluoren-9-ylmethoxycarbonyl)-N-methylamino)-3-(2-thienyl)propionic acid; Fmoc-C2, (2R)-2-(N-((9H-fluoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(2-fluorophenyl)propionic acid; Fmoc-C3, (2R)-2-(N-(((9H-fluoren-9-yl)methoxy)carbonyl)-N-methylamino)-3-phenyl-propionic acid; Fmoc-C4, (2R)-2-(N-((9H-fluoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(4-methoxy-phenyl)propionic acid.

All building blocks have been purchased from Syn-thetech Inc. (Albany, Oregon 97321) or prepared aspreviously described [8–10].

The RP-HPLC analysis was performed using UVdetection at 214, 254, 276 and 301 nm on a Vydac218TP54 4.6× 250 mm 5µ C-18 silica column, whichwas eluted at 1 mL/min at 42 °C. Two different elutionconditions were used:

System A: the column was equilibrated with 5%acetonitrile in a buffer consisting of 0.1 M ammoniumsulphate, which was adjusted to pH 2.5 with 4 Msulphuric acid. After injection the sample was eluted by agradient of 5–60% acetonitrile in the same buffer for50 min.

System B: the column was equilibrated with 5%acetonitrile/0.1% TFA/water and eluted by a gradient of5% acetonitrile/0.1% TFA/water to 60% acetonitrile/0.1% TFA/water for 50 min.

The LC-MS analyses were performed on a PE SciexAPI 100 LC/MS system using a Waterst 3 × 150 mm3.5µ C-18 symmetry column and positive ionspray witha flow rate of 20µL/min. The column was eluted with alinear gradient of 5–90% acetonitrile, 85–0% water and10% trifluoroacetic acid (0.1%)/water in 15 min at a flowrate of 1 mL/min.

5.1. 3-(1-Aminoethyl)-N-methyl-N-((1R)-1-(N-methyl-N-((1R)-1-(methylcarbamoyl)-2-(2-thienyl)ethyl)-carbamoyl)-2-(2-naphthyl)ethyl)benzamide

The N-Methyl-PAL-Resin (75 mg, 0.045 mmol, load:0.60) was washed with 5% diisopropylethylamine indichloromethane (2× 2 mL), dichloromethane (3× 2 mL)and dimethylformamide (3× 2 mL) and then swelled indimethylformamide (2 mL). Then 2-(N-((9H-fluoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(2-thienyl)pro-pionic acid (46 mg, 0.09 mmol) in dimethylformamide(1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl-uronium hexafluorophosphate (34 mg, 0.09 mmol) indimethylformamide (1 mL), 1-hydroxy-7-azabenzo-triazole (15 mg, 0.09 mmol) in dimethylformamide(1 mL) and diisopropylethylamine (31µL, 0.18 mmol) indimethylformamide (1 mL) were added and the mixture

789

was shaken overnight. The resin was filtered and washedwith dimethylformamide (3× 2 mL), dichloromethane (3× 2 mL) and dimethylformamide (2 mL). Then 20% pip-eridine in dimethylformamide (5 mL) was added and themixture was shaken for 20 min, filtered and washed withdimethylformamide (3× 2 mL), dichloromethane (3×2 mL) and dimethylformamide (2 mL). Then 2-(N-((9H-fluoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(2-naphthyl)propionic acid (41 mg, 0.09 mmol) in dimethyl-formamide (1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (34 mg,0.09 mmol) in dimethylformamide (1 mL), 1-hydroxy-7-azabenzotriazole (15 mg, 0.09 mmol) in dimethylforma-mide (1 mL) and diisopropylethylamine (31µL,0.18 mmol) in dimethylformamide (1 mL) were addedand the mixture was shaken overnight. The resin was fil-tered and washed with dimethylformamide (3× 2 mL),dichloromethane (3× 2 mL) and dimethylformamide(2 mL). Then 20% piperidine in dimethylformamide(5 mL) was added and the mixture was shaken for 20 min,filtered and washed with dimethylformamide (3× 2 mL),dichloromethane (3× 2 mL) and dimethylformamide(2 mL). Then 3-(1-(tert-butoxycarbonylamino)ethyl)-benzoic acid (22 mg, 0.09 mmol) in dimethylformamide(1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl-uronium hexafluorophosphate (34 mg, 0.09 mmol) indimethylformamide (1 mL), 1-hydroxy-7-azabenzo-triazole (15 mg, 0.09 mmol) in dimethylformamide(1 mL) and diisopropylethylamine (31µL, 0.18 mmol) indimethylformamide (1 mL) were added and the mixturewas shaken overnight. The resin was filtered and washedwith dimethylformamide (3× 2 mL), dichloromethane (3× 2 mL) and dimethylformamide (2 mL). The resin wascooled to 0 °C and 50% trifluoroacetic acid in dichlo-romethane (4 mL) was added and the mixture was shakenfor 10 min at 0 °C. The resin was filtered and washed with50% trifluoroacetic acid in dichloromethane (2× 0.5 mL)and the combined filtrates were concentrated under astream of nitrogen. The obtained product was dissolved inacetonitrile/water 1:20 (10 mL) and applied to a C-18Sep-Pak Classic© cartridge (0.25 g, purchased from Wa-tersy), which had been prewashed with acetonitrile(10 mL) and water (10 mL). Then water/trifluoroaceticacid 99.9:0.1 (5 mL), followed by water/acetonitrile/trifluoroacetic acid 89.9:20:0.1 (4 mL) was run throughthe Sep-Pak© and the filtrate was discarded. Then theSep-Pak© was washed with water/acetonitrile/trifluoroacetic acid 64.9:35:0.1 (4 mL) and the filtrate wasdiluted with water (11 mL) and lyophilised to 4 mg of thetitle product.

5.2. In vitro charaterisation in rat pituitary cell assay

The Sprague-Dawley male albino rats (250± 25 g)were purchased from Møllegaard, Lille Skensved, Den-mark. The rats were housed in group cages (four to eightanimals/cage) and placed in rooms with 12 h light cycle.The room temperature varied from 19–24 °C and thehumidity from 30–60%.

The cells were isolated from rat pituitaries and dis-persed into single cells using trypsin and cultured for 3 d.The cells were then washed and stimulated for 15 minwith different GH secretagogues. The supernatant was de-canted and assayed for GH content in a rat GH SPA-assay.

Acknowledgements

We are thankful to Peter Andersen, Annette Heewagen,Lotte G. Sørensen, Anne G. Christensen, Sanne Kold,Annette Nielsen and Nille Birkebæk Larsen for technicalassistance.

References

[1] Bowers C.Y., Chang J., Momany F.A., Folkers F., in: MacIntyre I.(Ed.), Molecular Endocrinology, Elsevier, Amsterdam, 1977, p. 287.

[2] Smith R., Van Der Ploeg L.H.T, Howard A.D., Feighner S.D., ChengK, Hickey G.J. et al., Endocr. Rev. 5 (1997) 621–645.

[3] Clemmesen B., Overgaard K., Riis B., Christiansen C., OsteoporosisInt. 3 (1993) 330–336.

[4] Brixen K., Nielsen H.K., Mosekilde L., Flyvbjerg A., Bone Miner.(1990) 609–618.

[5] Rudman D., Feller A.G., Nagraj H.S., Gergans G., Lalitha P.,Goldberg A.F. et al., N. Engl. J. Med. 323 (1990) 1–9.

[6] Patchett A.A., Nargund R.P., Tata J.R., Chen M.H., Baraket K.J.,Johnston D.B.R. et al., Proc. Natl. Acad. Sci. USA 92 (1995)7001–7005.

[7] Raun K., Hansen B.S., Johansen N.L., Thøgersen H., Madsen K.,Ankersen M., Andersen P.H., Eur. J. Endocrinol. 139 (1998)552–561.

[8] Ankersen M., Johansen N.L., Madsen K., Hansen B.S., Raun K.,Nielsen K.K. et al., J. Med. Chem. 41 (1998) 3699–3704.

[9] Hansen T.K., Ankersen M., Hansen B.S., Raun K., Nielsen K.K.,Lau J. et al., J. Med. Chem. 41 (1998) 3705–3714.

[10] Peschke B., Ankersen M., Hansen B.S., Hansen T.K., Johansen N.L.,Madsen K. et al., Eur. J. Med. Chem. 34 (1998) 361–380.

[11] Albericio F., Kneib-Cordonier N., Biancalana S., Gera L., MasadaR.I., Hudson D., Barany G., J. Org. Chem. 55 (1990) 3730–3743.

[12] PALaldehyde is commercially available from the BioSearch divisionof PerSeptive Biosystems(Framingham, MA).

[13] Heiman M.L., Nekola M.V., Murphy W.A., Lance V.A., Coy D.H.,Endocrinology 116 (1985) 410–415.

[14] Kubinyi H., Curr. Opin. Drug Discov. Develop. 1 (1998) 16–27.

[15] Balkenhohl F., von dem Bussche-Hünnefeld C., Lansky A., ZechelC., Angew. Chem. Int. Ed. Engl. 35 (1996) 2288–2337.

790

Original article

Heterocyclic congeners of PD 128,907 with a partially hydrogenatedbenzomorpholine moiety as potential dopamine D3-receptor ligands

Norbert Matzankea, Werner Löwea, Sylvie Perachonb, Pierre Sokoloffb, Jean-Charles Schwartzb,Holger Starka*

aFreie Universität Berlin, Institut für Pharmazie, Königin-Luise-Strasse 2+4, 14195 Berlin, GermanybUnité de Neurobiologie et Pharmacologie Moléculaire (U. 109), Centre Paul Broca de l’INSERM, 2ter, rue d≠Alésia, 75014 Paris, France

(Received 19 November 1998; revised 14 April 1999; accepted 26 April 1999)

Abstract – With a straightforward seven-step synthesis, racemic perhydro[1,4]benzoxazin-6-on was synthesized in overall good yields viaregioselective epoxid ring-opening to the correspondingâ-aminoalcohol. The oxazine derivative was the key intermediate for the preparationof heteroaromatic analogues of the dopamine D3-receptor preferring agonist PD 128,907. The morpholine moiety of PD 128,907 wasincorporated in diazole and diazine compounds obtained by different ring closure reactions. The target compounds obtained were structurallyrelated to non-ergot heteroaromatic dopamine agonists which display preferential activity at the D3 receptor, e.g., quinpirole, quinerolane, orpramipexole. The five membered aminothiazole, aminoselenazole, and pyrazole derivatives showed at least one order of magnitude higherbinding at the human D3 receptor than that at the D2L receptor. Although the novel compounds displayedKi values only in the micromolarconcentration range, the most active ones showed full agonist activity in a functional assay on mitogenesis. © 1999 Éditions scientifiques etmédicales Elsevier SAS

dopamine / D3-receptor / D2-receptor / PD 128,907 / agonist / mitogenesis / pramipexole / quinpirole / quinerolane

1. Introduction

Dopamine is among the most widely studied neu-rotransmitters of the mammalian central nervous system(CNS). Recent advances in the molecular biology ofdopamine receptors have resulted in the classification intoD1-like (D1 and D5) and D2-like (D2–D4) receptor sub-types [1, 2]. Each of the five characterized dopaminereceptor subtypes belongs to the superfamiliy of Gprotein-coupled receptors and contains approximately400 amino acids arranged in similar tertiary structures ofseven putative transmembrane domains. Although thereceptors were well characterized by molecular biologyand localized in different brain areas, their specificfunctions are poorly understood at present. The D3

receptor has restricted expression in limbic brain areas,associated with cognitive functions and motivated behav-iour [3, 4]. Due to the distinct neuroanatomical distribu-tion of different D2-like subreceptors, as well as pharma-

cological and behavioural studies, it is suggested that D3

receptors seem to be important targets for the develop-ment of drugs for the treatment of Parkinson’s dis-ease [5], schizophrenia [6], and drug abuse, e.g., ligandswith agonist activity at D3 receptors reduced the self-administered cocaine intake in rats [7]. The developmentof more structurally diverse and selective D3-receptoragonists is required to provide a better understanding ofpharmacology and pathophysiology of dopamine receptorsubtypes. Dopamine itself had 20 times higher affinity atthe D3 receptor than that at the D2 receptor [3]. Differentnon-ergot agonists were described possessing varyingdegrees of preference for the D3 receptor. Tetralinderivatives like 7-OH-DPAT (7-hydroxy-N,N-dipropyl-2-aminotetralin) [4] or PD 128,907 ((+)-4aR,10bR-3,4,4a,10b-tetrahydro-4-propyl-2H,5H-[1]benzopyrano[4,3-b]1,4-oxazin-9-ol) [8, 9] showed higher affinity atthe D3 receptor than that at other D2-like receptors. Thesame is true for some heterocyclic analogues like prami-pexole, quinerolane and quinpirole [10, 11] containing anaminothiazole, an aminopyrimidine and a pyrazole ring,*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 791−798 791© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

respectively. The reference agonist showing high activityin combination, with one of the highest degrees of D3

receptor preferring selectivity, is PD 128,907(figure 1).This compound contains a morpholine moiety as onestructural element which may be responsible for theincreased selectivity compared to the structurally related7-OH-DPAT. The afore mentioned heterocyclic agonistsalso displayed a preferential higher affinity at the D3

receptor compared to that at other D2-like receptors. Theaim of this study was the synthesis and pharmacologicalinvestigation of compounds which have unified the struc-tural morpholine element of PD 128,907 and the hetero-

cyclic moiety of other D3-receptor preferring agonists orrelated structures, e.g., pramipexole, quinerolane, andquinpirole.

2. Chemistry

The racemicN-substituted perhydro[1,4]benzoxazin-6-on ((±)-7) as key intermediate was synthesized by thestraightforward seven-step route outlined infigure 2.

The synthesis oftrans-configurated (±)-3 relied on amodification of chemistry originally described by Chengand co-workers [12] for the preparation of ligands for thej opioid receptor. Cheng et al. obtained the epoxid2 viaa four-step route starting from 1,4-cyclohexanedionemonoethylene ketal, but this can be done in a simpletwo-step procedure previously described [13, 14]. Thus,the enol ether of 1-methoxy-1,4-cyclohexadiene [15] wascleaved by a catalytic amount ofp-toluyl sulfonic acid(p-TsOH) and the resulting carbonyl group was subse-quently protected as ethylene ketal. Epoxidation of1 withm-chloroperoxybenzoic acid (m-CPBA) provided 2 in

Figure 1. Structure of PD 128,907.

Figure 2. Synthesis of precursor (±)-7. (a) ethylene glycol,p-TsOH, toluene,∆T; (b) m-chloroperoxybenzoic acid, CH2Cl2; (c)propylamine, water, ambient temperature; chloroacetyl chloride, Et3N, CH2Cl2; (e) n-tetrabutylammonium hydrogensulphate (cat.),NaOH; (f) LiAlH4, THF, ∆T; (g) 2 N HCl, ∆T.

792

good overall yield. Epoxid ring-opening of2 with propylamine resulted exclusively in the formation of the desiredtrans-configuratedâ-aminoalcohol (±)-3. The isomericaminoalcohol could not be isolated, as depicted infigure 2. The selectivity and rate of conversion of thisreaction is remarkable, since such nucleophilic epoxidering openings with amines normally require hetero-geneous catalysis [16] or amine activation with aluminia-organic reagents [17]. This observation of selectivity iscongruent to the experiments of Chen et al. [12], whoisolated, in the reaction of this epoxid with pyrrolidine,the two regioisomeric aminoalcohols in a ratio of 13:1(89% yield).

Reaction of3 with chloroacetyl chloride [18] led to theamide4, which is cyclized under phase-transfer condi-tions to the 1,4-benzoxazine5. In the1H NMR spectra of4 both rotameres are present in equal amounts. Thisamide was reduced with LiAlH4 in THF to give the amine6. Initial attemps to deprotect6 with PPTS/wet ac-etone [19] failed, but this could be achieved by 2 NHCl [20] and afforded the desired racemic aminoketone7.

The heterocyclic substituted 1,4-benzoxazines weresynthesized as shown infigure 3. The thiazolo- andselenazolo substituted compounds8 and9 were preparedin a well known manner [21] by treatment of7 withbromine in acidic medium, and subsequent reaction of the

α-haloketone intermediate either with thiourea or seleno-urea. Bromination of7 could have taken place on eitherside of the carbonyl group to produce either the desiredlinear compounds8 and 9 or the undesired angularproducts. However, bromination of position 5 is stericallyunfavoured and the evidence of substitution could bedone by an1H NMR off resonance decoupling experi-ment of compound8: decoupling of the 9a-proton re-sulted in a double doublet pattern of the protons in the9-position, which would be impossible if the moleculewas angularly annelated. Analogous observations weredescribed by Kornfeld et al. [22] in the course of thepreparation of rigid 3-(2-aminoethyl)pyrroles as dopam-ine agonists.

The precursor10 as a masked 1,3-dicarbonyl com-pound was synthesized by the treatment of the free baseof 7 with tris(dimethylamino)methane [23] in good yield.Treatment of this enaminone10 either with guanidinehydrochloride/K2CO3 [24] or hydrazine [20, 22] resultedin the formation of the 2-aminopyrimidine derivative12and the pyrrole derivative11 in its corresponding tauto-meric forms, respectively. All compounds which wereselected for pharmacological testing were converted withanhydrous HCl in their water soluble salts (see Experi-mental section). Due to the moderate pharmacologicalaffinity the compounds were not separated into theirenantiomers.

Figure 3. Synthesis of heterocyclic dopamine receptor ligands. (a) i: Br2, HBr, 0 °C, ii: H2N-CX-NH2, ∆T; (b) HC(NMe2)3, DMF,65 °C; (c) hydrazine, MeOH, ambient temperature; (d) guanidine x HCl, K2CO3, EtOH, ∆T.

793

3. Pharmacology

Displacement assays were performed with chinesehamster ovary (CHO) cell lines stably transfected withhuman D2L or D3 receptor cDNA using [125I]iodo-sulpiride at a concentration of 0.1 nM [25]. The nonspe-cific binding was determined in the presence of enom-apride. Binding data were analysed by computerizednonlinear regression for a one-site model. TheKi valueswere derived from IC50 values according to the Cheng-Prussoff equation [26].

Additional functional experiments were performed toclarify the mode of action for the most active compoundsof this series. NG 108-15 cells expressing the humandopamine D3 receptor were incubated with forskoline andthe drug in increasing concentrations [25]. Then, [3H]thy-midine was added, and increase in mitogenesis, as ameasurement of agonist activity, could be measured bycounting the radioactivity incorporated into the cells after2 h. A decrease in, or inhibition of, mitogenesis in relationto the effect of the reference compound quinpirole wouldshow a partial agonist or antagonist effect, respectively.

4. Pharmacological results and discussion

The target compounds (±)-8–12 displayed at least 10times higher binding affinities at the D3 receptor than thatat the D2 receptor(table I). Their affinity constants at theD3 receptor were found to be in the micromolar concen-tration range. Although the compounds (±)-8–11 werefound to possess higher D3-receptor affinity than theintermediate products of their chemical synthesis, thedifference between these heteroaromatic products andtheir chemical educts were not important due to their

overall moderate affinity constants. The aminopyrimidinederivative (±)-12 showed equipotency to the secondaryamine intermediate (±)-3 and even lower D3-receptoraffinity than the non-aromatic morpholine precursor mol-ecule (±)-6. Compounds possessing a five-memberedheteroaromatic moiety ((±)-8–11) were found in thisseries to be active in a lower concentration range than theother tested compounds. Interestingly, the seleno ana-logue (±)-9 of the pramipexole related derivative (±)-8showed about two times higher affinity than that ofcompound (±)-8. For dopamine agonists used in therapyfor Parkinson’s disease, an anti-oxidative or radicalscavenger effect seemed to be beneficial. These propertiesmay be enhanced in seleno heterocycles compared tosulfur heterocycles. Therefore, further investigations tostudy the anti-oxidative effects of aminothiazole andaminoselenazole derivatives are currently in progress.

Compared to PD 128,907, pramipexole, quinerolane,and quinpirole, the novel compounds possess the basicamino functionality on another side of the molecule(reversed orientation), i.e., the positions of the oxygenand the nitrogen in the morpholine moiety have to bechanged to be able to superimpose with the referenceagonists. Concerning the pharmacological activity, thisseemed to be the major drawback for the compoundsprepared. The position of the basic amino functionality inrelation to the aromatic hydrogen-bond area were de-scribed in different molecular modelling and structure-activity relationship studies to be essential for highaffinity D2-like receptor binding [27]. These findingscould be emphasized by the results of the present studyfor binding at D3-receptors. Compounds with an oppositeorientation of oxygen and nitrogen atom in the morpho-line moiety possessing a stronger relationship to PD

Table I. Binding and functional studies at Dopamine D2 and D3 receptors.

Inhibition of binding at Stimulation of D3 receptor

Compound D2 receptor D3 receptor Ki (D2)/ EC50 i.a.a

Ki [µM] Ki [µM] Ki (D3) [µM]

(±)-3 600 68± 3 13 ndb nd(±)-6 > 600 172± 28 > 3.4 nd nd(±)-8 99 ± 2 6.2 ± 0.5 16 7.1± 3.0 106(±)-9 49 ± 3 2.9 ± 0.4 17 1.3± 0.08 101(±)-11 99 ± 5 2.3 ± 0.3 43 0.54± 0.09 100(±)-12 > 600 46± 5 > 12 nd ndQuinpirolec,d 1.4 0.039 36 0.00086 100Quinerolanec,d 0.34 0.004 95 0.00015 96Pramipexolec 0.79 0.004 193 0.00023 98PD 128,907c 0.389 0.002 216 0.00064 96

(a) i.a. = intrinsic activity (i.a. (quinpirole) = 100); (b) nd = not determined; (c) ref. [32]; (d) ref. [33].

794

128,907 compared to the ones described in this paperhave already been reported as potent dopamine receptoragonists [28, 29].

Nevertheless, when the three most active compounds(±)-8–11 were tested for their mode of action in themitogenesis test it was found that all compounds wereable to induce a full mitogenesis rate indicating a fullagonist behaviour. Although the compounds displayedonly moderate affinity they were still able to induce achange in the receptor structure that is required for signaltransduction indicated by complete receptor activation.

5. Experimental protocols

Melting points were uncorrected.1H NMR spectrawere determined with a Bruker AC 300 (300 MHz) or aBruker Avance DPX 400 (400 MHz) instrument asindicated by spectrometer frequency. Chemical shiftswere reported inδ values (ppm) relative to an internalstandard of tetramethylsilane. Abbreviations used inNMR analysis were as follows: s, singlet; d, doublet; t,triplet; m, multiplet; br s, broad singlet; and dd, doublet ofdoublets. IR spectra were determined with a Perkin-Elmer 297 spectrophotometer and Mass spectra wereobtained on a CH-7A-Varian MAT (70 eV) instrument.Microanalyses were performed with a Perkin-Elmer 240B and C analyzer. Analyses indicated by the symbols ofthe elements or functions were within± 0.4% of theo-retical values. Thin layer chromatography was performedon Merck precoated TLC plates with silica gel 60-F254and visualized with UV light (254 nm), after treatmentwith iodine or after treatment with Fast Blue B salt.Column (flash) chromatography [30] was performed withMerck silica gel 60 (230–400 mesh). Solvents andreagents were used as purchased, except as noted. Ether,THF and toluene were distilled from sodium metal/benzophenone ketyl. [125I]Iodosulpiride (2 000 Ci/mmol)and [3H]thymidine (110 Ci/mmol) were obtained fromAmersham International, Ltd. (Buckinghamshire, UK)and emonapride (YM 09152) from the YamanouchiPharmaceutical Co. (Tokyo, Japan).

5.1. Chemistry

5.1.1. 1,4-Dioxaspiro[4,5]dec-7-ene(1)Compound1 was prepared according to the procedure

of Lambert et al. [13] from 1-Methoxy-1,4-cyclo-hexadiene [15] in 80% yield: b.p. 70–71 °C (13 mbar)(lit. [13] b.p. 62–64 °C, 7 Torr); IR (CHCl3 solution)3 027, 2 934, 2 879, 1 687 cm–1; 1H NMR (CDCl3, 300MHz) δ 1.74–1.81 (m, 2H), 2.27–2.28 (m, 4H), 3.99

(s, 4H), 5.60–5.74 (m, 2H); MSm/e140 (M, 35), 125 (8),112 (8), 99 (62), 86 (100).

5.1.2. (±)-Spiro[1,3-dioxolane-2,3≠-[7]oxabicyclo[4,1,0]heptane](2)

Compound2 was prepared by the procedure of Chenget al. [12] from 1 in 67% yield: b.p. 99 °C (9 mbar)(lit. [14] b.p. 100 °C, 16 Torr). IR (CHCl3 solution)2 942, 2 882 cm–1; 1H NMR (CDCl3, 300 MHz) δ1.41–1.70 (m, 2H), 2.00–2.20 (m, 4H), 3.15–3.18 (m,2H), 3.87–3.98 (m, 4H); MSm/e 156 (M, 2), 140 (22),112 (12), 99 (85), 86 (100).

5.1.3. (±)-trans-7-Propylamino-1,4-dioxaspiro[4,5]decan-8-ol ((±)-3)

To a stirred mixture of2 (6.25 g, 40.0 mmol) andpropylamine (6.58 mL, 80.0 mmol) was added 4 mL H2Odropwise at 0 °C. The reaction mixture was stirred atambient temperature under N2. After 40 h the mixturewas evaporated in vacuo to remove the water and theexcess of propylamine to give a red oil. This oil waspurified by flash chromatography (10% MeOH in CH2Cl2with 1% concentrated NH4OH) on silica gel to give (±)-3(5.148 g, 60%) as a light red oil which slowly crystallizedon standing: m.p. 65–66 °C; IR (KBr) 3 390, 3 300,3 166, 2 956, 2 876, 2 824, 1 635 cm–1; 1H NMR (CDCl3,400 MHz)δ 1.00 (m, 3H), 1.39–2.55 (m, 10H), 2.79 (m,1H), 3.29 (m, 1H), 4.02 (m, 4H) (no OH/NH signals weredetected); MSm/e 215 (M, 5), 186 (37), 156 (64), 129(6), 101 (10), 86 (100). Anal. C11H21NO3 (C, H, N).

5.1.4. (±)-trans-2-Chloro-N-(8-hydroxy-1,4-dioxaspiro[4,5]decan-7-yl)-N-propyl-acetamid((±)-4)

To a stirred solution of (±)-3 (5.08 g, 23.6 mmol) andtriethylamine (3.29 mL, 23.6 mmol) in 100 mL CH2Cl2was added chloroacetyl chloride (1.88 mL, 23.6 mmol)dropwise over a period of 10 min at 0 °C. After 30 minthe mixture was allowed to warm to ambient temperatureand stirred for a further 2 h. The organic layer wasseparated, washed with 20 mL 2 N HCl, 20 mL H2O,20 mL brine, dried (MgSO4), and concentrated in vacuowhereupon red crystals formed. Recrystallization fromCH2Cl2/n-hexane gave pure (±)-4 (4.50 g, 65%) as co-lourless crystals: m.p. 172 °C; IR (KBr) 3 408, 2 961,2 880, 1 647 cm–1; 1H NMR (CDCl3, 400 MHz) δ0.90–0.94 (m, 6H), 1.49–2.14 (m, 21H), 2.22 (s, 1H,collapse after D2O-exchange), 2.56 (1H, collapse afterD2O-exchange), 2.94 (m, 1H), 3.23–3.34 (m, 2H),3.65–3.76 (m, 2H), 3.97–4.13 (m, 9H), 4.33–4.35 (m,1H); MSm/e291 (M, 1), 256 (14), 232 (23), 156 (49), 86(100). Anal. C13H22ClNO4 (C, H, N).

795

5.1.5. (±)-trans-4-Propyl-2,3,4,4a,5,7,8,8a-octahydro-spiro[6H-1,4-benzoxazine-6,2≠-[1,3]dioxolane]-3-on((±)-5)

Compound (±)-4 (4.50 g, 15.4 mmol) and tetra-n-butylammonium hydrogen sulphate (525 mg, 1.54 mmol)were dissolved in 300 mL CH2Cl2 under an atmosphereof nitrogen. 61.7 mL of 0.5 N NaOH was added viasyringe, and the biphasic system was stirred vigerously atambient temperature for 20 h. The organic layer wasseparated, washed with water (2× 40 mL), brine (40 mL),dried (MgSO4), and concentrated in vacuo to give ayellow oil. This oil was purified by flash chromatography(10% MeOH in CH2Cl2) on silica gel to give (±)-5(3.804 g, 97%) as a light yellow oil which slowlycrystallized on standing: m.p. 67 °C; IR (KBr) 3 448,2 964, 2 880, 1 655 cm–1; 1H NMR (CDCl3, 400 MHz)δ0.91 (m, 3H), 1.43–2.13 (m, 8H), 3.04–3.74 (m, 4H), 3.98(s, 4H), 4.26 (dd,J = 16.0/21.2 Hz, 2H); MSm/e255 (M,92), 226 (31), 198 (53), 155 (68), 141 (47), 99 (100).Anal. C13H21NO4 (C, H, N).

5.1.6. (±)-trans-4-Propyl-2,3,4,4a,5,7,8,8a-octahydro-spiro[6H-1,4-benzoxazine-6,2≠-[1,3]dioxolane] hydro-chloride ((±)-6)

LiAlH 4 (1.61 g, 42.3 mmol) was suspended in 20 mLTHF under an atmosphere of nitrogen. The solution of(±)-5 (3.60 g, 14.1 mmol) in 30 mL THF was addeddropwise via syringe at 0 °C over a period of 20 min. Thereaction mixture was allowed to warm to ambient tem-perature and then heated at reflux for 2 h. After cooling,excess hydride was destroyed by the careful addition of10 mL water. The mixture was filtered and the solid waswashed with ether (5× 20 mL). The filtrate was washedwith brine (2× 30 mL), dried (K2CO3) and concentratedin vacuo. The resulting oil was purified by flash chroma-tography (10% MeOH in CH2Cl2 with 1% concentratedNH4OH) on silica gel to give the free base of (±)-6 as alight yellow oil. This oil was dissolved in dry ether and astream of dry gasous HCl was bubbled through thesolution at 0 °C to produce the hydrochloride (±)-6(3.793 g, 97%) as colourless crystals: m.p. 180–181 °C;IR (KBr) 3 431, 2 965, 2 881, 2 585, 2 475, 2 325 cm–1;1H NMR (DMSO-d6, 400 MHz) δ 0.90 (m, 3H),1.44–1.86 (m, 7H), 2.29 (m, 1H), 2.85–3.38 (m, 5H), 3.73(m, 1H), 3.92 (m, 6H), 11.41 (br s, 1H); MSm/e241 (M,12), 212 (100), 139 (29). Anal. C13H23NO3⋅HCl (C, H,N).

5.1.7. (±)-trans-4-Propyl-3,4,4a,5,6,7,8,8a-octahydro-2H-1,4-benzoxazin-6-on hydrochloride((±)-7)

Compound (±)-6 (3.793 g, 13.7 mmol) was dissolvedin 50 mL 2 N HCl and heated to reflux for 2 h, cooled,

neutralized with 10% NaOH, saturated with NaCl andextracted with CH2Cl2 (5 × 20 mL). The organic extractswere dried over K2CO3 and concentrated in vacuo toprovide a yellow oil. This oil was purified by flashchromatography (10% MeOH in CH2Cl2 with 1% con-centrated NH4OH) on silica gel to give the free base of(±)-7 as a light yellow oil, which was converted in thehydrochloric salt as described before to give (±)-7(2.954 g, 93%) as colourless crystals: m.p. 197 °C; IR(KBr) 3 419, 2 969, 2 880, 2 581, 2 479, 2 419,1 719 cm–1; 1H NMR (DMSO-d6, 400 MHz)δ 0.91 (t,J= 7.3 Hz, 3H), 1.63–1.69 (m, 3H), 2.06–3.47 (m, 10H),4.02–4.03 (m, 2H), 4.13–4.14 (m, 1H), 11.83 (br s, 1H);MS m/e 197 (M, 34), 168 (100), 140 (42). Anal.C11H19NO2⋅HCl (C, H, N).

5.1.8. (±)-trans-8-Propyl-4a,6,7,8,8a,9-hexahydro-4H-thiazolo[4,5-g][1,4]benzoxazin-2-amine dihydrochloride((±)-8)

To a stirred solution of (±)-7 (233.7 mg, 1 mmol) in10 mL 47% HBr was added Br2 (1.76 mL of a 10%solution in 47% HBr, 1.1 mmol) dropwise at 0 °C. After2 h, thiourea (83.7 mg, 1.1 mmol) was added as a solid,and the mixture was heated to reflux for further 2 h. Afterevaporation of the HBr in vacuo, the residue was treatedwith 10 mL 10% NaOH with cooling and extracted withCH2Cl2 (5 × 10 mL). The combined organic layers weredried (MgSO4) and concentrated in vacuo to provide acrude base of (±)-8 as a brown solid. This residue waspurified by flash chromatography (10% MeOH in CH2Cl2with 1% concentrated NH4OH) on silica gel to give thefree base of (±)-8 as a foam. This foam was dissolved indry ethanol (15 mL), treated with saturated HCl inethanol (2 mL) and evaporated in vacuo again. Recrys-tallization from dry 2-propanol/ether provided the titlecompound (185.7 mg, 57%) as a grey solid: m.p.240–241 °C; IR (KBr) 3 375, 3 247, 3 064, 2 936, 2 720,2 621, 2 510, 1 627, 1 577 cm–1; 1H NMR (DMSO-d6,400 MHz)δ 0.97 (t,J = 7.1 Hz, 3H), 1.73–1.87 (m, 2H),2.51–3.47 (m, 9H), 4.30–4.44 (m, 3H), 9.30 (br s, 3H),12.67 (br s, 1H); MSm/e253 (M, 28), 224 (29), 151 (8),127 (100). Anal. C12H19N3OS⋅2 HCl⋅H2O (C, H, N).

5.1.9. (±)-trans-8-Propyl-4a,6,7,8,8a,9-hexahydro-4H-selenazolo[4,5-g][1,4]benzoxazin-2-amine dihydro-chloride ((±)-9)

This compound was prepared and purified in the samemanner as described in the synthesis of (±)-8 in a1.5 mmol scale to provide (±)-9 (137.8 mg, 25%) as alight red solid: m.p. 220–221 °C; IR (KBr) 3 397, 3 260,3 073, 2 933, 2 722, 2 624, 1 623, 1 579 cm–1; 1H NMR(DMSO-d6, 400 MHz) δ 0.94 (t, J = 7.3 Hz, 3H),

796

1.73–1.77 (m, 2H), 2.89–3.52 (m, 9H), 4.03–4.19 (m,3H), 9.58 (br s, 3H), 12.12 (br s, 1H); MSm/e301 (M, Semain isotope, 12), 272 (13), 127 (100). Anal.C12H19N3OSe⋅2 HCl⋅H2O (C, H, N).

5.1.10. (±)-trans-(E)-7-Dimethylaminomethylidene-4-propyl-2,3,4,4a,5,6,8,8a-octahydro-1,4-benzoxazin-6-on((±)-10)

To a solution of (±)-7 (986.4 mg, 5 mmol, free base) in5 mL of dry DMF was added tris(dimethylami-no)methane (2.33 mL, 15 mmol). The solution wasstirred for 16 h at 65 °C under a nitrogen atmosphere andthen concentrated in vacuo to provide a brown oil. Thisoil was purified by flash chromatography (10% MeOH inCH2Cl2 with 1% concentrated NH4OH) on silica gel togive (±)-10 (1.066 g, 85%) as a yellow oil which veryslowly crystallized on standing. A portion of this materialwas recrystallized from cyclohexane/n-hexane to give(±)-10 as yellow crystals: m.p. 65 °C; IR (KBr) 3 451,2 959, 2 859, 2 803, 1 648, 1 559 cm–1; 1H NMR(DMSO-d6, 400 MHz) δ 0.83 (t, J = 7.2 Hz, 3H),1.29–1.46 (m, 2H), 1.89 (dd,J = 5.8/11.6 Hz, 1H),2.03–2.62 (m, 6H), 2.73 (d,J = 11.4 Hz, 1H), 2.96 (dd,J= 4.2/6.7 Hz, 1H), 3.05 (s, 6H), 3.23–3.27 (m, 1H), 3.56(t, J = 11.1 Hz, 1H), 3.77 (d,J = 11.1 Hz, 1H), 7.31 (s,1H); MS m/e252 (M, 29), 235 (7), 193 (25), 84 (84), 58(100). Anal. C14H24N2O2 (C, H, N).

5.1.11. (±)-trans-8-Propyl-1,4,4a,6,7,8,8a,9-octahydro-pyrazolo[3,4-g][1,4]benzoxazine dihydrochloride((±)-11) (cf. [28, 29])

To a solution of (±)-10 (309.5 mg, 1.23 mmol) in10 mL dry methanol was added hydrazine (0.39 mL,12.3 mmol) and the mixture was stirred for 19 h. Thesolvent was evaporated and the crude product was puri-fied by flash chromatography (10% MeOH in CH2Cl2with 1% concentrated NH4OH) on silica gel to give thebase of (±)-11 (206.8 mg, 76%) as a yellow oil. The HClsalt was formed in ethanol and crystallized from dry2-propanol/ether to provide (±)-11as a hygroscopic foam.IR (KBr) 3 395, 2 932, 2 877, 2 688, 2 615, 2 487, 2 360,1 676, 1 645, 1 570 cm–1; 1H NMR (DMSO-d6, 400MHz) δ 0.96 (m, 3H), 1.75 (m, 2H), 2.51–3.48 (m, 9H),4.03–4.11 (m, 3H), 7.59 (s, 1H), 7.75 (br s, 2H), 11.97 (brs, 1H); MSm/e221 (M, 48), 192 (100), 127 (39), 98 (22).Anal. C12H19N3O⋅2 HCl (C, H, N).

5.1.12. (±)-trans-9-Propyl-5a,7,8,9,9a,10-hexahydro-5H-pyrimidino[4,5-g][1,4]benzoxazin-2-amine dihydro-chloride ((±)-12)

To a solution of (±)-10 (293.1 mg, 1.16 mmol) in15 mL dry ethanol was added dry guanidine⋅HCl (1.11 g,11.6 mmol) and K2CO3 (1.61 g, 11.6 mmol). The suspen-

sion was refluxed for 2 h and then evaporated to drynessin vacuo. The residue was treated with water (30 mL) andextracted with CH2Cl2 (5 × 10 mL). The combinedorganic layers were dried (K2CO3) and concentrated invacuo. The solid residue was converted into its hydro-chloric salt and recrystallized from dry ethanol/ether toprovide (±)-12 (252.8 mg, 68%) as a yellow solid: m.p.196 °C. IR (KBr) 3 229, 3 137, 2 969, 2 935, 2 880,2 689, 2 542, 2 484, 2 349, 1 664 cm–1; 1H NMR(DMSO-d6, 400 MHz) δ 0.96 (t, J = 7.3 Hz, 3H),1.72–1.77 (m, 2H), 2.54–3.53 (m, 9H), 4.05–4.16 (m,3H), 7.99 (br s, 3H), 8.33 (s, 1H), 12.24 (br s, 1H); MSm/e248 (M, 21), 219 (100). Anal. C13H20N4O⋅2 HCl (C,H, N).

5.2. Pharmacological testing

5.2.1. Displacement studies [25]Transfected cell lines consisting of CHO cell lines

stably transfected with human dopamine D2L or D3

receptor DNA [11] were cultured in Dulbecco’s ModifiedEagle Medium supplemented with 10% foetal calf serumin a 5% CO2 humidified atmosphere. They were har-vested from culture dishes in the presence of 0.2%trypsin, centrifuged at 2 000g for 5 min and homog-enized in 10 mM Tris-HCl, pH 7.4 containing 5 mMMgCl2 using a Polytron. The homogenate was centri-fuged at 20 000g for 15 min at 4 °C, and the pellet wasresuspended by sonification in 50 mM Tris-HCl, pH 7.4,containing 120 mM: NaCl, 5 mM KCl, 2 mM CaCl2, and8 mM MgCl2 (incubation buffer). Membranes were eitherused immediately or after storage at –70 °C. Bindingassays were started by the addition of 200µL of mem-branes (1–20µg of protein) from transfected cells dilutedin incubation buffer 1 supplemented with 0.2% bovineserum albumin to polystyrene tubes containing, in100µL, 0.1 nM [125I]iodosulpiride and drug diluted in100µL of incubation buffer. Nonspecific binding (1–20%of total binding) was determined in the presence of 1µMenomapride. Incubations were run at 30 °C for 30 min.All reactions were stopped by vacuum filtration throughWhatman GF/B glass-fibre filters coated in 0.3% poly-ethylenimine with an automated cell harvester (Brandel-Beckmann, Gauthersburg, MD) and were rinsed 3 timeswith 5 mL of ice-cold incubation buffer. Filters werecounted by gamma spectrometry in 5 mL of ACS II(Amersham).

EC50 values (± SE) and maximal responses werecalculated from concentration-response curves. Intrinsicactivity was calculated relative to quinpirole, a fullagonist [31].

797

5.2.2. Mitogenesis [25]NG 108-15 cells expressing the human dopamine

D3-receptor [31] were cultured in Dulbecco’s ModifiedEagle Medium supplemented with 10% foetal calf serumin a 5% CO2 humidified atmosphere and plated incollagen-coated 96-well plates. After a 24 h culture, cellswere washed twice with culture medium without foetalcalf serum and incubated for 16 h with 1µM forskolinand the drug in increasing concentrations or quinpirole ascontrol. Then, [3H]thymidine (1µCi/well) was added for2 h and cells were harvested by vacuum filtration throughWhatman GF/C glass-fibre filters by using an automatedcell harvester and were rinsed 15 times with 200µL ofphosphate-buffered saline. Radioactivity was counted byliquid scintigraphy in 5 mL of ACS (Amersham).Ki

values were derived from IC50 values according to theCheng-Prussoff equation [26], taking into account theKd

of [125I]iodosulpiride for respective receptors. Data weremeans ofKi values from data obtained in at least threeseparate experiments, and the statistical error was ex-pressed as± SEM.

Acknowledgements

This work was supported by the Biomedical & HealthResearch programme (BIOMED) of the European Unionand by the National Institute on Drug Abuse (NIDA)(DAII534-01), USA.

References

[1] Schwartz J.-C., Giros B., Martres M.P., Sokoloff P., Semin. Neuro-sci. 4 (1992) 99–108.

[2] Gingrich J.A., Caron M.G., Annu. Rev. Neurosci. 16 (1993)731–745.

[3] Sokoloff P., Giros B., Martres M.P., Bouthenet M.L., Schwartz J.-C.,Nature (London) 347 (1990) 146–151.

[4] Lévesque D., Diaz J., Pilon C., Martres M.P., Giros B., Souil E.,Schott D., Morgat J.L., Schwartz J.-C., Sokoloff P., Proc. Natl. Acad.Sci. USA 89 (1992) 8155–8159.

[5] Hagan J.J., Midlemiss D.N., Sharpe P.C., Poste G.H., TrendsPharmacol. Sci. 18 (1997) 156–163.

[6] Shafer R.A., Levant B., Psychopharmacology 135 (1998) 1–16.

[7] Caine S.B., Koob G.F., Parsons L.H., Everitt B.J., Schwartz J.-C.,Sokoloff, P., Neuroreport 8 (1997) 2373–2377.

[8] Pugsley A., Davis M.D., Akunne H.C., Mackenzie R.G., Shih Y.H.,Damsma G. et al., J. Pharmacol. Exp. Ther. 275 (1995) 1355–1366.

[9] Bancroft G.N., Morgan K.A., Flietstra R.J., Levant B., Neuropsy-chopharmacology 18 (1998) 305–316.

[10] Mierau J., Schneider F.J., Ensinger H.A., Chio C.L., Lajiness M.E.,Huff R.M., Eur. J. Pharmacol. 290 (1995) 29–36.

[11] Sokoloff P., Andrieux M., Besançon R., Pilon C., Martres M.P.,Giros B., Schwartz J.-C., Eur. J. Pharmacol. 225 (1992) 331–337.

[12] Cheng C.Y., Wu S.C., Hsin L.W., Tam S.W., J. Med. Chem. 35(1992) 2243–2247.

[13] Lambert J.B., Xue L., Bosch R.J., Taba K.M., Marko D.E., Urano S.,Lebreton P.R., J. Am. Chem. Soc. 108 (1986) 7575–7579.

[14] Van der Louw J., Van der Baan J.L. Out G.J.J., De Kanter F.J.J.,Bickelhaupt F., Klumpp G.W., Tetrahedron (1992) 9901–9905.

[15] Wilds A.L., Nelson N.A., J. Am. Chem. Soc. 75 (1953) 5360–5365.

[16] Posner G.H., Rogers D.Z., J. Am. Chem. Soc. 99 (1977) 8208–8214.

[17] Gorzynski Smith J., Synthesis (1984) 629–656.

[18] Dijkstra D., Mulder T.B., Rollema H., Tepper P.G., Van der Weide J.,Horn A.S., J. Med. Chem. 31 (1988) 2178–2182.

[19] Sterzycki R., Synthesis (1979) 724–725.

[20] McQuaid L.A., Latz J.E., Clemens J.A., Fuller R.W., Wong D.T.,Mason N.R., J. Med. Chem. 32 (1989) 2388–2396.

[21] Berney D., Schuh K., Helv. Chim. Acta 65 (1982) 1304–1309.

[22] Bach N.J., Kornfeld E.C., Jones N.D., Chaney M.O., Dorman D.E.,Paschal J.W., Clemens J.A., Smalstig E.B., J. Med. Chem. 23 (1980)481–491.

[23] White J.D., Badger R.A., Kezar III H.S., Pallenberg A.J., SchiehserG.A., Tetrahedron 45 (1989) 6631–6644.

[24] De D., Mague J.T., Byers L.D., Krogstad D.J., Tetrahedron Lett. 36(1995) 205–208.

[25] Sautel F., Griffon N., Sokoloff P., Schwartz J.C., Launay C., SimonP. et al., J. Pharmacol. Exp. Ther. 275 (1995) 1239–1246.

[26] Cheng Y.C., Prussoff W.H., Biochem. Pharmacol. 22 (1973)3099–3108.

[27] Casagrande C., Bertolini G., in: Giardinà D., Piergentili A., PiginiM. (Eds.) Ser. Pharmacochemistry Library, Vol. 24, Elsevier, Am-sterdam, 1996, pp. 67–84.

[28] Booher R.N., Kornfeld E.C., Smalstig E.B., Clemens J.A., J. Med.Chem. 30 (1987) 580–583.

[29] Booher R.N., Cornfeld E.C., (Lilly, Eli and Co.) U. S. US 4,552,956.(1986); Chem. Abstr. 104 (1986) P186431k.

[30] Still W.C., Kahn M., Mitra A., J. Org. Chem. 43 (1978) 2923–2925.

[31] Pilon C., Lévesque D., Dimitriadou V., Griffon N., Martres M.P.,Schwartz J.-C., Sokoloff P., Eur. J. Pharmacol. 268 (1994) 129–139.

[32] Sautel F., Griffon N., Lévesque D., Pilon C., Schwartz J.-C.,Sokoloff P., Neuroreport 6 (1995) 329–332.

[33] Sokoloff P., Andrieux M., Besancon R., Pilon C., Martres M.P.,Giros B., Schwartz J.-C., Eur. J. Pharmacol. 225 (1992) 331–337.

798

Original article

Carbonic anhydrase inhibitors – part 70#. Synthesis and ocular pharmacologyof a new class of water-soluble, topically effective intraocular pressure lowering

agents derived from nicotinic acid and aromatic/heterocyclic sulfonamides

Claudiu T. Supurana*, Andrea Scozzafavaa, Luca Menabuonib, Francesco Mincionec,Fabrizio Brigantia, Giovanna Mincionea

aUniversità degli Studi, Laboratorio di Chimica Inorganica e Bioinorganica, Via Gino Capponi 7, I-50121, Firenze, ItalybOspedale San Giovanni di Dio, S.O. Oculistica, Via Torregalli 3, I-50123, Firenze, Italy

cUniversità degli Studi, Institute of Ophthalmology, Viale Morgagni 85, I-50134, Firenze, Italy

(Received 14 January 1999; revised 30 April 1999; accepted 6 May 1999)

Abstract – Reaction of twenty aromatic/heterocyclic sulfonamides containing a free amino, imino, hydrazino or hydroxyl group, withnicotinoyl chloride afforded a series of water-soluble (as hydrochloride or triflate salts) compounds. The new derivatives were assayed asinhibitors of three carbonic anhydrase (CA) isozymes, hCA I, hCA II (cytosolic forms) and bCA IV (membrane-bound form); h = human, b= bovine isozyme. Efficient inhibition was observed against all three isozymes, but especially against hCA II and bCA IV (in nanomolarrange), two isozymes known to play a critical role in aqueous humour secretion within the ciliary processes of the eye. Some of the bestinhibitors synthesized were applied as 2% water solutions directly into the eye of normotensive or glaucomatous albino rabbits. Very strongintraocular pressure (IOP) lowering was observed for many of them, and the active drug was detected in eye tissues and fluids. This resultprompted us to re-analyse the synthetic work done by other groups for the design of water soluble, topically effective antiglaucomasulfonamides. According to these researchers, the IOP lowering effect is due to the intrinsic nature of the specific heterocyclic sulfonamideconsidered, among which the thienothiopyran-2-sulfonamide derivatives represent the best studied case. Indeed, the first agents developed forsuch applications, such as dorzolamide, are derivatives of this ring system. In order to prove that the tail (in this case the nicotinoyl moiety)conferring water solubility to a sulfonamide CA inhibitor is critically important, similarly to the ring to which the sulfonamido group is grafted,we also prepared a dorzolamide derivative to which the nicotinoyl moiety was attached. This new compound is more water soluble thandorzolamide (as hydrochloride salt), behaves as a strong CA II inhibitor, and acts similarly to the parent derivative in lowering IOP inexperimental animals. Thus, it seems that the tail conferring water solubility is more important for topical activity as an antiglaucoma drugthan the heterocyclic/aromatic ring to which the sulfonamido moiety is grafted. © 1999 Éditions scientifiques et médicales Elsevier SAS

carbonic anhydrase / aromatic, heterocyclic sulfonamides / nicotinoyl chloride / antiglaucoma drugs / hydrochloride salts /dorzolamide

1. Introduction

The sulfonamides represent an important class ofbiologically active compounds, with at least five differentclasses of pharmacological agents that have been ob-tained from the sulfanilamide structure as lead, thederivative initially studied by Domagk [2] as the firstmodern chemotherapeutic drug. Indeed, the antibacterial

sulfonamides [3] continue to play an important role inchemotherapy, alone or in combination with otherdrugs [4], the sulfonamides that inhibit the zinc enzymecarbonic anhydrase (CA, EC 4.2.1.1) possess manyapplications as diuretic, antiglaucoma or antiepilepticdrugs among others [5–7], the hypoglycaemic sulfona-mides are extensively used in the treatment of someforms of diabetes [8], whereas the thiazides and high-ceiling diuretics might be considered as a fortunatedevelopment of the CA inhibitors [9], but these com-pounds possess a different pharmacological profile, inde-pendent of CA inhibition [10, 11]. Finally, some antithy-

# See [1].*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 799−808 799© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

roid drugs have also been developed starting from thesulfonamide structure as lead molecule [12].

The second class of the above mentioned pharmaco-logical agents, ie., the sulfonamides with CA inhibitoryaction, have been thoroughly investigated in the last 10years, mainly in the search for a topically effectiveantiglaucoma drug [13–19]. The possibility of adminis-tering a sulfonamide via the topical route directly into theeye, although investigated in the 1950s [20, 21], has beentotally unsuccessful, whereas the systemic administra-tion, quite useful in lowering intraocular pressure (IOP),was generally accompanied by undesired side effects, dueto CA inhibition in other tissues than the eye [21]. In1983, Maren’s group [13] postulated that a water-solublesulfonamide, also possessing a relatively balanced lipidsolubility, would be an effective IOP lowering drug viathe topical route, but at that moment no inhibitorspossessing such physico-chemical properties existed.They started to be developed in several laboratories soonthereafter [13–19], and in 1995 the first such pharmaco-logical agent, dorzolamide1 entered into clinical use inthe USA and Europe [22]. A second compound, brinzola-mide2, quite similar structurally to dorzolamide has alsorecently been approved for the topical treatment ofglaucoma in the USA(figure 1)[23].

Thus, in a series of interesting papers [15, 24–29], theMerck, Sharp and Dohme group has developed thesynthesis of a large series of generally bicyclic hetero-cyclic sulfonamides (derivatives of benzothiazole- [24];benzofuran- [25]; indole- [26]; benzo[b]-thiophene- [27,28]; thieno-thiopyran [15, 29], etc), which were thentested as IOP lowering agents, and which led to the abovementioned drug (dorzolamide). Still, the greatest majorityof the synthesized compounds proved to be potent aller-gens in vivo since their sulfonamido group was nucleo-philically displaced by reduced glutathione. More thanthat, the only compounds with acceptable water solubilityproved to be hydrochlorides of amino-derivatives of thethienothiopyran-sulfonamides of the dorzolamide type[15, 29]. Obviously, the approach followed by this group

was to explore as many heterocyclic rings as possible onwhich the sulfonamido moiety should be grafted, and thisapproach was extremely beneficial to the chemistry ofheterocyclic sulfonamides. Still, this approach seemed tous not a very fortunate one for the design of topicallyactive IOP lowering agents, and we decided to explorethe opposite one, i.e., to graft moieties that would ensurewater solubility (as salts of a strong acid) on the classicalring systems of the aromatic/heterocyclic sulfonamidespossessing CA inhibitory properties.

In this paper we report the reaction of twenty aromatic/heterocyclic sulfonamides containing a free amino,imino, hydrazino or hydroxyl group, with nicotinoylchloride, which afforded a series of water-soluble (ashydrochloride or triflate salts) sulfonamides with strongCA inhibitory properties. Moreover, dorzolamide hasbeen derivatized similarly, at its secondary amino group,and the obtained compound also possessed a good watersolubility as the hydrochloride salt. The new compoundsreported here were tested for the inhibition of three CAisozymes, hCA I, hCA II and bCA IV (h = human, b =bovine isozyme). Affinities in the nanomolar range weredetected for some compounds for isozymes II and IV. Themost active derivatives were assayed in vivo in normo-tensive and glaucomatous rabbits for their IOP loweringproperties. Very strong intraocular pressure (IOP) lower-ing was observed for many of them, and the active drugwas detected in eye tissues and fluids. Our conclusion isthat the water-solubilizing tail seems to be more impor-tant than the ring on which the sulfonamido moiety isgrafted, and that topically active antiglaucoma drugsmight be obtained from many other classes of sulfona-mides than the thienothiopyran-sulfonamides and theirderivatives.

2. Experimental protocols

Melting points were determined with a heating platemicroscope and are not corrected; IR spectra were ob-tained in KBr pellets with a Perkin-Elmer 16PC FTIR

Figure 1. Structures of dorzolamide1 and brinzolamide2.

800

spectrometer, whereas1H-NMR spectra were obtainedwith a Varian 300CXP apparatus in solvents specified ineach case. Chemical shifts are expressed asδ valuesrelative to Me4Si as standard. Elemental analyses weredone by combustion for C, H, N with an automated CarloErba analyser, and were± 0.4% of the theoretical values.

Sulfonamides3–22used in synthesis were either com-mercially available compounds (from Sigma, Acros orAldrich) or were prepared as described previously:4-hydrazino-benzenesulfonamide6 by diazotization ofsulfanilamide followed by reduction of the diazonium saltwith tin(II) chloride [30]; halogenosulfanilamides9–12by halogenation of sulfanilamide as reported in theliterature [31]; compound17 from 5-amino-1,3,4-thia-diazole-2-sulfonamide (obtained from acetazola-mide) [32] by acylation with the phthalimido-derivativeof â-alanine, followed by hydrazinolysis [33], whereasimine 16 by deprotection of methazolamide with concen-trated hydrochloric acid [34]. The benzothiazole-2-sulfonamide derivatives18–20 were prepared as de-scribed in ref. [35], whereas the alcohols21 and22 fromthe corresponding amines by diazotization followed byhydrolysis of the diazonium salts [31]. Dorzolamide1was prepared as described in the literature [36]. Nicotin-oyl chloride hydrochloride, triflic acid and triethylaminewere from Acros. Acetonitrile, acetone (Merck) or othersolvents used in the synthesis were doubly distilled andkept on molecular sieves in order to maintain them inanhydrous conditions.

2.1. Chemistry

2.1.1. General procedure for the preparation of nico-tinoyl derivatives of the aromatic/heterocyclic sulfon-amides23–43

An amount of 10 mM sulfonamide3–22 or 1 wasdissolved/suspended in 50 mL of anhydrous acetonitrileor acetone and then treated with 0.178 g (10 mM) ofnicotinoyl chloride hydrochloride. The stoichiometricamount (200µL) of triethylamine was then added and thereaction mixture was magnetically stirred at 4 °C for4–10 h. By means of TLC, the conversion of all thesulfonamides to the corresponding nicotinoyl derivativeshas been monitored. When the reaction was completed,the solvent was evaporated until a small volume of thereaction mixture was obtained. Generally the new com-pounds crystallized spontaneously by leaving the abovemixture at 4 °C overnight. In some cases, the concen-trated liquor obtained after the evaporation of the solventwas poured into 50 mL of cold water, then the reactionproducts precipitated and filtered. The prepared com-pounds were recrystallized from ethanol or ethanol-water

(1:1, v/v). Yields were in the range of 70–90%. Hydro-chlorides of derivatives23–43 were obtained from thefree bases and a methanolic HCl solution, in methanol assolvent. The hydrochlorides precipitated by leaving theabove mixtures at 4 °C overnight. The hydrochlorideswere analysed for the presence of Cl– by potentiometrictitrations. The obtained data were± 0.5% of the theoreti-cal values calculated for the proposed formulas (data notshown). Triflate salts were similarly obtained from thefree bases23–43and the stoichiometric amount of triflicacid, in water as solvent.

2.1.1.1. 2-(Nicotinoylamido)-benzenesulfonamide23White crystals, m.p. 250–252 °C; IR (KBr), cm–1:

1 140 (SO2sym), 1 290 (amide III), 1 370 (SO2

as), 1 550(amide II), 1 680 (amide I), 3 090 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H, ArHfrom nicotinoyl); 7.15–7.66 (m, 4H, ArH, 1,2-phenylene); 7.50 (br s, 2H, SO2NH2); 8.14 (br s, 1H,CONH); Anal., found: C, 51.90; H, 4.10; N, 14.96%;C12H11N3O3S requires: C, 51.98; H, 4.00; N, 15.15%.

2.1.1.2. 3-(Nicotinoylamido)-benzenesulfonamide24White crystals, m.p. 265–266 °C (dec.); IR (KBr),

cm–1: 1 135 (SO2sym), 1 290 (amide III), 1 370 (SO2

as),1 570 (amide II), 1 690 (amide I), 3 080 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H,ArH from nicotinoyl); 7.10–7.50 (m, 4H, ArH, 1,3-phenylene); 7.56 (br s, 2H, SO2NH2); 8.11 (br s, 1H,CONH); Anal., found: C, 51.78; H, 3.85; N, 15.07%;C12H11N3O3S requires: C, 51.98; H, 4.00; N, 15.15%.

2.1.1.3. 4-(Nicotinoylamido)-benzenesulfonamide25White crystals, m.p. 280–281 °C (dec.); IR (KBr),

cm–1: 1 150 (SO2sym), 1 290 (amide III), 1 345 (SO2

as),1 560 (amide II), 1 690 (amide I), 3 060 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H,ArH from nicotinoyl);δA 7.18,δB 7.75 (AA≠BB≠system,4H, JAB = 7.9 Hz, ArH from 4-sulfamoylphenyl); 7.56 (brs, 2H, SO2NH2); 8.19 (br s, 1H, CONH); Anal., found: C,51.67; H, 4.05; N, 14.88%; C12H11N3O3S requires: C,51.98; H, 4.00; N, 15.15%.

2.1.1.4. 4-(Nicotinoylhydrazido)-benzenesulfonamide26White crystals, m.p. 265–267 °C; IR (KBr), cm–1: 980

(N–N), 1 150 (SO2sym), 1 290 (amide III), 1 365 (SO2

as),1 555 (amide II), 1 690 (amide I), 3 090 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H,ArH from nicotinoyl);δA 7.18,δB 7.71 (AA≠BB≠system,4H, JAB = 7.8 Hz, ArH from 4-sulfamoylphenyl); 7.59 (brs, 2H, SO2NH2); 8.06 (br s, 2H, CONHNH); Anal.,found: C, 49.40; H, 4.16; N, 19.03%; C12H12N4O3Srequires: C, 49.31; H, 4.14; N, 19.17%.

801

2.1.1.5. 4-(Nicotinoylamidomethyl)-benzenesulfonamide27White crystals, m.p. 271–273 °C (dec.); IR (KBr),

cm–1: 1 170 (SO2sym), 1 290 (amide III), 1 372 (SO2

as),1 545 (amide II), 1 690 (amide I), 3 090 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 4.90 (s, 2H, CH2);7.05–7.98 (m, 4H, ArH from nicotinoyl);δA 7.22, δB

7.79 (AA≠BB≠system, 4H,JAB = 7.9 Hz, ArH from4-sulfamoylphenyl); 7.67 (br s, 2H, SO2NH2); 8.16 (br s,1H, CONH); Anal., found: C, 53.81; H, 4.78; N, 14.21%;C13H13N3O3S requires: C, 53.60; H, 4.50; N, 14.42%.

2.1.1.6. 4-(Nicotinoylamidoethyl)-benzenesulfonamide28White crystals, m.p. 278–280 °C (dec.); IR (KBr),

cm–1: 1 150 (SO2sym), 1 290 (amide III), 1 359 (SO2

as),1 540 (amide II), 1 690 (amide I), 3 080 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 3.10 (t, 2H,αCH2

from the CH2CH2 bridge); 3.70 (t, 2H,âCH2 from theCH2CH2 bridge); 7.05–7.98 (m, 4H, ArH from nicoti-noyl); δA 7.15,δB 7.62 (AA≠BB≠system, 4H,JAB = 7.9Hz, ArH from 4-sulfamoylphenyl); 7.67 (br s, 2H,SO2NH2); 8.17 (br s, 1H, CONH); Anal., found: C, 55.40;H, 5.03; N, 13.45%; C14H15N3O3S requires: C, 55.07; H,4.95; N, 13.76%.

2.1.1.7. 3-Fluoro-4-(nicotinoylamido)-benzenesulfon-amide29

White crystals, m.p. 234–235 °C. IR (KBr), cm–1:1 150 (SO2

sym), 1 290 (amide III), 1 348 (SO2as), 1 550

(amide II), 1 680 (amide I), 3 060 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 6.60 (br s, 2H, SO2NH2);7.05–7.98 (m, 4H, ArH from nicotinoyl); 7.05–7.89 (m,3H, Ar H from the F-substituted ring); 8.15 (br s, 1H,CONH); Analysis, found: C, 48.54; H, 3.61; N, 14.07%;C12H10FN3O3S requires: C, 48.81; H, 3.41; N, 14.23%.

2.1.1.8. 3-Chloro-4-(nicotinoylamido)-benzenesulfon-amide30

White crystals, m.p. 238–239 °C. IR (KBr), cm–1:1 155 (SO2

sym), 1 290 (amide III), 1 339 (SO2as), 1 550

(amide II), 1 690 (amide I), 3 090 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 6.70 (br s, 2H, SO2NH2);7.05–7.98 (m, 4H, ArH from nicotinoyl); 7.05–7.76 (m,3H, Ar H the 2-Cl-substituted ring); 8.15 (br s, 1H,CONH); Analysis, found: C, 46.27; H, 3.37; N, 13.39%;C12H10ClN3O3S requires: C, 46.23; H, 3.23; N, 13.48%.

2.1.1.9. 3-Bromo-4-(nicotinoylamido)-benzenesulfon-amide31

White crystals, m.p. 230–232 °C. IR (KBr), cm–1:1 160 (SO2

sym), 1 290 (amide III), 1 356 (SO2as), 1 540

(amide II), 1 690 (amide I), 3 060 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 6.65 (br s, 2H, SO2NH2);7.05–7.98 (m, 4H, ArH from nicotinoyl); 7.05–7.86 (m,

3H, Ar H the 2-Br-substituted ring); 8.14 (br s, 1H,CONH); Analysis, found: C, 40.55; H, 3.00; N, 11.50%;C12H10BrN3O3S requires: C, 40.46; H, 2.83; N, 11.80%.

2.1.1.10. 3-Iodo-4-(nicotinoylamido)-benzenesulfon-amide32

White crystals, m.p. 210–212 °C. IR (KBr), cm–1:1 145 (SO2

sym), 1 290 (amide III), 1 360 (SO2as), 1 545

(amide II), 1 690 (amide I), 3 070 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 6.60 (br s, 2H, SO2NH2);7.05–7.98 (m, 4H, ArH from nicotinoyl); 7.08–7.79 (m,3H, Ar H the 2-I-substituted ring); 8.14 (br s, 1H,CONH); Analysis, found: C, 35.87; H, 2.43; N, 10.36%;C12H10IN3O3S requires: C, 35.75; H, 2.50; N, 10.42%.

2.1.1.11. 4,5-Dichloro-6-nicotinoylamido-benzene-1,3-disulfonamide33

White crystals, m.p. 267–268 °C. IR (KBr), cm–1:1 140 (SO2

sym), 1 290 (amide III), 1 370 (SO2as), 1 550

(amide II), 1 690 (amide I), 3 080 (NH); 3 360 (NH2);1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H, ArHfrom nicotinoyl); 7.54 (s, 1H, ArH from the pentasubsti-tuted benzene ring); 7.68 (br s, 4H, 2 SO2NH2); 8.10 (brs, 1H, CONH); Analysis, found: C, 33.69; H, 2.40; N,13.08%; C12H10Cl2N4O5S2 requires: C, 33.89; H, 2.37;N, 13.17%.

2.1.1.12. 6-Chloro-4-nicotinoylamido-benzene-1,3-disul-fonamide34

White crystals, m.p. 290–294 °C (dec.). IR (KBr),cm–1: 1 150 (SO2

sym), 1 290 (amide III), 1 330 (SO2as),

1 540 (amide II), 1 680 (amide I), 3 060 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm:δA 7.05–7.98 (m,4H, ArH from nicotinoyl); 7.35 (s, 1H, ArH from disul-famoylphenyl); 7.59 (s, 1H, ArH from disulfamoylphe-nyl); 7.75 (br s, 4H, 2 SO2NH2); 8.14 (br s, 1H, CONH);Analysis, found: C, 36.59; H, 2.90; N, 14.21%;C12H11ClN4O5S2 requires: C, 36.88; H, 2.84; N, 14.34%.

2.1.1.13. 5-Nicotinoylamido-1,3,4-thiadiazol-2-sulfon-amide35

White crystals, m.p.> 310 °C; IR (KBr), cm–1: 1 180(SO2

sym), 1 295 (amide III), 1 340 (SO2as), 1 545 (amide

II), 1 690 (amide I), 3 060 (NH), 3 375;1H-NMR(DMSO-d6), δ, ppm: 6.94 (br s, 2H, SO2NH2); 7.05–7.98(m, 4H, ArH from nicotinoyl); 8.26 (br s, 1H, CONH);Anal., found, C, 33.60; H, 2.60; N, 24.45%; C8H7N5O3S2

requires: C, 33.68; H, 2.47; N, 24.55%.

2.1.1.14. 5-Nicotinoylimido-4-methyl-2-sulfonamido-δ2-1,3,4-thiadiazoline36

White crystals, m.p.> 310 °C; IR (KBr), cm–1: 1 182(SO2

sym), 1 290 (amide III), 1 366 (SO2as), 1 540 (amide

802

II), 1 690 (amide I), 3 080 (NH), 3 380 (NH2);1H-NMR

(DMSO-d6), δ, ppm: 3.90 (s, 3H, Me); 6.96 (br s, 2H,SO2NH2); 7.05–7.98 (m, 4H, ArH from nicotinoyl);Anal., found, C, 35.99; H, 3.12; N, 23.25%; C9H9N5O3S2

requires: C, 36.11; H, 3.03; N, 23.40%.

2.1.1.15. 5-(Nicotinoylamidoethylcarboxamido)-1,3,4-thiadiazol-2-sulfonamide37

White crystals, m.p. 287–289 °C (dec.), IR (KBr),cm–1: 1 150 (SO2

sym), 1 270 and 1 290 (amide III), 1 330(SO2

as), 1 450, 1 570 (amide II), 1 690 and 1 710 (amideI), 3 090 (NH); 3 360 (NH2);

1H-NMR (DMSO-d6), δ,ppm: 2.25–2.60 (m, 4H, CH2CH2); 6.88 (br s, 3H, CONH+ SO2NH2); 7.05–7.98 (m, 4H, ArH from nicotinoyl);8.24 (br s, 1H, CONH from nicotinoylamido moiety);Analysis, found: C, 37.15; H, 3.19; N, 23.46%;C11H12N6O4S2 requires: C, 37.07; H, 3.39; N, 23.58%.

2.1.1.16. 6-Nicotinoylamido-benzothiazol-2-sulfonamide38White crystals, m.p. 290–294 °C (dec.), IR (KBr),

cm–1: 1 165 (SO2sym), 1 290 (amide III), 1 344 (SO2

as),1 540 (amide II), 1 680 (amide I), 3 060 (NH); 3 360(NH2);

1H-NMR (DMSO-d6), δ, ppm: 7.05–7.98 (m, 4H,ArH from nicotinoyl); 6.94 (dd, 1H,J = 9 Hz; J = 3 Hz,H-5 of benzothiazole); 7.10 (d, 1H,J = 3 Hz, H-7 ofbenzothiazole); 7.78 (d, 1H,J = 9 Hz, H-4 of benzothia-zole); 8.10 (br s, 2H, SO2NH2); 8.18 (br s, 1H, CONH);Analysis, found: C, 46.85; H, 2.94; N, 16.47%;C13H10N4O3S2 requires: C, 46.70; H, 3.01; N, 16.76%.

2.1.1.17. 6-Nicotinoyloxy-benzothiazol-2-sulfonamide39White crystals, m.p. 281–283 °C (dec.), IR (KBr),

cm–1: 1 030 (CO–O), 1 160 (SO2sym), 1 350 (SO2

as),1 450, 1 775 (COO), 3 360 (NH2);

1H-NMR (DMSO-d6),δ, ppm: 7.05–7.98 (m, 4H, ArH from nicotinoyl); 6.90(dd, 1H,J = 9 Hz; J = 3 Hz, H-5 of benzothiazole); 7.11(d, 1H,J = 3 Hz, H-7 of benzothiazole); 7.79 (d, 1H,J =9 Hz, H-4 of benzothiazole); 8.10 (br s, 2H, SO2NH2);Analysis, found: C, 46.40; H, 2.90; N, 12.38%;C13H9N3O4S2 requires: C, 46.56; H, 2.71; N, 12.53%.

2.1.1.18. 6-Nicotinoyloxyethyloxy-benzothiazol-2-sulfon-amide40

White crystals, m.p. 245–246 °C, IR (KBr), cm–1:1 030 (CO–O), 1 175 (SO2

sym), 1 341 (SO2as), 1 450,

1 770 (COO), 3 360 (NH2);1H-NMR (DMSO-d6), δ,

ppm: 2.89 (t, 3H, CH2); 3.14 (t, 3H, CH2); 6.95 (dd, 1H,J = 9 Hz;J = 3 Hz, H-5 of benzothiazole); 7.05–7.98 (m,4H, ArH from nicotinoyl); 7.10 (d, 1H,J = 3 Hz, H-7 ofbenzothiazole); 7.79 (d, 1H,J = 9 Hz, H-4 of benzothia-zole); 8.15 (br s, 2H, SO2NH2); Analysis, found: C,47.58; H, 3.66; N, 10.89%; C15H13N3O5S2 requires: C,47.49; H, 3.45; N, 11.07%.

2.1.1.19. 4-(Nicotinoyloxymethyl)-benzenesulfonamide41White crystals, m.p. 244–246 °C; IR (KBr), cm–1:

1 040 (CO–O), 1 155 (SO2sym), 1 325 (SO2

as), 1 780(COO), 3 310 (NH2);

1H-NMR (DMSO-d6), δ, ppm: 4.90(s, 2H, CONHCH2); 7.05–7.98 (m, 4H, ArH from nico-tinoyl); 7.08–7.41 (m, AA≠BB≠, J = 7.2 Hz; 4H, ArH,phenylene); 7.49 (s, 2H, SO2NH2); Anal., found, C,53.19; H, 4.21; N, 9.37%; C13H12N2O4S requires: C,53.42; H, 4.14; N, 9.58%.

2.1.1.20. 4-(Nicotinoyloxyethyl)-benzenesulfonamide42White crystals, m.p. 240–243 °C. IR (KBr), cm–1:

1 040 (CO–O), 1 157 (SO2sym), 1 332 (SO2

as), 1 760(COO), 3 300 (NH2);

1H-NMR (DMSO-d6), δ, ppm: 3.10(t, 2H, αCH2 from the CH2CH2 bridge); 3.70 (t, 2H,âCH2 from the CH2CH2 bridge); 6.95 (br s, 2H,SO2NH2); 7.05–7.98 (m, 4H, ArH from nicotinoyl);7.05–7.52 (m, AA≠BB≠, J = 7.3 Hz, 4H, ArH, phe-nylene); Anal., found, C, 54.95; H, 4.67; N, 8.97%;C14H14N2O4S requires: C, 54.89; H, 4.61; N, 9.14%.

2.1.1.21. 5,6-Dihydro-4-[N-nicotinoyl-(ethylamido)]-6-methyl-4H-thieno-[2,3-b]thiopyran-2-sulfonamide 7,7-dioxide43

White crystals, m.p. 270–272 °C; IR (KBr), cm–1:1 135 (SO2

sym), 1 290 (amide III), 1 345 (SO2as), 1 545

(amide II), 1 680 (amide I), 3 366 (NH2);1H-NMR

(DMSO-d6), δ, ppm: 1.29 (d, 3H, Me); 1.39 (t, 3H, Mefrom ethyl); 2.55 (m, 1H, CH); 2.80 (m, 1H, CH);3.05–3.20 (m, 2H, CH2 from ethyl); 4.37 (m, 2H, CH2);7.05–7.98 (m, 4H, ArH from nicotinoyl); 8.03 (s, 1H, CH,ArH from thienyl); 8.25 (br s, 2H, SO2NH2); Anal.,found, C, 44.57; H, 4.39; N, 9.66%; C16H19N3O5S3

requires: C, 44.74; H, 4.46; N, 9.78%.

2.2. Pharmacology

2.2.1. Enzyme assayHuman CA I and CA II cDNAs were expressed in

Escherichia colistrain BL21 (DE3) from the plasmidspACA/hCA I and pACA/hCA II described by Forsman etal. [37] (the two plasmids were a gift from Prof. SvenLindskog, Umea University, Sweden). Cell growth con-ditions were those described by Lindskog’s group [38]and enzymes were purified by affinity chromatographyaccording to the method of Khalifah et al. [39]. Enzymeconcentrations were determined spectrophotometricallyat 280 nm, utilizing a molar absorptivity of49 mM–1.cm–1 for CA I and 54 mM–1.cm–1 for CA II,respectively, based on Mr = 28.85 kDa for CA I, and29.30 kDa for CA II, respectively [40, 41]. CA IV wasisolated from bovine lung microsomes as described by

803

Maren et al., and its concentration has been determinedby titration with ethoxzolamide [42].

Initial rates of 4-nitrophenyl acetate hydrolysis cataly-sed by different CA isozymes were monitored spectro-photometrically, at 400 nm, with a Cary 3 instrumentinterfaced with an IBM compatible PC [43]. Solutions ofsubstrate were prepared in anhydrous acetonitrile; thesubstrate concentrations varied between 2× 10–2 and1 × 10–6 M, working at 25 °C. A molar absorption coef-ficient e of 18 400 M–1.cm–1 was used for the4-nitrophenolate formed by hydrolysis, in the conditionsof the experiments (pH 7.40), as reported in the litera-ture [43]. Non-enzymatic hydrolysis rates were alwayssubtracted from the observed rates. Duplicate experi-ments were done for each inhibitor concentration, and thevalues reported throughout the paper are the mean of suchresults. Stock solutions of inhibitor (1 mM) were pre-pared in distilled-deionized water with 10–20% (v/v)DMSO (which is not inhibitory at these concentrations)and dilutions up to 0.01 nM were done thereafter withdistilled-deionized water. Inhibitor and enzyme solutionswere preincubated together for 10 min at room tempera-ture prior to assay, in order to allow for the formation ofthe E-I complex. The inhibition constant KI was deter-mined as described by Pocker and Stone [43]. Enzymeconcentrations were 3.5 nM for hCA II, 12 nM for hCAI and 36 nM for bCA IV (this isozyme has a decreasedesterase activity [44] and higher concentrations had to beused for the measurements).

2.2.2. Measurement of tonometric IOPAdult male New Zealand albino rabbits weighing

3–3.5 kg were used in the experiments (three animalswere used for each inhibitor studied). The experimentalprocedures conform to the Association for Research inVision and Ophthalmology Resolution on the use ofanimals. The rabbits were kept in individual cages withfood and water provided ad libitum. The animals weremaintained on a 12 h:12 h light/dark cycle in a tempera-ture controlled room, at 22–26 °C. Solutions of inhibitors(2%, as hydrochlorides, by weight) were obtained indistilled deionized water. The pH of these solutions wasaround 5.50–6.40.

IOP was measured using a Digilab 30R pneumatonom-eter (BioRad, Cambridge, MA, USA) as described byMaren’s group [45, 46]. The pressure readings werematched with two-point standard pressure measurementsat least twice each day using a Digilab calibration verifier.All IOP measurements were done by the same investiga-tor with the same tonometer. One drop of 0.2% oxybup-rocaine hydrochloride (novesine, Sandoz) diluted 1:1with saline was instilled in each eye immediately before

each set of pressure measurements. IOP was measuredthree times at each time interval, and the means reported.IOP was measured first, immediately before drug admin-istration, then at 30 min after the instillation of thepharmacological agent, and then every 30 min for aperiod of several hours. For all IOP experiments, drugwas administered to only one eye, leaving the contralat-eral eye as an untreated control. The ocular hypotensiveactivity is expressed as the average difference in IOPbetween the treated and control eye, in this way mini-mizing the diurnal, seasonal and interindividual varia-tions commonly observed in the rabbit [45, 46]. All dataare expressed as mean± SE, using a one-tailedt test.Ocular hypertension was elicited in the right eye of albinorabbits by the injection ofα-chymotrypsin (from Sigma)as described by Melena et al. [48]. The IOP of operatedanimals was checked after approximately four weeks, andanimals with an elevated pressure of 30–36 mm Hg wereused at least one month after the injection ofα-chymotrypsin.

2.2.3. Drug distribution in ocular fluids and tissuesThe general procedure of Maren’s group has been

followed [45, 46]. The animals were killed with anintracardiac air injection. Aqueous humour (both poste-rior and anterior chamber fluids) were withdrawn. Then,the cornea and anterior uvea (iris plus attached ciliarybody) were dissected, rinsed well with water, blotted,weighed and put into 1–2 mL of water. For isolation ofthe ciliary processes, intact anterior uvea rings wereplaced on a parafilm covered piece of polystyrene foam ina Petri dish. The tissue was wetted with normal saline anddissected under a microscope, then ciliary processes wereliberated from their attachment to the iris, cut, weighedand put in 0.5 mL of distilled water. The tissue from 4eyes (average weight of 8 mg/eye) was pooled for druganalysis. Samples were boiled for 5 min (in order todenature CA, and free drug from the E-I complex),diluted and then incubated with a known amount ofenzyme. The activity of the free enzyme and in thepresence of the inhibitor were determined as describedabove. A calibration curve has been used in order todetermine the fractional inhibition in the different tissues,as described in [45, 46].

3. Results

Compounds prepared by reaction of nicotinoyl chlo-ride with aromatic/heterocyclic sulfonamides, of type23–43, are shown below(figure 2). Inhibition data againstthree CA isozymes, hCA I, hCA II and bCA IV withcompounds1–43 are presented intable I. In vivo IOP

804

lowering data with some of the most active CA inhibitorsreported here, in normotensive and glaucomatous rabbits,after topical administration of the drug, are shown intables IIandIII , respectively. Ex vivo distribution data ofcompound35 in ocular tissues and fluids after the topicaladministration in normotensive rabbits, are shown intable IV.

4. Discussion

Reaction of sulfonamides3–22 or 1 with nicotinoylchloride afforded a series of new compounds of type

Figure 2. Structure of derivatives3–43.

Table I. CA inhibition data with standard inhibitors1–2, theparent sulfonamides3–22and the new derivatives23–43reportedin the present study, against isozymes I, II and IV.

Inhibitor KI* (nM)

hCA Ia hCA IIa bCA IVb

Dorzolamide1 50 000 9 45Brinzolamidec 2 – 3.2 45.33 45 400 295 1 3104 25 000 240 2 2005 28 000 300 3 0006 78 500 320 3 2157 25 000 170 2 8008 21 000 160 2 4509 8 300 60 18010 9 800 110 32011 6 500 40 6612 6 000 70 12513 6 100 28 17514 8 400 75 16015 8 600 60 54016 9 300 19 35517 455 3 12518 70 9 1919 55 8 1720 50 7 1521 24 000 125 56022 18 000 110 45023 21 800 310 57024 20 500 285 31025 16 000 142 16526 23 200 324 40027 1 200 79 11528 1 100 66 10229 547 38 7030 630 49 7831 650 45 8032 650 42 6933 540 36 7734 628 39 6535 36 2 936 28 3 1137 21 5 1238 13 5 1939 9 3 840 9 2 741 2 160 79 14042 2 100 64 11643 2 000 5 11

*Standard error for the determination of KI-s was of 5–10% (fromtwo different assays).aHuman (cloned) isozyme;bIsolated frombovine lung microsomes;cFrom [47].

805

23–43. The reaction was generally performed in acetoneor acetonitrile as solvent, in the presence of triethylamineas base. In the case of compounds15 and16, the aboveprocedure led to very low yields of nicotinoylamidoderivatives, and Schotten-Baumann conditions were ap-plied for obtaining35 and36 in good yields. Hydrochlo-rides of the new derivatives were then prepared byreacting the free bases23–43 with a methanolic HClsolution. Similarly were obtained the triflate salts, byreaction of bases23–43 with triflic acid in water assolvent. These salts possess a very good water solubility,generally in the range of 3–5% (data not shown). The pHof such solutions were generally around 5.5–6.0, makingthem appropriate for topical application directly into theeye.

Compounds 3–43 were characterized by standardchemical and physical methods that confirmed theirstructure (see Experimental protocols for details) andwere assayed for the inhibition of isozymes hCA I, hCAII and bCA IV (table I).

Inhibition data against the three CA isozymes, hCA I,hCA II and bCA IV with the new derivatives(table I)prove that the nicotinoylamido-sulfonamides23–43 re-ported here generally behave as strong inhibitors, withgreatly increased efficiencies as compared to the parentcompounds from which they were prepared (the sulfon-

amides3–22). The efficiency of the obtained inhibitorgenerally varied in the following way, based on the parentsulfonamide from which it was prepared: the derivativeof p-hydrazino-benzenesulfonamide26 < the orthanil-amide23 ≅ the metanilamide24 < the sulfanilamide25 <the homosulfanilamides27 < the p-aminoethyl-benzene-sulfonamides28 < the 1,3-benzene-disulfonamides33and34≅ the halogeno-substituted sulfanilamides29–32<the 1,3,4-thiadiazole-2-sulfonamides35 and 37 ≅4-methyl-δ2-1,3,4-thiadiazoline-2-sulfonamide36 ≅ thedorzolamide derivative 43 < the benzothiazole-2-sulfonamides38–40. All three CA isozymes investigatedhere were susceptible to inhibition with this type ofsulfonamide, with hCA II and bCA IV the most inhib-itable, followed by hCA I, generally less susceptible toinhibition as compared to the first two isozymes.

The promising in vitro CA inhibitory activity of someof the newly prepared compounds prompted us to inves-tigate their effect in vivo, on the intraocular pressure(IOP), after topical application directly into the eye, innormotensive and glaucomatous rabbits, frequently usedas an animal model of glaucoma [13–15, 22, 23, 45].Some of these results are shown intables II and III .

The inhibitors selected for in vivo studies were amongthe most active against hCA II and IV, in the preparedseries, such as compounds35–40, and43. The following

Table II. Fall of IOP of normotensive rabbits (20.1± 2.0 mm Hg), after treatment with one drop (50µL) of a solution 2 % of CA inhibitor(as hydrochloride salt, with the pH value shown below) directly into the eye, at 30, 60 and 90 min after administration.

Inhibitor pH ∆IOP (mm Hg)*

t = 0 t = 30 min t = 60 min t = 90 min1 (dorzolamide) 5.5 0 2.2± 0.10 4.1± 0.15 2.7± 0.0835 5.5 0 5.4± 0.12 10.9± 0.11 12.5± 0.1736 5.8 0 6.2± 0.10 12.4± 0.14 14.1± 0.1237 5.5 0 5.1± 0.12 8.1± 0.11 8.6± 0.1239 5.7 0 2.1± 0.05 4.0± 0.10 4.5± 0.1040 5.5 0 2.4± 0.05 4.3± 0.11 3.5± 0.0843 5.5 0 2.5± 0.06 4.5± 019 7.0± 0.14

* ∆IOP = IOPcontrol eye– IOPtreated eye; Mean± average spread (n = 3).

Table III. Fall of IOP of glaucomatous rabbits (34.1± 2.0 mm Hg), after treatment with one drop (50µL) of a solution 2 % of CA inhibitor(as hydrochloride salt, with the pH value shown below) directly into the eye, at 30, 60 and 90 min after administration.

Inhibitor pH ∆IOP (mm Hg)*

t = 0 t = 30 min t = 60 min t = 90 min1 5.5 0 4.3± 0.25 7.1± 0.30 5.0± 0.2535 5.5 0 10.3± 0.20 15.2± 0.20 19.1± 0.1536 5.8 0 8.5± 0.10 13.2± 0.20 20.4± 0.2043 5.5 0 4.8± 0.10 6.6± 0.10 11.6± 0.15

* ∆IOP = IOPcontrol eye– IOPtreated eye; Mean± average spread (n = 3).

806

facts should be noted regarding the data oftables II andIII . Some of the new compounds assayed in vivo, such as35, 36, 37 and 43, showed much more effective IOPlowering effects as compared to dorzolamide1, both after30 min from the administration of the inhibitor within therabbit eye, as well as at other times (1, 1.5 and 2 h,respectively), in normotensive as well as glaucomatousanimals. A second group of inhibitors, such as39 and40,showed IOP lowering effects of the same order ofmagnitude as those of dorzolamide, both after half anhour or longer periods after the administration. Mentionshould be made that the pH of the solutions administeredin these experiments was in the range of 5.0–5.9 for allinhibitors used.

In table IV, the drug distribution in ocular fluids andtissues of normotensive rabbits is shown, after the topicaladministration of one of the most active topical inhibitorsin the prepared series, i.e., compound35.

It is seen from the above data that at 1 and 2 h aftertopical administration of the drug, high levels of35 werefound in the cornea, aqueous humour and ciliary pro-cesses. Based on the inhibition constant of this compound(2 nM for CA II, and 9 nM for CA IV, respectively), thefractional inhibition estimated in these tissues/fluids is of99.5–99.9%, proving the fact that the IOP decrease isindeed due to CA inhibition [45, 46].

In conclusion, we report here a general approach forthe preparation of water-soluble, topically effective anti-glaucoma sulfonamides, by attaching water-solubilizingmoieties (such as isonicotinoyl) to well-known aromatic/heterocyclic sulfonamides. The new compounds reportedhere might lead to the development of more efficientantiglaucoma drugs.

Acknowledgements

This research was financed by the EU grant ERBCIPDCT 940051. Thanks are addressed to Drs M.A. Iliesand M. Barboiu for expert technical assistance with thepreparation of several sulfonamide intermediates used inthis work.

References

[1] Menabuoni L., Scozzafava A., Mincione F., Briganti F., MincioneG., Supuran C.T., J. Enzyme Inhib. 15 in press. (1999)

[2] Domagk G., Dt. Med. Wocheschr. 61 (1935) 250–254.

[3] Northey E.H., in: The Sulfonamides and Allied Compounds, Rein-hold, New York, 1948, pp. 1–267.

[4] Mandell G.L., Sande M.A., in: Gilman A.G., Rall T.W., Nies A.S.,Taylor P. (Eds.), The Pharmacological Basis of Therapeutics, 8thEdition, Pergamon Press, New York, 1990, pp. 1047–1064.

[5] Supuran C.T., in: Puscas I. (Ed.), Carbonic Anhydrase and Modu-lation of Physiologic and Pathologic Processes in the Organism,Helicon Timisoara, 1994, pp. 29–111.

[6] Weiner I.M., in: Gilman A.G., Rall T.W., Nies A.S., Taylor P. (Eds.),The Pharmacological Basis of Therapeutics, 8th Edition, PergamonPress, New York, 1990, pp. 713–732.

[7] Supuran C.T., Scozzafava A., Ilies M.A., Iorga B., Cristea T.,Briganti F., Chiraleu F., Banciu M.D., Eur. J. Med. Chem. 33 (1998)577–595.

[8] Boyd A.E., Diabetes 37 (1988) 847–850.

[9] Beyer K.H., Baer J.E., Pharmacol. Rev. 13 (1961) 517–562.

[10] Maren T.H., Drug Dev. Res. 10 (1987) 255–276.

[11] Supuran C.T., Conroy C.W., Maren T.H., Eur. J. Med. Chem. 31(1996) 843–846.

[12] Maren T.H., Annu. Rev. Pharmacol. Toxicol. 16 (1976) 309–327.

[13] Maren T.H., Jankowska L., Edelhauser G.F., Sanyal G., Exp. EyeRes. 36 (1983) 457–480.

[14] Katritzky A.R., Caster K.C., Maren T.H., Conroy C.W., Bar-Ilan A.,J. Med. Chem. 30 (1987) 2058–2062.

[15] Ponticello G.S., Freedman M.B., Habecker C.N., Lyle P.A., SchwamH., Varga S.L. et al., J. Med. Chem. 30 (1987) 591–597.

[16] Boriack P.A., Christianson D.W., Kingery-Wood J., WhitesidesG.M., J. Med. Chem. 38 (1995) 2286–2291.

[17] Supuran C.T., Popescu A., Ilisiu M., Costandache A., Banciu M.D.,Eur. J. Med. Chem. 31 (1996) 439–448.

[18] Supuran C.T., Briganti F., Scozzafava A., J. Enzyme Inhib. 12(1997) 175–190.

[19] Supuran C.T., Scozzafava A., Popescu A., Bobes-Tureac R., BanciuA., Creanga A., Bobes-Tureac G., Banciu M.D., Eur. J. Med. Chem.32 (1997) 445–452.

[20] Becker B., Am. J. Ophthalmol. 39 (1955) 177–183.

[21] Maren T.H., Physiol. Rev. 47 (1967) 595–782.

[22] Ponticello G.S., Sugrue M.F., Plazonnet B., Durand-Cavagna G.,Pharm. Biotechnol. 11 (1998) 555–574.

[23] Silver L.H., Am. J. Ophthalmol. 126 (1998) 400–408.

[24] Graham S.L., Shepard K.L., Anderson P.S., Baldwin J.J., Best D.B.,Christy M.E. et al., J. Med. Chem. 32 (1989) 2548–2554.

[25] Graham S.L., Hoffman J.M., Gautheron P., Michelson S.R., ScholzT.H., Schwam H. et al., J. Med. Chem. 33 (1990) 749–754.

[26] Graham S.L., Scholz T.H., Synthesis (1986) 1031–1033.

[27] Prugh J.D., Hartmann G.D., Mallorga P.J., McKeever B.M., Mich-elson S.R., Murcko M.A. et al., J. Med. Chem. 34 (1991)1805–1818.

[28] Hartmann G.D., Halczenko W., Prugh J.D., Smith R.L., SugrueM.F., Mallorga P.J. et al., J. Med. Chem. 35 (1992) 3027–3033.

[29] Baldwin J.J., Ponticello G.S., Anderson G.S., Christy M.E., MurckoM.A., Randall W.C. et al., J. Med. Chem. 32 (1989) 2510–2513.

[30] Crippa G.B., Maffei S., Gazz. Chim. Ital. 71 (1941) 97–99.

[31] Cingolani E., Gazz. Chim. Ital. 78 (1948) 275–282.

Table IV. Ocular tissue concentrations (µM) after 1 and 2 h,following corneal application of one drop (50µL) of a 2 % solutionof the compound35 in normotensive albino rabbits.

Time (h) Drug concentration (µM)*

Cornea Aqueous humour Ciliary process1 h 150± 5 283± 10 51± 32 h 47± 4 39 ± 3 10 ± 1

* Mean ± standard deviation (n = 3).

807

[32] Jitianu A., Ilies M.A., Scozzafava A., Supuran C.T., Main GroupMet. Chem. 20 (1997) 147–153.

[33] Barboiu M., Scozzafava A., Supuran C.T., J. Enzyme Inhib. 15 inpress. (1999)

[34] Scozzafava A., Supuran C.T., J. Enzyme Inhib. 13 (1998) 419–442.

[35] Schoenwald R.D., Eller M.G., Dixson J.A., Barfknecht C.F., J. Med.Chem. 27 (1984) 810–812.

[36] Blacklock T.J., Sohar P., Butcher J.W., Lamanec T., GrabowskiE.J.J., J. Org. Chem. 58 (1993) 1672–1679.

[37] Forsman C., Behravan G., Osterman A., Jonsson B.H., Acta Chem.Scand. B42 (1988) 314–318.

[38] Behravan G., Jonasson P., Jonsson B.H., Lindskog S., Eur. J.Biochem. 198 (1991) 589–592.

[39] Khalifah R.G., Strader D.J., Bryant S.H., Gibson S.M.,Biochemistry 16 (1977) 2241–2247.

[40] Nyman P.O., Lindskog S., Biochim. Biophys. Acta 85 (1964)141–151.

[41] Henderson L.E., Henriksson D., Nyman P.O., J. Biol. Chem. 251(1976) 5457–5463.

[42] Maren T.H., Wynns G.C., Wistrand P.J., Mol. Pharmacol. 44 (1993)901–906.

[43] Pocker Y., Stone J.T., Biochemistry 6 (1967) 668–678.

[44] Baird T.T., Waheed A., Okuyama T., Sly W.S., Fierke C.A.,Biochemistry 36 (1997) 2669–2678.

[45] Maren T.H., Brechue W.F., Bar-Ilan A., Exp. Eye Res. 55 (1992)73–79.

[46] Brechue W.F., Maren T.H., Invest. Ophthalmol. Vis. Sci. 34 (1993)2581–2587.

[47] Stams T., Chen Y., Boriack-Sjodin P.A., Hurt J.D., Liao J., May J.A.et al., Protein Sci. 7 (1998) 556–563.

[48] Melena J., Santafe J., Segarra-Domenech J., Purras G., J. Ocul.Pharmacol. Ther. 15 (1999) 19–27.

808

Original article

5-Substituted UTP derivatives as P2Y2 receptor agonists#

Bernd H.A. Knoblaucha,d, Christa E. Müllera,c*, Leif Järlebarkd, Grace Lawokod, Thomas Kottkeb,Martin A. Wikströme, Edith Heilbronnd

aInstitut für Pharmazie und Lebensmittelchemie, Pharmazeutische Chemie,Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany

bInstitut für Anorganische Chemie, Universität Würzburg, Am Hubland, D-97074 Würzburg, GermanycUniversität Bonn, Pharmazeutisches Institut, Pharmazeutische Chemie Poppelsdorf, Germany

dStockholm University, Department of Neurochemistry and Neurotoxicology,Arrhenius Laboratories of Natural Sciences, Stockholm, Sweden

eKarolinska Institute, Nobel Institute for Neurophysiology, Department of Neuroscience, Stockholm, Sweden

(Received 8 February 1999; revised 30 April 1999; accepted 6 May 1999)

Abstract – A series of 5-alkyl-substituted UTP derivatives, which had been synthesized previously with a moderate degree of purity, wasresynthesized, purified, and characterized. Synthetic and purification procedures were optimized. New spectroscopic data, including13C- and31P NMR data, are presented. Phosphorylation reactions yielded a number of side products, such as the 2≠-, 3≠-, and 5≠-monophosphates, the2≠,3≠-cyclic monophosphates, and the 2≠,3≠-cyclic phosphates of the 5≠-triphosphates. Furthermore, raw products were contaminated withinorganic phosphates, including cyclometatriphosphate, phosphate, and pyrophosphate. The uracil nucleotides were investigated for theirpotency to increase intracellular calcium concentrations by stimulation of P2Y2 receptors (P2Y2R) on NG108–15 cells, a mouseneuroblastoma× glioma cell line, and in human basal epithelial airway cells, including a cystic fibrosis (CF/T43) cell line. UTP exhibited EC50values of ca. 1µM (in NG108–15 cells) and of 0.1µM (in CF/T43 cells), respectively. 5-Substituted UTP derivatives were agonists at theP2Y2R, but were less potent than UTP. 5-Ethyl-UTP, for example, exhibited an EC50 value of 99µM at P2Y2R of NG108–15 cells and provedto be a full agonist. With increasing volume of the 5-substituent of UTP derivatives, P2Y2 activity decreased. © 1999 Éditions scientifiqueset médicales Elsevier SAS

P2Y2 receptor agonists / P2U receptor agonists / UTP analogues / pyrimidine nucleotides / cystic fibrosis

1. Introduction

Membrane receptors for endogenous nucleotides, suchas ATP and UTP(figure 1) belong to the purine/pyrimidine receptor family. Purine receptors are sub-divided into two separate families, the adenosine recep-

tors, or P1-receptors (P1R), and the purine and pyrimi-dine nucleotide receptors or P2-receptors (P2R) [1, 2].The growing family of known P2 receptors comprisessubfamilies of ionotropic (P2X) and metabotropic (P2Y)receptors. While P2XR subtypes appear to be activatedexclusively by adenine nucleotides, subtypes of P2YRexist, including P2Y2, P2Y4, P2Y6, at which uracilnucleotides, namely UTP or UDP, show activity and mayact as physiological agonists. The P2Y2R is a ubiqui-tously expressed UTP-sensitive P2YR subtype, alsoknown as P2UR [3]. ATP and UTP are equipotent asagonists at the P2Y2R. It has recently been found that theimpaired chloride transport in the bronchi of patientssuffering from cystic fibrosis can be bypassed by stimu-lation of a P2Y2R to activate an alternative chloridechannel [4]. UTP is currently under development as a

# Preliminary results were presented at the Joint Meeting of theGerman and Swiss Pharmaceutical Societies in Zürich, Swit-zerland, 1997, and at the International Symposium on Adeno-sine and Adenine Nucleotides in Ferrara, Italy (abstract inDrugDev. Res.43 (1998) 34).* Correspondence and reprintsDr Christa Müller, Pharmazeutisches Institut der UniversitätBonn, Pharmazeutische Chemie Poppelsdorf, Kreuzbergweg26, D-53115 Bonn.

Eur. J. Med. Chem. 34 (1999) 809−824 809© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Figure 1. Structures of standard nucleotides.

810

novel therapeutic agent for the symptomatic treatment ofcystic fibrosis.

Only few P2Y2R agonists are known so far andstructure-activity relationships for P2Y2R agonists arevirtually unknown [5, 6]. In the present study we synthe-sized a series of 5-substituted UTP derivatives. Weselected UTP as a lead structure to develop P2Y2Rligands, rather than ATP, because of the higher selectivityof UTP versus other P2R subtypes, e.g., P2XR. Inaddition, degradation products of ATP and analogues(adenosine and analogues) may interact with P1R (ad-enosine receptors), but not those of UTP and ana-logues [7]. The synthesized nucleotides were investigatedin vitro for their potency to increase intracellular calciumconcentration ([Ca2+]i) caused by a P2Y2R-mediatedstimulation of phospholipase C [8].

2. Chemistry

5-Alkyl-substituted UTP derivatives10a–ewere syn-thesized according to published procedures with modifi-cations (figure 2)[9–12]. Uracil derivatives6a–e weresilylated using hexamethyldisilazane (HMDS) in thepresence of trimethylsilylchloride, or ammonium sulfate,

respectively. Derivatives with longer side chains dis-solved faster in HMDS due to their higher lipophilicity ascompared to those with smaller substituents. After com-plete dissolution, which indicated completion of thereaction, silylated uracil derivatives7a–ewere character-ized by their1H and13C NMR spectra(table I). Conden-sation with 1-O-acetyl-2,3,5-tri-O-benzoyl-â-D-ribofura-nose in the presence of tin(IV) chloride yieldednucleosides8a–e, essentially as described [9]. For 5-ethylderivative 7b, as an example, different reaction condi-tions were investigated. As solvents, 1,2-dichloroethaneor acetonitrile were used. As Lewis acid catalysts tin(IV)chloride, or trimethylsilyltriflate, were applied. All differ-ent combinations gave very good yields of8b (80–90%).In 1,2-dichloroethane as a solvent, traces of N3-nucleoside were formed, which could be removed bysilica gel column chromatography. Trimethylsilyltriflatewas found not to be superior to tin(IV) chloride.

Hydrolysis was performed using a saturated aqueoussodium hydrogencarbonate solution. In our hands, potas-sium hydrogencarbonate showed no advantage over thesodium salt, in contrast to results reported by Szemo etal. [9]. Filtration over silica gel efficiently removedtin(IV) oxide hydrate formed during the reaction. The

Figure 2. Preparation of 5-substituted UTP derivatives.

811

benzoylated nucleosides8a–ewere deprotected via trans-esterification using methanolic sodium methylate solu-tion. The formed benzoic acid methyl ester was conve-niently removed by extraction with diethyl ether andsubsequent lyophilization to yield the nucleosides9a–e.A recently described direct method for nucleoside syn-thesis [13], which does not require protection of thesugar, was investigated for the preparation of the 5-propylderivative10c, but was not successful in our hands. Atlow temperatures no reaction occurred and the startingmaterial was recovered. At temperatures over 80 °C onlydegradation was observed.

Nucleotides10a–e were prepared from nucleosides9a–eaccording to published procedures with slight modifi-cations [9–12]. The lyophilized nucleosides were dis-solved in dry trimethylphosphate and reacted with phos-phorus oxychloride, with or without the addition of1,8-bis(dimethylamino)naphthalene (‘proton sponge’).The reactive intermediate phosphorodichloridate wascoupled with a mixture of one equivalent of tri-n-butylamine and a five-fold excess of tri-n-butyl-ammonium diphosphate in dimethylformamide to yield10a–e(figure 2).

Nucleosides8a–eand9a–ewere characterized by their1H NMR spectra, which were in accordance with pub-lished data [9]. In addition,13C NMR spectra wererecorded(tables II and III) .

Crystals suitable for X-ray crystallography could beobtained from nucleosides8b and 8d. The structure ofnucleoside8d is shown infigure 3 (see alsotable VI).

Compound8d showed a nucleosidic torsion angle (O3-C5-N1-C1) of –133.22°. Thus, the nucleoside crystallizedin the anti-conformation despite bulky 5- and 5≠-substituents. The torsion angleγ about the exocyclicC8–C9 bond(figure 3)may be described as +sc (gauche/gauche). A vibrational disorder of the isopropyl sidechain could be seen, probably facilitated by the low datacollection temperature of 173 K. As an interesting fea-ture, the sugar puckering in8d deviated from the confor-mation observed in unprotected 5-methyluridine [14],which adopted a C3≠-endo conformation, while8dshowed a sugar puckering that can be described asC1≠-exo.

The synthesized nucleotides were purified by anionexchange chromatography using diethylaminoethyl-(DEAE-) Sephadex A25. The triethylammonium saltswere converted to the corresponding sodium salts bydissolution in absolute methanol and subsequent precipi-tation by the addition of sodium iodide in acetone. Thesodium salts were obtained as amorphous powders andtherefore better to handle for biological tests than thetriethylammonium salts, which were obtained as oily,transparent liquids.

Nucleotides were characterized by13C and31P NMRspectroscopy and plasma desorption mass spectrometry.Purity was determined by HPLC/UV and31P NMRspectroscopy(tables IIIandIV). The combination of bothmethods proved to be important for purity control(ta-ble IV).

Table I. 13C-NMR data of silylated uracil derivatives (in CDCl3) δ (ppm).

Compound O-Si(CH3)3 C2 C4 C5 C6 R5

7a 0.32 168.15 161.78 112.57 158.87 12.207b 0.29 167.84 161.69 118.33 157.90 13.42, 20.307c 0.26 167.87 161.66 116.75 158.63 13.67, 22.27, 28.917d 0.29 167.39 167.39 122.30 158.63 21.63, 26.247e –0.41 167.26 161.08 116.08 157.83 13.08, 21.54, 25.91, 30.76

Table II. 13C-NMR data of protected nucleosides (in DMSO-d6), δ (ppm).

Compound C5≠ C3≠ C2≠ C4≠ C1≠ Caromat. benzoyl C5 C6 C2 C4 R5

8a 63.7 70.7 73.1 78.8 88.5 128.4, 128.5, 128.7, 129.2,129.3, 133.5, 133.7

164.4, 164.6, 165.5 110.1 136.9 150.9 162.9 11.9

8b 63.6 70.6 73.0 78.8 88.8 128.4, 128.6, 128.7, 129.2,129.3, 133.5, 133.8

164.5, 164.6, 165.5 115.7 136.7 150.2 163.2 12.9, 19.5

8c 63.6 70.7 72.9 78.8 88.4 128.4, 128.5, 128.6, 129.3,133.5, 133.6, 133.7

164.5, 164.6, 165.5 114.1 137.0 150.2 163.2 13.3, 21.1, 28.1

8d 63.7 70.6 73.0 78.8 89.4 128.4, 128.6, 128.7, 129.2,129.3, 133.5, 133.8

164.5, 164.6, 165.5 119.9 136.1 150.0 162.9 21.0, 21.1, 25.5

8e 63.8 70.7 73.0 78.9 88.6 128.4, 128.5, 128.7, 129.2,129.3, 133.5, 133.7

164.6, 164.7, 165.5 114.4 137.1 150.3 163.3 13.6, 21.7, 25.9, 30.2

812

Table III. 13C-NMR data of nucleosides and nucleotides (in DMSO-d6), δ (ppm).

Compound C5≠ C3≠ C2≠ C4≠ C1≠ C5 C6 C2 C4 R5 3JC4≠-P2JC5≠-P

9a 60.9 69.8 73.4 84.8 87.6 109.2 136.4 150.8 163.8 12.29b 60.8 69.9 73.5 84.8 87.7 115.1 135.8 150.7 163.4 12.9, 19.79c 60.8 69.8 73.4 84.7 87.8 113.2 136.4 150.6 163.3 12.9, 21.0, 28.49d 60.7 69.9 73.8 84.7 88.1 119.4 134.9 150.4 163.0 21.2, 21.4, 25.49e 60.8 69.8 73.6 84.8 87.8 113.5 136.4 150.7 163.5 12.9, 21.7, 25.9, 30.1

10a 66.6 74.9 71.5 85.0 89.2 113.2 138.7 153.4 167.8 14.2 9.2 6.110b 66.7 74.6 71.6 85.1 89.0 118.9 138.2 153.2 167.1 12.1, 21.2 9.2 4.910c 66.8 64.8 71.6 85.1 89.2 117.2 139.0 153.3 167.3 14.4, 22.8, 29.6 9.2 6.110d 67.0 75.0 71.8 85.3 89.4 123.5 138.7 153.2 166.9 22.2, 22.5, 21.2 9.2 6.110e 66.9 74.8 71.6 85.1 89.3 117.7 138.9 153.8 167.5 14.9, 23.1, 27.5, 31.9 9.2 6.1

Figure 3. X-ray structure of nucleoside8d.

813

Phosphorylation reactions were investigated in somedetail. It had been reported that reaction of nucleosideswith phosphorus oxychloride was somewhat faster intrimethyl phosphate as compared to triethyl phos-phate [15]. We investigated the reaction of ethyl deriva-tive 9b with phosphorus oxychloride under various con-ditions. Reaction time was always virtually identical inboth solvents. Proton sponge accelerated reaction timefrom 7 to 2 h, while pre-heating of reagents before theaddition of phosphorus oxychloride had no effect.

Phosphorylation of9a–eto obtain triphosphates10a–eyielded a number of side products (figure 4). For butylderivative 10e, structure elucidation of the major sideproducts was performed as preparation of10eyielded thelargest quantity of side products. A fraction isolated at aconcentration of 0.13–0.16 M triethyl ammonium bicar-bonate (TEAB) buffer showed five spots in thin layerchromatography on cellulose-coated plates, all of whichcontained phosphorus as shown by spraying with aphosphate-specific reagent. Four of the spots showed UVabsorption at 266 nm (table V). The mass spectra (anionicmode) confirmed triphosphate10eas the major product.Several pH-dependent signals were detected in the31PNMR spectrum. According to Cozzone and Jar-detzky [16], the signals could be assigned to the 2≠-monophosphate (11), the 3≠-monophosphate (12), the5≠-monophosphate (13), and the 2≠,3≠-cyclomonophos-phate (14, figure 5). The identification of the side prod-ucts is shown intable V.

A fraction isolated at 0.43 M TEAB buffer concentra-tion contained two more side products and a smallamount of10e. The major side product showed a mass of600.7. The 31P NMR spectrum in dimethylsulfoxideshowed four signals, a triplet at –23.87 ppm (Pâ, JPαPâ =

21.8 Hz,JPγPâ = 22.9 Hz), a doublet at –11.93 ppm (Pα,JPαPâ = 21.8 Hz), a doublet at –11.39 ppm (Pγ, JPγPâ =22.9 Hz), and a singlet at 16.41 ppm. These data indicatethat the compound (15) contains a 2≠,3≠-cyclophosphategroup in addition to the 5≠-triphosphate residue of10e.The second side product with a molecular mass of 619.1is a ring-open analogue of15, since the 31P NMRspectrum clearly shows a signal in the range of amonophosphate in addition to the 5≠-triphosphate struc-ture. Bisphosphorylation at the 2≠- and 3≠-hydroxylgroups was not observed. The ratio of formed10e:15wasca. 2:1.

The particularly large amount of side products ob-served in the preparation of the 5-butyluridine triphos-phate10e may be due to steric interaction between thebulky butyl residue and the 5≠-hydroxyl function in polaraprotic solvents, which leads to an increased phosphory-lation of the secondary hydroxyl groups at C2≠ and C3≠.The preparation of other nucleotides yielded the analo-gous side products, but in smaller quantities. Use ofproton sponge as reaction accelerator was not the causefor the formation of 2≠- and 3≠-phosphoric acid estersand 2≠,3≠-cyclophosphates as side products, as we ob-served the formation of the same compounds in theabsence of proton sponge.

In addition to nucleotide side products, large amountsof inorganic phosphates, including cyclometatriphosphate(P3O9

3–), phosphate (PO43–), pyrophosphate (P2O7

4–)and linear triphosphate (P3O10

5–) were obtained(table V).The stable cyclometatriphosphate was particularly diffi-cult to remove by standard chromatographic methods dueto partial coelution with the target UTP derivatives(figure 4and table V).

Table IV. 31P-NMR (in D2O), MS data and HPLC retention times of nucleotides10a–e, δ (ppm).31P-NMR MSa HPLC (UV; λ =266 nm)

Compound pH Pα Pâ Pγ Mole Peak Ion System Ab System Bc

Retention time (Purity) Retention time (Purity)

10a 5.9 –10.69 d –22.56 t –10.04 d 496.9 [M – H]– 13.51 (94 %) 7.56 (95 %)J = 19.6 Hz J = 19.5 Hz J = 19.5 Hz

10b 6.5 –10.26 d –21.90 t –9.26 d 511.1 [M – H]– 21.88 (95 %) 7.82 (96 %)J = 19.6 Hz J = 19.6 Hz J = 19.5 Hz

10c 6.7 –10.38 d –22.30 t –9.05 d 524.9 [M – H]– 25.97 (93 %) 8.01 (94 %)J = 19.6 Hz J = 19.5 Hz J = 19.5 Hz

10d 7.7 –10.58 d –22.03 t –7.14 d 524.8 [M – H]– 17.36 (94 %) 9.40 (92 %)J = 20.0 J = 20.2 Hz J = 19.8 Hz

10e 5.8 –10.98 d –22.56 t –10.04 d 539.1 [M – H]– 31.15 (92 %) 8.07 (93 %)J = 20.2 Hz J = 20.0 Hz J = 19.8 Hz

a Negative ion plasma desorption MS.b System A: Nucleosil RP-18 column with eluent A, 0.1 M triethylammonium acetate buffer and eluentB, acetonitrile (gradient: 0–15%, 30 min).c System B: Macherey-Nagel ET125/4 Nucleosil 4000-7 PEI column with eluent A, 0.01 MTris/HCl (pH = 8.4) and eluent B, Tris/HCl (pH = 8.4), 1.0 M NaCl.

814

Table V. Analysis of isolated fractions obtained from ion exchange chromatography of crude10e (comparefigure 3).

Fraction I II III IV V VI

TEAB concentration[%]a

0 13–16 31–36 41–44 70–80 80

Yield [mg] (of 1 820mg crude product)b

810 241 250 67 114 33

Spectroscopic methodsUV maximum at [nm] 218c 266 266 266 266 26631P-NMR (in D2O) –7.3 (s); 2.4 (s); 3.0

(s); 3.1 (s)(s); 2.8 (s); 2.3 (s);0.4 (s); 0.0 (s)

–9.8 (d); –11.0 (d);–20.8 (s); –22.4 (t)

21.8 (s); –5.2 (d);–9.8 (d);

4.8 (s); 4.2 (s);2.2(s);

21.8 (s); 4.5. (s); –8.1(s)d; –10.3 (s)d;

(pH value) (pH = 7.2) (pH = 4.5) (pH = 6.2) –20.8 (t) –7.7– –10.3 (m); –21.7 (s)d

19.2 (s); 4.0 (s); 3.7(s); 3.5 (s); 2.7 (s)

(pH = 8.0) –20.0 (s); –21.7 (s)d

(pH = 7.4)(pH=8.0)

(pH = 9.5)MS (neg. Mode) n.d.e 379f 538f 601f, [619] [540, 619], 700f 760f

Cellulose-TLC (solvent: 2-propanol:NH3(25 %):H2O = 6:3:1)UV detection at 266nm

3 spots 4 spots 1 spots 2 spots 4 spots 3 spots

Phosphate detection 3 spots 5 spots 2 spots 2 spots 5 spots 5 spots(FeCl3/5-sulfosalicylicacid reagent)

Deduced structureInorganic phosphate asdetermined by 31P-NMR

PO43– PO4

3– P3O93– P3O9

3–, PO43–,

P3O105–

PO43–, P3O10

5–

P2O74–

Nucleotidesg and otherorganic compounds

phosphates of 1,8-bis-(dimethyl)-ammonium-naphthalene

11 10e 15 structure could notbe determinedh

structure could not bedeterminedi12

1314

a 100% corresponds to 1 M TEAB buffer.b Unidentified: 305 mg.c Typical spectrum for aromatic compounds.d Signal not resolved.e n.d. = not determined.f Mainpeak.g For structures seefigures 2and 5. h Structure contains a total of 5 phosphate groups but no 2′3′-cyclic phosphate groups.i Structure contains a total of6 phosphate groups (one of them is a 2′3′-cyclic phosphate group).

815

3. Biological evaluation

P2Y2 receptor activity was determined in a mouseneuroblastoma× glioma hybrid cell line (NG 108-15) byfluorimetric measurement of the increase in intracellularcalcium concentration caused by stimulation of P2Y2

receptors via activation of phospholipase C. Selectedcompounds were additionally tested in a human cysticfibrosis epithelial airway cell line (CF/T43) and a humanbasal epithelial airway cell line (BEA).

4. Results and discussion

4.1. Neuroblastoma× glioma cells (NG108-15 cell line)

The putative physiological P2Y2 receptor agonistsUTP (1) and ATP (4), the ATP analogue ZTP (5), uridinediphosphate UDP (2) and the corresponding monophos-phate UMP (3) were investigated at NG108-15 cells forcomparison. Initially, single concentrations (50 and/or500 µM) of UTP derivatives and standard nucleotideswere used(table VII). Before testing it was shown thatnot even at high concentrations (500µM) did the inor-

ganic phosphates pentasodium triphosphate (Na5P3O10),trisodium tricyclometaphosphate (Na3P3O9), and sodiumpyrophosphate (Na4P2O7), which could be present as sideproducts in some of the nucleotides, show any effects. Aspreviously shown, UTP acted as a potent P2Y2 receptoragonist, exhibiting an EC50 value of 1.25µM and amaximal increase in intracellular calcium of 69%(ta-ble VIII). The observed potency for UTP is in accordancewith published data [5, 17]. ATP was somewhat lesspotent, exhibiting an EC50 value of 17.7µM, but showedabout the same maximal stimulation as UTP. The loweractivity of ATP may be due to faster enzymatic degrada-tion of ATP as compared to UTP under test condi-tions [18, 19]. The ring-open, base-modified ATP ana-logue ZTP (5) was found to be nearly equipotent to ATPin our assay system.

In the system described, addition of UDP (2) alsoresulted in an increase in intracellular calcium concentra-tion, with an EC50 value of 3.2µM (table VIII). Efficacy,however, was only 61% of that of UTP. UDP haspreviously been shown to be inactive at P2Y2R [19, 20],but it might have been contaminated with UTP, or couldhave been phosphorylated enzymatically, and the result-

Figure 4. Separation of crude10eon DEAE-Sephadex A25 by FPLC.

816

ing UTP may have been responsible for the effectsmeasured. Uridine monophosphate (UMP,3) exhibitedonly very weak activity at high concentrations of 500µM,which might, again, be due to contamination by UTPand/or enzymatic conversion to UTP [18, 19].

5-Alkyl-substituted UTP derivatives showed loweractivity at P2Y2 receptors than UTP. Activity decreasedwith increasing size of the 5-substituent exhibiting thefollowing rank order of potency: UTP> 5-methyl-UTP>5-ethyl-UTP > 5-isopropyl-UTP > 5-propyl-UTP >5-butyl-UTP(table VII). A recorded dose-response curvefor 5-ethyl-UTP (10b) showed this compound to be a fullagonist at P2Y2R of NG108-15 cells exhibiting an EC50

value of 99µM (table VIII).

4.2. CF/T43 and basal epithelial airway (BEA) cells

Selected compounds were further tested in CF/T43cells [4], a human epithelial airway cell line containingthe defective chloride channel (CFTR), that causes cysticfibrosis, and compared with a basal epithelial airway(BEA) cell line. Results from measurements of P2Y2R-mediated increase [Ca2+]i were virtually identical inCF/T43 and BEA cells(tables IX and X). Activity ofnucleotides and maximal increase in [Ca2+]i was higher inCF/T43 and BEA cells as compared to NG108-15 cells.ATP was found to be nearly equipotent to UTP in CFT/43and BEA cells(table IX), while ZTP was somewhat lesspotent (table X). 5-Methyl-UTP (10a) showed a dose-

Table VI. Crystal data of compound8d.

Empirical formula C33H30N2O9

Formula weight 598.59Temperature 173 (2) KWavelength MoKα, 71.073 pmCrystal system orthorhombicSpace group C2221

Unit cell dimensions a = 1 482.37 (7) pmα = 90 deg.b = 2 322.16 (14) pmâ = 90 deg.c = 1 780.29 (13) pmγ = 90 deg.

Volume 6.1283 (6) nm3

Z 8Reflections used for cell refinement 25Density (calculated) 1.298 Mg/m3

Absorption coefficient 0.095 mm-1

F(000) 2 512Crystal size 0.3× 0.2 × 0.2 mmTheta range for data collection 2.29–25.05 deg.Index ranges –17≤ h ≤ 17, –27≤ k ≤ 27, –21≤ l ≤ 21Reflections collected 20 046Independent reflections 5 425 [R(int) = 0.0561]No. of standard reflections / decay 3 / 0 %Completeness to 2Theta = 25.05 99.6 %Absorption correction N/aRefinement method Full-matrix least-squares on F2

Treatment of hydrogen atoms mixeda

Data / restraints / parameters 5 425 / 203 / 436Goodness-of-fit on F^2 0.912Final R indices [I> 2sigma(I)]b R1 = 0.0392, wR2 = 0.0835R indices (all data)b R1 = 0.0638, wR2 = 0.0913Absolute structure parameter –0.5(9)Extinction coefficient 0.0017(2)Largest diff. peak and hole 157 and –201 e.nm–3

Diffractometer: ENRAF-NONIUS CAD4Program: SHELXS-93 [28], SHELXL-97 [29]

a See Experimental.b Definition of R values:R1 = R i Fo|–|Fc i /R| andwR2 = $Rw~ Fo

2 − Fc2!

2/Rw~ Fo2!

2%1/2; w = 1/$r2~ Fo

2! + ~ 0.0494⋅ P !

2}; P = ~ Max~ 0, Fo2! + 2 Fc

2!/3.

817

dependent increase in [Ca2+]i and appeared to be nearlyequipotent to ATP and UTP(table X), while 5-ethyl-UTP(10b, 10–100 nM) only slightly increased [Ca2+]i in BEAcells (table X).

The higher potency of nucleotides in epithelial airwaycells as compared to NG108-15 cells may reflect species

differences (human, mouse). Differences in enzyme pat-tern and enzymatic activity of nucleotidases and phos-phatases may also contribute to the differences mea-sured [18, 19, 21]. A further factor, which has to be takeninto account, may be differences in techniques used forthe measurements. Thus, NG108-15 cells were used in

Figure 5. Structures of identified side products in the phosphorylation of9e (preparation of10e).

Table VII. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in NG108-15 cells, results fromtesting of single concentrations of compounds.

Compounda Percent increase in [Ca2+]i ± SEM (number (n) of independent experiments)50 µM concentration of test compound 500µM concentration of test compound

1 UTP 69.3± 8.2 (n = 10) (179.0± 26.9)b (n = 5)2 UDP 42.4± 4.5 (n = 4) 31.3 (n = 1)3 UMP n.d.c 10.6 ± 1.3 (n = 2)4 ATP 44.6± 5.9 (n = 3) 67.2± 17.9 (n = 2)5 ZTP 54.3± 13.5 (n = 2) n.d.10a 5-Methyl-UTP n.d. n.d.

52.0 ± 29.7 (n = 2) at 100µM10b 5-Ethyl-UTP 28.9± 7.8 (n = 3) 71.7± 12.0 (n = 3)10c 5-Propyl-UTP 0 (n = 2) 38.6± 5.3 (n = 3)10d 5-Isopropyl-UTP n.d. 46.2± 25.0 (n = 2)10e 5-Butyl-UTP n.d. 25.0± 0.7 (n = 3)

a For structures seefigures 1and 2. b High values are due to lysis of cells observed at a concentration of 500 µM of UTP.c n.d. = notdetermined.

818

suspension while CF/T43 and BEA cells were used asattached cell layers in culture dishes (see Experimentalprotocol). It has been shown that stressing of cells, e.g,.by agitation, may result in a massive release of endog-enous nucleotides [22], a fact which clearly affects thetest results obtained with exogenously applied nucle-otides. Such variations in EC50 values for nucleotides,e.g., UTP and ATP, at P2Y2R in different test systemshave been described [5]. In the present study, degradation

of compounds was not investigated, though enzymaticand chemical hydrolysis undoubtedly play a role [21].Calcium measurements were, however, performed veryfast, within seconds up to a few minutes, and therefore,only moderate degradation will have occurred. It hadbeen shown that in airway epithelial cells, where strongUTP degradation is observed, after 5 min, about 70% ofintact UTP is still present [21].

4.3. Structure-activity relationships

The P2Y2R is unique in that it is stimulated by purine(ATP) as well as pyrimidine nucleotides (UTP). Weconfirmed previous findings [5, 20, 22, 23] that UTP (1)and ATP (4) activate P2Y2R in low micromolar tosubmicromolar concentrations (EC50 of UTP: 1.25µM atNG108-15 cells, 0.10µM at CF/T43 cells, and 0.16µMat BEA cells; EC50 of ATP: ca. 18µM at NG108-15 cells,and ca. 0.2µM at CF/T43 and BEA cells, seetables VIIIand IX). Efficacy was similar for both physiologicalagonists. A ring-open analogue of ATP, ZTP (5) exhibitedsimilar potency as ATP indicating that purine base modi-fication was possible without loss of activity. Truncation

Table VIII. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in NG108-15 cells, results fromdose-response curves.

Compound EC50 [µM]a Maximal effect in % of max. UTP effect (= 100 %)(95 % confidence intervals) (number (n) of independent experiments)

1 UTP 1.25 (0.11–14) 100± 12 (n = 10)2 UDP 3.21 (0.28–36) 61± 7 (n = 4)4 ATP 17.7 (5.6–56) 95± 12 (n = 3)5 ZTP 17.1b 78 ± 19 (n = 2)10b 5-Ethyl-UTP 99 (20–480) 104± 17 (n = 3)

a Results from three independent experiments unless otherwise noted.b Single dose-effect curve.

Table IX. Effects of standard nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in CF/T43cells and basal epithelial airway (BEA) cells, results from dose-response curves.

Compound Cell line EC50 [µM](95 % confidence intervals)a

1 UTP CF/T43cells 0.10 (0.01–0.93)BEA cells 0.16 (0.005–5.6)

4 ATP CF/T43cells 0.17 (0.01–2.9)BEA cells 0.20b

a Determined in 3–4 separate experiments unless otherwise noted.b Two separate experiments.

Table X. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in CF/T43 cells and BEA cells,results from testing of single concentrations of compounds

Compound Cell line Percent increase in [Ca2+]i ± SEM(number of independent experiments)10 nM concentration of test compound 100 nM concentration of test compound

1 UTP CF/T43cells 108± 80 (n = 3) 197± 58 (n = 3)BEA cells 110± 49 (n = 3) 173± 113 (n = 3)

4 ATP CF/T43cells 103± 20 (n = 3) 266± 94 (n = 3)BEA cells 199± 94 (n = 3) 208± 42 (n = 3)

5 ZTP CF/T43cells 36± 29 (n = 2) 117± 55 (n = 2)BEA cells 80± 35 (n = 2) 138± 22 (n = 2)

10a 5-Methyl-UTP CF/T43cells 155± 46 (n = 2) 269± 24 (n = 2)BEA cells 113± 32 (n = 2) 312± 15 (n = 2)

10b 5-Ethyl-UTP CF/T43cells 39± 11 (n = 2) 43 ± 26 (n = 2)BEA cells 17± 24 (n = 2) 41 ± 7 (n = 2)

819

of the triphosphate chain in UTP to UDP (2) and furtherto UMP (3), led to a decrease or loss in activity asreported [5, 20, 22], suggesting that the four negativecharges are important for electrostatic interactions withthe receptor protein. Only few synthetic UTP analogueshave previously been investigated as P2Y2R ligands [5, 6,21, 22]. Exchange of the bridgingâ-γ-oxygen atom in thetriphosphate chain by NH, CH2, or CF2, respectively, inorder to enhance stability towards nuceotidases, have ledto a decrease in activity [23]. The only substitutiontolerated in the phosphate chain was the replacement of aγ-oxygen atom by sulfur, as in UTPγS [21]. The com-pound was nearly as potent as UTP itself. A series of UTPderivatives in which the oxygen in the 4-position of theuracil moiety was replaced by various substituents, in-cluding (alkyl)amino, (alkyl)thio or alkoxy, had beeninvestigated, but all of the derivatives were less potentthan UTP [6].

A 5-substituted UTP derivative that had been investi-gated was 5-bromo-UTP. It was found to be less potentthan UTP or ATP at P2Y2R [22]. The present study showsthat alkyl substituents in the 5-position of UTP are nottolerated by the receptors either. Introduction of an ethylgroup in the 5-position of UTP, for example, decreasedactivity at P2Y2R of NG108-15 cells ca. 80-fold (from anEC50 for UTP of 1.25 µM to 99 µM for 5-ethyl-UTP10b). With increasing volume of the 5-substitutent P2Y2

activity was decreased(table VII). Since both polar5-substituents, such as bromo, and non-polar, especiallybulky alkyl substituents, led to decreased activity, it canbe concluded, that steric reasons, i.e., lacking bulktolerance of the receptor in that area, are responsible forthis effect.

5. Conclusion

Methods and conditions for the preparation and puri-fication of 5-substituted UTP derivatives were investi-gated and optimized. New spectroscopic data on UTPderivatives are presented, including13C and 31P NMRdata. 5-Substituted UTP derivatives were found to be fullagonists at P2Y2 receptors of NG108-15 cells, and basalepithelial airway cells without or with defective CFTRchannel (CF/T43 cell line). Potency of 5-substituted UTPderivatives decreased with increasing volume of the5-substitutent.

It is planned to investigate the potency of 5-substitutedUTP derivatives at other P2Y receptor subtypes, such asP2Y4 or P2Y6. The presented SAR of UTP derivativesmay contribute to the design of selective ligands forsubtypes of the uracil nucleotide-sensitive P2Y receptors.

6. Experimental protocols

6.1. Chemistry

Melting points were determined with a Büchi 530apparatus and are uncorrected. Nuclear magnetic reso-nance spectra were determined using a Bruker AC-200spectrometer (1H: 200 MHz, 13C: 50.3 MHz). Chemicalshifts are given in ppm downfield from tetramethylsilaneas internal standard.31P-NMR spectra were recorded ona Bruker AMX 400 spectrometer at 161 MHz with H3PO4

as an internal standard. UV spectra were recorded with aPerkin Elmer Lambda 12 spectrometer. HPLC was per-formed using the following equipment: Pharmacia 2249gradient pump, 2141 variable wavelength monitor set to266 nm, and a 2221 integrator. Purity controls wereperformed on two systems: system A: Nucleosil RP-18column with eluent (A) 0.1 M triethylammonium acetatebuffer, and eluent (B) acetonitrile/H2O (gradient: 0–15%,30 min). System B: Macherey-Nagel ET125/4 Nucleosil4000-7 PEI column with eluent (A) 0.01 M Tris/HCl (pH8.4) and eluent (B) 0.01 M Tris/HCl (pH 8.4), 1.0 MNaCl. Negative ion plasma desorption mass spectra wereobtained on Applied Biosystem BIO-ION 20 plasmadesorption mass spectrometer, using252Cf as source offission fragments. Nucleotides were purified on a Phar-macia High Load system using a triethyl ammoniumhydrogen carbonate buffer gradient (0–80%). Thin layerchromatography was performed on silica gel F254

(Merck), and in some cases on cellulose TLC plates(Merck). The following solvent systems were used fordevelopment of the TLC:

(S1) dichloromethane:methanol = 9:1 for compounds8a–e;

(S2) dichloromethane:methanol = 3:1 for compounds9a–e;

(S3) 2-propanol:NH4OH:water = 6:3:1 for compounds10a–e, 11–15.

Phosphate spraying agent (0.1% FeCl3 x 6 H2O; 7%5≠-sulfosalicylic acid in 75% ethanol) according to Wadeand Morgan [24] was used for identification of organicand inorganic phosphates. RF-values of phosphates werecompared to data published by Ebel [25].

6.1.1. General procedure for silylationA suspension of 5 mmol of uracil derivatives6a–e in

20 mL of 1,1,1,3,3,3-hexamethyldisilazane (HMDS) and1 mL of trimethylsilyl chloride, or a few crystals of(NH4)2SO4 respectively, was refluxed until a clear solu-tion was obtained. The solution was allowed to cool, andexcess HMDS was removed under reduced pressure. Thesilylated uracil derivatives7a–ewere kept under nitro-gen. A small sample of the oily sirup was dissolved in

820

CDCl3 and checked for complete silylation by1H NMRspectroscopy. Yields ranged from 96–98%.

Selected1H-NMR data in CDCl3, δ (ppm):7d: 0.30 (s, 9H, C-O-Si(CH3)3); 0.32 (s, 9H, C-O-

Si(CH3)3); 1.14 (d, 6H, C5-CH-(CH3)2, J = 6.9Hz); 2.85(sept., 1H, C5-CH-(CH3)2, J = 6.9Hz); 7.99 (s, 1H,C6-H).

6.1.2. General procedure for the synthesis of nucleosidesSilylated uracil derivatives7a–e(5 mmol) under nitro-

gen were dissolved in 5 mL of 1,2-dichloroethane, and1.30 g (5.5 mmol) of SnCl4 (10% excess) was added withvigorous stirring. 1-O-Acetyl-2,3,5-tri-O-benzoyl-â-D-ribofuranose (2.2 g, 0.4 mmol) in 40 mL of 1,2-dichloroethane was added dropwise to the slightly yel-lowish solution. The mixture was stirred as indicatedbelow, completion of reaction was controlled by TLC(dichloromethane:methanol = 9:1). Reaction times werebetween 4–5 h. The reaction solution was poured into asaturated aqueous NaHCO3 solution (ca. 50 mL), or onhumidified solid NaHCO3 with successive adding ofsaturated NaHCO3 solution, respectively, under vigorousstirring, and then allowed to stand overnight. The suspen-sion was filtered over silica gel and the gel was washedtwice with 50 mL of dichloromethane and twice with100 mL of ethyl acetate. The organic phase was sepa-rated, dried (Na2SO4), filtered over silica gel and evapo-rated to dryness. The glassy residue was crystallized frommethanol, or ethanol respectively, to yield8a–e. Ifnecessary, the removal of unreacted sugar and sideproducts was achieved by column chromatography onsilica gel using dichloromethane:methanol = 9:1 as elu-ent. Yields ranged from 81–87%.

Selected1H-NMR data in DMSO-d6, δ (ppm):8b: 0.93 (t, 3H, C5-CH2-CH3, J = 7.3 Hz); 2.08 (q, 2H,

CH3, J = 7.3 Hz); 4.64–4.77 (m, 3H, C4≠H, C5≠H2); 5.96(m, 2 H, C2≠H, C3≠H); 6.21 (d, 1H, C1≠H, J1≠2≠ = 3.8Hz); 7.39–8.04 (m, 16H, C6-H, Haromat.); 11.48 (br s, 1H,N1-H, exchangeable).

8c: 0.76 (t, 3H, propyl CH3, J = 7.1 Hz); 1.33 (sext.,2H, propyl CH3CH2-, J = 7.3Hz); 2.05 (t, 2H, propylCH3CH2CH2-, J = 7.2 Hz); 4.61–4.78 (m, 3H, C4≠H,C5≠H2); 5.90 (m, 2H, C2≠-H, C3≠-H); 6.23 (d, 1H,C1≠-H, J = 3.7 Hz); 7.38–8.06 (m, 16H, C6H, Haromat);11.48 (br s, 1H, N3-H, exchangeable).

8d: 1.00 (d, 6H, CH(CH3)2, J = 6.9Hz); 2.85 (sept., 1H,CH(CH3)2, J = 6.6Hz); 4.65–4.75 (m, 3H, C4≠H,C5≠H2); 6.00 (d, 2H, C3≠H, C2≠H); 6.25 (d, 1H, C1≠H,J = 5.1Hz); 7.01 (d, 1H, C6H,J = 0.9Hz); 7.38–8.05 (m,16H, C6H, Haromat); 11.47 (br s, 1H, N3-H).

6.1.3. Deprotection of nucleosidesTribenzoyl nucleoside (8a–e, 0.92 mmol) was dis-

solved in a mixture of 20 mL of absolute methanol and2.76 mL of 5% methanolic sodium methylate solution.The solution was stirred for 5–8 h and completion of thereaction was determined by TLC (dichloromethane:methanol = 3:1). Neutralisation of the solution wasachieved by adding DOWEX-50 WX-8 ion exchangeresin washed previously with methanol. After filtering theresin off, 20 mL of water was added and methanol wasevaporated. The benzoic acid methyl ester was extractedwith diethyl ether and the water fraction was lyophilizedto yield the nucleosides9a–ein 94–98% yield.

Selected1H-NMR data in DMSO-d6, δ (ppm):5-n-Propyluridine (9c): 0.84 (t, 3H, C5-CH2-CH2-CH3,

J = 7.3Hz); 1.42 (sext., 2H, C5-CH2-CH2-CH3, J = 7.5Hz); 2.14 (dt, C5-CH2-CH2-CH3, J1≠2≠ = 7.5 Hz);3.49–3.68 (m, 2H, C5≠H2); 3.82 (q, 1H, C4≠H, J =3.46/2.94Hz); 3.96–4.03 (m, 2H, C2≠H, C3≠H); 5.06,5.33 (br s, 3H, OH (exchangeable)); 5.77 (d, 1H, C1≠H,J = 5.3 Hz); 7.73 (s, 1H, C6H); 11.24 (br s, 1H, N3-H(exchangeable)).

6.1.4. PhosphorylationSynthesis of nucleoside 5≠-triphosphate synthesis was

adapted from literature procedures [10, 26, 27].

6.1.5. Preparation of tri-n-butylammonium diphosphateSodium diphosphate decahydrate (6.69 g; 15 mmol)

was dissolved in 150 mL of water (twice distilled).Excess of Dowex ion exchange resin 50× 8, 20–50 mesh,proton form, prewashed several times with water, wasadded to the solution of sodium diphosphate, and themixture was gently stirred for 60 min. A mixture of60 mL of ethanol and 7.14 mL of tributylamine in a flaskwas placed in an ice-water bath, and the diphosphatesolution was filtered directly into the flask. The resin wasrepeatedly washed with water until the filtrate was nolonger acidic. The solvent was then evaporated undervacuum at 40 °C, yielding a thick, nearly colourlesssyrup. This residue was treated twice with 100 mL ofethanol and then evaporated. The residue was taken up in40 mL of dimethylformamide (DMF, anhydrous grade,from Fluka) and evaporated again. This residue was takenup in 30 mL of anhydrous DMF, yielding 30 mL of 0.5 Mtri-n-butylammonium diphosphate in DMF. The solutionwas stored sealed and cooled at 4 °C until used.

6.1.6. Preparation of triethylammoniumhydrogen-carbonate (TEAB) buffer

A 1 M solution of TEAB was prepared by bubblingCO2 through a 1 M triethylamine solution in water at0–4 °C for several hours (pH approx. 7.4–7.6).

821

6.1.7. Preparation of nucleotides10a–eLyophilized nucleoside (9a–e, 1 mmol) was dissolved

in 5 mL of trimethyl phosphate (dried over 10 Å molecu-lar sieve). The mixture was stirred at room temperatureunder nitrogen and then cooled to 4 °C. Dry 1,8-bis(dimethylamino)naphtalene (‘proton sponge’, 0.32 g,1.5 mmol) was added, followed by 1.3 mmol (0.20 g) ofPOCl3, 5 min later. After several hours of stirring at0–4 °C, tri-n-butylamine (0.1 mL, 0.72 mmol) was addedto the solution followed by 10 mL (5 mmol) of 0.5 Mtri-n-butylammonium diphosphate solution in DMF. After2–5 min the mixture was poured into 0.5 M cold aqueousTEAB solution (30 mL, pH 7.5) and stirred at 0–4 °C forseveral minutes.

The solutions were allowed to reach room temperaturewith stirring and were then left standing for 1 h. Trim-ethylphosphate was extracted with t-butylmethyl etherand the aqueous solutions were evaporated and lyo-philized to yield glassy colourless oils. The reactionswere controlled by TLC using freshly prepared solventsystem S3. TLC plates were dried before UV absorptionwas detected and the plates were subsequently sprayedwith phosphate reagent.

6.1.8. Purification of nucleotidesThe crude nucleoside 5≠-triphosphates were purified

by ion exchange column chromatography using DEAE-Sephadex A25 (Pharmacia) HCO3

–-form swelled in 1 MNH4HCO3 solution at 4 °C. After equilibrating the col-umn with deionized water, the crude product was dis-solved in 5 mM aqueous triethylammonium carbonatebuffer. The column was washed with deionized water,followed by a solvent gradient of 0–800 mM TEAB inapproximatly 3 000 mL of solvent to elute the triphos-phates. Fractions were collected and appropriate fractionspooled, evaporated, diluted in water and lyophilized toyield a product with a ‘clear glass’ appearance. Puritywas assessed by HPLC/UV and31P NMR spectroscopy.Isolated yields: 33% (10a), 24% (10b), 37% (10c), 40%(10dd), 26% (10e).

6.1.9. Preparation of sodium saltsNucleotides (50 mmol) were dissolved in absolute

methanol (dried over Mg) and evaporated. This procedurewas repeated twice. The solid was then dissolved in 5 mLof absolute methanol with stirring. A 1.5-fold excess of a1 M sodium iodide solution in acetone was addeddropwise. The solution was then diluted with absoluteacetone. The formed precipitate was filtered off andwashed several times with absolute acetone. The whitesolid was dried under high vacuum.

6.2. Determination of X-ray structure

Single pure crystals of8d were obtained by slowcrystallization from ethanol at room temperature overseveral weeks. Low temperature X-ray intensity datawere collected on an Enraf-Nonius CAD4 diffractometerusing MoKα radiation. The structure was resolved withdirect methods using SHELXS-97 [28] and refined byfull-matrix least squares iteration against F2, usingSHELXL-97 [29]. All non-hydrogen atoms were refinedanisotropically. The coordinates of hydrogen atomsbonded to C5, C6, C7, and C8 were freely refined usingisotropic displacement parameters. For all other hydrogenatoms a riding model was employed in the refinementwith their Uiso constrained to equal 1.2 times the Ueq ofthe parent atoms (1.5 times Ueq in the case of thehydrogen atoms of the methyl groups). Crystallographicdata were deposited at the Cambridge CrystallographicData Centre as supplementary publication (depositionnumber 120084). Free copies of data are available from:CCDC; 12 Union Road, Cambridge CB21EZ (fax:+ 44 1223 336 033; e-mail: deposit@ccdc.cam.ac.uk).

6.3. Biological studies

ATP, UTP, UDP, UMP, and ZTP were purchased fromSigma Chemical Co. in the highest available purity grade.Nucleotides were checked for autofluorescence prior totesting.

6.3.1. NG108-15 cell cultureNG108-15 cells were kindly provided by Berit Färjh,

Dept. of Medical Neurochemistry, Lund University Hospi-tal, Sweden. NG108-15 cell line (hybrid cell line generatedby fusion of mouse neuroblastoma and C6 glioma cells)was cultured in tissue culture flasks (75 or 150 cm2). Cellswere grown adherent in tissue culture flasks in a 5% CO2

atmosphere at 37 °C until confluent. The growth mediumconsisted of a mixture of 500 mL of Dulbecco’s modifiedEagle medium (DMEM) supplemented with 4.5 g/L ofglucose and pyruvate, 55 mL of foetal calf serum, 10 mL ofHAT (hypoxanthine, aminopterin, thymidine) supplement(Gibco article No. 15140-114) and 5.5 mL of penicillin/streptomycin (Gibco article no. 21060-017). Medium waschanged nearly every day depending on cell density asdetermined by visual inspection.

6.3.2. Preparation of fura2-loaded cells for measure-ments

NG108–15 cells were examined under a microscopeand cells were used when confluent. The medium wasremoved from the flask, growth medium (10 mL at 37 °C)was added and the cells were dispersed into a single cell

822

suspension by gentle, brief, repeated pipetting. The cellswere centrifuged and the pellet was resuspended inKrebs-Ringer-Hepes (KRH) buffer (conc. in mM: NaCl125; KCl 5.0; MgSO4 1.2; KH2PO4 1.2; CaCl2 2.0;glucose 6.0 Hepes 25; pH adjusted to 7.4, with NaOH) ina concentration of 106 cells per mL. The cells wereincubated with 2µM of fura-2/AM for 45 min at 37 °C inthe dark, spun down (at 1 000g) and washed twice withthe same amount of KRH buffer as used before.

6.3.3. Measurement of [Ca2+] i using NG108-15 cells insuspension

[Ca2+]i was determined as described by Lin et al. [30]measuring fluorescence of fura-2 loaded cells (2 mL ofcell suspension containing 106 cells per experiment,37 °C) using a Hitachi Model F-2000 dual wavelengthfluorescence spectrophotometer. Test compound (10µL)was added. The additon of P2Y2R agonists resulted in atransient increase in intracellular calcium concentration.After baseline was reached again, 10µL of Triton X 100(10% aqueous solution) was added, followed by theaddition of 10µL of a saturated solution of EGTA and10 µL of a saturated solution of Tris base. [Ca2+]i wascalculated as described [31].

6.3.4. CF/T43 and BEA cell culturesAirway epithelial cell lines CF/T43 and BEA (kindly

provided by Dr J.R. Yankaskas), which express an en-dogenous P2Y2R [4] were routinely grown in 75 cm2

tissue culture flasks (Costar) using keratinocyte growthmedium (KGM; Clonetics, catalog No. CC-3115) at37 °C in a humidified 4% CO2 atmosphere. Medium waschanged 2–3 times a week depending on cell density, andcells were passaged every 7–10 d. For passaging, cellswere washed twice with phosphate-buffered saline (PBS)and then treated with 0.1% trypsin/EDTA for 5–10 min todetach the cells. Soybean trypsin inhibitor (three-foldexcess) was used to neutralize trypsin prior to passaging.

6.3.5. Measurement of intracellular calcium in CF/T43and BEA cell monolayers

Cell culture dishes (20 cm2) with a confluent mono-layer of CF/T43 or BEA cells were washed twice with2 mL of Krebs-Ringer-Hepes (see above). After theaddition of fura-2/AM (2µM), cells were incubated at37 °C in the dark for 1 h. The medium was changed to4 mL of fresh KRH buffer and cells were washed 3 timesbefore being incubated for another 15 min to removenon-hydrolysed dye. Immediately before measurement,the experimental incubation medium was replaced again(2 mL) and the culture dishes were placed in the spec-trofluorometer (photon counter; Curt Lindmark Innova-tion AB, Sweden). After base line stabilization, the test

compound was rapidly added with careful shaking, cul-ture dishes were re-inserted within 10 s, and changes influorescence were registered until base line levels werereached again (3–10 min). Then, Ca2+ -ionophore iono-mycin (10µM) and subsequently MnCl2 (20 mM) wereadded in order to determine maximum and minimumvalues of fura-2 fluorescence. Calculations of [Ca2+]iwere made according to Grynkiewicz et al. [31].

Acknowledgements

We thank Gaby Huhurez for skillful technical assis-tance. C.E. Müller is grateful for support given by theFonds der Chemischen Industrie and the Industrie- undHandelskammer Würzburg-Schweinfurt. This projectwas supported by grants from the Deutscher Akademis-cher Austauschdienst (DAAD) to B.K. and C.E.M. andfrom the Svenska Institutet to E.H.

References

[1] Fredholm B., Abbracchio M.P., Burnstock G. et al., Pharm. Rev. 46(1994) 143–156.

[2] Bhagwat S.S., Williams M., Eur. J. Med. Chem. 32 (1997) 183–193.

[3] Abbracchio M.P., Burnstock G., Pharmac. Ther. 64 (1994) 445–475.

[4] Parr C.E., Sullivan D.M., Paradiso A.M., Lazarowski E.R., BurchL., Olsen J.C., Erb L., Weisman G.A., Boucher R.C., Turner J.T.,Proc. Natl. Acad Sci. USA 91 (1994) 3275–3279.

[5] Heilbronn E., Knoblauch B.H.A., Müller C.E., Neurochem. Res. 22(1997) 1041–1050.

[6] Shaver S.R., Pendergast W., Siddiqi S.H., Yerxa B.R., Croom D.K.,Dougherty R.W., James M.D., Jones A.N., Rideout J.L., NucleosidesNucleotides 16 (1997) 1099–1102.

[7] Müller C.E., Stein B., Curr. Pharm. Design. 2 (1996) 501–530.

[8] Järlebark L., Erlandsson M., Uri A., King B.F., Ziganshin A.U.,Johansson C., Heilbronn E., Biochem. Biophys. Res. Comm. 229(1996) 363–369 .

[9] Szemõ A., Szabocs A., Sági J., Ötvös L.J., Carbohydr. NucleosidesNucleotides 7 (1980) 365–379.

[10] Zimmet J., Järlebark L., Hammarberg T. et al., Nucleosides. Nucle-otides 12 (1993) 1–20.

[11] Ludwig L., Acta. Biochim. Biophys. Acad. Sci. Hung. 16 (1981)131–133.

[12] Mishra N.C., Broom A.D., J. Chem. Soc. Chem. Commun. (1991)1276–1277.

[13] Bennua-Skalmowski B., Krolikiewicz K., Vorbrüggen H.,Tetrahedron. Lett. 36 (1995) 7845–7848.

[14] Hunt D.J., Subramanian E., Acta. Crystallogr. B25 (1969)2144–2152.

[15] Ikemoto T., Haze A., Hatano H., Kitamoto Y., Ishida M., Nara K.,Chem. Pharm. Bull. 43 (1995) 210–215.

[16] Cozzone P.J., Jardetzky O., Biochemistry 15 (1976) 4853–4859.

[17] Conigrave A.D., Jiang L., Cell. Calcium 17 (1995) 111–119.

[18] Harden T.K., Lazarowski E.R., Boucher R.C., Trends. Pharmacol.Sci. 18 (1997) 43–46.

823

[19] Nicholas R.A., Watt W.C., Lazarowski E.R., Qing L., Harden T.K.,Mol. Pharmacol. 50 (1996) 224–229.

[20] Erb L., Lustig K.D., Sullivan D.M., Turner J.T., Weisman G.A.,Proc. Natl. Acad. Sci. USA 90 (1993) 10449–10453.

[21] Lazarowski E.R., Watt W.C., Stutts M.J., Stutts M.J., Brown H.A.,Boucher R.C., Harden T.K., Br. J. Pharmacol. 117 (1996) 203–209.

[22] Lazarowski E.R., Watt W.C., Stutts M.J., Boucher R.C., HardenT.K., Br. J. Pharmacol. 116 (1995) 1619–1627.

[23] Pendergast W., Siddiqi S.H., Rideout J.L., James M.K., DoughertyR.W., Drug. Dev. Res. 37 (1996) 133 (abstract).

[24] Wade H.E., Morgan D.M., Biochem. J. 60 (1955) 264–270.

[25] Ebel J.P., Bull. Soc. Franc. (1953) 1089–1095.

[26] Moffat J.G., Can. J. Chem. 42 (1964) 599–604.

[27] Kovacs T., Ötvös L., Tetrahedron. Lett. 29 (1988) 4525–4528.

[28] Sheldrick G.M., Acta. Crystallogr. A46 (1990) 467.

[29] Sheldrick G.M., Program for crystal structure refinement, Universityof Göttingen, 1997.

[30] Lin T.A., Lustig K., Sportiello M.G., Weisman G., Sun G.Y., J.Neurochemistry 60 (1993) 1115–1125.

[31] Grynkiewicz G., Poenie M., Tsien R.Y., J. Biol. Chem. 260 (1985)3440–3450.

824

Original article

Investigation of SAR requirements of SR 142801 through an indexedcombinatorial library in solution

Luca F. Ravegliaa*, Mauro Vitalic, Marco Articoa, Davide Graziania,Douglas W.P. Hayb, Mark A. Luttmannb, Renzo Menaa,

Giorgio Pifferic, Giuseppe A.M. Giardinaa

aDepartment of Medicinal Chemistry, SmithKline Beecham S.p.A., Via Zambeletti, 20021 Baranzate, Milano, ItalybDepartment of Pulmonary Pharmacology, SmithKline Beecham Pharmaceuticals, 709 Swedeland Road, King Of Prussia, PA, 194106,

USAcIstituto di Chimica Farmaceutica, V.le Abruzzi 42, 20133 Milano, Italy

(Received 3 May 1999; accepted 25 May 1999)

Abstract – To rapidly gain information on structure-activity relationship (SAR) requirements of the human neurokinin 3 (hNK-3) receptorantagonist SR 142801, an indexed combinatorial library was synthesised in solution and screened on the hNK-3 receptor. SAR considerationsdrawn from binding affinity of combinatorial mixtures were confirmed through the synthesis and biological evaluation of some individualcompounds. © 1999 E´ditions scientifiques et médicales Elsevier SAS

tachykinins / neurokinin 3 receptor antagonists / combinatorial chemistry / indexed libraries / structure-activity relationships

1. Introduction

The neurokinin 3 (NK-3) is one of the three receptorsfor the family of peptides named tachykinins or neuroki-nins and belongs to the G protein-coupled receptorsuperfamily [1]. In 1996, our group described the identi-fication of potent and selective non-peptide NK-3 recep-tor antagonists featuring the 2-phenyl-4-quinoline-carboxamide framework, including SB 223412 [2, 3];this is one of the three main chemical classes of non-peptide NK-3 receptor antagonists reported to date [4, 5].The other two are the peptide-derived structures reportedby Parke-Davis (PD 161182) [6] and the 3,4-dichloro-phenylpiperidines which were first described by Sanofi(exemplified by SR 142801,figure 1) [7] and subse-quently by Merck Sharp and Dohme [8]. Due to ourinvolvement in the NK-3 area, we were interested ininvestigating the SAR requirements of the 3,4-dichloro-phenylpiperidine NK-3 receptor antagonists, only mar-ginally described in the literature [8]. For this reason, the

synthesis of a small library of analogues of SR 142801was planned via a combinatorial chemistry approach.

Since the 3-(3,4-dichlorophenyl)-3-propylpiperidineframework A(figure 1)of SR 142801 appears to be a keyfeature of this class of NK-3 receptor antagonists, able todrive their selectivity for the NK-3 receptor with respectto the NK-1 and NK-2 [4, 8], we decided to keep this*Correspondence and reprints

Figure 1. Structure of SR 142801.

Eur. J. Med. Chem. 34 (1999) 825−835 825© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

moiety fixed and focus our attention on variations in thebasic head X and the acylating group Y.

In designing the library, 7 amines (X1–X7) and 7acylating groups (Y1–Y7) were selected, on the basis ofcommercial availability of corresponding reagents, togive a total of 49 compounds (X1–7AY1–7), as all thepossible combinations of X and Y substituents(figure 2).To simplify the synthetic pathway, it was planned to workon racemic mixtures.

Solution phase combinatorial chemistry was the tech-nique of choice for the synthesis of the library because asuitable point of attachment to the resin could not befound, ruling out a solid phase approach. Moreover, sincein the synthetic scheme the “combinatorial reaction” wasa well consolidated amide coupling at the last step(figures 4and5), and since the number of compounds tosynthesise was small (49), we reasoned that pools ofcompounds could easily be obtained by using mixtures ofreactants, without risk of getting too many by-products.In particular, to help the structural determination of activecomponents of the library, the strategy of indexed (ororthogonal) combinatorial libraries was utilised [9–11].

2. Chemistry

2.1. Synthesis of intermediate compounds

In figure 3, the synthesis of intermediate compounds isdescribed. Racemic aminol1 was obtained according to

literature procedure [12], protected at the nitrogen andactivated to nucleophilic substitution by converting thehydroxy group into the mesylate and this, in turn, into themore reactive iodo derivative2 [13]. Displacement ofiodine with the 7 aminesX1–X7 (CH3CN, K2CO3, 90 °C,18 h) afforded the desired intermediate compoundsX1A–X7A, which were purified by flash column chroma-tography.

2.2. Synthesis of the library: set 1

As depicted infigure 4, each intermediate compoundwas subsequently deprotected (20% TFA/CH2Cl2, r.t.,18 h) and reacted with a stoichiometric and equimolarmixture of the 7 acylating agentsY1–Y7, (CH2Cl2,K2CO3, r.t., 18 h). Simple washing of the reaction mix-tures with H2O and evaporation to dryness afforded thefirst set (set 1) of desired mixtures (or sublibraries)X1AY1–7–X7AY1–7 in satisfactory to high yields(60–80%) [14a] and high purity (95–98%, LC/UV) [14b].Peak identity was assessed by LC/MS and all expectedmolecular weights were detected in each mixture.

It is worth noting that the seven sublibraries of set 1contain all the desired 49 compounds; each sublibrary hasthe X substitution fixed and different from one another,while containing all possible Y substituents. Since the Yportion is common to all the sublibraries, binding affinitydata on this set will clarify the SAR related to the Xgroup.

Figure 2. Library design.

826

2.3. Synthesis of the library: set 2

In figure 5, the synthesis of the second set (set 2) ofsublibraries is described. An equimolar amount of each ofthe intermediate compoundsX1A–X7A (figure 3) wasmixed into the same reaction vessel and Boc-deprotected;the mixture was then split into seven equal portions(X1–7A), and each portion was reacted with a stoichio-metric amount of acyl chloride/phenylisocyanate/phenylsulfonyl chlorideY1–Y7 (CH2Cl2, K2CO3, r.t.,18 h). Again, simple washing of the reaction mixtures

with H2O and evaporation to dryness produced thedesired sublibrariesX1–7AY1–X1–7AY7 in high yields(74–95%) [14a] and high purity (90–98%, LC/UV, exceptfor X1–7AY4: 83%) [14b]. All expected molecularweights were detected in each mixture by LC/MS.

The seven sublibraries of set 2 contain the same 49compounds of set 1, but in a different combination: in thiscase, the Y substituent is fixed and peculiar to anysublibrary, while all possible X substitutions are commonto all. Binding affinity data on this second set will clarify

Figure 3. Synthesis of intermediates.

827

the SAR related to the Y group. By combining informa-tion coming from the two sets of sublibraries, the best X(set 1) and the best Y (set 2) substituents should beidentified, thus allowing structural determination of themost active compounds of the whole library.

2.4. Synthesis of individual compounds

To prove the validity of the technique and the predic-tions made from binding affinity data of mixtures, the sixindividual compounds reported intable II were synthe-sised, deprotecting the appropriate intermediate com-

pound (20% TFA/CH2Cl2, r.t., 18 h) and reacting it withthe appropriate acylating agent (CH2Cl2, K2CO3, r.t.,18 h).

3. Pharmacology

Receptor binding assays were performed with crudemembranes from CHO cells expressing the hNK-3 recep-tor as detailed previously [3, 15]. For NK-3 receptorcompetition binding studies, [125I]-[MePhe7]-NKB bind-ing to hNK-3-CHO membranes was performed using the

Figure 4. Synthesis of the library: set 1.

828

procedure of Sadowski and co-workers [16]. Specificbinding was determined by subtracting total binding fromnon-specific binding, which was assessed as the bindingin the presence of 0.5µM cold [MePhe7]-NKB. Percentinhibition of specific binding was determined for eachconcentration of the compounds and the IC50, defined asthe concentration required to inhibit 50% of the specificbinding, obtained from concentration-response curves.Values reported intables I and II are the apparentinhibition constant (Ki), which was calculated from theIC50 as described by Cheng and Prusoff [17].

4. Results and Discussion

Human NK-3 receptor binding affinity data for the 14sublibraries synthesised are reported intable I. Results ofsublibraries of set 1 revealed that pyrrolidinocarbonyl-methylpiperazine (X7, Ki = 113 nM) should be slightlymore active than phenylpiperidine (X1, Ki = 160 nM)which, in turn, should be more active than phenylpipera-zine (X2, Ki = 232 nM). A strongly reduced affinity isforseen for isopropylpiperazine (X3, Ki = 743 nM) andmorpholine (X4, Ki = 1235 nM).

Figure 5. Synthesis of the library: set 2.

829

As far as the Y substitution is concerned (set 2),benzoyl (Y2, Ki = 242 nM), also present in SR 142801,appears to be the best substituent. Increasing the distanceof the phenyl ring from the carbonyl with a methylenespacer (Y3, Ki = 394 nM) slightly decreases the bindingaffinity, while substitution at the phenyl ring with twomethoxy groups (Y5, Ki = 1 835 nM) or replacement ofthe phenyl with a methyl (Y1, Ki = 9 163 nM) resulted ina marked loss of activity. Dimethylacryloyl (Y4, Ki = 503nM) and phenylsulfonyl (Y7, Ki = 570 nM), although lesspotent than the benzoyl (Y2, Ki = 242 nM), maintain acertain activity, while the urea (Y6, Ki = 1497 nM) does

not appear to be a good replacement for the benzamide.To test the reliability of the predictions made frommixtures, six individual compounds (some predicted tobe the most active and some predicted to have mediumactivity) were synthesised and screened(table II). Therank order of potency of individual compounds wasfound to be in good agreement with the prediction,X7AY2 being the most active (hNK-3 binding affinity, Ki= 35.4± 10.4 nM) andX1AY5 the least active (hNK-3binding affinity, Ki = 513± 99.5 nM) amongst individualcompounds prepared. It is worth noting that the bestindividual compound synthesised,X7AY2, appears to

Table I. Binding affinities of library compounds.

Set 1 Set 2

X Y hNK-3 Ki(nM) a,b

X Y hNK-3 Ki(nM)a,b

Y1-7 160 X1-7 Y1 CH3CO 9163

Y1-7 232 X1-7 Y2 PhCO 242

Y1-7 743 X1-7 Y3 PhCH2CO 394

Y1-7 1235 X1-7 Y4 (CH3)2C=CHCO 503

Y1-7 525 X1-7 Y5 2,4-(MeO)2PhCO 1835

Y1-7 440 X1-7 Y6 PhNHCO 1497

Y1-7 113 X1-7 Y7 PhSO2 570

aInhibition of [125I]MePhe7-NKB binding in hNK-3-CHO cell membranes.bsingle determination (n = 1).

830

have an hNK-3 binding affinity about 30 times lowerthan SR 142801 (hNK-3 binding affinity, Ki = 1.2±0.3 nM) [18]. SinceX7AY2 differs from SR 142801 onlyfor the basic head X, this SAR study outlines theimportance of the X substitution in conferring highbinding affinity to the 3,4-dichlorophenylpiperidine NK-3receptor antagonists. Additional work is needed to under-stand the requirements of this portion of the molecule andX7 could be a good starting point for further chemicalmodifications in order to obtain novel 3,4-dichlorophenyl-

piperidine NK-3 receptor antagonists with an improvedbiological profile compared to SR 142801 [19].

5. Conclusions

In conclusion, an indexed combinatorial library ofanalogues of SR 142801 was synthesised in solution andscreened for its hNK-3 receptor binding affinity. Synthe-sis and biological evaluation of some individual com-pounds proved the reliability of this very simple, acces-sible and rather under-utilised technique and allowed usto gain information on SAR requirements of 3,4-dichlorophenylpiperidine NK-3 receptor antagonists, sig-nificantly decreasing the number of reactions and samplesto test.

6. Experimental protocols

6.1. Chemistry

Melting points were determined with a Büchi 530 hotstage apparatus and are uncorrected. Proton NMR spectrawere recorded on a Bruker ARX 300 spectrometer at303 K unless otherwise indicated. Chemical shifts wererecorded in parts per million (δ) downfield from tetra-methylsilane (TMS); NMR spectral data are reported as alist. IR spectra were recorded in Nujol mull or neat onsodium chloride disks or in KBr with a Perkin-Elmer1420 spectrophotometer; mass spectra were obtained on aFinnigan MAT TSQ-700 spectrometer. Silica gel used forflash column chromatography was Kiesegel 60 (230–400mesh) (E. Merck AG, Darmstadt, Germany). All evapo-rations were performed at reduced pressure. Combustionelemental analyses were performed by Redox s.n.c.,Milan, Italy and analyses indicated by the symbols of theelements were within± 0.4% of the theoretical values. Allreagents utilised infigures 3–5 are commercially avail-able compounds and were used without further purifica-tion. Racemic aminol1 and SR 142801 were synthesisedaccording to Giardina et al. [12] and to Chen et al. [13].

6.2. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-iodopropyl)piperidine2

Racemic aminol1 [12] (4.7 g, 16.3 mmol) and di-tert-butyl dicarbonate (4.3 g, 19.6 mmol) were dissolved,under nitrogen atmosphere, in dry CH2Cl2 (75 mL). Thereaction mixture was stirred at room temperature for 8 hand left standing overnight. The solvent was evaporatedto dryness and the crude material was purified by flashcolumn chromatography, eluting with a mixture of

Table II. Binding affinities of individual compounds.

Individual compounds

X Y hNK-3 Ki(nM)a,b

PhCO 46.2± 9.2

2,4-(MeO)2PhCO

513 ± 99.5

PhCO 50.9± 9.2

PhCO 35.4± 10.4

PhCH2CO 289± 68.0

PhSO2 175 ± 50.2

PhCO SR 142801c,d 1.2 ± 0.3

aInhibition of [125I]MePhe7-NKB binding in hNK-3-CHO cellmembranes.bmean± SEM for three determinations (n = 3).cinhouse data.dSR 142801 is a single enantiomer [18].

831

CH2Cl2/MeOH 98:2, to obtain 5.4 g (85%) of N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-hydroxy-propyl)piperidine as a pale yellow oil. This compound(5.4 g, 13.9 mmol) was dissolved in CH2Cl2 (50 mL);triethylamine (TEA) (2.2 mL, 15.8 mmol) was added andthe solution was cooled to 0 °C. Methanesulfonyl chlo-ride (1.2 mL, 15.4 mmol), dissolved in CH2Cl2 (7 mL),was added dropwise and the reaction mixture was al-lowed to reach room temperature and stirred for addi-tional 2 h. The reaction was quenched with water and theextracted organic layer was washed with 10% citric acid,5% NaHCO3 and brine, dried over MgSO4, filtered andevaporated to dryness to yield 6.0 g (93%) of N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-methane-sulfonyloxypropyl)piperidine as a pale yellow oil. Thiscompound (6.0 g, 12.9 mmol) was dissolved in acetone(120 mL) and KI (3.5 g, 20.9 mmol) was added to thesolution which was refluxed for 16 h. Insoluble materialwas filtered off and the filtrate was evaporated to drynessto yield 6.4 g (100%) of the title compound as a red oil,which was used in the following reactions without furtherpurification. IR (neat) 2 940, 2 885, 1 692, 1 470,1 428 cm–1. 1H-NMR (CDCl3) δ 7.44 (d, 1H), 7.38 (d,1H), 7.17 (dd, 1H), 3.91 (d, 1H), 3.58–3.47 (m, 1H), 3.31(d, 1H), 3.31–3.20 (m, 1H), 3.01 (t, 2H), 2.07–1.98 (m,1H), 1.80–1.40 (m, 7H), 1.49 (s, 9H).

6.3. General procedure for the synthesis of intermediatecompoundsX1A–X7A

K2CO3 (0.303 g, 2.2 mmol) and the appropriate amine(1.1 eq.) were added to a solution of N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-iodopropyl)piperidine2 (0.99 g, 1.98 mmol) in CH3CN (7 mL). Thereaction mixture was heated to 80 °C under magneticstirring for 18 h; then it was filtered and the filtrate wasevaporated to dryness. The crude solid obtained wasdissolved in EtOAc and washed with H2O. The organiclayer was dried over MgSO4, filtered and evaporated todryness to yield the title compounds which were purifiedby flash column chromatography. Yields after chromatog-raphy were in the range: 75–96%. Spectroscopic data forcompoundsX1A–X7A are reported below.

6.3.1.N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidineX1A

IR (Nujol) 2 926, 1 690, 1 674, 1 604, 1 554 cm–1.1H-NMR (CDCl3–343 K) δ 7.47 (d, 1H), 7.37 (d, 1H),7.30–7.14 (m, 6H), 4.00 (d, 1H), 3.56 (dt, 1H), 3.26 (d,1H), 3.26–3.18 (m, 1H), 2.90–2.83 (m, 2H), 2.50–2.40(m, 1H), 2.21 (t, 2H), 2.09–1.91 (m, 3H), 1.80–1.13 (m,

11H), 1.49 (s, 9H). ESI-MS (positive, solvent: methanol,spray 4.5 keV, skimmer: 60 eV, capillary 220 °C)m/z531(MH+).

6.3.2. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperazin-1-yl)propyl]piperidineX2A

IR (KBr) 2 942, 2 816, 1 692, 1 602, 1 554,1 502 cm–1. 1H-NMR (CDCl3) δ 7.44 (d, 1H), 7.38 (d,1H), 7.24 (dd, 2H), 7.18 (d br, 1H), 6.90 (d, 2H), 6.84 (dd,1H), 4.05 (m br, 1H), 3.60 (m, br, 1H), 3.20–3.10 (m,6H), 2.43 (m, 4H), 2.21 (t, 2H), 2.05 (m br, 1H),1.75–1.10 (m, 7H), 1.46 (s, 9H). ESI-MS (positive,solvent: methanol, spray 4.5 keV, skimmer: 60 eV,capillary 220 °C)m/z532 (MH+), 554 (MNa+).

6.3.3. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-isopropylpiperazin-1-yl)propyl]piperidineX3A

IR (neat) 2 942, 2 812, 1 690, 1 554, 1 470 cm–1.1H-NMR (CDCl3) δ 7.42 (d, 1H), 7.31 (d, 1H), 7.15 (d br,1H), 4.05 (m br, 1H), 3.60 (m br, 1H), 3.13 (d, 1H), 3.12(m, 1H), 2.64 (m, 1H), 2.50 (m, 4H), 2.35 (m, 4H), 2.20(t, 2H), 2.05 (m, 1H), 1.70–1.00 (m, 7H), 1.45 (s, 9H),1.00 (d, 6H). ESI-MS (positive, solvent: methanol, spray4.5 keV, skimmer: 60 eV, capillary 220 °C)m/z 498(MH+).

6.3.4. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(morpholin-4-yl)propyl]piperidineX4A

IR (neat) 2 944, 2 858, 2 810, 1 690, 1 590,1 554 cm–1. 1H-NMR (CDCl3) δ 7.44 (d, 1H), 7.35 (d,1H), 7.14 (d br, 1H), 4.02 (m br, 1H), 3.63 (m br, 4H),3.59 (m br, 1H), 3.13 (d, 1H), 3.12 (m, 1H), 2.30 (m, 4H),2.18 (t, 2H), 2.05 (m br, 1H), 1.71–1.00 (m, 7H), 1.46 (s,9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV,skimmer: 60 eV, capillary 220 °C)m/z 457 (MH+), 479(MNa+).

6.3.5. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(1,2,3,4-tetrahydroisoquinolin-2-yl)propyl]piperidineX5A

IR (neat) 2 932, 1 692, 1 554 cm–1. 1H-NMR(CDCl3–343 K) δ 7.48 (d, 1H), 7.36 (d, 1H), 7.20 (dd,1H), 7.10–7.03 (m, 3H), 6.94 (m, 1H), 4.01 (d, 1H), 3.58(dt, 1H), 3.47 (s, 2H), 3.25 (d, 1H), 3.20 (ddd, 1H), 2.83(t, 2H), 2.59 (t, 2H), 2.34 (t, 2H), 2.09–2.00 (m, 1H),1.78–1.13 (m, 7H), 1.49 (s, 9H). ESI-MS (positive,solvent: methanol, spray 4.5 keV, skimmer: 60 eV,capillary 220 °C)m/z503 (MH+).

832

6.3.6. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(1-phenyl-1,3,8-triazaspiro[4,5]decan-4-on-8-yl)propyl]piperidineX6A

IR (neat) 2 938, 2 856, 1 708, 1 690, 1 602,1 502 cm–1. 1H-NMR (CDCl3–343 K) δ 7.46 (d, 1H),7.36 (d, 1H), 7.29 (dd, 2H), 7.19 (dd, 1H), 6.99–6.89 (m,3H), 5.61 (s br, 1H), 4.70 (s, 2H), 3.90 (d br, 1H), 3.56(m, 1H), 3.29–3.19 (m, 2H), 2.75–2.42 (m, 6H), 2.23 (m,2H), 2.05 (m, 1H), 1.78–1.05 (m, 9H), 1.47 (s, 9H).ESI-MS (positive, solvent: methanol, spray 4.5 keV,skimmer: 60 eV, capillary 220 °C)m/z601 (MH+), 545.

6.3.7. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidineX7A

IR (neat) 2 944, 2 874, 2 812, 1 690, 1 642,1 554 cm–1. 1H-NMR (CDCl3–343 K) δ 7.44 (d, 1H),7.36 (d, 1H), 7.18 (dd, 1H), 3.90 (d, 1H), 3.55 (dt, 1H),3.54–3.43 (m, 4H), 3.23 (d, 1H), 3.20 (ddd, 1H), 3.09 (s,2H), 2.55 (m, 4H), 2.36 (m, 4H), 2.19 (t, 2H), 2.08–1.99(m, 1H), 1.98–1.80 (m, 4H), 1.78–1.00 (m, 7H), 1.49 (s,9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV,skimmer: 60 eV, capillary 220 °C)m/z567 (MH+).

6.4. General procedure for the synthesis of library (set 1)compoundsX1AY1-7–X7AY1–7

6.4.1 Synthesis of sublibraryX1AY1–7The synthesis of sublibraryX1AY1–7 is reported as an

example of the synthesis of all sublibraries of the first setof the library. N-tert-butoxycarbonyl-3-(3,4-dichloro-phenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidineX1A (0.301 g, 0.55 mmol) was dissolved in dry CH2Cl2(8 mL); TFA (2 mL) was added and the reaction mixturewas stirred at room temperature overnight. After evapo-ration to dryness, the residue was taken up with CH2Cl2,washed with 5% NaHCO3, dried over MgSO4, filteredand evaporated to dryness, to obtain 3-(3,4-dichloro-phenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine.This compound was dissolved in dry CH2Cl2 (10 mL),and K2CO3 (165.3 mg, 1.2 mmol) was added; then asolution of acetyl chloride (564µL, 7.9 mmol), benzoylchloride (918µL, 7.9 mmol), phenylacetylchloride(1 047µL, 7.9 mmol), dimethylacryloyl chloride(880µL, 7.9 mmol), 2,4-dimethoxybenzoyl chloride(1 585 mg, 7.9 mmol), phenylisocianate (859µL,7.9 mmol) and phenylsulfonyl chloride (1 013µL,7.9 mmol), brought to 100 mL with dry CH2Cl2, wasprepared and 1 mL of this solution was added to thereaction mixture, which was stirred at room temperatureovernight. The reaction was quenched with H2O (5 mL)and the extracted organic layer was washed with H2O,dried over MgSO4, filtered and evaporated to dryness to

obtain 0.233 g (78%) of the title sublibrary. All sevencompounds forming the sublibrary were characterised byLC/MS.

Purity of the seven sublibraries of set 1, calculated asthe sum of the areas of the seven peaks of the LC/UVchromatogram attributed by LC/MS to the seven com-pounds forming each sublibrary, is reported below:

X1AY1–7 98%; X2AY1–7 98%; X3AY1–7 95%;X4AY1–7 98%; X5AY1–7 97%; X6AY1–7 98%; X7AY1–797%.

6.5. General procedure for the synthesis of library (set 2)compoundsX1–7AY1–X1–7AY7

6.5.1. Synthesis of sublibraryX1–7AY1

The synthesis of sublibraryX1–7AY1 is reported as anexample of the synthesis of all sublibraries of the secondset of the library. N-tert-butoxycarbonyl-3-(3,4-di-chlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine X1A (116 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenyl-piperazin-1-yl)propyl]piperidine X2A (116 mg,0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichloro-phenyl)-3-[3-(4-isopropylpiperazin-1-yl)propyl]piperidineX3A (108 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(morpholin-4-yl)propyl]piperidine X4A (100 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(1,2,3,4-tetrahydroisoquinolin-2-yl)propyl]piperidine X5A(110 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(1-phenyl-1,3,8-triazaspiro[4,5]decan-4-on-8-yl)propyl]piperidine X6A (131 mg,0.217 mmol) and N-tert-butoxycarbonyl-3-(3,4-dichloro-phenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidine X7A (124 mg, 0.217 mmol) weredissolved, under magnetic stirring, in CH2Cl2 (36 mL);TFA (9 mL) was added and the reaction mixture wasstirred at room temperature overnight. After evaporationto dryness, the crude material was taken up with CH2Cl2and washed with 5% NaHCO3, dried over MgSO4,filtered and evaporated to dryness to yield 0.826 g of amixture of Boc-deprotectedX1A–X7A. This mixture wassplit into seven equal portions (one for each sublibrary)and K2CO3 (65 mg, 0.47 mmol) and acetyl chloride(15.49µL, 0.217 mmol) were added to one of theseportions, dissolved in CH2Cl2 (15 mL). After stirringovernight, the reaction was quenched with H2O (5 mL),the organic layer was separated, washed with H2O, driedover MgSO4, filtered and evaporated to dryness to yield75 mg (74%) of the title sublibrary. All seven compoundsforming the sublibrary were characterised by LC/MS.

833

Purity of the seven sublibraries of set 2, calculated asthe sum of the areas of the seven peaks of the LC/UVchromatogram attributed by LC/MS to the seven com-pounds forming each sublibrary, is reported below:

X1–7AY1 98%; X1–7AY2 94%; X1–7AY3 98%;X1–7AY4 83%; X1–7AY5 98%; X1–7AY6 94%; X1–7AY798%.

6.6. Synthesis of individual compounds reported intable II

6.6.1. Synthesis of N-benzoyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine hydrochlorideX1AY2

The synthesis of this compound is reported as a generalprocedure for the synthesis of all six individual com-pounds prepared. N-tert-Butoxycarbonyl-3-(3,4-di-chlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine X1A (0.23 g, 0.43 mmol) was dissolved inCH2Cl2 (8 mL); TFA (2.2 mL), dissolved in CH2Cl2(2 mL), was added, and the reaction was stirred at roomtemperature overnight. After evaporation to dryness, thecrude material was taken up with CH2Cl2 and washedwith 5% NaHCO3, dried over MgSO4, filtered andevaporated to dryness to yield crude 3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine. Half of this crude material (≈ 0.21 mmol) wasdissolved in CH2Cl2 (5 mL), K2CO3 (80 mg, 0.58 mmol)and benzoyl chloride (24.4µL, 0.21 mmol) were addedand the reaction was stirred at room temperature over-night and then quenched with H2O (2 mL). The organiclayer was separated, washed with H2O, dried overMgSO4, filtered and evaporated to dryness to yield140 mg (61%) of N-benzoyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine as a yellowoil. This compound was transformed into its hydrochlo-ride by dissolving in MeOH and treating with HCl/Et2O.Evaporation to dryness and trituration with Et2O affordedthe title compound as a solid. IR (KBr) 3 425, 2 937,2 532, 1 622, 1 444 cm–1. 1H-NMR (CDCl3–333 K) δ7.40–7.10 (m, 13H), 4.25 (m, 2H), 3.58–3.30 (m, 4H),2.90–2.81 (m, 2H), 2.45 (m, 1H), 2.22 (t, 2H), 2.20–2.09(m, 1H), 2.00–1.55 (m, 9H), 1.35–1.10 (m, 2H). ESI-MS(positive, solvent: methanol, spray 4.5 keV, skimmer: 60eV, capillary 220 °C) m/z 535 (MH+). Anal.C32H37Cl3N2O (C, H, N, Cl).

6.6.2. N-(2,4-dimethoxy)benzoyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine hydro-chlorideX1AY5

IR (KBr) 3 447, 2 939, 2 668, 1 607, 1 467 cm–1.1H-NMR (CDCl3–333 K) δ 7.50–7.20 (m, 9H),6.50–6.40 (m, 2H), 4.25 (m, 2H), 3.80 (s, 6H), 3.70–3.10

(m, 4H), 2.90–2.81 (m, 2H), 2.45 (m, 1H), 2.20 (m, 2H),2.15–2.00 (m, 1H), 2.00–1.60 (m, 9H), 1.30–1.10 (m,2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV,skimmer: 60 eV, capillary 220 °C)m/z595 (MH+). Anal.C34H41Cl3N2O3 (C, H, N, Cl).

6.6.3. N-benzoyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenyl-piperazin-1-yl)propyl]piperidine dihydrochlorideX2AY2

IR (KBr) 3 447, 2 947, 2 403, 1 614, 1 440 cm–1.1H-NMR (CDCl3–333 K)δ 7.42–7.20 (m, 10H), 6.90 (d,2H), 6.80 (dd, 1H), 4.25 (m, 1H), 3.53 (d, 1H), 3.53–3.30(m, 2H), 3.15 (m, 4H), 2.45 (m, 4H), 2.24 (t, 2H),2.18–2.08 (m, 1H), 1.88 (ddd, 1H), 1.79–1.42 (m, 4H),1.40–1.10 (m, 2H). ESI-MS (positive, solvent: methanol,spray 4.5 keV, skimmer: 60 eV, capillary 220 °C)m/z536(MH+). Anal. C31H37Cl4N3O (C, H, N, Cl).

6.6.4. N-benzoyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrol-idinocarbonylmethyl)piperazin-1-yl]propyl}piperidinedihydrochlorideX7AY2

IR (KBr) 3 436, 2 954, 2 437, 1 662, 1 618,1 438 cm–1. 1H-NMR (CDCl3–333K) δ 7.40–7.10 (m,8H), 4.40 (m, 1H), 3.52–3.30 (m, 7H), 3.09 (s, 2H), 2.51(m, 4H), 2.31 (m, 4H), 2.18 (t, 2H), 2.15–2.05 (m, 1H),1.95–1.80 (m, 5H), 1.70–1.40 (m, 4H), 1.35–1.05 (m,2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV,skimmer: 60 eV, capillary 220 °C)m/z571 (MH+). Anal.C31H42Cl4N4O2 (C, H, N, Cl).

6.6.5. N-phenylacetyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidinedihydrochlorideX7AY3

IR (KBr) 3 435, 2 954, 2 546, 1 653, 1 630,1 451 cm–1. 1H-NMR (CDCl3–333 K) δ 7.40–7.05 (m,8H), 4.40 (m, 1H), 3.70 (s, 2H), 3.49 (m, 5H), 3.23 (m,2H), 3.09 (s, 2H), 2.52 (m, 4H), 2.31 (m, 4H), 2.11 (t,2H), 2.10–2.00 (m, 1H), 1.98–1.75 (m, 4H), 1.75–1.00(m, 7H). ESI-MS (positive, solvent: methanol, spray 4.5keV, skimmer: 60 eV, capillary 220 °C)m/z585 (MH+).Anal. C32H44Cl4N4O2 (C, H, N, Cl).

6.6.6. N-benzenesulfonyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidine dihydrochlorideX7AY7

IR (KBr) 3 427, 2 943, 2 582, 1 658, 1 469, 1 340,1 162 cm–1. 1H-NMR (CDCl3–333 K) δ 7.78 (m, 2H),7.55 (m, 3H), 7.40 (m, 2H), 7.21 (m, 1H), 3.47 (m, 5H),3.10 (m, 1H), 3.10 (s, 2H), 2.79 (m, 2H), 2.51 (m, 4H),2.32 (m, 4H), 2.17 (t, 2H), 2.00–1.50 (m, 10H), 1.30–1.10(m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5keV, skimmer: 60 eV, capillary 220 °C)m/z607 (MH+).Anal. C30H42Cl4N4O3S (C, H, N, Cl).

834

Acknowledgements

The authors wish to thank Dr Alberto Cerri for NMRspectroscopic determinations.

References

[1] Pritchard M.C., Boden P., Drugs Fut. 20 (1995) 1163–1173.

[2] Giardina G.A.M., Sarau H.M., Farina C., Medhurst A.D., GrugniM., Foley J.J., Raveglia L.F. et al., J. Med. Chem. 39 (1996)2281–2284.

[3] Giardina G.A.M., Sarau H.M., Farina C., Medhurst A.D., GrugniM., Raveglia L.F. et al., J. Med. Chem. 40 (1997) 1794–1807.

[4] Giardina G.A.M., Raveglia L.F., Exp. Opin. Ther. Patents 7 (1997)307–323.

[5] von Sprecher A., Gerspacher M., Anderson G., Idrugs 1 (1998)73–91.

[6] Boden P., Eden J.M., Hodgson J., Horwell D.C., Pritchard M.C.,Raphy J., Suman–Chauhan N., Bioorg. Med. Chem. Lett. 5 (1995)1773–1778.

[7] Edmonds-Alt X., Bichon D., Ducoux J.P., Heaulme M., Miloux B.,Poncelet M. et al., Life Sci. 56 (1995) PL27–PL32.

[8] Harrison T., Korsgaard M.P.G., Swain C.J., Cascieri M.A., SadowskiS., Seabrook G.R., Bioorg. Med. Chem. Lett. 8 (1998) 1343–1348.

[9] Smith P.W., Lai J.Y.Q., Whittington A.R., Cox B., Houston J.G.,Stylli C.H., Banks M.N., Tiller P., Bioorg. Med. Chem. Lett. 4(1994) 2821–2824.

[10] Pirrung M.C., Chen J., J. Am. Chem. Soc. 117 (1995) 1240–1245.

[11] Deprez B., Williard X., Bourel L., Coste H., Hyafil F., Tartar A., J.Am. Chem. Soc. 117 (1995) 5405–5406.

[12] Giardina G.A.M., Grugni M., Rigolio R., Vassallo M., Erhard K.,Farina C., Bioorg. Med. Chem. Lett. 6 (1996) 2307–2310.

[13] Chen H.G., Chung F.Z., Goel O.P., Johnson D., Kesten S., Knobels-dorf J., Lee H.T., Rubin J.R., Bioorg. Med. Chem. Lett. 7 (1997)555–560.

[14] (a) Calculated utilising the average molecular weight. (b) Calculatedas the sum of the areas of the seven peaks of the LC/UVchromatogram, attributed by LC/MS to the seven compoundsforming each mixture.

[15] Sarau H.M., Griswold D.E., Potts W., Foley J.J., Schmidt D.B.,Webb E.F. et al., J. Pharmacol. Exp. Ther. 281 (1997) 1303–1311.

[16] Sadowski S., Huang R.R.C., Fong T.M., Marko O., Cascieri M.A.,Neuropeptides 24 (1993) 317–319.

[17] Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973)3099–3108.

[18] Being SR 142801 a single enantiomer about 2-fold more potent thanthe corresponding racemate (in-house data), the real difference inbinding affinity for the hNK-3 receptor between X7AY2 and (±) SR142801 is about 15-fold.

[19] In house data demonstrate that SR 142801 has high systemic plasmaclearance (36.9± 10.6 mL/min/kg) and low oral bioavailability (11± 2%) in rats; in addition, it showed significant interactions withsome isoenzymes of P450 family (specifically, P2D6 and P3A4).

835

Preliminary communication

Synthesis and antimicrobial activity of heterocyclic ionone-like derivatives

Maria Anzaldia, Enzo Sottofattoria, Rolando Rizzettob,Barbara Granello di Casaletob, Alessandro Balbia*

aDipartimento di Scienze Farmaceutiche, Università degli Studi di Genova, Viale Benedetto XV, 3, 16132 Genova, ItalybIstituto di Igiene e Prevenzione, Università degli Studi di Genova, Viale Pastore, 1, 16132 Genova, Italy

(Received 2 November 1998; revised 2 March 1999; accepted 3 March 1999)

Abstract – A number of heterocyclic ionone-like derivatives5 were prepared with appropriate bifunctional reagents by one-pot cyclisationof 3-dimethylamino-5-(2,6,6-trimethyl-2-cyclohexen-1-yl)-2,4-pentadienal3a, which was, in turn, obtained fromα-ionone with N,N-dimethylformamide/phosphorus oxychloride. All compounds5 possess remarkable activity against the selected Gram-positive, Gram-negative bacterial strains and againstCandida albicans. Derivatives 5a2 and 5a6, acting at a high level especially against bothPropionibacterium acnesandStaphylococcus aureus, could play a potential role in the treatment of acne and related skin disorders. © 1999Editions scientifiques et médicales Elsevier SAS

ionone-like derivatives / short retinoids / antimicrobial activity

1. Introduction

Since their introduction nearly two decades ago [1], theuse of retinoids for topical and systemic treatment ofpsoriasis and other hyperkeratotic and parakeratotic skindisorders has increased. They are also used for thetreatment of severe acne and acne-related dermatoses.Nonetheless, various side effects, the most serious ofwhich being the teratogenicity, coupled with their longelimination half life, have limited their use [2]. Further-more, while the recovery ofPropionibacterium acnesandother anaerobic bacteria in the skin is markedly reduced,there is an increase in the colonisation ofStaphylococcusaureuson the skin and, in general, a significant rise in theincidence of cutaneous staphylococcal infection in pa-tients treated with oral retinoids [3].

While searching for more biologically active and lesstoxic compounds, the structure of retinoic acid 1(fig-ure 1) has been modified both in the cyclohexenyl ringand in the polyene side chain.

Major transformations have led to retinoids whichbarely resemble the original retinoic acid [4, 5]. All thesemodifications, which lead to the three generations of

retinoids known today (non-aromatic, mono-aromaticand poly-aromatic), are often associated with reducedtoxicity while biological activity is maintained or evenenhanced. Among these, Etretinate and Acitretin arecurrently prescribed in systemic use, while Adapalene [4,5] and Tazarotene [4, 6] have recently been introduced onthe market for topical use.

Having a method to synthesise new heterocyclicionone-like derivatives on hand [7] and because of thesimilarity of these compounds to the so-called “shortheteroretinoids” [8–11], we focused our biological atten-tion both on the reduction of theP. acnesand the decreaseof staphylococcal infections. In fact, retinoids would beof interest for topical use if they were found to possess*Correspondence and reprints

Figure 1. The structure of retinoic acid 1.

Eur. J. Med. Chem. 34 (1999) 837−842 837© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

the following two properties: counteracting both theproliferation ofP. acnesand of other bacteria (includingS. aureus) which are responsible for the majority ofinfections observed.

2. Chemistry

Our laboratory recently began to investigate the syn-thesis of new short retinoid-like derivatives [7]. In fact,while studying the Vilsmeier reaction (VR) onα- andâ-ionones, we outlined both the formation of the classicalâ-chlorovinylaldehydes2 and the unexpected products3(figure 2), which is unprecedented in a Vilsmeier reaction.

Surprisingly, the chlorine of compounds2 had diffi-culty in reacting, while the dialkylamino group of3 actedas an outstanding leaving group, allowing for the forma-tion of new heterocyclic retinoids. In fact, all attempts tocyclise compounds2 with bifunctional reagents failed.We were only able to isolate the classical formylderiva-tives as the semicarbazones4aand4b. On the other hand,compound3a reacted rapidly with bifunctional reagentssuch as hydrazine and guanidine. The reaction, whichinvolved both the formyl and the dimethylamino groupsgave a new series of heterocyclic ionone-like derivatives.

To complete both the previously reported study and forbiological purposes, we have re-synthesised compounds5a1, 5a3–a5, thus improving the yield, and have ex-tended the ring closure on key intermediates3aand3b byother bifunctional reagents, therefore obtaining new“short heteroretinoids”(figure 3).All reagents used re-acted quickly, giving the expected compounds in high

yield and without side products, confirming the particularbehaviour of the enamines3.

3. Results and discussion

Table Isummarises the germicidal (GE log values) andkilling percentage effects of compounds5a1–a6, 5b andthe synthon3a against two Gram-positive, two Gram-negative bacterial strains andCandida albicans(seeExperimental for biological methods). Pyrazole deriva-tives 5a1–a3 exhibited a good to excellent activityagainst bothC. albicansandP. acnes. Introduction of aphenyl ring in the position–1 of the pyrazole led toconsiderable enhancement of the antibacterial activityversus S. aureus, while the presence of the 2,4-dinitrophenyl group in the same position decreased theactivity againstPseudomonas aeruginosa, Escherichiacoli and S. aureuscompared to the other two, althoughthe activity againstP. acnesand C. albicansremainedalmost unchanged. The substitution of one nitrogen withoxygen, which led to the isoxazole derivative5a4,decreased almost all the activity against all the microbesbut for P. acnes. Moreover, in this case the activity wasonly second to phenyl-pyrazole derivative5a2. Introduc-tion of a pyrimidine in place of a pyrazole or an isoxazolering did not improve the activity againstC. albicans, P.aeruginosaandE. coli. On the contrary, they showed agood to excellent level of potency againstS. aureusandP. acnes.

From a general point of view it could be stated that thealready notable activity of the key intermediate3a versusC. albicans, S. aureusand P. acnes, is enhanced andspecifically directed towards one of the three abovebacterial strains when a particular heterocyclic moiety isintroduced in synthon3a: namely pyrazole versusC.albicans, phenylpyrazole versusP. acnes, methyltiopyri-midine versusS. aureus. Moreover, lipophilicity, whilenot increasing selectivity, enhances activity towards allthe species. In fact, the more lipophilic5a2 and 5a6present the highest activity towardsP. acnes andS. aureus, respectively, and also possess a good toexcellent activity versus the other species.

4. Conclusion

From a structural-activity point of view, certain fea-tures can be noted. Compounds5 could be classified as“short-heteroretinoids” [8–11], related to the natural onesthrough the presence of the cyclohexenyl ring which is afeature of α- and â-ionones. On the other hand, thepotential therapeutic use of these compounds (in partic-

Figure 2. The formation of the classicalâ-chlorovinyl-aldehydes2 and the unexpected products3.

838

ular 5a2 and 5a6) is closer to the new Adapalene andTazarotene [4–6] than classical retinoids. The latter, infact, may not only causeS. aureusinfections whenadministered orally, but at the same time do not have asignificant effect on aerobic and anaerobic bacteria invitro [3].

Compounds5a2and5a6, acting both onS. aureusandon P. acnes, could be considered as candidates for topicaltreatment of acne and/or a coadjuvant in the oral admin-istration of classical retinoids.

5. Experimental protocols

5.1. Chemistry

Melting points were determined with Fisher-Johnsapparatus and are uncorrected. The IR spectra wererecorded in chloroform or in potassium bromide disks on

a Perkin-Elmer 398 spectrometer. The1H and13C-NMRspectra were recorded on a Bruker AC 300 (300 MHz,1H; 75 MHz,13C) or a Varian Gemini 200 (200 MHz,1H;50 MHz, 13C) spectrometers in deuteriochloroform solu-tions with tetramethylsilane as the internal standard (δ =0). The purity of all compounds was checked by thin-layer chromatography on silica gel 60-F-254 pre-coatedplates and the spots were located in UV light or byvanillin in sulphuric acid. Elemental analyses were per-formed on a Carlo Erba 1106 Elemental Analyser in theMicroanalysis Laboratory in our Institute and the resultswere within± 0.4% of theoretical values. For elementalanalysis and spectral data of5a1, 5a3–a5 see [7].

5.1.1. 5-(2, 6, 6-Trimethyl-2-cyclohexen-1-yl)ethenyl-1H-pyrazole5a1

A solution of 0.25 g (1 mmol) of3a and 1 mL ofhydrazine hydrate (20 mmol) in 10 mL of ethanol was

Figure 3. Synthesis of compounds4a, 4b, 5a and5b.

839

allowed to stand at –10 °C for 30 min and 1 h at roomtemperature. The resulting oil, purified by chromatogra-phy on silica gel (toluene/ethyl acetate 1:1), gave5a1 in88% yield as a thick oil.

5.1.2. 1-Phenyl-5-[(2,6,6-trimethyl-2-cyclohexen-1-yl)ethenyl]-1H-pyrazole5a2

A solution of 0.4 g (1.6 mmol) of3a and phenylhydra-zine hydrochloride [0.5 g (3.2 mmol) in 5 mL of waterand 0.8 g of sodium acetate trihydrate] in 10 mL ofethanol was refluxed for 15 min and allowed to stand atroom temperature for 1 d. The resulting oil was extractedwith chloroform. The combined extracts were dried(sodium sulphate) and evaporated to afford a red oilwhich was purified by chromatography on silica gel(toluene/ethyl acetate 1:1), giving5a2 in 85% yield as athick oil; IR (film): ν 3 100, 2 920, 1 670, 1 600, 1 500,1 450, 1 390 cm–1; 1H-NMR (200 MHz): δ 0.88 (3H, s,CH3), 0.92 (3H, s, CH3), 1.20 (1H, m, H-5′), 1.45 (1H, m,H-5′), 1.60 (3H, s, CH3), 2.03 (2H, m, H-4′), 2.25 (1H, d,H-1′ J 8.98 Hz), 5.45 (1H, m, H-3′), 6.07 (1H, dd, ethene,J 8.98; 15.76 Hz), 6.20 (1H, d, ethene,J 15.76 Hz), 6.47(1H, d, H-3,J 1.98 Hz), 7.47 (5H, m, arom-H) 7.61 (1H,d, H-4, J 1.98 Hz); 13C-NMR (50 MHz, CDCl3) 23.44(CH3), 23.53 (CH2), 27.39 (CH3), 28.23 (CH3), 32.00(CH2), 33.11 (C), 55.35 (CH), 104.37 (CH), 119.44 (CH),122.17 (CH), 125.72 (CH), 125.72 (CH), 128.17 (CH),129.50 (CH), 129.50 (CH), 133.69 (C), 136.62 (CH),140.17 (C), 140.56 (CH), 141.71 (C). Anal. C20H24N2 (C,H, N).

5.1.3. 1-(2,4-Dinitrophenyl)-5-[(2,6,6-trimethyl-2-cyclo-hexen-1-yl)ethenyl]-1H-pyrazole5a3

A solution of 0.25 g (1 mmol) of3a and 2,4-dinitro-phenylhydrazine [0.2 g (1 mmol) in ethanol/H2SO4] in10 mL of ethanol was allowed to stand at –10 °C for 30min and 2 h at room temperature. The precipitated solidwas collected by filtration and washed with ethanol,yielding 5a3 as already pure yellow crystals (90% yield,m.p. 116–117 °C from ethanol).

5.1.4. 5-(2,6,6-Trimethyl-2-cyclohexen-1-yl)ethenyl-iso-xazole5a4

A solution of 0.4 g (1.6 mmol) of3a and hydroxy-lamine hydrochloride [0.5 g (7.0 mmol) in 3 mL of water]in 30 mL of ethanol and 2 mL of a 10% water solution ofsodium hydroxide was allowed to stand at room tempera-ture for 3 d. The resulting oil, purified by chromatographyon silica gel (toluene/ethyl acetate 1:1), gave5a4 in 88%yield as a thick oil.

5.1.5. 6-(2,6,6-Trimethyl-2-cyclohexen-1-yl)ethenyl-2-ami-nopyrimidine5a5

A solution of 0.5 g (2.0 mmol) of3a and guanidinecarbonate [0.36 g (2.0 mmol) in 3 mL of water] in 30 mLof ethanol was refluxed for 48 h. The resulting oil wasextracted with chloroform. The combined extracts weredried (sodium sulphate) and evaporated to afford an oilwhich was purified by chromatography on silica gel(toluene/ethyl acetate 1:1), giving5a5 in 85% yield as athick oil.

Table I. Germicidal effect (GE) and killing percentage effect (KE %) of “short-retinoids”5 and3a against the five species mentioned.

Compound C. albicans P. aeruginosa E. coli S. aureus P. acnes mean GE± DS

3a GE 2.447 0.286 0.456 2.058 2.041 1.46± 1.01KE % 99.64 48.18 65.00 99.13 99.09

5a1 GE 3.56 0.26 0.56 0.98 2.35 1.54± 1.38KE % 99.97 45.46 72.50 89.41 99.55

5a2 GE 2.05 1.16 1.50 1.94 4.91 2.31± 1.19KE % 99.11 93.13 96.87 98.85 99.99

5a3 GE 2.263 0.176 0.301 0.503 2.981 1.25± 1.28KE % 99.46 33.33 50.00 68.57 99.90

5a4 GE 1.21 0.36 0.60 1.38 3.01 1.31± 1.04KE % 93.90 53.67 75.00 95.79 99.90

5a5 GE 0.75 0.30 0.58 1.93 2.01 1.12± 0.80KE % 82.31 50.00 73.91 98.82 99.03

5a6 GE 1.83 1.15 1.29 3.02 2.17 1.89± 0.75KE % 98.52 92.94 94.82 99.91 99.33

5b GE 0.75 0.27 0.58 1.90 2.06 1.11± 0.81KE % 82.50 45.45 74.40 97.42 99.13mean % 94.43 57.77 75.31 93.49 99.49DS 7.67 22.54 15.13 10.61 0.4

840

5.1.6. 2-Methylthio-4-[(2,6,6-trimethyl-2-cyclohexen-1-yl)ethenyl]-pyrimidine5a6

To the solution of 0.5 g (2.0 mmol) of3a in 10 mL ofethanol, 0.55 g (2.0 mmol) of S-methylisourea hydrogensulphate dissolved in 2 mL of sodium hydroxide 2 N and10 mL of water was added and the mixture refluxed for48 h. The resulting oil was extracted with chloroform.The combined extracts were dried (sodium sulphate) andevaporated to afford an oil which was purified by chro-matography on silica gel (toluene/ethyl acetate 1:1),giving 5a6 in 85% yield as a thick yellow oil; IR (film):ν 3 100, 2 900, 2 860, 1 680, 1 560, 1 450, 1 375,1 200 cm–1; 1H-NMR (200 MHz): δ 0.85 (3H, s, CH3),0.92 (3H, s, CH3), 1.24 (1H, m, H-5′), 1.44 (1H, m, H-5′),1.59 (3H, s, CH3), 2.04 (2H, m, H-4′), 2.27 (1H, d, H-1′J 8.98 Hz), 2.29 (3H, s, SCH3), 5.47 (1H, m, H-3′), 6.19(1H, d, ethene,J 8.65 Hz), 6.59 (1H, d, H-4,J 5.30 Hz),6.70 (1H, dd, ethene,J 15.15, 8.65 Hz), 8.15 (1H, d, H-5,J 5.30 Hz); 13C-NMR (50 MHz) 18.95 (CH3S), 23.41(CH3), 23.56 (CH2), 27.81 (CH3), 28.35 (CH3), 31.78(CH2), 33.11 (C), 55.18 (CH), 108.82 (CH), 122.53 (CH),130.17 (CH), 133.27 (C), 141.89 (CH), 158.34 (CH),164.408 (C), 164.55 (C). Anal. C16H22N2S (C, H, N, S).

5.1.7. 6-(2,6,6-Trimethyl-1-cyclohexen-1-yl)ethenyl-2-aminopyrimidine5b

By use of the procedure in 5.1.5, compound5b wasprepared from3b. The product was finally separated bychromatography on silica gel in 70% yield as a thick oil;IR (film): ν 3 300, 3 180, 2 900, 1 680, 1 610, 1 550,1 440 cm–1; 1H-NMR (200 MHz): δ 1.02 (6H, s, CH3),1.48 (2H, m, H-5′), 1.60 (2H, m, H-4′), 1.74 (3H, s, CH3),2.02 (2H, t, H-3′), 5.19 (2H, s, NH2), 5.92 (1H, d, ethene,J 16.28 Hz), 6.48 (1H, dd, ethene,J 16.28 Hz), 6.58 (1H,d, H-4, J 5,26 Hz), 8.19 (1H, d, H-5,J 5.26). Anal.C15H21N3 (C, H, N).

5.1.8. Semicarbazone4aAn ethanolic solution of 0.5 g of 3-chloro-5-(2,6,6-

trimethyl-2-cyclohexen-1-yl)-2,4-pentadienal2a, 0.7 g ofsodium acetate trihydrate and 0.5 g of semicarbazidehydrochloride was heated at 90 °C with stirring. Aftercooling, the precipitate obtained was filtered off andcrystallised from ethanol. The yellow crystals formedwere pure4a, (35% yield; m.p. 209–210 °C from etha-nol). IR (KBr): ν 3 480, 3 140, 2 900, 1 700, 1 580,1 420, 1 160 cm–1; 1H-NMR (200 MHz, DMSO-d6): δ0.80 (3H, s, CH3), 0.90 (3H, s, CH3), 1.20 (1H, m, H-5′),1.40 (1H, m, H-5′), 1.55 (3H, s, CH3), 2.02 (2H, m, H-4′),2.30 (1H, d, H-1′, J 9.06 Hz), 5.48 (1H, m, H-3′), 5.95(1H, dd, ethene,J 9.06 andJ 9.30 Hz), 6.35 (2H, s, NH2),6.38 (1H, d, ethene,J 9.30 Hz), 6.50 (1H, d, CH–CH=N,

J 10.66 Hz), 7.92 (1H, d, CH–CH=N,J 10.66 Hz), 10.45(1H, s, NH); Anal. C15H22N3OCl (C, H, N, Cl).

5.1.9. Semicarbazone4bBy use of the above procedure, semicarbazone4b was

obtained in 38% yield as yellow powder from2b, m.p.196–198 °C from ethanol. IR (KBr):ν 3 480, 3 120,2 850, 1 690, 1 580, 1 420, 1 160 cm–1; 1H-NMR (200MHz, DMSO-d6): δ 1.05 (6H, s, CH3), 1.45 (2H, m,H-5′), 1.55 (2H, m, H-4′), 1.70 (3H, s, CH3), 2.05 (2H, m,H-3′), 6.35 (2H, s, NH2), 6.45 (1H, d, ethene,J 15.78 Hz),6.60 (1H, d, CH–CH=N,J 10.52 Hz), 6.65 (1H, d, ethene,J 15.78 Hz), 7.98 (1H, d, CH–CH=N,J 10.52 Hz), 10.48(1H, s, NH); Anal. C15H22N3OCl (C, H, N, Cl).

5.2. Biological methods

The micro-organisms used in this study werePseudomonas aeruginosaATCC 9027 (16× 107 CFU/mL), Escherichia coliATCC 8739 (11× 107 CFU/mL),Staphylococcus aureusATCC 6538 (40× 107 CFU/mL),Candida albicansATCC 10231 (77× 107 CFU/mL),Propionibacterium acnesATCC 11827 (1.54× 107 CFU/mL).

The microbicidal activity has been determined by a“Reybrouck in vitro test” [12, 13] which was carried outusing 5 min of medication time at 25± 1 °C and 30 minof contact time with LPHT inactivator (0.3% lecithin, 3%polysorbate 80, 0.1% histidine, 0.5% sodium thiosulfate).

The bacterial strains were obtained by the 4th subcul-ture at 37± 1 °C/48h on TSA slants from freeze-driedstock cultures.

Bacterial suspension was obtained in TSB after anincubation time of 24 h at 37± 1 °C followed by centrifu-gation at 2 000g for 15 min and resuspension in tryptonedilution water (TDW).

The number of viable organisms in the inoculum weredetermined by the plating technique, mixing 1.0 mLsamples from the dilutions 10–6, 10–7 and 10–8 with20 mL of TSA, melted and tempered to 45 °C. Allinoculations were carried out in triplicate.

The test was carried out in a water bath at 25± 1 °C. Attime zero, 0.1 mL volumes of the different bacterialsuspensions were added to 10 mL of the retinoid solutionto be tested. After 5 min, 1 mL volumes were transferredfrom the medication mixtures into 9 mL of inactivatorsolution (LPHT). After a contact time of 30 min, 0.1 mLportions of the undiluted inactivated mixture and of the10–1, 10–2, 10–3 and 10–4 dilutions in TDW were spreadin triplicate on plates containing 20 mL of TSA. Thecolony forming units were counted after incubation at37 ± 1 °C for 24 h.

841

For the negative controls, which were submitted to thesame medication procedure, distilled water was usedinstead of the retinoid derivatives.

The germicidal activity is numerically expressed usingthe Germicidal Effect (GE, the decimal reduction time),meaning the ratio, expressed as logarithm, between thenumber of colony forming units (CFU) per mL in thecontrol mixture without the tested retinoid (Nc) to thenumber of CFU after the exposure to the retinoids (Ns) orGE = log Nc – log Ns. The corresponding killing effectpercentage is obtained by the following mathematicalexpression:

KE % = Nc − NsNc ⋅ 100

Acknowledgements

We wish to thank CNR and MURST (Rome) forfinancial support.

References

[1] Orfanos C.E., Schuppli R., Dermatologica Suppl. (1) (1978) 1–64.

[2] Orme M., Back D.J., Shaw M.A., Allen W.L., Tjia J., Cunliffe W.J.,Jones D.H., Lancet 2 (8405) (1984) 752–753.

[3] Flemetakis A.C., Tsambaos D.G., J. Chemother. 1 (6) (1989)374–376.

[4] Gollnick H., Schramm M., Dermatology (Basel) 196 (1998)119–125.

[5] L’Oreal S.A., Fr. Demande 2761600, 1998.

[6] Sefton G., Lew-Kaya D.A., Allergan Sales Inc., USA, PCT Int.Appl., WO 9856375, 1998.

[7] Sottofattori E., Anzaldi M., Balbi A., J. Heterocyclic Chem. 35(1998) 1377–1380.

[8] Coquelet C., Roussillon S., Sincholle D., Bonne C., Alzatet A., PCTInt. Appl. (1985) WO 8504652; Drugs of the future (1986) 557–558.

[9] Nastruzzi C., Simoni D., Manfredini S., Barbieri R., Feriotto G.,Baraldi P.G., Spandidos D., Guarneri M., Gambari R., AnticancerRes. 9 (1989) 1377–1384.

[10] Simoni D., Manfredini S., AghazadehTabrizi M., Bazzanini R.,Baraldi P.G., Ferroni R., Traniello F., Nastruzzi C., Feriotto G.,Gambari R., Drug Design and Discovery 8 (1992) 165–177.

[11] Cortesi R., Esposito E., Gambari R., Menegatti E., Nastruzzi C., Eur.J. Pharmacol. Sci. 2 (1994) 281–291.

[12] Reybrouck G., Werner H.P., Zbl. Bakteriol. ParasitenkundeInfektions-krankh. Hyg. Orig. B 165 (1977) 126–137.

[13] Reybrouck G., Borneff J., Van De Voorde H., Werner H.P., Zbl.Bakteriol. Parasitenkunde Infektionskrankh. Hyg. Orig. B 168(1979) 463–479.

842

Preliminary communication

4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidineand its derivatives: synthesis and affinity at 5-HT2A, 5-HT2B and 5-HT2C

serotonin receptors

Francesco Claudia*, Loredana Scocciaa, Gianfabio Giorgionia, Beatrice Accorronia,Gabriella Maruccia, Stefania Gessib, Anna Siniscalchib, Pier Andrea Boreab

aDepartment of Chemical Sciences, University of Camerino, Via S. Agostino 1, 62032 Camerino, ItalybDepartment of Experimental and Clinical Medicine, Section of Pharmacology, University of Ferrara, Via Fossato di Mortara 17–19,

44100 Ferrara, Italy

(Received 24 July 1998; accepted 19 May 1999)

Abstract – 4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine (7) and its derivatives modified at the carbonyl group ofthe fluorobenzoyl moiety were prepared and evaluated for affinity at 5-HT2A, 5-HT2C (rat cortex) and 5-HT2B (rat stomach fundus) serotoninreceptors. Compound7 bound the 5-HT2A sites with higher affinity (Ki = 8.2 nM) than the 5-HT2B (Kb = 1 290 nM) and 5-HT2C ones (Ki= 54.2 nM). Modification of the benzoyl carbonyl group decreased the 5-HT2A and 5-HT2C affinities but did not significantly influence 5-HT2Baffinity. This suggests that the carbonyl group is the determinant for the interaction with 5-HT2A and 5-HT2C receptor subtypes. Compound7 was found to be a 5-HT2A receptor antagonist. © 1999 Éditions scientifiques et médicales Elsevier SAS

4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine derivatives / synthesis / 5-HT2A, 5-HT2B, and 5-HT2C seroto-nin receptor affinity / 5-HT 2A antagonistic activity

1. Introduction

Serotonin (5-hydroxytryptamine, 5-HT) mediates anumber of neuronal processes both in the central nervoussystem and peripheral tissues. During the last decademultiple 5-HT receptor subtypes have been characterizedand grouped in seven classes (5-HT1–5-HT7) [1]. The5-HT2 class includes the subtypes 5-HT2A, 5-HT2B, and5-HT2C which are grouped together considering theirhigh degree of transmembrane sequence homology andsecond messenger coupling system. The 5-HT2A subtypeis present in the brain (cortical regions) [2] and periphery(gastrointestinal tract, cardiovascular system) [3], and isinvolved in various cardiovascular and mental disorders,such as depression and schizophrenia [4]. The 5-HT2B

receptor is expressed in rat stomach fundus, where itmediates a contractile response to 5-HT. Its mRNAtranscript is present in the human brain [5], and it hasbeen suggested to be involved in the pathophysiology of

migraine [6]. The 5-HT2C receptor was initially charac-terized in the choroid plexus, is widely distributed in thebrain [2], and has been suggested as playing a role inmigraine, obsessive compulsive disorders, and anxi-ety [7].

To date, few antagonists discriminating between the5-HT2 subtypes are available [8–14]. With the lack ofselective agents has come the realization that manypharmacological and physiological effects once attributedto 5-HT2A receptors (previously 5-HT2) may, in fact,involve more than one receptor subtype. Hence, there is aneed to discover new agents able to bind with highselectivity at one of the three subtypes.

The 3-{2-[4-(4-fluorobenzoyl)-1-piperidinyl]ethyl}-2,4(1H,3H)-quinazolinedione (ketanserin,1, figure 1) is aprototypical 5-HT2A antagonist reported to bind with aslittle as 15-fold [9] to as much as 140-fold selectivity for5-HT2A versus 5-HT2C receptors [10]. It was used toexplain the molecular structural prerequisites for the5-HT2A antagonist binding [9, 10], and to build the*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 843−852 843© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

three-dimensional models of 5-HT2A and 5-HT2C recep-tors [15–17].

Several studies on ketanserin analogues showed thatthe 4-(4-fluorobenzoyl)piperidine (2) fragment is en-dowed with 5-HT2A antagonistic activity and seems to beessential for binding at the 5-HT2A receptor [9, 18].

Among the 5-HT2 agonists, the most extensivelystudied is the 1-(4-iodo-2,5-dimethoxyphenyl)-2-amino-propane (3 R(-)-DOI), currently reported as a 5-HT2C/2A

agonist [19]. Investigations on 1-(2,5-dimethoxyphenyl)-2-aminopropane derivatives have established that themethyl groupα to the amine in3 does not influence thein vitro receptor affinity. Thus, the 1-(4-iodo-2,5-dimethoxyphenyl)-2-aminoethane4 binds with high af-finity (K i = 1.53 nM) the 5-HT2 receptors labelled byR-[125I]-DOI [20], and has been reported to act as agonistat 5-HT2 receptors [21].

Site-directed mutagenesis and molecular modellingstudies indicate that the piperidine nitrogen atom of1 andthe amino group of3 have a strong elecrostatic interac-tion with the carboxylate anion of the aspartate residueAsp 155 in transmembrane helix 3 (TM 3) of the 5-HT2A

receptor [15, 17]. Therefore, it would appear that Asp-155 is a common binding site for agonists3 and4, and forantagonist1.

Some time ago, Ariens suggested that an agonist can beturned into a competitive antagonist by appending to theagonist structure a hydrophobic bulky group able to bindthe accessory binding sites of the receptor [22].

On the basis of the Ariens strategy and of the abovementioned observations, we decided to connect the 4-(4-fluorobenzoyl)piperidine with the 2-(4-iodo-2,5-dimetho-xyphenyl)ethyl fragment of4 to obtain 4-(4-fluoro-benzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]pipe-ridine7 (figure 2). This compound was chosen in order toobtain a new 5-HT2 antagonist and to investigate how amodified tail tied to 4-(4-fluorobenzoyl)piperidine couldinfluence selectivity for the 5-HT2A, 5-HT2B, and 5-HT2C

subtypes.Studies with ketanserin and related 5-HT2 antagonists

have suggested that the carbonyl group of 4-(4-fluorobenzoyl)piperidine is involved in a hydrogen-bonding interaction with the 5-HT2A/2C receptors [23]and may have a prominent role in anchoring ketanserin to5-HT2A receptors [10, 15]. Thus, in order to obtainfurther information about the role of the carbonyl groupin the affinity and selectivity for the three 5-HT2 sub-types, two series of compounds were synthesized. Thefirst series keeps an sp2 carbon atom placed between thepiperidine and the fluorophenyl ring (8–11, figure 3),while in the other (12–16, figure 4) the same carbon hasan sp3 hybridization.

2. Chemistry

The desired 4-(4-fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine7 was synthesized bycondensation of 4-(4-fluorobenzoyl)piperidine2 with thetosyl ester of 2-(4-iodo-2,5-dimethoxyphenyl)ethanol6(figure 2). Iodination of 2-(2,5-dimethoxyphenyl)ethanolwith iodine in the presence of silver trifluoroacetate gavethe 2-(4-iodo-2,5-dimethoxyphenyl)ethanol5 which wasesterified with tosyl chloride. Oxime8 and theO-alkyloximes10and11were obtained by reaction of theketone7 in pyridine and absolute ethanol with hydroxyl-amine or O-ethylhydroxylamine or O-benzylhydroxyl-amine(figure 3). The oxime acetate9 was obtained from8 by reaction with acetic anhydride. Oxime isomers werenot separated. Ketone7 was also reduced by sodiumborohydride to alcohol12 or transformed into the diox-olane13. From 12, by reaction with acetic anhydride orbenzoyl chloride, the esters14 and 15 were obtained(figure 4).

In order to obtain the derivative16 with a methylenereplacing the carbonyl group, the tosylate ester6 wasreacted with 4-(4-fluorobenzyl)piperidine.

Figure 1. Structures of ketanserin, 4-(4-fluorobenzoyl)piperidine (2), DOI, and 1-(4-iodo-2,5-dimethoxyphenyl)-2-aminoethane (4).

844

3. Pharmacology

The affinity of compounds for 5-HT2A and 5-HT2C

receptors was assessed in vitro in the cerebral cortexpreparations. [3H]Ketanserin as radiolabelled ligand forthe 5-HT2A receptors, and [3H]mesulergine for the5-HT2C receptors were used. The antagonistic affinity atthe 5-HT2B receptors was determined by the inhibition of5-HT-induced contractions of rat stomach fundus. Theresults are reported intable I. The 5-HT2B receptoraffinities were obtained from functional data; however,caution should be exercised when comparing them withdata from binding assays.

Derivatives7, 9, 10, and13 were evaluated for theirantagonist activity at central 5-HT2A receptors by testingtheir ability to antagonize the facilitatory effect of 5-HTon basal acetylcholine release from guinea-pig striatalslices [24]. Acetylcholine release induced by 5-HT couldbe attributed to 5-HT2A receptor activation. In fact, it wasconcentration-dependently antagonized by 5-HT2A an-tagonists added to the superfusion medium from the

beginning of the experiment, while the 5-HT2C antagonistmesulergine was unable to counteract it, except at10 µmol, a high concentration able to block 5-HT2A

receptors as well.

4. Results and discussion

The data in table I indicate that the 4-(4-fluoro-benzoyl)piperidine2 binds with the same affinity at5-HT2A and 5-HT2C sites. Moreover, 2-(4-iodo-2,5-dimethoxyphenyl)ethylamine4 binds with lower affinityto the 5-HT2A than to the 5-HT2C sites (Ki = 300± 33 nMvs. 2.5 ± 0.14 nM). Accordingly, in the assays ofacetylcholine release from guinea-pig striatal slices, its5-HT2A agonistic action is negligible (net [3H]cholineefflux increase at 30µM = 0.11± 0.03%,n = 5), as is theantagonistic one (5 HT 30µM-induced net [3H]cholineefflux increase in the presence of4 = 0.55± 0.06%, notsignificantly different from 5-HT 30µM alone (0.89±0.11%). Functional assays on rat stomach fundus indicatethat4 acts as a full agonist (pD2 = 6.54± 0.12, less potent

Figure 2. Synthesis of7.

845

than 5-HT: pD2 = 8.20 ± 0.09). Introduction on thepiperidine nitrogen of the 2-(4-iodo-2,5-dimethoxy-phenyl)ethyl fragment increases the affinity for all 5-HT2

receptor subtypes, and this effect is more pronounced for5-HT2A sites (80-fold) than for 5-HT2B and 5-HT2C

(13-fold) ones. On the other hand, when the amino groupof 4 is substituted by the 4-(4-fluorobenzoyl)piperidinemoiety, the 5-HT2A and 5-HT2B affinities increase and the5-HT2C affinity decreases. Compound7 binds the 5-HT2A

sites with lower affinity than does the reference com-

Figure 3. Synthesis of compounds9–11.

846

pound 1 (8.2 nM vs. 0.24 nM) and, under our assayconditions, displays the same selectivity for 5-HT2A

versus 5-HT2C receptors (about 7-fold). Compound7 isalso 157-fold more selective for 5-HT2A versus 5-HT2B

receptors. These results suggest that the 2-(4-iodo-2,5-dimethoxyphenyl)ethyl fragment enhances to a greaterextent the affinity for 5-HT2A than that for 5-HT2B or5-HT2C receptors.

Figure 4. Synthesis of compounds12–16.

847

The modification of the carbonyl group in7 alwaysdecreases the 5-HT2A and 5-HT2C affinities and does notaffect significantly the 5-HT2B affinity. In the compoundshaving an sp2 carbon atom (8–11), the carbonyl substitu-tion with the oxime group decreases the 5-HT2A and5-HT2C affinities. Both acetylation and etherification ofthe oxime hydroxyl group increases the affinities. Theseare influenced in an opposite way: the 5-HT2A affinityincreases and 5-HT2C affinity decreases following theorder9, 10, 11. This could suggest a different interactionof 9–11with the 5-HT2A and 5-HT2C sites. It can also beseen fromtable I that the replacement of the carbonylgroup with an sp3 carbon atom (compounds12–16)decreases the 5-HT2A and 5-HT2C affinities. In the serieswith an sp3 carbon atom between the piperidine and thefluorophenyl ring, only the dioxolane13 presents mod-erate 5-HT2A affinity. The highest decrease in affinity isobserved when the carbonyl group is reduced to alcohol12. Of importance is that the 5-HT2A and 5-HT2C

affinities decrease to a greater extent in derivatives8 and12 bearing a hydroxyl group. The results indicate that thecarbonyl oxygen participates in a key binding interactionand that a hydrogen-bond acceptor seems to be requiredby the 5-HT2A and 5-HT2C receptors. Moreover, it shouldbe observed that also the carbonyl group reduction tomethylene (compound16) decreases the affinities. Withregard to the 5-HT2B receptor, it is evident that thecarbonyl group modification does not significantly influ-ence the affinity.

Derivatives7, 9, 10, and13show antagonist activity atcentral 5-HT2A receptors counteracting the release of

acetylcholine induced by 5-HT(figure 5). They behave ascompetitive antagonists and are less active than thestandard1. The potency order was1 > 7 > 9 > 10 > 13,in good agreement with the binding data. All the antago-nists, when tested at the highest concentration, proved to

Table I. 5-HT2A, 5-HT2C (rat cortex) receptor binding affinities, and 5-HT2B (rat stomach fundus) affinity of compounds1, 2, 4, 7–16a.

Compound Ki (nM)b Kb (nM)c

5-HT2A 5-HT2C 5-HT2B

1 0.24 ± 0.01 1.76± 0.05 8 660± 8102 667 ± 44 710± 59 > 10 0004 300 ± 33 2.5± 0.14 > 10 000d

7 8.2 ± 0.4 54.2± 2.1 1 290± 608 666 ± 44 2 270± 176 1 300± 1509 100 ± 6 1 450± 87 1 350± 17010 150 ± 6 887± 59 1 000± 13011 275 ± 14 598± 37 NDe

12 616 ± 60 1 220± 78 807± 11513 118 ± 14 432± 19 1 080± 17014 325 ± 14 687± 59 ND15 275 ± 14 1 660± 137 7 140± 98016 238 ± 7 497± 20 ND

aAll values represent means± SEM; n ≥ 3 determinations.bBinding affinity (rat cortex; 5-HT2A [3H]keranserin, 5-HT2C [3H]mesulergine).cApparent antagonist dissociation constant, rat stomach fundus. The compounds were tested at 10–5 M and behave as competitive antagonists.dAffinity constant (Kb) determined as reported in [28].eNot determined.

Figure 5. Basal tritium efflux from guinea-pig striatal slicesprelabelled with [3H]choline. Relationship between 5-HT2Aantagonist concentrations (µmol, abscisses, log scale) andinhibition of the facilitatory effect of 5-HT 30µmol (%,ordinate). Points represent the means± SEM of 4–9 experi-ments.■ Ketonserin,m 7, . 9, ◆ 10, ★ 13.

848

be devoid of intrinsic activity, except13, which at 30µMevoked a net increase in [3H]choline efflux of 0.67±0.11% (n = 3).

In conclusion, compound7 represents a successfulapplication of the Ariens strategy in the design of a5-HT2A antagonist by substituting the amino group of2-(4-iodo-2,5-dimethoxyphenyl)ethylamine with thebulky lipophilic moiety of 4-(4-fluorobenzoyl)piperidine.Comparing1 with 7, it is evident that the 3-ethyl-2,4-quinazolinedione moiety gives higher 5-HT2A and5-HT2C affinities than does the 2-(4-iodo-2,5-dimethoxy-phenyl)ethyl fragment.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Buchi 510apparatus and are uncorrected. Microanalyses were per-formed on a 1106 Carlo Erba CHN Analyzer, and theresults were within± 0.4% of the theoretical values.1HNMR spectra were recorded on a Varian VXR 200 MHzspectrometer. Chemical shifts are reported in parts permillion (δ) downfield from the internal standard tetram-ethylsilane (Me4Si). The identity of all new compoundswas confirmed both by elemental analysis and NMR data;homogeneity was confirmed by TLC on silica gel Merck60 F254. Chromatographic purifications were performedby Merck-60 silica gel columns 70–230 mesh ASTMfrom Merck with a reported solvent.

5.1.1. 2-(4-Iodo-2,5-dimethoxyphenyl)ethanol5A solution of iodine (0.92 g, 3.6 mmol) in chloroform

(20 mL) was added dropwise, under stirring at roomtemperature, to a slurry of silver trifluoroacetate (0.89 g,3.6 mmol) and 2-(2,5-dimethoxyphenyl)ethanol (0.66 g,3.6 mmol) in chloroform (5 mL). The mixture was stirredfor 30 min. The insoluble material was filtered. Thefiltrate was washed with aqueous 10% Na2SO3, brine anddried (Na2SO4). The solvent was evaporated and theresidue was crystallized from isopropyl ether, m.p.:93–94 °C; yield 76%.1H-NMR (CDCl3) δ 1.59 (1H, bs,OH), 2.88 (2H, t,J = 6.3 Hz, ArCH2), 3.80 (8H, m,OCH3, CH2O), 6.68 (1H, s, ArH), 7.21 (1H, s, ArH).Anal. C10H13IO3 (C, H).

5.1.2. 2-(4-Iodo-2,5-dimethoxyphenyl)ethyl-p-toluensul-fonate6

p-Toluenesulfonyl chloride (0.63 g, 3.3 mmol) wasadded portionwise to a solution of5 (0.92 g, 3 mmol) indry pyridine (1 mL). After stirring for 5 h at roomtemperature, 2 N HCl (9 mL) was added to yield a

precipitate. This was filtered and crystallized from cyclo-hexane, m.p.: 110–112 °C; yield 82%.1H-NMR (CDCl3)δ 2.46 (3H, s, CH3), 2.89 (2H, t,J = 6.3 Hz, ArCH2), 3.63(3H, s, OCH3), 3.88 (3H, s, OCH3), 4.22 (2H, t,J = 6.3Hz, CH2O), 6.57 (1H, s, ArH), 7.08 (1H, s, ArH), 7.23(2H, d,J = 8.5 Hz, ArH), 7.58 (2H, d,J = 8.5 Hz, ArH).Anal. C17H19IO5S (C, H).

5.1.3. 4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidine hydrochloride7

A mixture of 6 (3 g, 6.5 mmol), 4-fluoro-benzoylpiperidine (1.35 g, 6.5 mmol) and K2CO3 (1.28 g,9.3 mmol) in acetone (30 mL) was warmed to reflux for20 h. The solvent was removed by evaporation, and theresidue was partitioned between H2O and CHCl3. Theorganic extracts were dried (Na2SO4) and evaporated.The residue was purified by chromatography eluting withCHCl3:acetone:MeOH (6.7:3:0.3, v/v) and crystallizedfrom isopropyl ether; m.p.: 98–100 °C; yield 84%.1H-NMR (CDCl3) δ 1.88 (4H, m, H3,5pip), 2.21 (2H, m,H2,6ax-pip), 2.58 (2H, m, ArCH2), 2.80 (2H, m, NCH2),3.08 (2H, m, H2,6eq-pip), 3.21 (1H, m, H4pip), 3.78 (3H, s,OCH3), 3.83 (3H, s, OCH3), 6.70 (1H, s, ArH), 7.14 (2H,m, ArH), 7.20 (1H, s, ArH), 7.98 (2H, m, ArH). Hydro-chloride: crystallized from absolute EtOH, m.p.:232–234 °C. Anal. C22H25FINO3.HCl (C, H, N).

5.1.4. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methanone oxime hydro-chloride8

A mixture of 7 (1 g, 2 mmol), hydroxylamine hydro-chloride (0.35 g, 5 mmol), and pyridine (2 mL) in abso-lute EtOH (5 mL) was warmed to reflux for 3 h. Afterevaporation of the solvent, water was added. The precipi-tate was filtered and crystallized from absolute EtOH;m.p.: 239–241 °C; yield 92%.1H-NMR (DMSO-d6) δ1.88 (4H, m, H3,5pip), 2.23 (1H, m, H4pip), 2.96 (4H, m,ArCH2, H2,6ax-pip), 3.11 (2H, m, NCH2), 3.55 (2H, m,H2,6eq-pip), 3.75 (6H, s, OCH3), 6.93 (1H, s, ArH), 7.38(5H, m, ArH), 10.24 (1H, bs, NH+), 10.90 e 11.45 (1H,2s, OH). Anal. C22H26FIN2O3.HCl (C, H, N).

5.1.5. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methanone O-acetyloximehydrochloride9

A mixture of 8 (0.8 g, 1.6 mmol) and acetic anhydride(2 mL) was stirred at 60 °C for 1.5 h. The mixture wascooled and partitioned between 5% aqueous Na2CO3 andEt2O. The organic extracts were washed with H2O, dried(Na2SO4) and evaporated. The residue was crystallizedfrom isopropyl ether; m.p.: 78–80 °C, yield 77%. Theresidue was dissolved in Et2O and treated with etherealHCl. The solid obtained was crystallized from acetone;

849

m.p.: 168–170 °C. NMR (DMSO-d6) δ 2.0 (7H, m,H3,5pip, CH3), 3.11 (7H, m, ArCH2, H4pip, H2,6ax-pip), 3.61(2H, m, H2,6eq-pip), 3.78 (6H, s, OCH3), 6.96 (1H, m,ArH), 7.42 (5H, m, ArH), 10.35 (1H, bs, NH+). Anal.C24H28FIN2O4.HCl (C, H, N).

5.1.6. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphe-nyl)ethyl]piperidin-4-yl}methanone-O-ethyloxime hydro-chloride10

This was made in the same way as8 from 7 andO-ethylhydroxylamine hydrochloride. The precipitate ob-tained after addition of water was filtered and treated witha saturated solution of Na2CO3. The mixture was ex-tracted with Et2O. The organic extracts were dried(Na2SO4) and evaporated. The residue was dissolved inabsolute EtOH, and EtOH saturated with HCl gas wasadded. The solvent was evaporated and the residue wascrystallized from 2-propanol; m.p.: 203–205 °C, yield75%. 1H-NMR (CDCl3) δ 1.1 (3H, t,J = 6.7 Hz, CH3),2.0 (2H, m, H3,5ax-pip), 2.39 (2H, m, H3,5eq-pip), 2.68 (3H,m, H2,6ax-pip, H4pip), 3.19 (4H, m, ArCH2, NCH2), 3.66(2H, m, H2,6eq-pip), 3.78 (3H, s, OCH3), 3.80 (3H, s,OCH3), 4.08 (2H, q,J = 6.7 Hz, OCH2), 6.82 (1H, s,ArH), 7.10 (2H, m, ArH), 7.25 (3H, m, ArH), 12.42 (1H,bs, NH+). Anal. C24H30FIN2O3.HCl (C, H, N).

5.1.7. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methanone O-benzyloximehydrochloride11

This was made in the same way as8 from 7 andO-benzylhydroxylamine hydrochloride. The precipitateobtained after addition of water was filtered and treatedwith a saturated solution of Na2CO3. The mixture wasextracted with Et2O. The organic extracts were dried(Na2SO4) and evaporated. The residue was dissolved inabsolute EtOH, and EtOH saturated with HCl gas wasadded. The solvent was evaporated and the residue wascrystallized from 2-propanol; m.p.: 145–147 °C, yield71%. 1H-NMR (DMSO-d6) δ 1.82 (3H, m, H3,5ax-pip,H4pip), 2.94 (6H, m, ArCH2, H2,6ax-pip, H3,5eq-pip), 3.12(2H, m, NCH2), 3.45 (2H, m, H2,6eq-pip), 3.76 (6H, s,OCH3), 5.02 (2H, s, OCH2), 6,92 (1H, s, ArH), 7.30(10H, m, ArH), 9.97 (1H, bs, NH+). Anal.C29H32FIN2O3.HCl (C, H, N).

5.1.8. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methanol fumarate12

To a solution of the ketone7 (0.5 g, 1 mmol) inCH3OH (12 mL), stirred at room temperature, NaBH4

(0.038 g, 1 mmol) was added. The mixture was stirred for12 h. The solvent was evaporated and the residue waspartitioned between H2O and CH2Cl2. The organic ex-tracts were dried (Na2SO4) and evaporated. The residue

was purified by chromatography eluting withCHCl3:acetone:MeOH (6.7:3:0.3, v/v) to give an uncrys-tallizable oil; yield 84%.1H-NMR (CDCl3) δ 1.43 (4H,m, H3,5pip), 2.0 (4H, m, OH, H4pip, H2,6ax-pip), 2.49 (2H,m, ArCH2), 2.74 (2H, m, NCH2), 3.02 (2H, m, H2,6eq-pip),3.70 (3H, s, OCH3), 3.78 (3H, s, OCH3), 4.35 (1H, m,HCO), 6.65 (1H, s, ArH), 7.01 (2H, m, ArH), 7.18 (1H, s,ArH), 7.25 (2H, m, ArH). The oil was dissolved inabsolute EtOH and treated with a solution of fumaric acidin absolute EtOH. The fumarate was filtered and crystal-lized from absolute EtOH, m.p.: 189–191 °C. Anal.C22H27FINO3.C4H4O4 (C, H, N).

5.1.9. 4-[2-(4-Fluorophenyl)-[1,3]dioxolan-2-yl]-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine hydro-chloride13

A mixture of ketone7 (0.5 g, 1 mmol) and ethyleneglycol (3 mL) was saturated with HCl gas and heated at90 °C for 1 h. The resulting solid was collected andcrystallized from absolute EtOH; m.p.: 263–265 °C, yield93%.1H-NMR (DMSO-d6) δ 1.64 (4H, m, H3,5pip), 2.09(1H, m, H4pip), 2.88 (4H, m, ArCH2, H2,6ax-pip), 3.11 (2H,m, NCH2), 3.50 (2H, m, H2,6eq-pip), 3.70 (8H, m, OCH3,OCH2), 3.98 (2H, m, OCH2), 6.92 (1H, s, ArH), 7.21 (2H,m, ArH), 7.38 (3H, m, ArH), 9.93 (1H, bs, NH+). Anal.C24H29FINO4.HCl (C, H, N).

5.1.10. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methylacetate fumarate14

A mixture of 12 (0.5 g, 1 mmol), dry pyridine (5 mL)and acetic anhydride (0.47 mL, 5 mmol) was stirred atroom temperature for 12 h, and then poured into ice.EtOH (2 × 30 mL) was added and the mixture waspartially evaporated. The aqueous concentrate was madebasic with 2 N NaOH, and the mixture was extracted withEt2O. The organic extracts were dried (Na2SO4) andevaporated to give an uncrystallizable oil; yield 76%.1H-NMR (CDCl3) δ 1.40 (3H, m, H3,5ax-pip, H4pip), 1.85(4H, m, H2,6ax-pip, H3,5eq-pip), 2.07 (3H, s, CH3), 2.52 (2H,m, ArCH2), 2.77 (2H, m, NCH2), 3.01 (2H, m, H2,6eq-pip),3.75 (3H, s, OCH3), 3.81 (3H, s, OCH3), 5.48 (1H, d,J =8.8 Hz, OCH), 6.68 (1H, s, ArH), 7.0 (2H, m, ArH), 7.18(1H, s, ArH), 7.25 (2H, m, ArH). The oil was dissolved inabsolute EtOH and treated with a solution of fumaric acidin absolute EtOH. The fumarate was filtered and crystal-lized from absolute EtOH, m.p.: 188–190 °C. Anal.C24H29FINO4.C4H4O4 (C, H, N).

5.1.11. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidin-4-yl}methylbenzoate fumarate15

A mixture of 12 (0.5 g, 1 mmol), dry pyridine (5 mL)and acetic anhydride (0.52 mL, 4.5 mmol) was stirred atroom temperature for 12 h, and then poured into ice.

850

EtOH (2 × 30 mL) was added and the mixture waspartially evaporated. The aqueous concentrate was madebasic with 2 N NaOH and the mixture was extracted withEt2O. The organic extracts were dried (Na2SO4) andevaporated to give an uncrystallizable oil; yield 74%.1H-NMR (CDCl3) δ 1.51 (3H, m, H3,5ax-pip, H4pip), 2.0(4H, m, H2,6ax-pip, H3,5eq-pip), 2.57 (2H, m, ArCH2), 2.79(2H, m, NCH2), 3.07 (2H, m, H2,6eq-pip), 3.75 (3H, s,OCH3), 3.81 (3H, s, OCH3), 5.78 (1H, d,J = 8.1 Hz,OCH), 6.70 (1H, s, ArH), 7.04 (2H, m, ArH), 7.19 (1H, s,ArH), 7.42 (5H, m, ArH), 8.08 (2H, m, ArH). The oil wasdissolved in absolute EtOH and treated with a solution offumaric acid in absolute EtOH. The fumarate was filteredand crystallized from absolute EtOH, m.p.: 181–183 °C.Anal. C29H31FINO4.C4H4O4 (C, H, N).

5.1.12. 4-(4-Fluorobenzyl)-1-[2-(4-iodo-2,5-dimethoxy-phenyl)ethyl]piperidine hydrochloride16

A mixture of 6 (0.46 g, 1 mmol), 4-fluoro-benzoylpiperidine (0.19 g, 1 mmol) and K2CO3 (0.2 g,1.5 mmol) in acetone (10 mL) was warmed to reflux for20 h. The precipitate was filtered and the solution wasevaporated. The residue was partitioned between H2Oand CHCl3. The organic extracts were dried (Na2SO4)and evaporated. The residue was purified by chromatog-raphy eluting with CHCl3:acetone:MeOH (6.7:3:0.3, v/v)and crystallized from acetone; m.p.: 96–98 °C; yield66%. 1H-NMR (CDCl3) δ 1.47 (5H, m, H3,5pip, H4pip),1.97 (2H, m, H2,6ax-pip), 2.50 (4H, m, H2,6eq-pip,ArCH2),2.75 (2H, m, ArCH2), 2.98 (2H, m, NCH2), 3.77 (3H, s,OCH3), 3.82 (3H, s, OCH3), 6.70 (1H, s, ArH), 6.98 (2H,m, ArH), 7.10 (2H, m, ArH), 7.20 (1H, s, ArH). Thepurified residue was dissolved in Et2O and treated withethereal HCl. The solid obtained was crystallized fromabsolute EtOH; m.p.: 204–205 °C. Anal.C22H27FINO2.HCl (C, H, N).

5.2. Pharmacology

5.2.1. MaterialsKetanserin tartrate was purchased from Research Bio-

chemicals International (Natick, MA, USA). [3H]Ket-anserin (64.1 Ci/mmol) was purchased from New En-gland Nuclear, Boston, Mass., USA. [3H]Mesulergine (76Ci/mmol) and [3H]choline (81 Ci/mmol) were purchasedfrom Amersham Radiochemical Centre (Buckingham-shire, UK). All substances employed in the bindingassays were dissolved in distilled water.

5.2.2. AnimalsIn the radioligand-binding studies rat cortex was ob-

tained from male Wistar rats (250–300 g body weight)

purchased from Nossan (Varese, Italy). Sections of stom-ach fundus were obtained from male CD Outbred rats(Charles River, Calco, Italy) weighing 125–150 g.

5.2.3. Binding assaysCerebral cortices of male Wistar rats (150–200 g) were

dissected on ice. The tissue was homogenized in 50 mmolTris-HCl buffer (pH = 7.7 at 25 °C). The homogenate wascentrifuged at 40 000g for 10 min. The supernatant wasdiscarded and the pellet was resuspended in the samevolume of Tris-HCl buffer and incubated at 37 °C for10 min prior to a second centrifugation. Binding experi-ments [23] with [3H]ketanserin (64.1 Ci/mmol) and[3H]mesulergine (76 Ci/mmol) were performed in 250µLof buffer, which contained 1 nmol [3H]ketanserin or[3H]mesulergine, membranes from 10 mg (wet weight) oftissue and the compounds to be tested. After 30 min ofincubation at 25 °C, separation of bound from freeradioligand was performed by rapid filtration throughWhatman GF/B glass fibre filters, which were washedthree times with ice-cold buffer, dried and counted in5 mL of Aquassure (Packard, Downers Grove, USA).Non-specific binding was measured in the presence of10 µmol 5-HT for 5-HT2A sites and 10µmol cinanserinfor 5-HT2C sites with specific binding defined as the totalbinding minus the non-specific binding. Ki values werecalculated from the Cheng-Prusoff equation [25] Ki =IC50/1 + (ligand/Kd), where Kd = 0.8 nmol/L for [3H]ket-anserin and Kd = 1.9 nmol/L for [3H]mesulergine [26].

5.2.4. Determination of apparent 5-HT2B receptorantagonist dissociation constant

Experiments were performed as described by Nozulaket al. [27]. Male CD Outbred rats were sacrificed by CO2,and longitudinal sections of the stomach fundus wereprepared for in vitro examination. Strips were set up inorgan baths of 10 mL containing Krebs solution (compo-sition in mmol: NaCl, 118; KCl, 4.7; CaCl2, 1.25;KH2PO4, 1.2; MgSO4, 1.2; glucose, 11; NaHCO3, 25)constantly bubbled with 5% CO2 in oxygen. Contractionswere measured isotonically under a resting tension of 1 g.Prior to testing, the strips were allowed to equilibrate for1 h, during which time the bath was replaced every15 min.

After control cumulative contractile responses to sero-tonin were obtained in the stomach fundus, the tissueswere incubated with an appropriate concentration ofantagonist for 1 h. Contractile responses to serotoninwere then repeated in the presence of the antagonist. Onlyone antagonist concentration was examined in eachtissue. Apparent dissociation constants (Kb) were deter-

851

mined for each concentration of antagonist according tothe following equation:

Kb = [B]/(dose ratio – 1)

where [B] is the concentration of the antagonist, and thedose ratio is the ED50 of the agonist in the presence of theantagonist divided by the control ED50.

5.2.5. Inhibition of acetylcholine releaseInhibition of the facilitatory effect of serotonin on basal

acetylcholine release from guinea-pig striatal slices wasdetermined as previously described [24]. Caudate nucleusslices (400µm thick) were incubated with 0.1µmol[3H]choline (81 Ci/mmol) for 30 min and superfused at0.25 mL/min with Krebs solution (compositon in mmol:NaCl, 118.5; KCl, 4.8; CaCl2, 2.5; KH2PO4, 1.2; MgSO4,1.2; NaHCO3 25, glucose, 11; hemicholinium-3, 0.01),bubbled with 95% O2 and 5% CO2. The radioactivity ofthe 5 min superfusate samples was determined by liquidscintillation. The effect of 5-HT, both in absence and inpresence of antagonists, was quantified as the net increaseof tritium efflux over the basal one, calculated as frac-tional rate (FR), i.e. as percent of tissue tritium content.The net increase of [3H]choline efflux, induced by 5-HT30 µM, added to the Krebs solution from the 45th min ofsuperfusion, was 0.89± 0.11% (n = 11).

Acknowledgements

This work was supported by Consiglio Nazionale delleRicerche (CNR, Rome) grant 97.02771.CT03, and by theUniversity of Camerino.

References

[1] Hoyer D., Martin G., Neuropharmacology 36 (1997) 419–428 .

[2] Pazos A., Probst A., Palacios J.M., Neuroscience 21 (1987)123–139.

[3] Peroutka S.J., Snyder S.H., Mol. Pharmacol. 16 (1979) 687–699.

[4] Glennon R.A., Neurosci. Biobehav. Rev. 14 (1990) 35–47.

[5] Kursar J.D., Nelson D.L., Wainscott D.B., Baez M., Mol. Pharma-col. 46 (1994) 227–234.

[6] Fozard J.R., Kalkman H.O., Naunyn-Schmiedeberg’s Arch. Pharma-col. 350 (1994) 225–229.

[7] Curzon G., Kenneth G.A., Trends Pharmacol. Sci. 11 (1990)181–182.

[8] Kennett G.A., Curr. Opin. Invest. Drugs 2 (1993) 317–362.

[9] Herndon J.L., Ismaiel A., Ingher S.P., Teitler M., Glennon R.A., J.Med. Chem. 35 (1992) 4903–4910.

[10] Ismaiel A.M., Arruda K., Teitler M., Glennon R.A., J. Med. Chem.38 (1995) 1196–1202.

[11] Forbes I.T., Jones G.E., Murphy O.E., Holland V., Baxter G.T., J.Med. Chem. 38 (1995) 855–857.

[12] Audia J.E., Evrard D.A., Murdoch G.R., Droste J.J., Nissen J.S.,Schenck K.W. et al., J. Med. Chem. 39 (1996) 2773–2780.

[13] Bromidge S.M., Duckworth M., Forbes I.T., Ham P., King F.D.,Thewlis K.M. et al., J. Med. Chem. 40 (1997) 3494–3496.

[14] Weinhardt K.K., Bonhaus D.W., De Souza A., Bioorg. Med. Chem.Lett. 6 (1996) 2687–2692.

[15] Kristiansen K., Edvardsen Ø., Dahl S.G., Med. Chem. Res. 3 (1993)370–385.

[16] Kristiansen K., Dahl S.G., Eur. J. Pharmacol. 306 (1996) 195–210.

[17] Westkaemper R.B., Glennon R.A., Med. Chem. Res. 3 (1993)317–334.

[18] Holtje H.D., Jendretzki U.K., Arch. Pharm. 328 (1995) 577–584.

[19] Glennon R.A., Raghupathi R., Bartyzel P., Teitler M., Leonhardt S.,J. Med. Chem. 35 (1992) 734–740.

[20] Johnson M.P., Mathis C.A., Shulgin A.T., Hoffmann A.J., NicholsD.E., Pharmacol. Biochem. Behav. 35 (1990) 211–217.

[21] Nichols D.E., Frescas S., Marona-Lewicka D., Huang X., Roth B.L.,Gudelsky G.A., Nash J.F., J. Med. Chem. 37 (1994) 4346–4351.

[22] Ariens E.J., Beld A.J., Rodrigues de Miranda J.F., Simonis A.M., in:O’Brien R.D. (Ed.), The Receptors: A Comprehensive Treatise, Vol.1, Plenum, New York, 1979, pp. 33–91.

[23] Pierce P.A., Kim J.Y., Peroutka J., Naunyn-Schmiedeberg’s Arch.Pharmacol. 346 (1992) 4–11.

[24] Siniscalchi A., Beani L., Bianchi C., Neuropharmacology 29 (1992)1091–1093.

[25] Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973)3099–3108.

[26] Pazos S., Hoyer D., Palacios J.M., Eur. J. Pharmacol. 106 (1985)531–538.

[27] Nozulak J., Kalkman H.O., Floersheim P., Hoyer D., Schoeffter P.,Buerki H.R., J. Med. Chem. 38 (1995) 28–33.

[28] Furchgott R.F., Bursztyn P., Ann. NY Acad. Sci. 144 (1967)882–899.

852

Short communication

Synthesis, physicochemical characterization and cytotoxic screening of newcomplexes of cerium, lanthanum and neodymium with Niffcoumar sodium salt

Ilia Manolova*, Irena Kostovab, Spiro Konstantinovc, Margarita Karaivanovac

aDepartment of Industrial Pharmacy, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Sofia, BulgariabDepartment of Chemistry, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Sofia, Bulgaria

cDepartment of Pharmacology and Toxicology, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Sofia, Bulgaria

(Received 20 August 1998; accepted 20 January 1999)

Abstract – The complexes of the types Ln(HN)3.nH2O [where Ln = Ce, La, Nd; HN = Niffcoumar] have been synthesized by reaction ofNiffcoumar sodium salt and the appropriate lanthanide nitrates. The solid complexes have been characterized and identified on the basis ofelemental analysis, conductivities, IR- and1H-NMR-spectroscopy. We suppose that the lacton- and keto-carbonyl oxygen atoms of Niffcoumarare bonded to the metal ion probably as a bidentate ligand. The experimental data obtained by MTT-assay show that all tested compoundsmanifest a similar low cytotoxic activity on lymphoma derived P3HR1 cells (20% maximum cell growth inhibition). The complexes of Laand Nd have been shown to be more cytotoxic than the Ce complex against K-562 cells (about 75% maximal growth inhibition). Low tomoderate cytotoxic activity has been established for the complexes of La and Ce on TPH-1 cells (25% maximum inhibition of cell growth).Our data indicate that the lanthanide complexes may have some antitumour potential which depends on the metal and the cell line used. © 1999Éditions scientifiques et médicales Elsevier SAS

metal complexes of Niffcoumar / lanthanides / cytotoxic screening

1. Introduction

4-Hydroxycoumarin derivatives are of interest becauseof their physiological, photodynamic and bacteriostaticactivity [1]. 4-Hydroxycoumarin and its derivatives areknown to exhibit an inclination for complexation [2–5].

Recently, some interesting lanthanide complexes ofcoumarin derivatives like bis (4-hydroxy-3-cou-marinyl)acetic acid [6], N,N≠-bis(8-aceto-7-hydroxy-4-methylcoumarin)ethylenediamine [7, 8], coumarin-3-carb-oxylic acid [9, 10] have been reported.

As a part of our investigations on the synthesis andcharacterization of solid lanthanide complexes, we reportherein the synthesis and characterization of lanthanidecomplexes with Niffcoumar sodium salt. The ligand is ofspecial interest as it contains several potential donor sites.Niffcoumar sodium salt has not been used as a ligand upto now. Cytotoxic screening was carried out for the newlysynthesized complexes with lanthanides (III). The presentpaper is a continuation of the authors’ earlier work on the

complexes of lanthanides (III) with Warfarin sodium saltand Coumachlor sodium salt and their cytotoxic screen-ing [11].

Literature data show that coumarins have an antitu-mour activity [12, 13]. The polycyclic coumarins holdpromise as agents for the chemoprevention of can-cer [14].

Lanthanides are a subject of increasing interest inbioinorganic and coordination chemistry. Furthermore,the literature data show that lanthanides manifest anantitumour activity [15, 16], which is in accordance withour previous investigations.

2. Chemistry

2.1. Preparation of the complexes

The compounds used for preparing the solutionswere Merck products, p.a. grade: Ce(NO3)3.6H2O,La(NO3)3.6H2O and Nd(NO3)3.6H2O. The sodium salt of4-hydroxy-3[1-(4-nitrophenyl)-3-oxobutyl]-2H-1-benzo-*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 853−858 853© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

pyran-2-one (Niffcoumar sodium salt (figure 1) was pre-pared as follows: 3.53 g (10 mmol) 4-hydroxy-3[1-(4-nitrophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one (Niff-coumar) were added to 0.38 g (9.5 mmol) sodium hy-droxide in 30 mL water. The mixture was stirred vigor-ously at room temperature for 1 h until it was clear. Thesolution was filtered and the filtrate was evaporated todryness. The viscous residue was recrystallized fromethylacetate. Yield 3.2 g (86%). TLC: Al-foils Kieselgel60 F254, Merck (FRG), developed by cyclohexane-chloroform-acetic acid (10:10:4, vol. parts), detection:UV 254 nm. M.p. 132–134 °C, Büchi 510 apparatus(Switzerland), uncorrected. Niffcoumar sodium salt wasused for the preparation of metal complexes as a ligand.

General method of synthesis: the complexes weresynthesized by mixing a water solution of the ligand witha water solution of the corresponding metal (III) salts inamounts equal to a metal/ligand molar ratio of 1:3. At themoment of mixing of the solutions a precipitate wasobtained. The reaction mixture was stirred with anelectromagnetic stirrer at 25 °C for 1 h. The products thusobtained were separated from the solutions at pH 4–5,filtered off, washed three times with water and dried in adesiccator to constant weight.

The complexes were insoluble in water, methanol andethanol but had good solubility in DMSO. The complexeswere characterized by elemental analysis. They wereanalysed for their metal content using standard proce-dures after destroying the organic matter. The ligandcontained sodium ions. The complexes obtained wereanalysed for sodium ions by means of flame photometryand the analyses revealed a lack of these ions. Theprobable reason for this phenomenon was the hydrolysisof the salt in an aqueous solution. The water content inthe complexes was determined by Karl Fisher analysis.The formation of the complexes was confirmed by IR-and1H-NMR-spectroscopy.

Table I shows the results of the Karl Fisher analysis.All complexes behaved as non-electrolytes (λm < 14Ohm–1 cm2 mol–1).

3. Pharmacology

3.1. Colorimetric MTT (tetrazolium) assay

This assay is based on the cellular reduction of MTT[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide] by the mitochondrial dehydrogenase of viablecells to a blue formazan product. The production offormazan can be measured spectrophotometrically fol-lowing solubilization. We use the method described byMossmann [17] with some modifications.

P3HR1 Burkitt lymphoma, CML derived K-562 andAML derived THP-1 cells were seeded in 96-well plates(100µL/well at a density of 1× 105 cells/mL) andexposed to various concentrations of the tested com-pounds. After incubation for 48 or 72 h, the 5 mg/mLMTT solution in PBS was added to each well and wasfurther incubated for 4 h at 37 °C. The formazan crystalsformed were dissolved by adding 100µL/well of 5%formic acid in 2-propanol. After a few minutes at roomtemperature to ensure that all crystals were dissolved,absorption was measured by an ELISA reader, using atest wavelength of 580 nm. For each concentration atleast 8 wells were used (tables II–IV). Cell growthinhibition was calculated according to [OD of drugtreatment/OD of control]× 100. Data processing wasexecuted with Microsoft Excel, Microsoft Word andSigma Plot Windows.

4. Results and discussion

The analytical data of the complexes (table I) confirmthe composition Ln (HN)3.nH2O. The mode of bondingof the ligands to Ce(III), La(III) and Nd(III) ions waselucidated by recording the IR-spectra of the complexesas compared to those of the free ligands. IR-spectra of thecompounds were recorded on solid state in Nujol in therange 3 600–400 cm–1 (table V).

Figure 1. Structure of Niffcoumar sodium salt.

Table I. Elemental analyses of the complexes of Niffcoumar.

Complex M.p. (°C) H2O

Ce(HN)3.4H2O 270 5.935.68

La(HN)3.4H2O > 300 6.055.68

Nd(HN)3.6H2O > 300 7.958.26

HN = C19H15NO6

854

Table II. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on P3HR1 cells in comparison with theinorganic salts.

Compound MTT-formazan absorption at 580 nm

untreated control 25µM 100 µM 400 µM

Ce(NO3)3.6H2O 0.3936± 0.0672 0.4835± 0.0419 0.4216± 0.0511 0.3649± 0.0823La(NO3)3.6H2O 0.3936± 0.0672 0.4429± 0.0579 0.4258± 0.0801 0.4124± 0.0775Nd(NO3)3.6H2O 0.3936± 0.0672 0.4875± 0.0885 0.4843± 0.0245 0.3613± 0.0955Ce(HN)3.4H2O 0.8167± 0.0565 0.7270± 0.0270 0.7346± 0.1086 0.6410± 0.0163La(HN)3.4H2O 0.8167± 0.0565 0.7454± 0.0363 0.7154± 0.0939 0.6494± 0.1234Nd(HN)3.6H2O 0.8167± 0.0565 0.8485± 0.0732 0.7710± 0.0712 0.6531± 0.1327

Table III. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on K-562 cells in comparison with theinorganic salts.

Compound MTT-formazan absorption at 580 nm

untreated control 25µM 100 µM 400 µM

Ce(NO3)3.6H2O 0.4694± 0.030 0.3846± 0.040 0.3502± 0.010 0.3283± 0.020La(NO3)3.6H2O 0.4694± 0.030 0.3906± 0.020 0.3342± 0.050 0.2068± 0.010Nd(NO3)3.6H2O 0.4694± 0.030 0.3895± 0.030 0.3485± 0.030 0.2987± 0.020Ce(HN)3.4H2O 0.3168± 0.020 0.3791± 0.030 0.2897± 0.030 0.2390± 0.020La(HN)3.4H2O 0.3168± 0.020 0.4971± 0.030 0.3975± 0.070 0.1382± 0.020Nd(HN)3.6H2O 0.3168± 0.020 0.3720± 0.040 0.4192± 0.030 0.1538± 0.020

Table IV. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on THP-1 cells in comparison with theinorganic salts.

Compound MTT-formazan absorption at 580 nm

untreated control 25µM 100 µM 400 µM

Ce(NO3)3.6H2O 0.3745± 0.080 0.4416± 0.010 0.4312± 0.030 0.4230± 0.020La(NO3)3.6H2O 0.3745± 0.080 0.4078± 0.010 0.4217± 0.030 0.4026± 0.040Ce(HN)3.4H2O 0.3745± 0.080 0.3855± 0.010 0.3828± 0.030 0.2746± 0.020La(HN)3.4H2O 0.3745± 0.080 0.3968± 0.020 0.3763± 0.060 0.2871± 0.008

Table V. IR spectral data of Niffcoumar, Niffcoumar sodium salt and La(HN)3.4H2O.

Compound νOH νC=O νC=O νC=C νarom. νC–OH

(lacton)

C19H15NO6 3 295 1 686 1 619 1 572 1 512, 1 456 1 173C19H14Na NO6 3 582 1 715 1 653 1 595 1 516, 1 456 1 163La(HN)3.4H2O 3 432 1 696 1 651 1 597 1 514, 1 446 1 181

Table VI. 1H-NMR spectral data (100 MHz; DMSO-d6) of Niffcoumar sodium and its complexes.

Substance H5–H8 H9 H10 H12 H14, 15

NaN 6.95–7.40 m 5.05 t 3.40 d 3.05 s 7.60–8.10 dCe(HN)3.4H2O 6.90–7.40 m 4.40 s 3.35 b s 1.90 s 7.50–8.00 dLa(HN)3.4H2O 6.90–7.35 m 4.35 t 3.30 b d 2.05 s 7.50–8.10 dNd(HN)3.6H2O 7.20–7.60 m 6.80 s 3.50 b s 1.90 s 7.50–7.90 d

NaN = C19H14NaNO6 (Niffcoumar sodium salt); HN = C19H15NO6

855

4.1. IR-spectra of Niffcoumar sodium salt complexes

The bands appearing in the IR spectrum of the freeligand at 3 582, 1 715, 1 653, 1 595, 1 516, 1 456, 1 163were shifted in the complexes (table V). The weak bandobserved at 3 582 cm–1 in the spectra of the free ligandshifted to a lower wave number in the complexes. A broadband characteristic ofνOH of crystalline or coordinatedwater was observed at 3 432 cm–1 in the spectra of all thecomplexes and had higher intensity than the one of thefree ligand. This was attributed to the presence ofcrystalline water. A band at 1 715 cm–1 in the spectra ofthe ligand could be attributed to the stretching vibrationsof the carbonyl group in the lacton ring. The band wasstretched from 20–30 cm–1 to lower wave number values.This might be taken as evidence for the participation ofthe>C=O group in the coordination process. The band at1 653 cm–1 corresponded to the keto-group. The coordi-nation with oxygen was likely, but we couldn’t say whichoxygens were coordinated to Ln3+ ions. We supposed thatthe lacton- and keto-carbonyl oxygen atoms of Niffcou-mar were bonded to the metal ion probably as a bidentateligand. The band at 1 595 cm–1 could be related to thestretching vibrations of the conjugated olefinic system.The vibrations in the region 1 516–1 456 cm–1 corre-sponded to the aromatic system. The band at 1 163 cm–1

assigned asνC-OH was observed at more or less the sameposition in the complexes and it might be due to theexistence of an OH group in all cases. The data of theIR-analyses confirmed the compositions presented intable I.

4.2. 1H-NMR analysis

Proton spectra of the compounds, recorded at 100 MHzin DMSO-d6 as a solvent, confirmed the formation of thecomplexes. The chemical shifts are given inδ-scale. Thetypical chemical shifts of the1H-NMR spectra inDMSO-d6 solvent are shown intable VI. The typicalvalues of coupling constants in Hz were: H9, 7.0; H10,7.0; H14, 8.1 and H15, 8.1. The chemical shifts of the H9and H10 protons varied in the lanthanide complexesbecause of the shift properties of these metals. It wasevident that concerning the La(III) and Ce(III) complexesthere was an observable weak negative shift effect on theH9 and H10. The effect in the Nd(III) complex was morecomplicated. A strong positive shift effect for H9 and aweaker one for H10 existed.

5. Conclusion

The experimental data presented infigures2–4 showthat the lanthanide complexes possess certain antitumour

Figure 2. A. Cytotoxic effect of the cerium complex ofNiffcoumar on P3HR1 cells after 48 h incubation.B. Cytotoxiceffect of the lanthanum complex of Niffcoumar on P3HR1 cellsafter 48 h incubation.C. Cytotoxic effects of the neodymiumcomplex of Niffcoumar on P3HR1 cells after 48 h incubation.

A

B

C

856

potential against the three leukaemic cell lines used. Infigure 2 the cytotoxic profile of cerium, lanthanum andneodimium complexes is presented. All three compoundscause similar low cytotoxic effects, even at the highestconcentration applied (400µM, about 20% growth inhi-bition) after 48 h incubation. Infigure 3the effects of thecomplexes against the relatively resistant cell line K-562after 72 h incubation are summarized. It is noteworthythat the lanthanum and neodimium compounds exertmore pronounced cytotoxic effects in comparison tocerium (figure 3B and 3C). Furthermore the cerium andlanthanum complexes were tested on AML derivedTHP-1 cells. Both compounds induce similar low cyto-toxic effects, but only at the highest applied concentration(figure 4A and 4B). The corresponding lanthanide saltsare found to have very low or no activity as shown in

Figure 3. A. Cytotoxic effect of the cerium complex ofNiffcoumar on K-562 cells after 72 h incubation.B. Cytotoxiceffect of the lanthanum complex of Niffcoumar on K-562 cellsafter 72 h incubation.C. Cytotoxic effect of the neodymiumcomplex of Niffcoumar on K-562 cells after 72 h incubation.

A

B

CFigure 4. A. Cytotoxic effects of the cerium complex ofNiffcoumar on AML derived THP-1 cells after 48 h incubation.B. Cytotoxic effect of the lanthanum complex of Niffcoumar onAML derived THP-1 cells after 48 h incubation.

A

B

857

tables II–IV. So far we can conclude that the metaldetermines the antitumour spectrum of the newly synthe-sised and characterised complexes.

The LD50 has been estimated for Niffcoumar and itscerium complex. When administered i.p. (single dose) theLD50 of Niffcoumar is 600.0–800.0 mg/kg b.w. and theLD50 of its complex is 312.5–625.0 mg/kg b.w. In both ofthe cases the highest dose induces 100% mortality. Itmeans that the molar toxicity of the complex is approxi-mately 4.5-fold higher than Niffcoumar toxicity.

6. Experimental protocols

6.1. Chemistry

The carbon, hydrogen and nitrogen content of thecompounds were determined by elemental analysis. TLC:Al-foils Kieselgel 60 F254, Merck (FRG), developed bycyclohexane-chloroform-acetic acid (10:10:4, vol. parts),detection: UV 254 nm. M.p., Büchi 510 apparatus (Swit-zerland), uncorrected. The water content was determinedusing a Metrohn Herizall E55 Karl Fisher Titrator and thepresence of sodium ions was checked by means of flamephotometry. Conductometric measurement was carriedout at 25 °C on 10–3 M solutions in DMSO, by using aMetrohm 660 AG CH-9101 Herisau conductometer witha platinum electrode and a cell having a constant of0.79 cm–1. IR spectra (Nujol) were recorded on anIR-specrometer FTIR-8101M Shimadzu.1H-NMR spec-tra were recorded at room temperature on a Brucker WP100 (100 MHz) spectrometer in DMSO-d6. Chemicalshifts are given in ppm.

6.2. Biological evaluation

A laminar flow cabinet (Haereus HV 2436) and a CO2

gassed incubator (Haereus BB16) were used for cellculture. An ELISA-reader (Labsystems Uniskan I) wasused for MTT-assay measurements.

References

[1] Patonay T., Litkei G., Bognar R., Erdei J., Miszi C., Pharmazie 39(2) (1984) 84–91.

[2] Sharada L.N., Ganorkar M.C., Indian J. Chem. 27A (6) (1988)542–544.

[3] Bharat K.B., Jaya T.R.V., Ranabaore V., J. Archaeol. Chem. 4 (1986)35–37.

[4] Sharada L.N., Ganorkar M.C., Indian J. Chem. 27A (2) (1988)163–165.

[5] Kumar B.B., Raju V.J., Ranabaore V., Ganorkar M.C., Orient J.Chem. 3 (1) (1987) 34–39.

[6] Deng R., Wu J., Long L., Bull. Soc. Chim. Belg. 101 (6) (1992)439–443.

[7] Pokhariyal G.P., Proc. Natl. Acad. Sci. India 58A (3) (1988)369–373.

[8] Pokhariyal G.P., Indian J. Chem. 28A (10) (1989) 922–923.

[9] Castellani C.B., Carugo D., Inorg. Chim. Acta 159 (2) (1989)157–161.

[10] Dutt N.K., Sarma U.U.M., Indian J. Chem. 15A (4) (1977) 361–362.

[11] Kostova I., Manolov I., Konstantinov S., Karaivanova M.H., Eur. J.Med. Chem. 34 (1999) 63–68.

[12] Meyer B.N., Wall M.E., Wani M.C., Teylor H.L., J. Nat. Prod. 48 (6)(1985) 952–956.

[13] Baraldi P.G., Barco A., Benetti S., Guarner M., Manfredini S.,Pollini G.P., Simoni D., Tetrahedron Lett. 26 (43) (1985)5319–5322.

[14] Harvey R.G., Corter C., Ananthanarayan T.P., Schmolka S., J. Org.Chem. 53 (17) (1988) 3936–3943.

[15] Haiduc I., Silvestru C., Coord. Chem. Rev. 99 (1990) 253–296.

[16] Anghileri L.J., Crone-Escanye M.C., Anticancer Res. 7 (1987)1205–1208.

[17] Mossmann T., J. Immunol. Methods 65 (1983) 55–63.

858

Short communication

Vasodilating and antiarrhythmic activity of heteryl lactones

Ludmila Leitea*, Daina Jansonea, Maris Veverisa, Helena Cirulea, Yuris Popelisa,Gagik Melikyanb, Anna Avetisyanb, Edmunds Lukevicsa

aLatvian Institute of Organic Synthesis, 21 Aizkraukles Str., LV-1006 Riga, LatviabDepartment of Organic Chemistry, Yerevan State University, Yerevan 375049, Armenia

(Received 27 October 1998; revised 22 March 1999; accepted 23 March 1999)

Abstract – A new series of unsaturatedγ- andδ-lactones with pyridyl, quinolyl and nitrophenyl substituents (9, 10) have been synthesizedby the condensation of unsaturated methyl lactones with heteryl aldehyde or nitrobenzaldehyde in the base-catalysed aldol reaction. Theantiarrhythmic, vasodilating and cardiotonic activities of the synthesized compounds have been studied in vivo and ex vivo. 3-Cyano-5,5-dimethyl-4-[4'-(4-pyridyl)-1',3'-butadienyl)]-2(5H)-furanone (9e) displayed a significant vasodilating activity. The antiarrhythmic activity ofthis compound was higher, but its toxicity lower than that of the procainamide reference drug. Five-membered lactones, particularly3-cyano-4-(4-pyridylvinyl)-5,5-dimethyl-2(5H)-furanone (9c), exhibited a remarkable cardiotonic activity. The replacement of a pyridylsubstituent by a nitrophenyl group in the pyranone derivative did not change the cardiovascular activity and toxicity. © 1999 E´ditionsscientifiques et médicales Elsevier SAS

heteryl lactones / antiarrythmic activity / vasodilating activity / cardiotonic activity

1. Introduction

Some of the synthetic and natural compounds contain-ing an unsaturated five-membered lactone moiety exhibita cardiotonic activity [1]. For example, the heart glyco-sides contain aγ-lactone unit. On the other hand, itshould be noted that the majority of cardiotonic sub-stances of a novel generation is derived from N-hetero-cyclic compounds [2]. To determine the role of lactoneunits and heteryl substituents as pharmacophores respon-sible for antiarrhythmic and cardiotonic activity, pyridyl,quinolyl and nitrophenyl derivatives of unsaturatedγ-andδ-lactones have been prepared and tested in vivo andex vivo.

2. Chemistry

Pyridyl, quinolyl and nitrophenyl derivatives of unsat-uratedγ- and δ-lactones were synthesized according tothe procedure elaborated for the pyridyl derivatives offuranone and pyranone [3]. Methyl lactones7 and8 were

condensed with aldehydes (1–6) in the presence of NaOHas a catalyst(figure 1). The unusual reactivity of 2-, 3-and 4-pyridinecarboxaldehyde withγ- andδ-lactones in abase-catalysed aldol reaction was shown earlier [3]. Itwas found that the synthesis of heteryl lactones9a–cand10a–cwas accompanied by the formation of the [bis(2-oxo-3-cyano-5,5-dimethyl-2(5H)-furanyl-4-methyl)-methyl]-pyridines and [bis(2-oxo-3-cyano-6,6-dimethyl-5,6-di-hydropyranyl-4-methyl)-methyl]pyridines. To avoid thisaddition we used a 2-fold excess of pyridine carboxalde-hyde. The aldol condensation reaction of aldehydes4–6with lactones occurred traditionally and yielded thecorresponding unsaturated derivatives of lactones only.Lactone 9d was synthesized by the reaction of theethoxycarbonyl derivative of methyl furanone7b withaldehyde2. The ester was converted into the sodium saltby the additional amount of sodium hydroxide. Thereaction of aldehyde4 with furanone7a was carried outin the presence of a catalytic amount of NaOH at roomtemperature to give furanone9e in 21% yield. This wassignificantly lower than in the case of aldehyde3 con-densation reaction (yield 54%). Raising the temperaturedid not increase the yield. In all cases the values of*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 859−865 859© 1999 Editions scientifiques et médicales Elsevier SAS. All rights reserved

spin-spin coupling constants of the double bond CH=CHprotons (16–17 Hz) indicated that onlyE- or E,E-isomerswere isolated. The coupling constant ofα,â-protons inchain CH=CHα–CHâ=CH (compound 9e) equalled10.7 Hz. A similar value of coupling constant (10.4 Hz) isin agreement with theE-isomer of butadiene [4]. Toobtain an additional proof for the configuration of theunsaturated synthesized compounds,E-Z photoisomer-ization caused by UV irradiation of the condensationproducts9a and10b, d ande was studied. The isomer-ization process was controlled by electron absorptionspectroscopy. When the solution in ethanol was irradiatedwith UV light the intensity of theE-isomer absorptionband at 332–344 nm decreased and a band at 220–285 nmappeared. According to NMR data the formation of theE- andZ-isomer mixture caused such changes in spectra.

The spin-spin coupling constants of the double bondCH=CH protons in isomers appeared during irradiationand were 11–13 Hz. This means that the compoundsdescribed in the present article really haveE-configuration. The characteristics of the compounds arelisted in tables Iand II.

3. Results and discussion

The proposed cardiovascular activity of newly synthe-sized compounds was tested in ex vivo experiments onthe isolated rabbit ear artery and in vitro on anaesthetizedlaboratory animals. These models were chosen in ordernot to miss the compounds with antiarrhythmic, vasodi-lating and/or cardiotonic activities.Table III summarizesthe data of the antiarrhythmic screening test and acute

Figure 1. Synthesis of heteroaryl lactones.

860

toxicity obtained for heteryl lactones. The acute toxicityof the studied compounds was low (LD50 was over400 mg/kg). Only the acute toxicity of compound9c(LD50 180 mg/kg) and9d (200 mg/kg) was comparablewith that of lidocaine (LD50 238 mg/kg). The five-membered pyridyl lactones9 a–ehad a toxicity similar toits six-membered analogues except 3- and 4-pyridylfuranones, the toxicity of which was about 2-foldhigher than that of the substituted pyranones. The re-placement of pyridyl substituents for a quinolyl groupappeared slightly less toxic (see9c and f, 10c andd).

Heteryl derivatives9c, e and10d injected i.p. in dosesof 15 and 30 mg/kg caused the statistically significantprotection against CaCl2-induced arrhythmia. Their ac-tivity was higher but toxicity lower than that of thereference drug procainamide. The antiarrhythmic activitywas more marked for heteryl lactones with 4-pyridylsubstituent within the furanone series.

Compounds9b andc were also investigated on anaes-thetized rats. It was shown(table IV) that these pyridylfuranones decreased the number of animals with lethalarrhythmia induced by calcium chloride, but to a lowerextent than lidocaine. Compound9c (dose = 1 mg/kg,

i.v.) also protected against ventricular ectopic beats. So,we established that the 4-isomer of pyridyl furanone,9c,revealed the most pronounced antiarrhythmic activity andthe highest toxicity among all compounds studied(ta-ble III).

Almost all of the studied heteryl derivatives of unsat-uratedγ-and δ-lactones caused a more or less markedrelaxation of the previously contracted vessels of therabbit ear (table V). Compound 9e exhibited a pro-nounced vasodilating activity. At the same time, pyridyllactone 9a demonstrated two phases of vasodilation/vasoconstriction but9b showed a trend towards vasocon-striction. In this test, six-membered lactones (10a, b andc) possessed more pronounced vasodilation than five-membered ones (9a, b and c) and 10a (2-pyridyl) wasmore active than10b and c (3- and 4-isomers). In theseries of pyridyl lactones, we found a tendency of2-pyridyl substituents to induce a more selective vasodi-lation in comparison with 3- and 4-pyridyl substituents,as already stated with calcium channel modulators [5].Compounds9e, 9f and10d induced a significant vasodi-lation at a concentration of 50µM (9e being the mostactive). This means that these compounds may show a

Table I. Reaction conditions and characteristics of compounds9 d–f and10 d ande.

Compound Reaction temperature (°C) Reaction time (h) Yield (%) M.p. (°C) Formulaa

9d 80 4 52 275-280 C14H12NO4Na9e 80 2 21 161-163 C16H14N2O2

9f 20 1.5 41 229-230 C18H14N2O2

10d 20 3 44 222-224 C19H16N2O2

(dec)10e 80 1.2 61 224-226 C16H14N2O4

a all compounds were analysed for C, H and N.

Table II. 1H NMR chemical shifts (δ, ppm) and spin-spin coupling constants of compounds9d–f and10d ande.

Compound CH=CH (JHH Hz) Pyrone CH2 Pyrone orfuranone CH3

Phenyl, pyridyl or quinolyl CH (JHH Hz)

9d* 7.52, d (16.6); 7.31, d (16.6) - 1.54, s 8.74, d, H-2, (1.4); 8.49, dd (4.2, 1.4); 8.00,dt, H-4, (8.2, 1.4); 7.40, dd, H-5 (8.2, 4.2)

9e** 6.54, d (15.7); 7.01, d (15.7); 7.14, dd(10.7, 15.7); 7.55, dd (10.7, 15.7)

- 1.65, s 7.34, m, H-3, 5 (6.5); 8.66, m, H-2, 6 (6.1)

9f** 7.06, d (16.4); 8.60, d (16.4) - 1.75, s 7.65, d, H-3, (4.4); 7.70, m, H-6 (1.4; 6.9;8.5); 7.82, m, H-7 (1.4; 6.9; 8.5); 8.10, d, H-5(8.5); 8.20, d, H-8 (8.5); 9.01, d, H-2 (4.4)

10d* 7.54, d (15.7); 8.42, d (15.7) 3.35, s 1.50, s 7.74, m, H-6, (1.2, 7.8, 7.9); 7.86, m, H-7(1.2, 7.8, 7.9); 7.93, d, H-3 (4.6); 8.11, d, H-5(7.8); 8.50, d, H-8 (8.0); 9.00, d, H-8 (8.0);9.00, d, H-2 (4.6)

10e** 7.24, d (16.0); 7.59, d (16.0) 2.86, s 1.55, s 7.70-8.35, m, H-4

DMSO-d6,** CDCl3

861

strong TXA2 receptor antagonist activity in the rabbitblood vessels. The replacement of a 4-pyridyl substituentof furanone and pyranone (9c, 10c) for a 4-quinolyl group(9f, 10d) patently increased vasodilating properties inboth cases. Unlike pyridyl derivatives of furanone andpyranone a nitrophenyl pyranone10edid not possess thevasodilating activity.

Cardiotonic and vasodilating properties of heteryl lac-tones were also studied on anaesthetized cats. Maximumvariations in the haemodynamic parameters were regis-tered about 1–5 min after i.v. administration. The resultsobtained are shown intable VI. The compounds9b andc possess a significant cardiotonic activity. Compound9c(dose = 0.5 mg/kg) increased the systolic pressure of theleft ventricle by 30%, but dP/dt by 32%, and the mean

arterial blood pressure by 18%. At the same time, theheart rate remained unchanged. The positive inotropiceffect was observed with9b and 9c at doses of0.1–2.0 mg/kg (duration of action was slightly shorter for

Table III. Antiarrhythmic activity and acute toxicity of pyridyllactones in mice.

Compound Dose(mg/kg, i.p.)

Antiarrhythmicactivity(Y/N)a

Acute toxicityLD50(mg/kg, i.p.)

9a 30 1/5 > 40090 2/5

9b 30 1/5 > 40090 1/5

9c 15 3/5* 180 (138.5-234)30 3/5*

9d 15 1/5 200 (153.8-260)30 2/5

9e 15 3/5* 480 (347.8-662.4)30 2/590 2/5

9f 30 1/5 > 60090 2/5

10a 30 1/5 > 40090 2/5

10b 30 1/5 > 40090 1/5

10c 30 1/5 > 40090 1/5

10d 15 1/5 > 60030 3/5*90 2/5

10e 30 1/5 ≥ 60090 1/5

Procainamide 30 1/5 360 (257-504)90 3/5*

Lidocaine 15 2/5 238 (180.3-314)30 4/5*

Control 0/10

a Y, the number of mice protected against CaCl2-induced arrhyth-mia; N, total number of experimental animals; *P < 0.05 vs. controlgroup.

Table IV. Effect of compounds9b and c on CaCl2-induced ar-rhythmia and lethality in rats.

Compound Dose(mg/kg, i.v.)

Arrhythmia-scores(M ± SD)

Survival(%)

9b 0.3 3.3± 0.4 201.0 3.0± 0.5 403.0 2.8± 0.5 60*

9c 0.3 2.9± 0.5 401.0 2.3± 0.3* 60*3.0 2.6± 0.5 60*

Lidocaine 0.3 3.0± 0.4 201.0 2.1± 0.3* 60*3.0 1.8± 0.4* 80*- 3.8± 0.5 0

*P < 0.05 vs. control.

Table V. Vasodilating activity of the investigated compounds inrabbit ear artery.

Compound Concentration (µM) Relaxationa (%)

9a 10 5÷(-3)50 10÷(-5)

9b 10 -250 -8

9c 10 -350 -5

9e 10 30*50 50*

9f 10 2050 30*

10a 10 1250 27*

10b 10 1550 20

10c 10 250 15

10d 10 1950 32*

10e 10 550 15

Papaverine 1 810 35*50 76*

Control - 0Solvent - 8

a positive value means relaxation; negative one, vasoconstriction;*P < 0.05 vs. control.

862

9b). However,9b caused a more marked change in heartrate (within +5 to –10%), compared with9c. Both pyridylfuranones9b andc similarly increased the blood flow inthe femoral and carotid arteries, but this increase was lessthan the change in the blood pressure.

The vasodilating effect of a nitrophenyl substituent wassimilar to that of a pyridyl one. The vasodilating activityof pyranone with a nitrophenyl substituent is similar tothat of pyridyl derivatives of furanone and pyranone.

4. Conclusions

New derivatives ofγ- and δ-lactones with pyridyl,quinolyl and phenyl substituents withE-configurationwere synthesized, their antiarrhythmic, vasodilating andcardiotonic activities investigated. The acute toxicity ofthe compounds studied was low. On the whole, thefive-membered pyridyl lactones had a toxicity similar totheir six-membered analogues (except 3- and4-pyridylfuranones). The pyridyl substituents of pyridyllactones may be responsible for a higher toxicity incomparison with quinolyl lactones.

Pyridyl and quinolyl lactones caused a significantprotection against CaCl2-induced arrhythmia. The activ-ity depended on the position of a substituent in the

heterocycle. The most active were 4-isomers – 3-cyano-4-(4-pyridylvinyl)-5,5-dimethyl-2(5H)-furanone9c and3-cyano-4-(4-quinolylvinyl)-6,6-dimethyl-(5,6-dihydro)-2-pyranone10d. The replacement of a 4-pyridyl substitu-ent of furanone for a 4-quinolyl group decreased theantiarrhythmic activity. The similar replacement in thecase of 4-pyridylpyranone caused the activity increase.

The vasodilating activity depended on the ring size ofthe lactone, on the type of heteryl substituent, and on theposition of the substituent in the ring. Six-memberedlactones revealed more pronounced vasodilation thanfive-membered ones. 2-Pyridylfuranone and 2-pyridyl-pyranone demonstrated more marked vasodilation than 3-and 4-isomers. Besides, this 2-pyridylfuranone also pos-sessed a vasoconstricting action. The insertion of thesecond vinyl group in the aliphatic chain between thepyridine and lactone rings increased the activity ofpyridyl furanone. Thus 3-cyano-5,5-dimethyl-4-[4'-(4-pyridyl)-1',3'-butadienyl)]-2(5H)-furanone9e was themost active among the compounds studied.

The vasodilating effect of a nitrophenyl substituent issimilar to that of a pyridyl substituent. Cardiotonicactivity is more pronounced for 3- and 4-isomers ofpyridyl furanones. Cardiotonic activity of the correspond-ing 6-membered lactones is lower.

Table VI. Haemodynamic effects. Results are in % (in reference to the control initial value).

Compound Dose (mg/kg, i.v.) MABPa HRa LVSP dP/dt FBFa CBFa

9a 1.0 5 0 - - - 123.0 -20÷5 -5 - - - -5÷12

9b 0.1 0 -5 6 8 0 00.5 12 3÷(-10) 23 26 10 122.0 29 5 45 33 12 22

9c 0.1 3 -2 10 15 -5 30.5 18 0 30 32 12 162.0 30 6 50 38 15 20

10b 0.1 0 0 0 0 0 00.5 0 0 0 5 0 32.0 6 -3 5 8 5 5

10c 0.1 -3 0 0 0 0 00.5 5 0 5 7 0 62.0 10÷(-5) 5÷(-5) 7 10 5÷(-5) 3

Verapamil (hydrohloride) 0.02 -8 -5 -3 -3 18 150.1 -35 -12 -9 -16 45 80

MABP, mean arterial blood pressure; HR, heart rate; LVSP, left ventricular systolic pressure; dP/dt, left ventricular contractility; FBF, femoralartery blood flow; CBF, carotid artery blood flow.aHaemodynamic parameters are expressed as % change: predrug/postdrug; increase/decrease; positive value means increase; negative one,decrease.

863

5. Experimental protocols

5.1. Chemistry

5.1.1. MethodsThe 1H NMR spectra were recorded on a Bruker

WH-90/DS and on an AM-360 (360 MHz) spectrometersin CDCl3 or DMSO-d6, TMS internal standard. The massspectra were obtained on a Kratos MS-25chromatograph-mass-spectrometer with an ionizing en-ergy of 70 eV. Elemental analysis was performed on anElemental Analyzer Carlo Erba 1108. Silufol UV-254plates were used for TLC analysis, eluents – benzene-acetone 3:1. The melting points were determined on aBoetius stage and reported without corrections.E-Zphotoisomerization has been studied in ethanol solutionof compounds9a, 10b, d and e by irradiation with UVlight at 336, 332, 340 and 344 nm respectively for45 min. Electronic absorption spectra were recorded on aSpecord UV-VIS spectrometer. Starting furanones7 andpyrone 8 were synthesized according to [6] and [7],respectively. Pyridine aldehydes1–3were synthesized byvapour phase oxidation of methyl pyridines over V2O5-MoO3 catalyst [8]. Starting 1'-(4-pyridyl)acroleine4 and4-quinolinealdehyde5 were synthesized according to themethods described in [9] and [10]. All solvents were of ananalytical grade and used without further purification.

5.1.2. General procedure for the synthesis of compounds9d–f, 10d ande

A mixture of methyl lactone7 or 8 (20 mmol), thealdehyde 4–6 (20 mmol) and NaOH (1.25 mmol) inMeOH (20 mL) was stirred at room temperature orrefluxed for 1.2–4 h. The condensation products precipi-tated, and were filtered off after cooling to room tempera-ture. The precipitates of compounds9e, f, 10dandewererecrystallyzed from EtOH.

Compound9d was synthesized as described above butwhen the condensation was completed NaOH (21 mmol)was added to the cooled mixture. The resulting precipitatewas filtered off, washed with EtOH and air-dried. Thereaction conditions, yields and NMR data are shown intables I and II. Procedures for the synthesis of com-pounds9a–cand10a–cand their chemical and physicaldata are given in [3].

5.2. Pharmacological methods

5.2.1. Experiments ex vivoThe modified method for the experiments on the

isolated perfused rabbit ear blood vessels was used [11,12]. Male and female rabbits (2.6–3.3 kg) were eutha-nazed by i.v. injection of Na pentabarbital (80 mg/kg).

The central ear artery was dissected free at the ear baseand cannulated with polyethylene tube and perfused at aconstant flow from a 4-channel peristaltic pump, Geminy(Italy). The perfusion fluid was (mmol): NaCl 136.9; KCl2.68; CaCl2 1.8; MgCl2 1.05; NaHCO3 11.9;NaH2PO4 0.42; glucose 5.6 (pH 7.35). Intraluminal in-flow perfusion pressure was measured with a StathamP23I transducer and recorded on a physiograph DMP-4B(Narco Bio-System, Huston, USA). As flow remainedconstant, the alterations in perfusion pressure reflectedchanges in the blood vessel resistance, i.e., the degree ofvasoconstriction or relaxation. Vasoconstriction wascaused by intraluminal infusion of U-46619 (thrombox-ane A2 receptor agonist). The relaxant responses to theinvestigated compounds used in different concentrationswere tested. The responses were expressed as a per centrelaxation (% changes in perfusion pressure) without andwith the investigated compounds or solvent.

5.2.2. Experiments in vivo

5.2.2.1. Antiarrhythmic activity

5.2.2.1.1. Antiarrhythmic screening testAntiarrhythmic activity was tested in the experimental

antiarrhythmic screening model on male ICR:JCL mice(19–23 g) as was described earlier [13]. The tested com-pounds or solvents were administered i.p. 15 min beforei.v. infusion of 2% CaCl2 solution at a constant rate(0.02 mL/sec) in a dose of 180 mg/kg. The number ofanimals protected from CaCl2-induced lethal arrhythmiawas defined. The investigated compounds were dissolvedin NaCl 0.9% solution or in dimethyl acetamide and thendiluted with the NaCl 0.9% solution.

5.2.2.1.2. Calcium chloride-induced arrhythmia.The experiments were performed according to the

classical method [14]. In brief, Wistar male rats(180–220 g) were anaesthesized with urethan (1.20 mg/kg, i.p.). ECG was registered in II standard lead on aphysiograph DMP-4B (Narco Bio-Systems, Huston,USA). The 5% CaCl2 solution was i.v. injected at a doseof 180 mg/kg. Heart rhythm disturbances in scores [15]and lethality of animals were estimated. Solutions of thecompounds to be tested were administrated into thefemoral vein 3 min prior to CaCl2. Every dose was testedon five rats.

5.2.2.2. Cardiotonic action and hypotensive activityThe method used was described already in [16]. Adult

mongrel male and female cats (3.0–4.2 kg) were anaes-thesized withα-glucochloralose and urethan (80 and200 mg/kg, i.p.). The trachea was catheterized and con-

864

nected to an intermittent positive-pressure respirator(DP-8, SU). Arterial blood pressure from the right femo-ral artery, left ventricular pressure (electromanometer P23 ID) and the first derivative (dP/dt) were registered ona physiograph DMP-4B (Narco Bio-Systems, Houston,USA). The blood flow from the left femoral and leftcommon carotid arteries was measured with an electro-magnetic flowmeter MFV-1200 (Nihon Kohden, Tokyo,Japan). The heart rate was calculated using the bloodpressure wave. The solutions of the compounds investi-gated were injected through a catheter placed in thefemoral vein.

5.2.2.3. Toxicological examination

Acute toxicity of the unsaturated five- and six-membered lactones was studied in albino male andfemale ICR:JCL mice (18–22 g). Solutions or suspen-sions of the compounds (prepared with 0.6% Tween-80)were injected i.p. The experimental animals were ob-served for 10 days. Common state, mobility of animals,toxic symptoms and survival were estimated. To reducethe number of used animals, the maximal dose was400 mg/kg. If possible, LD50 was calculated when 50%of the animals died.

All references drugs (Procainamide, Lidocaine, Vera-pamil, Papaverine) were from commercial sources.

Acknowledgements

We are grateful to the Latvian Science Council for thefinancial support of the programme ‘The Development ofModern Directions in the Organic Chemistry forthe Production of New Medical Agents’. The authors

acknowledge with gratitude and affection the financialhelp they have also received from the Taiho LatvianFoundation.

References

[1] Avetisyan A.A., Tokmadzhyan G.G., Chem. Heterocycl. Compds. 23(1987) 595–609.

[2] Yuzhakov S.D., Mastafanova L.I., Mashkovski M.D., YahontovL.N., Khim-Farm. Zh. 26 (1992) 4–17; CA 117 (1992) 123832q.

[3] Jansone D., Leite L., Fleisher M., Popelis J.U., Mazheika I.,Lukevics E., Melikyan G., Avetisyan A., Chem. Heterocycl. Com-pds. 34 (1998) 267–270.

[4] Pretsch E., Clerc T.H., Simon W., Tables of Spectral Data forStructure Determination of Organic Compounds, Springer, 1989, p.H210.

[5] Ramesh M., Matowe W.C., Akula M.R., Vo D., Dagnio L., Moy-Cheong L.K.K., Wolowyk M.W., Knaus E.E., J. Med. Chem. 41(1998) 509–514.

[6] Pavlova L.A., Belgorodski V.V., Venus-Danilova E.D., Zh. Obsh.Khim. 36 (1966) 1386–1391; CA 66 (1967) 10797f.

[7] Avetisyan A.A., Kasparyan B.K., Dzandzapanyan A.N., DangyanM.T., Arm. Khim. Zh. 36 (1983) 341–343; CA 99 (1983) 139700r.

[8] Shymanska M.V., Leitis L.J., Iovele I.G., Goldberg Ju.Sh.,Skolmeistere R.A., Golender L.O., in: Krilov O. (Ed.), Problems ofKinetics and Catalysis. XIX. Partial Oxidation of Organic Com-pounds, Nauka, Moscow, 1985, pp. 175–187.

[9] Leitis L.J., Rubina K.I., Goldberg Yu.Sh., Jansone D.P., ShymanskaM.V., Izv. AN LatvSSR. Ser. khim. 4 (1980) 469–472; CA 94 (1981)14734v.

[10] Achremowicz L., Synthetic Commun. 26 (1996) 1681–1684.

[11] Blattner R., Classen H.G., Dehnert H., Doring H.G., Experiments onIsolated Smooth Muscle Preparations, Hugo Sachs Elektronik KG(BRD), 1980.

[12] Wilson S.K., Steinsland O.S., Nelson S.H., J. Cardiovasc. Pharma-col. 23 (1994) 127–135.

[13] Lukevics E., Veveris M., Dirnens V., Appl. Organomet. Chem. 11(1997) 805–811.

[14] Yoshidomi M., Sukamoto T., Morita T., Ito K., Nose T.,Arzneim.-Forsch./Drug Res. 32 (1982) 1056–1059.

[15] Rees S.A., Tsuchihashi K., Hearne D.J., Curtis M.J., J. Cardiovasc.Pharmacol. 22 (1993) 343–349.

[16] Veveris M., Syomin R., Proc. Latv. Acad. Sci. B 9 (1992) 90–96; CA118 (1993) 139512s.

865

Short communication

New amino derivatives of 1,2,3-triazolo[4,5-d]pyrimidines andtheir affinity towards A 1 and A2A adenosine receptors

Laura Bettib, Giuliana Biagia, Gino Giannaccinib, Irene Giorgia, Oreste Livia*, Antonio Lucacchinib,Clementina Maneraa, Valerio Scartonia

aDipartimento di Scienze Farmaceutiche, Facoltà di Farmacia, Università di Pisa, via Bonanno 6, 56126 Pisa, ItalybDipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Università di Pisa, via Bonanno 6, 56126 Pisa, Italy

(Received 19 December 1998; revised 7 April 1999; accepted 21 April 1999)

Abstract – Starting from the appropriate azides (4-chlorobenzyl-, 2-thiophenemethyl-, 2-fluorobenzyl-, and 4-fluorobenzylazides) in whichthe variation of the substituent is at the basis of the four series of derivatives (a–d), the 7-aminosubstituted 1,2,3-triazolo[4,5-d]pyrimidines4 were prepared by a well known synthetic route. The biological activity of compounds4 was expected on the basis of the presence ofparticular substituents on N(7), and these substituents were introduced by the reaction of the 7 lactamic carbonyl function, present onprecursors3, with cycloalkyl-, aralkyl- and arylamines. Radioligand binding assays at bovine brain adenosine A1 and A2A receptors showedthat some compounds possessed a high affinity and selectivity for the A1 receptor subtype. Furthermore, biological results indicated that thep-chlorobenzyl substituent lowered receptor binding, compared with the previously prepared benzyl and 2-chlorobenzyl derivatives,suggesting certain particular steric requirements of the lipophilic region which interacts with the benzyl substituent. The thiophenemethylsubstituent conferred more activity than the benzyl one. The presence of a fluorine atom on the benzyl group determined a high affinity,especially when it was in theorthoposition. Compounds4c.1(R = 2-fluorobenzyl, R≠ = cyclopentyl, Ki = 10.5 nM),4c.2(R = 2-fluorobenzyl,R≠ = cyclohexyl, Ki = 19.5 nM) and4d.1 (R = 4-fluorobenzyl, R≠ = cyclopentyl, Ki = 26 nM) were the most active for A1 receptors. © 1999Éditions scientifiques et médicales Elsevier SAS

1,2,3-triazoles / 1,2,3-triazolo[4,5-d]pyrimidines / A1-adenosine and A2A-adenosine receptor antagonists

1. Introduction

Our previous studies on 7-aminosubstituted 1,2,3-triazolopyrimidines [1] indicated that certain compounds,bearing a benzyl or 2-chlorobenzyl as the lipophilicsubstituent in the 3 position, showed a high affinity andselectivity towards the A1 receptor subtype.

These and other studies of ours on 1,2,3-triazolo-pyrimidines [2–6], in accordance with the receptor bind-ing site models proposed for adenosine agonists andantagonists [7], revealed three lipophilic binding regions(corresponding to the N-6, C-2 and N-9 positions of theazapurine ring) in the A1 adenosine receptors, arranged as

a fan-shape with respect to the NH function, which wasengaged as a hydrogen bond donor.

On the basis of our previous results, the investigationsinto these structures were continued in order to study themode of binding of these compounds with A1-receptorsby comparative SAR analysis. We report here the synthe-sis and biological evaluation of a series of 1,2,3-triazolo[4,5-d]pyrimidines (4) characterized by newbenzyl-type substituents (R = 4-chlorobenzyl, 2-thenyl,2-fluorobenzyl and 4-fluorobenzyl). In the 7 position,amino substituents which conferred a high biologicalactivity to previously prepared molecules (cyclopentyl-and cyclohexylamino, meta- and para-toluidino,α-methylbenzylamino, amphetamino) as well as otheramino substituents (isopropylamino, 2-butylamino,2-pentylamino, 3-pentylamino,p-bromoanilino, furfury-lamino, 2-pyridylamino and 2-hydroxy-cyclohexyl-amino) were introduced.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 867−875 867© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

The scheme for the preparation of compounds4 startedfrom the appropriate azide, which determined the Rsubstituent (series a–d) and followed a well-experimented synthetic route (figure 1). All the azideshave been reported in the literature (4-chlorobenzylazide, [8]; 2-thiophenemethylazide, [9];2-fluorobenzylazide [10]; and 4-fluorobenzylazide [11])but preparation of 2-thiophenemethylazide has not beendescribed.

The appropriate azide was reacted with cyanoaceta-mide in the presence of sodium ethoxide in refluxingethanol to give the 1-substituted-4-carboxamido-5-amino-1H-1,2,3-triazoles1a–d in good yields. Thesecompounds, by heating with formamide, easily cyclizedto the corresponding 7-hydroxy-triazolopyrimidines2a–dwhich, by reaction with thionyl chloride in chloroform,provided the expected chloroderivatives3a–d. Com-pounds 1a, 2a and 3a have been described in theliterature [8]. The chlorine atom in the 7 position wasreactive enough to undergo a nucleophilic displacementby primary aliphatic or aromatic amines (consequently,several triazolopyrimidine derivatives4 were prepared,corresponding to the four series characterized by the

lipophilic substituent in the 3 position:a (4-chlorobenzyl),b (2-thiophenemethyl),c (2-fluorobenzyl)andd (4-fluorobenzyl) (table I).

The structures of all the new compounds were assignedon the basis of the well-known reaction mechanismsalready regularly carried out in our laboratory: 1,3-dipolar cycloaddition of azides to activated methyleniccompounds, formation of the pyrimidine ring, chlorina-tion and nucleophilic displacement of the halogen byamines. The structures were also confirmed by analyticaland spectroscopic data.

3. Biochemistry

The 7-(aminosubstituted)-1,2,3-triazolo[4,5-d]pyrimi-dines were tested in radioligand binding assays foraffinity at A1 and A2A adenosine receptors in bovine braincortical membranes and in bovine brain striatal mem-branes respectively. [3H]R-(-)-N6-cyclohexyl-adenosine(CHA) was used as the A1 radio-ligand and [3H]-2-{[[p-(2-carboxyethyl)phenyl]ethyl]amino}-5≠-(N-ethylcarba-moyl)adenosine (CGS 21680) as the A2A radioligand.The experimental details of the receptor binding assayswere reported in a previous paper [1].

4. Results and discussion

The results of the A1 and A2A adenosine receptorbinding assays, expressed as inhibition constants (Ki,nM) in table II show that introduction of thep-chlorobenzyl substituent on the triazole ring in the 3position (compounds4a.1–7) reduces A1 and A2A ad-enosine receptor binding compared with the2-chlorobenzyl substituent [1]. The azapurine compound4a.1, a cyclopentylamino derivative, is an exception, withan A1 affinity constant of Ki = 128 nM, followed by thecyclohexylamino derivative4a.2 with a Ki = 1 360 nM.Both compounds showed a high receptor selectivity(table II); the inhibition percentages of the A2A adenosinereceptor binding assays were so low that the correspond-ing Ki values were not calculated. The shift of thechlorine from theortho to thepara position suggested amoderate and well-defined extent of the lipophilic regionwhich receives the benzyl substituent. Whereas theortho-chlorine atom on the benzyl group increased the affinitytowards the A1 receptor [1], the results of the4a seriesindicated that the chlorine atom in thepara positiondecreased the capacity to bind with the receptor site. Thisfact could mean that thepara-chloro substitution gener-ally caused a steric repulsion within the receptor, owingto the limited depth of the lipophilic site.

Figure 1. Synthetic route for compounds4.

868

Table I. Physicochemical data of triazolopyrimidine derivatives and intermediates.

Compound Yield% crystall. M.p. °C Molecular formula(molecular weight)

MS m/zsolvent M+ base peak

1b 56 EtOH 197–198 C8H9N5OS (223.25) 223 971c 73 EtOH 186–188 C10H10N5OF (235.22) 235 1091d 87 EtOH 221–223 C10H10N5OF (235.22) 235 1092b 59 EtOH 225–230 C9H7N5OS (233.25) 233 2042c 56 EtOH 221–224 C11H8N5OF (245.22) 216 (M+-29) 1092d 52 EtOH 225–228 C11H8N5OF (245.22) 245 1093b 27 60–80 °C petr. ether 79–80 C9H6N5SCl (251.69) 251 973c 63 60–80 °C petr. ether 85–88 C11H7N5FCl (263.66) 263 1093d 61 60–80 °C petr. ether 74–76 C11H7N5FCl (263.66) 263 1094a.1 18 MeOH 115–116 C16H17N6Cl (328.80) 328 1254a.2 36 60–80 °C petr. ether 109–111 C17H19N6Cl (342.83) 342 1254a.3 65 EtOH 110–112 C19H17N6Cl (364.84) 364 1254a.4 50a EtOH 158–161 C20H19N6Cl-HCl (415.33) 379 1254a.5 83 EtOH 175–176 C18H15N6Cl (350.81) 350 1254a.6 84 EtOH 163–164 C18H15N6Cl (350.81) 350 1254a.7 16 MeOH 115–117 C17H12N7O2Cl (381.78) 316 (M+-65) 1254b.1 49 MeOH 120–122 C14H16N6S (300.38) 300 2034b.2 42 MeOH 132–135 C15H18N6S (314.41) 314 974b.3 10a EtOH 148–152 C18H18N6S-HCl (386.90) 259 (M+-91) 974b.4 18a MeOH 139–142 C17H16N6S-HCl (372.88) 336 974b.5 56 EtOH 149–151 C16H14N6S (322.39) 322 974b.6 73 EtOH 170–172 C16H14N6S (322.39) 322 974c.1 46 MeOH 109–112 C16H17N6F (312.35) 312 1094c.2 71 MeOH 107–109 C17H19N6F (326.38) 326 1094c.3 27a EtOH 159–162 C20H19N6F-HCl (398.87) 362 1094c.4 b 42a EtOH 181–185 C20H19N6F-HCl (398.87) 271 (M+-91) 1094c.5 54 EtOH 139–142 C18H15N6F (334.36) 334 1094c.6 71 MeOH 139–141 C18H15N6F (334.36) 334 1094c.7 44 EtOH 153–156 C17H19N6OF (342.38) 342 1094c.8 84 EtOH 170–173 C16H13N6OF (324.32) 324 1094c.9 52 EtOH 196–198 C17H12N6BrF (399.22) 399 1094c.10 45a EtOH 182–185 C16H12N7F-HCl (357.78) 321 1094c.11 68 MeOH 127–129 C14H15N6F (286.31) 286 1094c.12 41a EtOH 160–165 C15H17N6F-HCl (336.80) 300 1094c.13 23a EtOH 167–170 C16H19N6F-HCl (350.83) 314 1094c.14 42 MeOH 107–109 C16H19N6F (314.37) 314 1094d.1 59 MeOH 121–124 C16H17N6F (312.35) 312 1094d.2 31 60–80 °C petr. ether 115–118 C17H19N6F (326.38) 326 1094d.3 67a EtOH 154–157 C20H19N6F-HCl (398.87) 362 1094d.4 78 EtOH 184–186 C18H15N6F (334.36) 334 1094d.5 54 EtOH 139–141 C18H15N6F (334.36) 334 1094d.6 45 100–140 °C petr. ether 142–144 C17H19N6OF (342.38) 342 1094d.7 88 EtOH 195–198 C16H13N6OF (324.32) 324 1094d.8 84 EtOH 205–208 C17H12N6BrF (399.22) 399 1094d.9 71 Toluene 214–217 C16H12N7F (321.32) 321 1094d.10 66 EtOH/H2O 99–101 C15H17N6F (300.34) 300 1094d.11 73 60–80 °C petr. ether 82–84 C16H19N6F (314.37) 314 1094d.12 65 60–80 °C petr. ether 90–92 C16H19N6F (314.37) 314 109

aIsolated as a hydrochloride;bSpecific rotation of the free base [α]D28 = – 42.3° (c = 1.30, CHCl3).

869

Table II. A1 and A2A adenosine receptor binding of triazolopyrimidines4a–d.

Compound R R1 A1 A2A Ki A2A/Ki A1

Ki nM Ki nM

4a.1 128± 12 > 10 000 > 78

4a.2 ″ 1 360± 82 > 10 000 > 7.3

4a.3 ″ > 10 000 > 10 000 –

4a.4 ″ > 10 000 > 10 000 –

4a.5 ″ > 10 000 > 10 000 –

4a.6 ″ > 10 000 > 10 000 –

4a.7 ″ > 10 000 > 10 000 –

-----------------------------------------------------------------------------------------------------------------------------------------

4b.1 37 ± 3 1 460± 131 39.4

4b.2 ″ 75 ± 5 1 740± 156 23.2

4b.3 ″ 39 ± 3 > 10 000 > 256

4b.4 ″ 68 ± 4 798± 56 11.7

4b.5 ″ 42 ± 3 616± 43 14.6

4b.6 ″ 172± 7 1 305± 104 7.5

870

Table II. continued.

Compound R R1 A1 A2A Ki A2A/Ki A1

Ki nM Ki nM

4c.1 10.5± 0.7 3 422± 205 325.9

4c.2 ″ 19.5± 1.2 3 973± 198 203.7

4c.3 ″ 73.4± 6.6 2 181± 153 29.7

4c.4 ″ 91 ± 6 932± 83 10.2

4c.5 ″ 56 ± 4 1 906± 172 34.0

4c.6 ″ 402± 32 > 10 000 > 24.8

4c.7 ″ 1 188± 95 > 10 000 > 8.4

4c.8 ″ 1 650± 149 9 800± 686 5.9

4c.9 ″ 112± 7 > 10 000 > 89

4c.10 ″ 45 ± 3 2 613± 183 58.0

4c.11 ″ 204± 18 > 10 000 > 49

4c.12 ″ 81 ± 6 > 10 000 > 123

4c.13 ″ 262± 23 > 10 000 > 38

4c.14 ″ 55 ± 4 4 000± 240 72.7

871

Table II. continued.

Compound R R1 A1 A2A Ki A2A/Ki A1

Ki nM Ki nM

4d.1 26 ± 2 > 10 000 > 384

4d.2 ″ 72 ± 6 > 10 000 > 138

4d.3 ″ 287± 26 > 10 000 > 34

4d.4 ″ 254± 18 > 10 000 > 39

4d.5 ″ 1 190± 83 > 10 000 > 8

4d.6 ″ 1 137± 102 > 10 000 > 8

4d.7 ″ > 10 000 > 10 000 –

4d.8 ″ 377± 26 > 10 000 > 26

4d.9 ″ 429± 30 > 10 000 > 23

4d.10 ″ 119.5± 9.5 > 10 000 > 83

4d.11 ″ 268± 24 > 10 000 > 37

4d.12 ″ 84 ± 6 7 586± 683 90.3

The 2-thiophenemethyl (2-thenyl) substituent (seriesb) was chosen because it could partially imitate the2-chlorobenzyl substituent in view of its steric andelectronic characteristics. This substituent appeared to beactively involved in A1 adenosine receptor binding: thetriazolopyrimidines4b.1–5 (table II) showed a high A1adenosine affinity (Ki < 100 nM) except for the

m-toluidino derivative4b.6 (Ki = 172 nM). The mostactive compound of this series was again the cyclopen-tylamino derivative4b.1 (Ki = 37 nM) but its affinity waslower than those of the 2-chlorobenzyl (Ki = 21 nM) [1]and 2-fluorobenzyl4c.1 (Ki = 10.5 nM) derivatives.These results further confirmed the ability of the A1

receptor to interact with lipophilic substituents bearing a

872

sulphur or a chlorine atom in the 2 position. Theamphetamino4b.3 and thep-toluidino 4b.5 derivativespossessed equivalent affinities (Ki = 39 and 42 nMrespectively), confirming the effectiveness of the amphet-amino substituent and of thepara-methyl position, com-pared with themeta-methyl one (4b.6, Ki = 173 nM).This effect was also recorded in thepara-fluorobenzylseries (4d.4, Ki = 254 nM; 4d.5, Ki = 1 190 nM). Amongcompounds of theb series, 4b.3 showed the bestselectivity (Ki A2A/Ki A1 > 256) and the most activecompound4b–1 followed with a ratio Ki A2A / Ki A1 =39.5.

The selection of the fluorobenzyl substituents, to com-pare with the corresponding chlorobenzyl substituents,was based upon the consideration that the strong elec-tronegativity of the fluorine, which is capable of accept-ing a hydrogen bond, allowed us to evaluate the effect ofpossible electronic, as well as steric factors on receptorbinding. As regards the 2-fluorobenzyl substituent, usedfor the preparation of the triazolopyrimidines4c.1–14(table II), a generalized increase in affinity was observedtowards the A1 adenosine receptors, compared with the2-chlorobenzyl derivatives [1], excluding the amphet-amino derivatives4c.3 and 4c.4. The most active com-pounds were once again the cyclopentylamino4c.1(Ki =10.5 nM) and the cyclohexylamino4c.2 (Ki = 19.5 nM)derivatives, which also showed the best selectivity to-wards the A1 adenosine receptors (Ki A2A/Ki A1 = 326and 203, respectively); the introduction of an alcoholicOH function on the cyclohexyl ring (compound4c.7) (seeGR-79326, GlaxoWellcome A1 adenosine agonist) mark-edly decreased binding affinity. The amphetamino sub-stituent maintained a good affinity (Ki < 100 nM), but itis worth noting that the racemic derivative4c.3was foundto be slightly more effective than the levorotatory isomer4c.4. The new furfurylamino substituent (4c.8) markedlylowered receptor binding. As regards the known7-arylamino substituents, the binding affinity was con-firmed to be high with thepara-toluidino derivative4c.5(Ki = 56 nM), while it decreased with themeta-toluidinoderivative4c.6. Also the new pyridylamino substituent of4c.10 induced a high receptor affinity (Ki = 45 nM) anda good selectivity. As for the four derivatives bearingbranched aliphatic amines not previously tested by us,coming from the simplification or opening of the cyclo-pentyl ring, it is interesting to point out the high affinityof 4c.14 (3-pentylamino derivative, Ki = 55 nM) and4c.12(2-butylamino derivative, Ki = 81 nM).

Finally, as regards the 4-fluorobenzyl substituent usedfor the preparation of the triazolopyrimidine derivatives4d.1–12, the results reported intable II show that inhibi-tion constants Ki < 100 nM were obtained for4d.1

(cyclopentyl derivative, Ki = 26 nM), 4d.2 (cyclohexylderivative, Ki = 72 nM) and4d.12 (3-pentyl derivative,Ki = 84 nM). Thus in this series, a high affinity towardsA1 adenosine receptors was conferred by aliphatic amino-substituents, which also showed a high receptor selectiv-ity. The other aminosubstituents induced a noticeabledecrease in binding affinity compared with the corre-sponding derivatives of the 2-chlorobenzyl [1] or2-fluorobenzyl series.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Kofler hot-stageand are uncorrected. IR spectra in nujol mulls wererecorded on a Perkin-Elmer Mod. 1310 spectrometer.Mass spectra were performed with a Hewlett PackardMS/System 5988. Elemental analyses (C, H, N) werewithin ± 0.4% of theoretical values and were performedon a Carlo Erba Elemental Analyzer Mod. 1106 appara-tus. Optical rotations were measured with a Violet AA-5polarimeter.

5.2. 2-Azidomethyl-thiophene

NaN3 (7.68 g, 118 mmol) was added to a solution of2-chloromethyl-thiophene (5.24 g, 39.5 mmol) in 25 mLof MeOH and 4 mL of H2O, and the suspension wasstirred at room temperature for 24 h. The inorganicprecipitate was removed by filtration and the filtrate wasconcentrated, treated with H2O and extracted withCHCl3. The chloroform extract was dried and evaporatedin vacuo to give the title compound as a yellow oil whichwas used without purification: 5.19 g, 94.5% yield. Theazide was short distilled in a tubular oven at 40–45 °C,1.8 mm Hg; IR (cm–1): 2 105 (N3).

5.2.1. 1-Substituted-4-carboxamido-5-amino-1H-1,2,3-triazoles1b–d

0.934 g (11 mmol) of cyanacetamide was added to astirred solution of sodium ethoxide (0.253 g, 0.011 gatom of Na) in 10 mL of absolute EtOH, and stirring wascontinued for 30 min. A solution of 10.0 mmol of thesuitable azide (2-thenyl-, 2-fluorobenzyl- or4-fluorobenzylazide) in 10 mL of absolute EtOH wasslowly added to the suspension, and then the mixture washeated under reflux for 1.5–2 h. The reaction mixture wasconcentrated in vacuo and treated with H2O, and theinsoluble material, consisting of the title compounds, wascollected and purified by crystallization (table I).

873

5.2.2. 3-Substituted-7-hydroxy-1,2,3-triazolo[4,5-d]pyri-midines2b–d

A solution of 10.0 mmol of the suitable triazole deriva-tive (1b, 1cor 1d) in 8 mL of formamide was refluxed for2 h. After cooling the reaction mixture was diluted withH2O and the solid precipitate was collected by filtrationand recrystallized (table I).

5.2.3. 3-Substituted-7-chloro-1,2,3-triazolo[4,5-d]pyri-midines3b–d

0.8 ml of DMF and 4.5 mL of SOCl2 were added to asuspension of 5.0 mmol of the suitable triazolopyrimidine(2b, 2cor 2d) in 22 mL of boiling anhydrous CHCl3. Thereaction mixture was refluxed for 2 h, the solvent wasevaporated in vacuo (temperature≤ 35 °C), and theresidue, after cooling at 0 °C, was triturated with crushedice. The solid formed was collected by filtration, driedand extracted repeatedly with boiling 60–80 °C petro-leum ether. The combined extracts were evaporated invacuo to give the title compounds as white solids (ta-ble I).

5.2.4. 3-(4-Chlorobenzyl)-7-(substituted amino)-1,2,3-triazolo[4,5-d]pyrimidines4a.1–7

A mixture of 3-(4-chlorobenzyl)-7-chloro-1,2,3-triazolo[4,5-d]pyrimidine 3a (0.40 g, 1.43 mmol), tri-ethylamine (0.24 mL, 1.70 mmol) and the suitable amine(1.70 mmol of cyclopentyl-, cyclohexyl-, (±)-α-methyl-phenethylamine, 3.70 mmol of (±)-α-methylbenzyl-amine, 5.0 mmol of para- and meta-toluidine or1.0 mmol of para-nitroaniline) in 10 mL of absoluteEtOH was refluxed for 2.5 h. For the isolation of com-pounds4a.1 and 4a.2, the reaction mixture was evapo-rated in vacuo, the residue was treated with H2O and 5%HCl (pH ≅ 3) and the insoluble material was collectedand purified by crystallization (4a.1) or by extractionwith 60–80 °C petroleum ether (4a.2) respectively. Forthe isolation of compounds4a.3, 4a.5, 4a.6, and4a.7, thereaction mixture was allowed to cool and the crystallizedprecipitate was collected by filtration. Isolation of4a.7required further treatment of the precipitate with 10%HCl to remove unreactedp-nitroaniline. For the isolationof 4a.4, the reaction mixture was evaporated in vacuo, theresidue was dissolved in MeOH and the derivative wasprecipitated as a hydrochloride by addition of Et2O-HCl(table I).

5.2.5. 3-(2-Thenyl)-7-(substituted amino)-1,2,3-tri-azolo[4,5-d]pyrimidines4b.1–6

A mixture of 3b (0.25 g, 1.0 mmol), triethylamine(0.17 mL, 1.2 mmol) and the suitable amine (1.2 mmol of

cyclopentyl-, cyclohexyl-, (±)-α-methylphenethyl-, (±)-α-methylbenzylamine,para- or meta-toluidine) in 10 mLof absolute EtOH was heated under reflux for 2.5 h. Thereaction mixture was evaporated in vacuo and the residuewas treated with H2O and purified by crystallization togive compounds4b.1, 4b.2, 4b.5, and 4b.6. For theisolation of4b.3and4b.4, the residue was extracted withEt2O, and then Et2O-HCl was added to the ether solutionto precipitate the derivatives as hydrochlorides whichwere collected and crystallized (table I).

5.2.6. 3-(2-Fluorobenzyl)-7-(substituted amino)-1,2,3-triazolo[4,5-d]pyrimidines4c.1–14

A mixture of 3c (0.30 g, 1.14 mmol), triethylamine(0.16 mL, 1.15 mmol) and the suitable amine (1.40 mmolof cyclopentyl-, cyclohexyl-, (±)-α-methylphenethyl- or(–)-α-methylphenethylamine; 4.0 mmol ofpara or meta-toluidine; 2.2 mmol oftrans-2-cyclohexanol-, furfuryl-,4-bromophenyl-, 2-pyridyl-, 2-butyl-, 2-pentyl- or3-pentylamine) in 10 mL of absolute EtOH (10 mL oftoluene for the pyridylamino derivative4c.10) was heatedunder reflux for 2.5 h. For the isolation of4c.5, 4c.6, 4c.8,4c.9and4c.14, the reaction mixture was allowed to cooland the crystallized precipitate was collected by filtration(table I). For the isolation of4c.1, 4c.2, 4c.7 and4c.10,the reaction mixture was evaporated in vacuo, the residuewas treated with H2O and 10% HCl and the insolublematerial was collected by filtration and crystallized (ta-ble I). For the isolation of4c.3, 4c.4, 4c.12and4c.13, thereaction mixture was evaporated in vacuo, the residuewas washed with H2O and 5% HCl and dissolved in Et2Oor absolute EtOH or MeOH. Addition of Et2O-HCl to thesolution precipitated the compounds as hydrochlorideswhich were purified by crystallization (table I). Theisopropylamino derivative4c.11was obtained by heatingthe 7-chloro-triazolopyrimidine3c (0.39 g, 1.48 mmol) in2 mL of isopropylamine at 100–110 °C in a closed tubefor 2.5 h and treating the reaction mixture with H2O toprecipitate the expected derivative (table I).

5.2.7. 3-(4-Fluorobenzyl)-7-(substituted amino)-1,2,3-triazolo[4,5-d]pyrimidines4d.1–12

A mixture of 3d (0.30 g, 1.14 mmol), triethylamine(0.16 mL, 1.15 mmol) and the suitable amine (1.40 mmolof cyclopentyl-, cyclohexyl-, (±)-α-methylphenethyl- orfurfurylamine; 4.0 mmol of para or meta-toluidine;2.3 mmol of trans-2-cyclohexanol-, 4-bromophenyl-,2-pyridyl-, 2-butyl-, 2-pentyl- or 3-pentylamine) in10 mL of absolute EtOH (10 mL of toluene for thepyridylamino derivative4d.9) was heated under refluxfor 2.5 h. The reaction mixture was evaporated in vacuo,

874

the residue was washed with H2O and 5% HCl (pH≅ 3)and the insoluble material was collected and purified bycrystallization (4d.1, 4d.6, 4d.10, 4d.11and4d.12) or byextraction with 60–80 °C petroleum ether (4d.2) (table I).For the isolation of4d.3, the liquid residue obtained afterthe rinses was dissolved in Et2O or MeOH, and Et2O-HClwas added to the solution to precipitate the derivative asa hydrochloride (table I). For the isolation of4d.4, 4d.5,4d.7, 4d.8 and4d.9 the reaction mixture was allowed tocool and the crystallized precipitate was collected byfiltration (table I).

Acknowledgements

We wish to thank the Consiglio Nazionale delleRicerche (C.N.R.) for financial support.

References

[1] Betti L., Biagi G., Giannaccini G., Giorgi I., Livi O., Lucacchini A.,Manera C., Scartoni V., J. Med. Chem. 41 (1998) 668–773.

[2] Biagi G., Giorgi I., Livi O., Lucacchini A., Martini C., Scartoni V.,Tacchi P., Il Farmaco 49 (1994) 183–186.

[3] Biagi G., Giorgi I., Livi O., Lucacchini A., Martini C., Scartoni V.,Tacchi P., Il Farmaco 49 (1994) 187–191.

[4] Biagi G., Giorgi I., Livi O., Scartoni V., Lucacchini A., Martini C.,Tacchi P., Il Farmaco 50 (1995) 13–19.

[5] Biagi G., Breschi C., Giorgi I., Livi O., Martini C., Scartoni V.,Scatizzi R., Il Farmaco 50 (1995) 659–667.

[6] Biagi G., Giorgi I., Livi O., Lucacchini A., Scartoni V., Il Farmaco51 (1996) 395–399.

[7] Dudley M.W., Peet N., Demeter D., Weintraub H.J.R., IjzermanA.D.P., Nordvall G., van Galen P.J.M., Jacobson K.A., Drug Dev.Res. 28 (1993) 237–243.

[8] Ried W., Laoutidis J., Chemiker-Ztg. 114 (1990) 246–248.

[9] Aubert T., Farnier M., Guilard R., Can. J. Chem. 68 (1990) 842–851.

[10] Kreher R., Bergman U., Tetrahedron Lett. 47 (1976) 4259–4262.

[11] Baraldi P.G., Manfredini S., Simoni D., Zappaterra L., Zocchi C.,Dionisotti S., Ongini E., Med. Chem. Lett. 4 (1994) 2539–2544.

875

Short communication

Synthesis and cardiovascular evaluationof N-substituted 1-aminomethyl-2-naphthols

A.Y. Shena*, C.T. Tsaib, C.L. Chena

aDepartment of Pharmaceutical Science, Foo Yin Institute of Technology, Ta-Liao, Kaohsiung County 831, TaiwanbDepartment of Biology, National Changhua University of Education, Changhua, Taiwan

(Received 9 December 1998; revised 15 April 1999; accepted 6 May 1999)

Abstract – A series of 1-alkylaminomethylnaphthols have been prepared. These compounds were readily prepared in good yields by additionof 2-naphthol to formalin and alkylamines. The hypotensive and bradycardiac effects of these compounds in normotensive rats as well as theirin vitro inotropic and aortic contraction effects in isolated rat left atria and aorta have been evaluated. A higher depressor and bradycardiacactivity was found for compounds substituted on nitrogen by naphthol with primary amines, i.e., ethylamine, propylamine, isopropylamine,or butylamine and with a cyclic secondary amine, i.e., pyrrolidinyl. These compounds produced biphasic changes in contractile force inisolated rat atria which was correlated to blood pressure and heart rate activity. 1-Isopropylaminomethyl-2-naphthol hydrochloride relaxedisolated rat aortic rings precontracted with high extracellular K+ (80 mM) and Ca2+ (1.9 mM). The suppressive effects of the compounds mayinvolve a direct inhibition of Ca2+ channels. The biological activity of these compounds can be explained in terms of substitution on nitrogen.© 1999 Éditions scientifiques et médicales Elsevier SAS

1-alkylaminomethyl-2-naphthol / blood pressure / heart rate / left atrial / aorta

1. Introduction

1-Pyrrolidinylmethyl-2-naphthol hydrochloride(TPY-â) has been shown to produce a reduction in bloodpressure (BP) and heart rate (HR) in anaesthetizedrats [1]. The ionic mechanism of the cardiovascularactivity of TPY-â has also been examined. The resultsindicated that suppressive effects of TPY-â involve adirect depressant action on heart cells and vascularsmooth cells [2]. The direct inhibition of voltage-dependent L-type Ca2+ channels is involved in the TPY-âmediated vasodilatory action. In addition, the inhibitoryeffect of TPY-â on cardiac contractibility through theblockade of L-type Ca2+ channels can be prevented byTPY-â mediated inhibition of the transient outward po-tassium current. Bril et al. [3] reported that a compound

with a combination of potassium and calcium channelantagonistic properties might consititute a novel anti-arrhythmic agent with reduced proarrhythmic risk. Someof the drugs used to treat ventricular arrhythmias havebeen shown to also act on the transient outward current attherapeutic concentrations [4–6]. The actions of TPY-âon the cardiovascular system encouraged us to search forand synthesize novel derivatives of aminomethylnaph-thol. This brought forth the modification of the pyrrolidi-nyl group in TPY-â with primary amines via the conden-sation of 2-naphthol with primary amines. The methodexists for attaching carbon substituents to the 1-positionof 2-naphthol, and many 2-naphthols bearing such sub-stituents are known in the literature [7]. A selection ofcompounds possessing a variety of 1-substituents wasdesired in order to broadly define the structure activityrelationship for this portion of the molecule. Thesecompounds, shown intable I, either were known in theliterature or were prepared according to well-establishedsynthetic methods [7, 8]. The purpose of the present studywas to evaluate the importance of the side substitution atthe 1-position of the aminomethylnaphthols.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 877−882 877© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

In the present work, reaction of 2-naphthol withformaldehyde and methylamine in a molar ratio of 1:2:1,respectively, in a methanol solution at 60 °C was found toproduce 2,3-dihydro-2-methyl-1H-naphth[1,2-e]-m-oxazine (1). Upon treating1 with hot aqueous hydrochlo-ric acid, formaldehyde was liberated and the hydrochlo-ride of 1-methylaminomethyl-2-naphthol (2) was formed(figure 1) [7]. The condensation, run in the same mannerexcept with ethylamine, propylamine, isopropylamine orbutylamine, resulted in the formation of 1-alkyl-aminomethyl-2-naphthols directly instead of naphthox-azine products. Compounds1, 2, 5, and6 were previouslysynthesized and reported [7]. The temperature and theparticular amine seem to be the important factors indetermining the course of the reaction of 2-naphthol withformaldehyde and primary amine as reported by Burke etal. [7].

The condensation of naphthols with formaldehyde andsecondary amines has been studied previously [8]. Reac-tion of piperidine and formaldehyde with 1- and2-naphthol has been shown to result in the introduction of

a piperidinomethyl group into the 1-position of2-naphthol and into the 2-position of 1-naphthol. Thisindicates that the substituentortho to the naphthylene

Table I. Hypotensive and bradycardic response following intravenous injection of aminomethylnaphthol derivatives (2.2µmol/kg) in rats.

1 2–13

Compound R1 R2 X HR (beats/min)a BP (mm Hg)a

maximum change % maximum maximum change % maximum

N.S. 4± 2 1 ± 1 4 ± 3 3 ± 21 -8 ± 4 2 ± 1 -5 ± 4 6 ± 52 CH3 H HCl -38 ± 15 11± 5 -10± 5 10± 53 C2H5 H HCl -240± 15* 62 ± 11* -75 ± 19* 69 ± 16*4 n-C3H7 H HCl -276± 31* 67 ± 13* -94 ± 21* 81 ± 18*5 i-C3H7 H HCl -268± 54* 65 ± 7* -82 ± 11* 70 ± 20*6 C4H9 H HCl -328± 43* 75 ± 21* -103± 16* 85 ± 15*7 CH3 CH3 -26 ± 3* 7 ± 2* -16 ± 5* 15 ± 4*8 C2H5 C2H5 -15 ± 4* 4 ± 2 -6 ± 4 5 ± 49 C3H7 C3H7 -13 ± 5* 3 ± 2 0 010 C4H9 C4H9 0 0 0 011 TPY- -258± 33* 64 ± 15* -81 ± 2* 77 ± 17*12 -(CH2)5- -137± 37* 32 ± 8* -59 ± 13* 53 ± 11*13 -(CH2)2-O-(CH2)2- -121± 17* 28 ± 8* -38 ± 8* 23 ± 9*

aPercent decrease in blood pressure and heart rate were calculated from the decrease in blood pressure and heart rate of the treatment groupand the blood pressure (normally 105–120 mm Hg) and heart rate (normally 385–405 beats/min) of the control group; 2.2µmol/kg usuallyamounts to 0.4–0.5 mg/kg; values are expressed as the mean± SEM. Asterisks indicate significant difference (t-test, calculated on the changes)from the N.S. (normal saline). *P < 0.05.

Figure 1. Synthetic routes for the 2,3-dihydro-2-methyl-1H-naphth[1,2-e]-m-oxazine (1) and aminomethylnaphthol deriva-tives (2–13).

878

hydroxyl group plays a role in determining the course ofthe reaction [7, 8].

3. Results and discussion

Aminonaphthols have been reported to exhibit anti-hypertensive, adrenoceptor blocking, and Ca2+ channelblocking activities [9–11]. As described in the experimen-tal section, the hypotensive and bradycardiac activity ofthe analogue series was tested in normotensive rats at adosage of 2.2µmol/kg. Figure 2 shows the chart record-ings of blood pressure (top traces) and heart rate (bottomtraces) of an anaesthetized rat. Intravenous injection ofcompound5 (2.2µmol/kg) induced a maximum reduc-tion of mean arterial pressure from 120 mm Hg to 25 mmHg within 30 s and the BP returned to the control value in20 min. Along with the sudden decrease in BP, the heartrate was also reduced from 400 beats min–1 to a minimumvalue of 130 beats min–1, and it took 40 min for fullrecovery to the control level. The acute hypotensive andbradycardiac responses induced by 2.2µmol/kg of thesecompounds are summarized intable I. Substitution of theamine hydrogen atom by alkyl groups of increasing chainlength altered the activity. In the dialkylaminomethyl-2-naphthol series, the straight chain derivatives (7–10) hadless hypotensive and bradycardiac activity than the cyclicanalogues (11–13). An increase in the size of theN-dialkyl group appeared to reduce hypotensive andbradycardiac activity. Despite the limited series fromwhich to evaluate the structure-activity relationship, pri-mary amine substitution (3–6) showed better potencythan secondary amine derivatives except methyl side

chain derivatives (1–2), which were not active even at ahigher dosage (data not shown).

The present results(table II) in vitro show that the1-alkylaminomethyl-2-naphthols (3–6) and the secondaryamines with cyclic side chains (11–13), at 30 µM,produced a biphasic effect on the contractile force inisolated rat left atria, i.e., an initial decrease and asustained increase. A representative inotropic response ofan electrically driven left atrium to 1-isopropyl-aminomethyl-2-naphthol hydrochloride (5) is depicted infigure 3. However, the straight chain derivatives (7–10)show the negative inotropic effect. The 4-Aminopyridine-sensitive transient outward current (Ito) has been shownto be present in rabbit, rat, cat, dog and human cardiacmyocytes, but not in guinea-pig cardiac myocytes [12–15]. It has been suggested that the compound11 (TPY-â)mediated inhibition of Ito would reverse the rat myocar-dial depressant effect caused by its initial inhibition ofL-type Ca2+ channels [3]. The similarity of the molecularstructure suggests that 1-isopropylaminomethyl-2-naphthol hydrochloride (5) and other analogues work inthe same manner as compound11 (TPY-â). In addition,Ca2+ (1.9 mM) elicited a 100% contraction in rat aorta inthe presence of high K+ (80 mM), 5 was more potent insuppressing the tone induced by high K+ and Ca2+ than7as shown intable III. High K+ has been shown to evokesmooth muscle contraction by promoting Ca2+ influxthrough voltage-sensitive Ca2+ channels which arereadily activated by membrane depolarization [16].Therefore, extracellular Ca2+ entry is thought to be themain cause of the high K+-induced contraction. Thepresent studies indicate that the suppressive effects of

Figure 2. Effects of intravenous injection of 1-isopropylaminomethyl-2-naphthol (5) and dipropylaminomethyl-2-naphthol (9) onblood pressure and heart rate in one anaestherized normotensive rat. Compounds were administered at the time indicated by an arrow.

879

these compounds could interfere with Ca2+ influx throughthe depolarized cell membrane to induce relaxation of rataorta. The results observed are consistent with previousfindings that11 (TPY-â) or aminonaphthols effectivelysuppressed the voltage-gated Ca2+ current [2, 17, 18].

In conclusion, the biological activity of these com-pounds can be explained in terms of substitution onnitrogen. The development of N-substituted 1 amino-methyl-2-naphthols with dual effects would be of poten-tial therapeutic advantage.

4. Experimental protocols

4.1. Chemistry

Melting points were determined on a YanagimotoMP-3 micromelting apparatus and are uncorrected.

Analyses indicated by the symbols of the elements werewithin ± 0.4% of the theoretical values. Infrared spectrawere obtained on a Shimadzu IR-408 spectrophotometer.Nuclear magnetic resonance spectra were recorded on aVarian Gemini T-300 spectrometer at the National SunYat-Sen University, Kaohsiung, and are expressed inparts per million (δ) with tetramethylsilane used as aninternal standard. Mass spectra recorded for the purposesof structure confirmation were obtained on a JeolJMS-HX 110 mass spectrometer at the National SunYat-Sen University, Kaohsiung. Elemental analysis wasperformed on a CHN-O-Rapid Heraeus Elemental Ana-lyzer at the National Cheng-Kung University, Tainan.Thin layer chromatography was carried out on precoatedsilica gel F254 chromatographic plates (20× 20 cm;0.2 mm). Methylamine, ethylamine, propylamine, isopro-pylamine and butylamine were all obtained from theTokyo Chemical Industry Co. (TCI). 2-Naphthol was theproduct of the Sigma Co. All other reagents used in thisstudy were EP grade products of E Merck.

Dialkylaminomethyl-2-naphthol derivatives were syn-thesized previously [8]. The synthesis and characteriza-tion of 1-alkylaminomethyl-2-naphthol will be reportedhere.

4.1.1. 2,3-Dihydro-2-methyl-1H-naphth[1,2-e]-m-oxa-zine(1)

Following the procedure of Burke et al. [7], to a cooledsolution of 12.4 g of 25% aqueous methylamine (0.1 mol)

Table II. Characteristics of aminomethylnaphthols (30µM) on thecontractile force in rat left atria stimulated at 1.0 Hz.

Compound Maximal force of contraction (%)

phase 1 phase 2

1 100 ± 18 (3) 100± 16 (3)2 100 ± 20 (3) 100± 19 (3)3 90 ± 16* (5) 220± 36* (4)4 85 ± 17* (5) 235± 42* (5)5 82 ± 13* (6) 238± 29* (5)6 83 ± 17* (5) 219± 23* (5)7 78 ± 19* (4) 78 ± 19* (4)8 80 ± 21* (4) 80 ± 21* (4)9 88 ± 15* (4) 88 ± 15* (4)10 97 ± 19* (4) 97 ± 19* (4)11 77 ± 12* (5) 225± 49* (4)12 90 ± 21* (5) 160± 23* (4)13 89 ± 16* (6) 120± 25* (5)

Values are means± SEM. Number of preparations (n) is given inparentheses.*P < 0.05 compared with the respective control bystudentt test.

Table III. Effects of dimethylaminomethyl-2-naphthol (7) and1-isopropylaminomethyl-2-naphthol hydrochloride (5) on the rataortic contraction induced by KCl and Ca.

Compound Contraction (% of Control)KCl (80 mM) + Ca (1.9 mM)

Control 100± 3.97 (45 µM) 79 ± 4.85 (45 µM) 50.4 ± 6.45 (150 µM) 15.3 ± 1.0

Percentages of the control contraction were calculated and presen-ted as mean± SEM (n = 4). *P < 0.05 as compared with therespective control.

Figure 3. Effect of 1-isopropylaminomethyl-2-naphthol (5)and dimethylaminomethyl-2-naphthol (7) on the contractileforce of isolated rat left atria. Stimuli were driven at 1 Hz with2 times the threshold. The presence of compound5 (30 µM)produced a biphasic response, while compound7 (30 µM)produced only a negative intropic effect.

880

in 60 mL of methanol was added 18.5 mL of 37%(0.2 mol) aqueous formaldehyde in 40 mL of methanoland 14.4 g of 2-naphthol (0.1 mol) in 50 mL of methanol.After 1.5 h of gentle refluxing at 60 °C, the reactionmixture was poured into 400 mL of cold water. Theproduct (95% yield) was collected and recrystallizedfrom methanol. M.p. 68–70 °C, IR (nuzol) cm–1: 2 994(aromatic C–H), 2 909 (CH3), 1 427 (C–O), 1 214 (C–N).1H NMR (300 MHz, DMSO):δ 7.82 (d,J = 8.4 Hz, 1H),7.65 (m, 2H), 7.48 (m, 1H), 7.38 (m, 1H), 7.03 (d,J = 9.0Hz, 1H), 4.82 (s, 2H), 4.22 (s, 2H), 2.53 (s, 3H). MS m/z(relative intensity): 128 M+ (100), 144 (72), 156 (42).Anal. C13H13NO (C, H, N).

4.1.2. 1-Methylaminomethyl-2-naphthol hydrochloride(2)

Following the procedure of Burke et al. [7], a solutionof 0.013 mol of 2,3-dihydro-2-methyl-1H-naphth[1,2-e]-m-oxazine and 1.0 mL of concentrated aqueous hydro-chloric acid in 60 mL of 85% aqueous propanol-1 wasdistilled for about 30 min. After the distillation wasinterrupted, 60 mL of acetone was added to the reactionmixture. Upon cooling and filtration the product wasobtained (91% yield). M.p. 187–189 °C (dec.). IR (nuzol)cm–1: 2 987 (aromatic C–H), 2 904 (CH3), 1 443 (C–O).1H NMR (300 MHz, D2O): δ 7.93 (m, 3H), 7.62 (dd,J =8.7, 7.2 Hz, 1H), 7.45 (dd,J = 7.5 Hz, 1H), 7.24 (d,J =9.6 Hz, 1H), 4.62 (s, 2H), 2.73 (s, 3H). MS m/z (relativeintensity): 128 M+ (100), 156 (42), 187 (26). Anal.C12H13NO HCl (C, H, N).

4.1.3. 1-Ethylaminomethyl-2-naphthol hydrochloride(3)To a solution of 3.6 g of 2-naphthol (0.025 mol) and

4.7 mL of 37% aqueous formaldehyde (0.06 mol) in30 mL of methanol 1.6 g of 70% ethylamine (0.025 mol)in 10 mL methanol was added dropwise. After 1.5 h ofgentle refluxing at 60 °C the oily residue was dissolved inabsolute alcohol. The hydrochloride was obtained bytreating a cold alcohol solution of the product withconcentrated hydrochloric acid (yield 60%). M.p.175–177 °C (dec.). IR (nuzol) cm–1: 3 454 (broad, NH,OH), 3 010 (aromatic C–H), 2 909 (CH3).

1H NMR (300MHz, D2O): δ 7.88 (m, 3H), 7.59 (dd,J = 8.7, 7.2 Hz,1H), 7.42 (dd,J = 7.5 Hz, 1H), 7.20 (d,J = 8.7 Hz, 1H),4.54 (s, 2H), 3.16 (q,J = 7.5 Hz, 2H). 1.29 (t,J = 7.5 Hz,3H). MS m/z (relative intensity): 128 M+ (100), 156 (57),201 (47). Anal. C12H13NO HCl (C, H, N).

4.1.4. 1-Propylaminomethyl-2-naphthol hydrochloride(4)

The product was prepared by a method similar to thatdescribed in the procedure for compound3. It wasrecrystallized from absolute alcohol (yield 65%). M.p.

188–190 °C (dec). IR (nuzol) cm–1: 3 486 (NH), 3 442(OH), 2 990 (aromatic C–H), 2 913 (CH2, CH3), 1 426(C–O). 1H NMR (300 MHz, D2O): δ 7.89 (m, 3H), 7.5(dd, J = 8.7, 7.2 Hz, 1H), 7.42 (dd,J = 7.5, 7.5 Hz, 1H),7.20 (d,J = 9.0 Hz, 1H), 4.54 (s, 2H), 3.04 (t,J = 7.2 Hz,2H), 1.70 (m, 2H). 0.92 (t,J = 7.5 Hz, 3H). MS m/z(relative intensity): 128 M+ (42), 215(30). Anal.C14H17NO HCl (C, H, N).

4.1.5. 1-Isopropylaminomethyl-2-naphthol hydrochloride(5)

The product was prepared by a method similar to thatdescribed in the procedure for compound3. It wasrecrystallized from absolute alcohol (yield 70%). M.p.183–185 °C (dec.). IR (nuzol) cm–1: 2 990 (aromaticC–H), 2 989 (C–C–C), 1 440 (C–O).1H NMR (300 MHz,DMSO): δ 8.07 (d,J = 8.7 Hz, 1H), 7.88 (d,J = 8.7 Hz,1H), 7.85 (d,J = 6.6 Hz, 1H), 7.55 (dd,J = 7.2, 7.5 Hz,1H), 7.37 (m, 2H), 4.09 (bs, 2H), 3.45 (m, 1H), 1.37 (s,3H). 1.34 (s, 3H). MS m/z (relative intensity): 128 M+

(42), 215(12). Anal. C14H17NO HCl (C, H, N, O).

4.1.6. 1-butylaminomethyl-2-naphthol hydrochloride(6)The product was prepared by a method similar to that

described in the procedure for compound3. It wasrecrystallized from absolute alcohol. M.p. 150–153 °C(dec.). IR (nuzol) cm–1: 3 473 (broad, OH, NH), 3 006(aromatic C–H), 2 916 (CH3), 1 436 (C–O), 1 299 (C–N).1H NMR (300 MHz, D2O): δ 7.86 (m, 3H), 7.59 (dd,J =8.1, 6.9 Hz, 1H), 7.42 (dd,J = 7.5, 7.5 Hz, 1H), 7.15 (d,J = 8.7 Hz, 1H), 4.43 (s, 2H), 3.31 (m, 1H), 1.84 (m, 1H),1.62 (m, 1H), 1.36 (d,J = 7.2 Hz, 3H), 0.96 (t,J = 7.5 Hz,3H). MS m/z (relative intensity): 128 M+ (100), 156 (52),229 (17). Anal. C15H19NO HCl (C, H, N).

4.2. Measurement of BP and HR by intravenous injection

The in vivo experiments have been described previ-ously [2]. In brief, Wistar rats (250–300 g) of either sexwere used in these studies. The animals were housed in ananimal room with a light/dark cycle of 12 h/12 h andwere fed rat chow and tap water ad libitum. The animalswere anaesthetized with urethane (Aldrich, 1.0 g/kg IP,supplemented with 300 mg/kg i.v. if necessary). Thetrachea was intubated to keep the airway patent. Femoralarterial blood pressure was measured through a PE 50tubing filled with heparin solution (25 units/mL) (Sigma)connected to a polygraph (Lectromed, U.K.) via a trans-ducer (PDCR 75, Lectromed, UK). The femoral vein wascannulated for drug administration. The HR was derivedby means of a cardiotachometer that was triggered by thearterial pressure pulse. The BP and HR were monitored

881

continuously. Rectal temperature was monitored andmaintained between 37 and 38 °C.

4.3. Measurement of contractile force in isolated rat leftatria

Experiments were performed following the methodsdescribed previously [2]. In brief, the hearts were quicklyremoved and rinsed in ice-cold Tyrode’s solution. Then,left atria were dissected and mounted at 0.5 g restingtension on stainless steel hooks in a 50 mL organ bath,and bathed at 37 °C in physiological saline solutioncontaining (mM): NaCl 118, KCl 4.8, CaCl2 2.5,MgSO4 1.2, KH2PO4 1.2, NaHCO3 24, glucose 11. Thebath was aerated with 95% O2 and 5% CO2 mixture. Oneend of the preparation was fixed to the bottom of the bathand the other end was connected by a hook to the level ofa force-displacement transducer (Ugo Basile, Comero,Italy). Stimuli were delivered as rectangular pulses of5 ms duration at 2 times the threshold via two platinumelectrodes placed on either side of the muscle. The tissueswere always allowed to equilibrate for 90 min before theexperiments were begun [2, 19].

4.4. Measurement of contractile force in isolated rataorta

Experiments were performed following the methodsdescribed previously [20]. In brief, the thoracic aorta wasisolated and excess fat and connective tissue were re-moved. Aortic rings (5 mm) were mounted in organ bathscontaining 5 mL and bathed at 37 °C in physiologicalsaline solution containing (mM): NaCl 118.4, KCl 4.7,CaCl2 1.9, MgSO4 1.2, KH2PO4 1.2, NaHCO3 25, glu-cose 11.7. The bath was aerated with a 95% O2 and 5%CO2 mixture. Two stainless steel hooks were inserted intothe aortic lumen, one was fixed while the other wasconnected to the transducer. The contractile effects ofcalcium were studied in rings stabilized in high-K+

solution without Ca2+. Calcium was then added to obtainthe desired concentrations. The high K+ solution wasprepared by substituting NaCl with KCl (80 mM) in anequimolar amount. Contractions were recorded isometri-cally via a force-displacement transducer connected to aGould polygraph (Model 2400).

4.5. Statistical analysis

Values are expressed as means± standard error of themean (SEM). Statistical analysis was performed with apairedt test, and aP value smaller than 0.05 was regardedas statistically significant.

Acknowledgements

This work was supported by a grant from the NationalScience Council (NSC 87-2314-B-242-001). The authorsthank Drs Sheng-Nan Wu, Chih-Tsao Chiu and Che-MingTeng for their helpful suggestions. The technical assis-tance of Hui-Ya Chang is also greatly appreciated.

References

[1] Shen A.Y., Chen C.L., Lin C.I., Chin. J. Physiol. 35 (1992) 45–54.

[2] Shen A.Y., Wu S.N., Drug. Dev. Res. 44 (1998) 87–96.

[3] Bril A., Gout B., Bonhomme M., Landais L., Faivre J.F., Linee P.,Poyser R.H., Ruffolo Jr. R.R., J. Pharmacol. Exp. Ther. 276 (1996)637–646.

[4] Nabauer M., Beuckelmann D.J., Erdmann E., Circ. Res. 73 (1993)386–394.

[5] Imaizumi Y., Giles W.R., Am. J. Physiol. 253 (1987) H704–H708.

[6] Nabauer M., Beuckelmann D.J., Circulation 86 (suppl 1) (1992)I-697.

[7] Burke W.J., Kolbezen M.J., Stephens C.W., J. Am. Chem. Soc. 74(1952) 3601–3605.

[8] Shen A.Y., Tsai M.I., Lien E.J., Acta. Pharm. 44 (1994) 117–126.

[9] Atwal K.S., O’Reilly B.C., Ruby E.P., Turk C.F., Aberg G., AsaadM.M., Bergey J.L., Moreland S., Powell J.R., J. Med. Chem. 30(1987) 627–635.

[10] Grundke M., Himmel H.M., Wettwer E., Borbe H.O., Ravens U., J.Cardiovasc. Pharmacol. 18 (1991) 918–925.

[11] Jim K.F., Matthews W.D., J. Pharmacol. Exp. Ther. 234 (1985)161–165.

[12] Josephson I.R., Sanchez-Chapula J., Brown A.M., Circ. Res. 54(1984) 157–162.

[13] Dukes I.D., Morad M., J. Physiol. (London) 435 (1991) 395–420.

[14] Furukawa T., Myerburg R.J., Furukawa N., Bassett A.L., Kimura S.,Circ. Res. 67 (1990) 1287–1291.

[15] Simurda J., Simurdova M., Christe G., Pflugers. Arch. 415 (1989)244–246.

[16] Van Breemen C., McNaughton E., Biochem. Biophys. Res. Com-mun. 39 (1970) 567–574.

[17] Atwal K.S., O’Reilly B.C., Ruby E.P., Turck C.F., Aberg G., AsaadM.M., Bergery J.L., Moreland S., Powell J.R., J. Med. Chem. 30(1987) 627–635.

[18] Grundke M., Himmel H.M., Wettwer E., Borbe H.O., Ravens U., J.Cardiovasc. Pharmacol. 18 (1991) 918–925.

[19] Wu S.N., Shen A.Y., Hwang T.L., Chin. J. Physiol. 39 (1996) 23–29.

[20] Ko F.N., Guh J.H., Yu S.M., Hou Y.S., Wu Y.C., Teng C.M., Br. J.Pharmacol. 112 (1994) 1174–1180.

882

Laboratory note

6-Thienyl and 6-phenylimidazo[2,1-b]thiazoles as inhibitors of mitochondrialNADH dehydrogenase§

Aldo Andreania*, Mirella Rambaldia, Alberto Leonia, Rita Morigia, Alessandra Locatellia,Gianluca Giorgib, Giorgio Lenazc, Anna Ghellid, Mauro Degli Espostie

aDipartimento di Scienze Farmaceutiche, Via Belmeloro 6, 40126 Bologna, ItalybCentro Interdipartimentale di Analisi e Determinazioni Strutturali, Via A. Moro, 53100 Siena, Italy

cDipartimento di Biochimica “G. Moruzzi”, Via Irnerio 48, 40126 Bologna, ItalydDipartimento di Biologia E.S., Via Irnerio 42, 40126 Bologna, Italy

eDepartment of Biochemistry and Molecular Biology, Monash University, Clayton, VIC, Australia

(Received 16 November 1998; accepted 10 February 1999)

Abstract – Starting from the potent inhibitory effect of the previously described 2-methyl-6-(2-thienyl)imidazo[2,1-b]thiazole on mitochon-drial complex I, we prepared a series of derivatives in order to study the effect of a different substitution at the positions 2, 5 and 6. Thereplacement of the thienyl group at position 6 with a phenyl group does not modify the biological behaviour of the lead compound, whereasthe shift of the methyl group from position 2 to position 5 yields a compound devoid of inhibitory effects. In both the 6-thienyl and 6-phenylseries, the lengthening of the chain at position 2 has provided useful information to outline the structural determinants of the ubiquinoneantagonist action in imidazothiazole derivatives. © 1999 Éditions scientifiques et médicales Elsevier SAS

imidazo[2,1-b]thiazole / mitochondrial complex I / ubiquinone / rotenone

1. Introduction

NADH-ubiquinone reductase, commonly known ascomplex I, is the most complicated and least understoodof the respiratory complexes of mitochondria and bacte-ria [1–3].

The interest in this intricate enzyme complex is in-creasing due to its possible involvement in the pathogen-esis of human neurodegenerative diseases such as Alzhe-imer’s [4], Parkinson’s [5], diabetes [6] and the fact thatcomplex I is becoming a preferred target of commercialpesticides, especially acaricides [7]. There is a plethora ofcomplex I inhibitors that act as antagonists of the hydro-phobic substrate ubiquinone, and generally differ in theiraction depending upon which quinone intermediate theypreferentially antagonize (quinone, semiquinone orquinol) [8]. Because the chemical determinants of the

different antagonistic actions are unclear [1, 8, 9], thestudy of a series of derivatives in which the chemicalstructures have been systematically modified is useful toelucidate how complex I inhibitors interact with theenzyme [10].

Recently, we have undertaken a systematic work tostudy the structure-activity relationships in complex Iinhibitors bearing an indole [11, 12] and imidazothiazolemoiety [13]. 2-Methyl-6-(2-thienyl)imidazo[2,1-b]thia-zole1 (figure 1)was the most potent compound and wasfound to have a mode of action overlapping that of theclassical inhibitor rotenone as well as that of the product-like inhibitor myxothiazol [8, 13]. Using1 as the leadcompound, we have evaluated first the effect of thesubstitution of the thienyl group in position 6 with aphenyl group (2) and the shift of the methyl group fromposition 2 to 5 (3). The results indicated that the thienylgroup may be replaced by a phenyl group without loss ofpotency, whereas the position of the methyl group iscritical (compound3 is inactive). We next planned thesynthesis of 6-thienyl and 6-phenyl derivatives with

§Presented in part at the First Italian-Swiss Meeting on Medi-cinal Chemistry (Torino, Italy, September 23–26, 1997).*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 883−889 883© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

various substituents at position 2, to identify whether thisposition in the imidazothiazole ring corresponded to theattachment point for mimics of the polyisoprenoid sidechain of ubiquinone, whose length is known to be criticalfor Complex I activity [14, 15].

If this were the case, 2-substituents with increasinghydrophobicity (13–17and19–24) were expected to havean enhanced inhibitory potency, proportional to theirhydrophobicity as previously reported for 2-substitutedacridones [16]. However, the structure-activity resultsindicated that, in imidazothiazoles, position 2 is stericallycritical for complex I inhibition, but is unlikely to be thejunction for mimics of the ubiquinone side chain. Duringthe synthesis, some unexpected compounds were isolated(18 and25–27) and two of them were tested (26 and27).

2. Chemistry

The 5-alkyl-2-aminothiazoles6 (figure 1) were pre-pared from the appropriate aldehydes4 which werebrominated at theα-position (5) and treated with thio-urea. When compounds6 (R = ethyl, propyl) were treatedwith the bromoacetylthiophens7–8, white solids wereisolated, corresponding to the intermediate salts10–11which were characterized only on the basis of their IRspectrum (νC=O ≈ 1 670 cm–1). These compounds wererefluxed with dilute hydrochloric acid in order to obtainthe imidazothiazoles13–17, whose spectroscopic data(table I) are in agreement with the assigned structures.Under the same experimental conditions, starting from6(R = butyl) and 3-(bromoacetyl)thiophene8, the resultingprecipitate had a structure different from10–11since the

Figure 1. Synthesis of 6-thienyl and 6-phenylimidazo[2,1-b]thiazoles.

884

IR showed an additional C=O stretching absorption. Thiscompound was considered as the iminothiazolone12which, after refluxing with dilute hydrochloric acid, gavethe imidazole18. The structure of this new derivative wasconfirmed by IR,1H-NMR (table I) and 13C-NMR, MS(see Experimental section).

The same reaction was repeated with 2-bromo-acetophenone9 and similar behaviour was observed:when R was ethyl or propyl, the intermediate salt wasanalogous to11, leading to compounds19–21, whereas,when R was a longer chain, the precipitate was analogousto 12 (it led to compounds25–27) and the filtrate

Table I. Imidazothiazoles13–27.

Compound Formula(mw) M.p., °C νmax, cm-1 δ (ppm); J (Hz) in DMSO-d6a

13 b C11H10N2S2

(234.3)111–113 1 265, 850, 725, 685 1.25 (3H, t, CH3, J = 7.5) 2.77 (2H, q, CH2, J = 7.5) 7.06 (1H, m, T)

7.35 (1H, m, T) 7.41 (1H, m, T) 7.70 (1H, s, th) 8.01 (1H, s, im)14 c C12H12N2S2

(248.4)105–106 1 270, 850, 720, 680 0.95 (3H, t, CH3, J = 7) 1.64 (2H, sex, CH2, J = 7) 2.73 (2H, t, CH2,

J = 7) 7.07 (1H, m, T) 7.36 (1H, m, T) 7.41 (1H, m, T) 7.71 (1H, s, th)8.02 (1H, s, im)

15 C12H12N2S2 108–112 1 660, 1 620, 1 560, 1 060 1.29 (6H, d, CH3, J = 7) 3.12 (1H, sep, CH,J = 7) 7.07 (1H, m, T) 7.36(1H, m, T) 7.41 (1H, m, T) 7.71 (1H, s, th) 8.00 (1H, s, im)(248.4)

16 C11H10N2S2 103–105 1 270, 1 070, 780, 715 1.25 (3H, t, CH3, J = 7.5) 2.76 (2H, q, CH2, J = 7.5) 7.46 (1H, m, T)7.55 (1H, m, T) 7.69 (1H, m, T) 7.71 (1H, s, th) 7.98 (1H, s, im)(234.3)

17 C12H12N2S2

(248.4)126–130 1 270, 1 260, 780, 725 0.95 (3H, t, CH3, J = 7.5) 1.63 (2H, sex, CH2, J = 7.5) 2.72 (2H, t,

CH2, J = 7.5) 7.46 (1H, m, T) 7.56 (1H, m, T) 7.69 (1H, m, T) 7.71(1H, s, th) 7.98 (1H, s, im)

18 C13H16N2O2S2

(296.4)176–180 3 500–2 200, 1 620, 1 590,

1 070, 7800.85 (3H, t, CH3, J = 7) 1.32 (4H, m, CH2) 1.76 (2H, m, CH2) 3.94(1H, t, CH,J = 7) 7.44 (1H, m, T) 7.56 (1H, s, im) 7.58 (1H, m, T) 7.67(1H, m, T)

19 d C13H12N2S(228.3)

136–140 1 600, 1 195, 1 065, 720 1.26 (3H, t, CH3, J = 7.5) 2.78 (2H, q, CH2, J = 7.5) 7.24 (1H, t, ar,J = 8) 7.38 (2H, t, ar,J = 8) 7.72 (1H, s, th) 7.81 (2H, d, ar,J = 8) 8.13(1H, s, im)

20 C14H14N2S(242.3)

142–145 1 600, 1 260, 1 065, 720 0.94 (3H, t, CH3, J = 7.5) 1.63 (2H, sex, CH2, J = 7.5) 2.72 (2H, t,CH2, J = 7.5) 7.23 (1H, t, ar,J = 8) 7.37 (2H, t, ar,J = 8) 7.71 (1H,s, th) 7.80 (2H, d, ar,J = 8) 8.12 (1H, s, im)

21 C14H14N2S(242.3)

136–140 1 595, 1 180, 770, 710 1.29 (6H, d, CH3, J = 7) 3.12 (1H, sep, CH,J = 7) 7.24 (1H, t, ar,J= 8) 7.38 (2H, t, ar,J = 8) 7.71 (1H, s, th) 7.81 (2H, d, ar,J = 8) 8.11(1H, s, im)

22 C15H16N2S(256.4)

128–130 1 595, 1 535, 1 260, 720 0.89 (3H, t, CH3, J = 7.5) 1.34 (2H, sex, CH2, J = 7.5) 1.58 (2H, qui,CH2, J = 7.5) 2.73 (2H, t, CH2, J = 7.5) 7.22 (1H, t, ar,J = 8) 7.36 (2H,t, ar, J = 8) 7.71 (1H, s, th) 7.79 (2H, d, ar,J = 8) 8.12 (1H, s, im)

23 C16H18N2S(270.4)

130–133 1 590, 1 250, 1 055, 710 0.88 (3H, t, CH3, J = 7) 1.32 (4H, m, CH2) 1.62 (2H, qui, CH2, J = 7)2.75 (2H, t, CH2, J = 7) 7.24 (1H, t, ar,J = 8) 7.38 (2H, t, ar,J = 8)7.73 (1H, s, th) 7.81 (2H, d, ar,J = 8) 8.13 (1H, s, im)

24 C17H20N2S(284.4)

109–111 1 590, 1 255, 760, 715 0.86 (3H, t, CH3, J = 7) 1.29 (6H, m, CH2) 1.60 (2H, qui, CH2, J = 7)2.74 (2H, t, CH2, J = 7) 7.23 (1H, t, ar,J = 8) 7.38 (2H, t, ar,J = 8)7.71 (1H, s, th) 7.80 (2H, d, ar,J = 8) 8.13 (1H, s, im)

25 C15H18N2O2S(290.4)

176–179 3 500–2 200, 1 615, 1 595,1 060, 750

0.85 (3H, t, CH3, J = 7) 1.33 (4H, m, CH2) 1.77 (2H, m, CH2) 3.93(1H, t, CH,J = 7) 7.20 (1H, t, ar,J = 8) 7.35 (2H, t, ar,J = 8) 7.65 (1H,s, im) 7.72 (2H, d, ar,J = 8)

26 C16H20N2O2S(304.4)

185–187 3 500–2 200, 1 610, 1 590,1 055, 750

0.84 (3H, t, CH3, J = 7) 1.26 (4H, m, CH2) 1.38 (2H, m, CH2) 1.77(2H, m, CH2) 3.94 (1H, t, CH,J = 7) 7.20 (1H, t, ar,J = 8) 7.36 (2H,t, ar, J = 8) 7.64 (1H, s, im) 7.75 (2H, d, ar,J = 8)

27 C17H22N2O2S(318.4)

185–188 3 500–2 200, 1 610, 1 595,1 060, 750

0.83 (3H, t, CH3, J = 7) 1.23 (6H, s, CH2) 1.38 (2H, t, CH2, J = 7) 1.79(2H, m, CH2) 3.94 (1H, t, CH,J = 7) 7.20 (1H, t, ar,J = 8) 7.36 (2H,t, ar, J = 8) 7.64 (1H, s, im) 7.73 (2H, d, ar,J = 8)

aAbbreviations: T = thiophene, th = thiazole, im = imidazole, ar = aromatic.bRef. [24], m.p. 116–118.cRef. [23], m.p. not reported.dRef. [24,25], m.p. 125–127.

885

contained the intermediate analogous to11 which, afterthe usual treatment with dilute hydrochloric acid, gavethe imidazothiazoles22–24.

For a deeper insight into the structure of the carboxylicacids 18 and 25–27, one of these, 2-(4-phenyl-2-imidazolylsulfanyl)heptanoic acid26, was subjected toDEPT, HETCOR, MS (see Experimental section) and tocrystallographic studies. Its crystal structure shows thatthe imidazole and the phenyl ring are almost planar asobserved in analogous compounds [17, 18], whereas theorientation of the carboxylic group is almost perpendicu-lar to them. Further crystallographic details will be givenelsewhere.

3. Biological results

Table II reports the biological activity of the referencecompounds (1–3) and of the newly synthesized imidazo-thiazole derivatives (13–17 and 19–24). Among theunexpected imidazole by-products, compounds26–27were quite inactive. Consequently, the analogues18 and25 were not tested.

The increase of hydrophobicity at position 2 in both6-(2-thienyl) (13–15) and 6-(3-thienyl) derivatives (16and 17) decreased the inhibitory potency on complex Iactivity in comparison with the 2-methyl derivative1.Therefore, position 2 seems to be very critical for thepotency of complex I inhibition.

Contrary to 6-thienyl derivatives, 6-phenyl derivativeswith 2-substituents of increasing hydrophobicity eithermaintained or reduced the inhibitory capacity of the leadcompound. There was no direct correlation between theincrease in length and hydrophobicity of the substituentand inhibition, since the 2-ethyl (19) and the 2-butyl (22)derivatives had a markedly reduced potency with respectto the 2-propyl derivative (20). Nevertheless, 2-pentyl(23) and mostly 2-esyl (24) derivatives showed an in-creased inhibitory effect, thus suggesting that the lengthof the chain is important for the interaction with thecomplex I binding site.

The different behaviour of the thienyl or phenyl deriv-atives indicated a slightly different interaction as inhibi-tors of complex I. In previous work we studied in detailthe inhibitory properties of compound1, which appearedto act as a quinol antagonist [13]. In order to clarify therelationship between the new phenyl derivatives andquinone interaction in complex I, compounds2 and 20were tested with two different substrates, namely thehydrophilic ubiquinone-1 (Q1) and the hydrophobicundecyl-benzoquinone (UBQ). The potency of2 washigher with Q1 (I50 = 75 µM) than with UBQ (I50 = 130µM), similarly to compound1 [13]. Conversely, com-pound20showed the same I50 for both Q1 and UBQ (120µM). Thus, the inhibitory effect of the more hydrophobic20 could depend on some internal reaction steps that arenot extensively influenced by the nature of the quinonesubstrate for complex I activity. Indeed, compound20acted as a classical non-competitive inhibitor with respectto both Q1 and UBQ, similarly to rotenone(figure 2)[8],while 2 showed mixed competition with UBQ, as previ-ously reported for1 [13]. Taken together, the biologicaldata suggest that the substitution of a thienyl group witha bulkier phenyl group at position 6 could shift theinhibitory action of imidazothiazole compounds fromantagonists of the quinol product to antagonists of thesemiquinone intermediate. The latter would correspond tothe action of rotenone in complex I [8].

4. Discussion

Structurally, most complex I inhibitors have a head-tailmodule that is critical for the antagonist action versus theubiquinone substrate [8, 10]. The results of the presentstudy clarify that complex I inhibition is strictly depen-dent on the substitutions at the extreme positions 2 and 6of the imidazothiazole ring. The different potency of thesubstituents at the 2 position, when position 6 is occupiedby either a thienyl or a phenyl group, indicates that stericconstraints in the binding site limit the length andbulkiness of imidazothiazoles for their optimal interac-

Table II. Biological activity of compounds1–3, 13–17, 19–24and26–27.

Compound Residual activity of NADH:UBQreductase (%)a

1 122 243 80

13 6314 9115 6316 6317 10019 8220 1721 10022 7323 4624 026 11427 107

aThe numbers are the average of at least three separate experi-ments. The assay conditions are described in the experimentalsection. UBQ (30µM) was used as electron acceptor. The finalconcentration of the compounds was 0.25 mM.

886

tion with complex I. In phenyl derivatives, the aliphaticchain substituents in position 2 might mimic the hydro-phobic tail of ubiquinone. A further lengthening of thechain which resembles the polyprenil tail of ubiquinoneshould better clarify if position 2 corresponds to theattachment point for hydrophobic substituents.

Future studies with systematic substitutions at otherpositions such as nitrogen 7 will define the determinantsto optimize the ubiquinone antagonist action of imida-zothiazoles and produce extremely potent inhibitors ofcomplex I.

5. Experimental

5.1. Chemistry

Compounds2–3 have been previously described [19].The aldehydes4 (butyraldehyde, valeraldehyde, hexanal,heptanal, octanal) and 2-bromoacetophenone9 are com-mercially available. 2-(Bromoacetyl)thiophene7 [20] and3-(bromoacetyl)thiophene8 [21] were prepared accord-ing to the literature. The bromoaldehydes5 were preparedunder experimental conditions analogous to those re-ported for the synthesis ofα-bromopropionaldehyde [22].The imidazo[2,1-b]thiazoles13 [23, 24], 14 [23] and19 [25] were already reported in the literature.

The melting points are uncorrected. Analyses (C, H, N)were within ± 0.4% of the theoretical values. TLC wasperformed on Bakerflex plates (Silica gel IB2-F): theeluent was a mixture of petroleum ether/acetone invarious proportions. The IR spectra(table I) were re-corded in nujol on a Perkin-Elmer 683. The1H-NMR(table I) and the13C-NMR spectra were recorded on aVarian Gemini (300 MHz), and were referenced tosolvent signals (DMSO). The EI mass spectra wererecorded on a VG 7070E. X-ray crystallography: the datawere collected on a Siemens P4 four-circle diffractome-ter.

5.1.1. 5-Alkyl-2-aminothiazoles6The appropriate bromoaldehyde5 (100 mmol) was

treated with 80 mmol of thiourea and stirred at 90–100 °Cfor 2 h. After cooling, the reaction mixture was treatedwith 2 N NaOH until basic and extracted with chloro-form. Compounds6 thus obtained were used, withoutfurther purification, in the following step.

5.1.2. Reaction of the 5-alkyl-2-aminothiazoles6 with thebromoacetylthiophenes7–8 (synthesis of13–18)

The appropriate compound6 (45 mmol) was dissolvedin acetone (100 mL) and treated with 2-(bromo-acetyl)thiophene 7 or 3-(bromoacetyl)thiophene8(45 mmol). The reaction mixture was refluxed for 3–6 h(according to a TLC test) and the resulting salt10–12wastreated, without further purification, with 200 mL of 2 NHCl. After 1 h reflux, the solution was cautiously basifiedby adding dropwise 15% NH4OH. The resulting base(13–18) was collected by filtration and crystallized fromethanol with a yield of 35–40%(table I).

The following spectra were also recorded:1) as an example of compound10, R = isopropyl was

chosen,1H-NMR: 1.21 (6H, d, CH3, J = 7); 3.01 (1H,sep, CH,J = 7); 5.69 (2H, s, CH2); 7.20 (1H, s, th); 7.38(1H, m, T); 8.11 (1H, m, T); 8.18 (1H, m, T); 9.59 (1H,s, NH).

Figure 2. Effect of compound2 and20on the Q1 (A) and UBQ(B) titration of the NADH:Q reductase activity. (●) Control.(∇) In the presence of 0.18 mM compound2. ([) In thepresence of 0.18 mM compound20.

A

B

887

2) compound12, IR: 3 500–2 200, 1 760, 1 685, 1 640,1 560 cm–1. 1H-NMR: 0.87 (3H, t, CH3, J = 7); 1.40 (4H,m, CH2); 2.05 (2H, m, CH2); 4.90 (1H, t, CH,J = 7); 5.37(2H, s, CH2); 7.58 (1H, m, T); 7.74 (1H, m, T); 8.78 (1H,m, T). MS: 296 (M+⋅, 11), 278 (M-H2O, 4), 268 (M-CO,9), 124 (C6H4SO, 30), 111 (C5H3SO, 100).

3) Compound18, MS: 296 (M+⋅, 13), 278 (M-H2O,100), 235 (278-C3H7, 85), 207 (235-CO, 40), 182 (49),123 (26), 109 (27), 55 (34).

5.1.3. Reaction of the 5-alkyl-2-aminothiazoles6 with2-bromoacetophenone9 (synthesis of19–27)

The appropriate compound6 (10 mmol) was dissolvedin acetone (50 mL) and treated with 10 mmol of2-bromoacetophenone9. The mixture was refluxed for2–5 h (according to a TLC test) and worked up with twoprocedures. For compounds19–21,acetone was evapor-ated under reduced pressure and the resulting residue wasrefluxed for 1 h with 100 mL of 2 N HCl. The solutionthus obtained was then basified with 15% NH4OH andthe resulting base19–21 was crystallized from ethanolwith a yield of 30–35%(table I). For Compounds22–27the precipitate was collected by filtration and worked upas above. Compounds25–27 were crystallized fromethanol with a yield of 15%. The filtrate, evaporated andworked up in the same manner, gave compounds22–24which were crystallized from ethanol with a yield of 3%.The following data were also recorded for compound26:13C-NMR (with DEPT and HETCOR): 13.80 (CH3),21.90 (CH2), 26.13 (CH2), 30.81 (CH2), 31.83 (CH2),50.29 (S-CH), 118.04 (im-5), 124.34 (ar-2 + 6), 126.56(ar-4), 128.59 (ar-3 + 5), 133.13 (im-4), 138.09 (ar-1),139.67 (im-2), 172.43 (COOH). MS: 304 (M+⋅, 45), 286(M-H2O, 8), 260 (M-CO2, 16), 229 (286-C4H9, 15), 203(48), 190 (70), 176 (100), 117 (56). Crystal data:C16H20N2O2S,M = 304.4, Monoclinic,a = 17.305 (2),b= 8.5690 (10),c = 10.8620 (10) Å,â= 99.67 (1)°,V =1587.8 (3) Å3 (by least-squares refinement on diffracto-meter angles for 43 randomly selected and automaticallycentred reflections), space group P21/c (n. 14), Z = 4, F(000) = 648,Dc = 1.27 g cm–3, µ (Mo-Kα) = 0.21 mm–1.

5.2. Biology

5.2.1. Mitochondrial preparationMitochondria from beef heart were prepared according

to standard procedures. They were used after at least twocycles of freezing and thawing in order to brake downmembranes.

5.2.2. Biochemical assaySubmitochondrial particles from beef heart were pre-

pared essentially as described by Hansen and Smith [26].

The redox activity of NADH:ubiquinone oxidoreductasewas measured by using Q1 or UBQ as electron acceptor.The test compound, dissolved in DMSO, was added to0.4 mg/mL of mitochondrial protein in 2 mL of 50 mMK2HPO4 buffer containing 1 mM EDTA, 2 mM KCN (pH7.4) and 0.1 mM NADH at room temperature. Thereaction was started by quinone and followed by decreasein absorbance of NADH in the dual wavelength mode at350 minus 410 nm with an extinction of 5.5 mM–1 cm–1

as previously described [27].

Acknowledgements

This work was supported by grants from the ItalianNational Research Council (CNR, Rome) and from theUniversity of Bologna (Funds for Selected ResearchTopics).

References

[1] Walker J.E., Q. Rev. Biophys. 25 (1992) 253–324.

[2] Videira A., Biochim. Biophys. Acta 1364 (1998) 89–100.

[3] Yagi T., Yano T., Di Bernardo S., Matsuno-Yagi A., Biochim.Biophys. Acta 1364 (1998) 125–133.

[4] Weyer G., Babej-Dolle R.M., Hadler D., Hofmann S., HermannW.M., Neuropsychobiology 36 (1997) 73–82.

[5] Schapira A.H.V., Biochim. Biophys. Acta 1364 (1998) 261–270.

[6] Degli Esposti M., Ngo A., Mayers M.A., Diabetes 45 (1996)1531–1534.

[7] Lummen P., Biochim. Biophys. Acta 1364 (1998) 287–296.

[8] Degli Esposti M., Biochim. Biophys. Acta 1364 (1998) 222–235.

[9] Degli Esposti M., Ghelli A., Ratta M., Cortes D., Estornell E.,Biochem. J. 301 (1994) 161–167.

[10] Miyoshi H., Biochim. Biophys. Acta 1364 (1998) 236–244.

[11] Andreani A., Rambaldi M., Locatelli A., Andreani F., Poglayen G.,Degli Esposti M., Eur. J. Med. Chem. 27 (1992) 729–734.

[12] Andreani A., Rambaldi M., Locatelli A., Leoni A., Ghelli A., DegliEsposti M., Pharm. Acta Helv. 69 (1994) 15–20.

[13] Andreani A., Rambaldi M., Leoni A., Locatelli A., Ghelli A., RattaM., Benelli B., Degli Esposti M., J. Med. Chem. 38 (1995)1090–1097.

[14] Lenaz G., Biochim. Biophys. Acta 1364 (1998) 207–221.

[15] Degli Esposti M., Ngo A., McMullen G.L., Ghelli A., Sparla F.,Benelli B., Ratta M., Linnane A.W., Biochem. J. 313 (1996)327–334.

[16] Oettmeier W., Masson K., Soll M., Reil E., Biochem. Soc. Trans. 22(1994) 213–216.

[17] Shilcrat S.C., Hill D.T., Bender P.E., Griswold D.E., Baures P.W.,Eggleston D.S., Lantos I., Pridgen L.N., J. Heterocycl. Chem. 28(1991) 1181–1187.

[18] Jain P.C., Acta Crystallogr. Sect. C: Cryst. Struct. Commun. 43(1987) 2415–2418.

888

[19] Pyl T., Giebelmann R., Beyer H., Liebigs Ann. Chem. 643 (1961)145–153.

[20] Kipnis F., Soloway H., Ornfelt J., J. Amer. Chem. Soc. 71 (1949)10–11.

[21] MacDowell D.W.H., Greenwood T.D., J. Heterocycl. Chem. 2(1965) 44–48.

[22] Reichel L., Jahns H.J., Liebigs Ann. Chem. 751 (1971) 69–76.

[23] Tanaka M., Otsuka K., Obata M., Tanabe T., Saida T., Nomura K.,Amemija K., Saga K., Kano S., Japan Kokai 74 70, 986; Chem.

Abstr. 81 (1974) 136142c.

[24] Meakins G.D., Musk S.R.R., Robertson C.A., Woodhouse L.S., J.Chem. Soc. Perkin Trans. I (1989) 643–648.

[25] O’Daly M.A., Hopkinson C.P., Meakins G.D., Raybould A.J., J.Chem. Soc. Perkin Trans. I (1991) 855–860.

[26] Hansen M., Smith A.L., Biochim. Biophys. Acta 81 (1964) 214–222.

[27] Degli Esposti M., Ghelli A., Crimi M., Estornell E., Fato R.,Lenaz G., Biochem. Biophys. Res. Commun. 190 (1993)1090–1096.

889

Original article

NO-Donors, part 3: nitrooxyacylated thiosalicylates and salicylates – synthesisand biological activities#

Stefan Endresa, Andreas Hackerb, Eike Noackb,Georg Kojdab, Jochen Lehmanna*

aPharmazeutisches Institut, Rheinische Friedrich-Wilhelms-Universität Bonn, An der Immenburg 4, D-53121 Bonn, GermanybInstitut für Pharmakologie, Heinrich-Heine-Universität Düsseldorf, Moorenstr. 5, D-40225 Düsseldorf, Germany

(Received 8 December 1998; revised 29 March 1999; accepted 31 March 1999)

Abstract – Organic nitrates release nitric oxide when incubated with thiosalicylic acid. S-Nitrooxyacylated esters and amides of thiosalicylicacid, together with the corresponding salicylates, were synthesized in order to perform a first in vitro evaluation of these newnitrate-thiol-hybrid prodrugs. These prodrugs might release NO in vivo after biotransformation without the use of endogenous reductives.None of these prodrugs released NO spontaneously when dissolved in buffer solution, but they did activate soluble guanylyl cyclase andinduced vasodilatation of phenylephrine-pretreated male Wistar rat aorta in a potency range between that of isosorbiddinitrate and glyceroletrinitrate. © 1999 Éditions scientifiques et médicales Elsevier SAS

organic nitrates / nitric oxide / thiosalicylates / salicylates / vasodilatation

1. Introduction

For more than a century organic nitrates have beenused in the treatment of angina pectoris and congestiveheart failure. The activity of these nitrates is mainlyattributed to the release of nitric oxide (NO), an endo-genous mediator with a rapidly growing list of physi-ological and pathophysiological functions [2]. The libera-tion of NO from organic nitrates in vivo is most likely anenzymatic metabolic reduction process [2, 3]. This doesnot however rule out the possibility of non-enzymaticreduction of nitrates by specific thiols. In general, allthiols reduce organic nitrates to inorganic nitrite, but onlya very few of them are able to produce NO as well. Thebasic structural feature of these special thiols is a carbo-nyl function located two carbons away from a thiol groupin a coplanar orientation as it is realized in cysteine,N-acetyl-cysteine and thiosalicylic acid [4–6]. Combin-ing organic nitrate and a NO-liberating thiol in one

molecule leads to prodrugs which might be endowed withtypical properties of organic nitrates and may show amore facilitated mechanism of NO-release, which alsomight reduce the development of nitrate tolerance. Somepromising approaches have already been made fornitrate-cysteine combinations, primarily N-nitrooxy-acylated cysteines [7] but also cysteinamides, nitrooxya-cylated at the amide nitrogen [8], and isosorbiddinitrateconnected with a cyclized L-cysteinamide [9]. We estab-lished S-nitrooxyacylation both for various cysteines andthiosalicylic acid derivatives. Here we give a first reporton the synthesis, stability and biological activities ofS-nitrooxyacylthiosalicylates. To evaluate the influenceof the thiosalicylate substructure the analogue salicylateswere synthesized as well.

2. Chemistry

The target compounds (SE) could not be obtained fromS-bromoacyl-thiosalicylates and silver nitrate. They wereobtained instead by acylation [10] of various thiosalicy-lates and salicylates with nitratoacids, catalysed by car-bonyldiimidazole (CDI)(figure 1).

# For part 2 see [1]*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 895−901 895© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

The nitratoacids3 were obtained by following twodifferent methods. Treatment of the hydroxyesters1 witha mixture of fuming nitric acid/acetanhydride followedby acid or base catalysed hydrolysis of the esters2 (routeA, figure 2), or treatment of the halogenated acids4 withsilver nitrate in dry acetonitrile (route B).

These nitratoacids were linked with thiosalicylates andsalicylates via S- resp. O-acylation. Not all of the plannedcombinations of nitratoacids with thioles or phenoleswere possible. Two nitratoacids (3a and d) underwentdecomposition when activated with carbonyldiimidazole.3b could be activated but decomposed as soon as a thiole

Figure 1. General route to the target compounds.

Figure 2. Synthesis of the nitratoacids.

896

was added and compound3e showed neither decompo-sition nor acylation. Thiosalicylic acid itself, as well asthe N-unsubstituted thiosalicylamide, suffered rapid de-composition with all nitratoacids. Obviously there is acompetition between reduction of the nitrate by thethiole [2–4] and formation of the thioesters.Figure 3gives these target compounds which we have obtainedsuccessfully as stable solid compounds.

O-Nitratoacylation of salicylamid did not yield theexpected compound8 but rather a mixture of theN-acylated10 and the diacylatedSE 161. A rearrange-ment producing N-acetyl- from O-acetyl-salicylamid is

described in the literature [11]. In order to obtain acompletely O-acylated target compound we treated9with two equivalents of3c and obtained pureSE 161.Finally, we nitrooxylated thiophenol (a thiole which isunable to release NO from organic nitrates) and obtainedSE 135(figure 4).

2.1. Stability

Stability of the target compounds under simulatedphysiological conditions was investigated. Substanceswere dissolved in 50 mM phophate-buffer pH =

Figure 3. Synthesis of the target compounds (part 1).

897

7/acetonitrile (5:1) and stored at 37 °C for 24 h. Beforeand after that period, HPLC analysis using an RP-8column with UV-detector was performed to detect pos-sible instabilities due to the combination of organicnitrates and NO releasing thioles resp. analogues. Thediminishment of the starting material at 37 °C/24 hranged between 0–3.7% with the exception ofSE 145(16%) andSE 161 (18%) By means of its lipophiliccharacter,SE 85 could not be chromatographed on thiscolumn. A sufficient thermal stability of this compoundwas certified by NMR-spectroscopy after 24 h at 37 °C inDMSO.

3. Biological results and discussion

3.1. Spontaneous liberation of NO

Liberation of NO was determined electrochemically bya Clark-type NO-sensitive electrode (Iso-NO, WorldPrecision Instruments Inc., Berlin, Germany). Measure-ments were performed with 10–8 up to 10–4 M solutionsof the compounds in Krebs-Henseleit buffer (pH 7.4)with constant stirring at 37 °C. No release of NO wasfound with theSE-compounds. This confirms that an invivo biotransformation of these prodrugs is indispen-sable.

3.2. Guanylyl cyclase activation

The most active compound,SE 175,was chosen toverify that the vasorelaxation observed with theSE

compounds was mediated by sGC-stimulation. Aftertreatment with 100 mMSE 175the cGMP level in a rataorta segment increased significantly from 2.8± 0.6 (n =10) to 8.8± 2.8 (n = 5) pmol cGMP/mg protein (P =0.012).

3.3. Vasorelaxation of isolated vessel segments

Vasorelaxation after precontraction with phenylephrinewas measured using aorta segments of rats. All of the nineinvestigated nitrates proved to be potent vasodilators.Figure 1 gives the relaxation curves for the three mostpotent, and the activities (EC50 values) for all compoundsin decreasing order. The benzylnitrate derivative showedto be the most active compound. The ‘carrier’ of thenitrato group (thiosalicylic acid ester, thiosalicylic amide,thiophenol, salicylic acid ester or salicylic amide) doesnot seem to have any influence on the biological activityin this in vitro assay. In vivo investigations, like decreas-ing blood pressure under repeated treatment, will benecessary to evaluate this influence(figure 5).

4. Experimental protocols

4.1. Vasorelaxation of isolated vessel segments

Method: after cervical translocation, the aorta of maleWistar rats (250–300 g) was dissected free and rapidlyimmersed in cold oxygenated Krebs-Henseleit solution(pH 7.4). Four ring segments (5 mm) of the thoracic aortawere suspended in individual organ chambers (10 mL)

Figure 4. Synthesis of the target compounds (part 2).

898

filled with Krebs buffer. The solution was aerated con-tinuously with 95% O2 and 5% CO2 maintained at 37 °C.A linear force transducer (Statham force displacementtransducer) recorded the actual tension of the aortic rings.After equilibration at a resting tension of 4 g (1 h)contractile function of segments was tested by applicationof KCl (60 mM). The vasorelaxing activity was studiedby cumulative application of the compounds after pre-contraction with a single dose of phenylephrine (1µM).

4.2. Activity of soluble guanylyl cyclase

Activity of soluble guanylyl cyclase was measured asreported earlier [12] by formation of [32P]-cGMP from[α-32P]-GTP.

4.3. Chemistry

Melting points were determined in open capillary tubeson a digital Gallenkamp melting point apparatus and arenot corrected. The IR spectra were recorded on a Perkin-Elmer 1420 using KBr pellets for solids and NaCl platesfor liquid substances. The1H-NMR spectra were ob-tained on a Bruker WM 200 (200 MHz) spectrometerusing DMSO-d6 as the solvent. Chemical shifts arereported inδ = ppm relative to tetramethylsilane as theinternal standard. Elemental analyses were carried outwith a Heräus CHN-O-Rapid or an Elementar Vario ELand were within± 0.4% of the calculated values.

4.3.1. Methyl thiosalicylate(6a)A mixture of 20.0 g (0.13 mol) thiosalicylic acid, 5.0 g

4-toluenesulfonic acid, 30 mL methanol and 500 mL

toluene was refluxed under argon for 48 h, washed withwater, saturated NaHCO3 solution, and water again(100 mL each), then dried (Na2SO4) and evaporated. Theresidue was distilled. 15.9 g (73%) colourless oil; b.p. =67–69 °C, 0.03 mm Hg ( [13]: 87–93 °C, 0.40 mm Hg);IR (KBr) cm–1: 1 700 (C=O);1H-NMR δ 3.83 (s, 3H,CH3), 5.26 (s, 1H, SH), 7.1–7.5 (m, 3H, aromat. H-3,4,5),7.92 (dd,J = 7.1/2.4 Hz, 1H, aromat. H-6).

4.3.2. Ethyl thiosalicylate(6b)As described for6a, using 15 mL ethanol. 16.7 g

(70%) colourless oil; b.p. = 69–72 °C, 0.04 mm Hg( [14]: b.p. not given); IR (KBr) cm–1: 1 700 (C=O);1H-NMR δ 1.28 (t,J = 8.5 Hz, 3H, CH3), 4.41 (q,J = 8.5Hz, 2H, CH2), 7.2–7.7 (m, 3H, aromat. H-3,4,5), 8.05(dd, J = 7.2/2.2 Hz, 1H, aromat. H-6).

The nitrooxycarboxylic acids were prepared as re-ported (3a [7, 15], 3c [7], 3d [15], 3e[15]), with theexception of the following:

4.3.3. 2-Nitrooxyisobutyric acid(3b)6.28 mL of fuming HNO3, followed by 14.3 mL acetic

acid anhydride were dropped into a stirred solution of5.0 g (76 mmol) ethyl 2-hydroxyisobutyrate and 0.2 gurea in 500 mL CH2Cl2, maintaining a temperature<10 °C. After stirring at room temperature for 24 h, themixture was poured into 800 mL of icewater. The organiclayer was separated, washed with water, saturatedNaHCO3 solution, and water again, then evaporated at<40 °C. Distillation of the oily residue produced 8.9 g(66%) of colourless ethyl 2-nitrooxyisobutyrate (b.p. =33–34 °C, 0.04 mm Hg). For hydrolysis, 5.0 g (28 mmol)of this ester were dissolved in 250 mL of CH3OH andadded at 10 °C to a solution of 3.17 g (56 mmol) KOH in20 mL of water. After stirring at room temperature for≅2 h (TLC control) the mixture was acidified with concen-trated hydrochloric acid and the solvent was evaporated.100 mL water were added and the mixture was extractedwith CH2Cl2 (2 × 200 mL). The organic layer was dried(Na2SO4), evaporated, and the residue crystallized fromn-hexane. 2.80 g (67%) colourless crystals; m.p. = 76 °C( [16]: 78–79 °C); IR (KBr) cm–1: 1 715 (C=O), 1 630and 1 290 (N=O);1H-NMR δ 1.59 (s, 6H, 2× CH3);13.50 (s, 1H, COOH). Anal. C4H7NO5 (C, H, N).

4.3.4. 11-Nitrooxyundecanoic acid(3g)5.0 g (18.85 mmol) 1l-bromoundecanoic acid in 30 mL

of dry acetonitrile were added to 3.52 g (20.72 mmol) ofAgNO3 dissolved in 30 mL acetonitrile and stirred for 3 hat 60 °C. The filtrate of the reaction mixture was pouredinto 250 mL of icewater and the precipitated product wasseparated, washed with water and dried. 3.38 g (73%)white powder; m.p. = 40–41 °C; IR (KBr) cm–1: 1 700

Figure 5. Relaxation-dose curves (n = 9–10) and halfmaximalvasorelaxing activity (EC50 values, given in mol/L,n = 2–10) ofnitrooxyacylated thiosalicylates and salicylates in thoracic aortavessels of rats.

899

(C=O), 1 620 and 1 280 (N=O);1H-NMR δ 1.20–1.80(m, 16H, 8× CH2); 2.20 (t, J = 7.8 Hz, 2H, CO-CH2);4.50 (t, J = 7.8 Hz, 2H, CH2-ONO2); 12.00 (s, 1H,COOH). Anal. C11H21NO5 (C, H, N).

4.3.5. 4-Nitrooxymethylbenzoic acid(3f)5.38 g 4-bromomethylbenzoic acid in 30 mL of dry

acetonitrile were added to 4.95 g (29.0 mmol) silvernitrate in acetonitrile and stirred overnight. The filtrate ofthe mixture was poured into 500 mL of icewater. Theprecipitated crude product was separated, dried in vacuoand recrystallized from diisopropylether. 4.12 g (84%)white powder; m.p. = 165 °C; IR (KBr) cm–1: 1 690(C=O), 1 670 and 1 270 (N=O);1H-NMR δ 5.65 (s, 2H,CH2-ONO2); 7.58 (d, J = 8.1 Hz, 2H, aromat. H-3,5);7.98 (d,J = 8.1 Hz, 2H, aromat. H-2,4). Anal. C8H7NO5

(C, H, N).

4.3.6. Ethyl S-(11-nitrooxyundecanoyl)-thiosalicylate(SE 85)

3.0 g (12.1 mmol) of 11-nitrooxyundecanoic acid (3g)and 2.21 g (12.1 mmol) ethyl thiosalicylate (6b) weredissolved in 30 mL of dry CH2Cl2. 2.73 g (13.23 mmol)DCC, dissolved in 20 mL CH2Cl2, were added withprotection by argon and stirred for 24 h. The solid wasseparated, the filtrate washed with 0.1 N hydrochloricacid (3× 30 mL) and dried (Na2SO4). After evaporation,the crude product was chromatographed on a silica gelcolumn with ethylacetate. 0.86 g (17%) yellow paste; IR(NaCl) cm–1: 1 710 (C=O), 1 630 and 1 725 (N=O);1H-NMR δ 1.25–1.75 (m, 19H, 8× CH2, CH2-CH3), 2.65(t, J = 7.2 Hz, 2H, CO-CH2-(CH2)8), 4.25 (q,J = 7.1 Hz,2H, CH2-CH3), 4.50 (t, J = 6.6 Hz, 2H, -CH2-ONO2),7.50–7.70 (m, 3H, aromat. H-3,4,5); 8.82 (dd,J = 6.8/2.2Hz, 1H, aromat. H-6). Anal. C20H29NO6S (C, H, N).

4.3.7. S-(3-Nitrooxypivaloyl)-thiophenole(SE 135)1.63 g (10.0 mmol) of3c in 50 mL of dry DMF were

cooled to –10 °C and 1.78 g (11.0 mmol) of CDI wereadded. After stirring (argon) for 2 h, 1.10 g (10 mmol) ofthiophenol were added and the reaction mixture wasstirred for another 2 h at –10 °C. 50 mL ethylacetate wereadded, the mixture washed with saturated NaCl solution(3 × 30 mL), dried over Na2SO4 and evaporated. Theresidue was chromatographed on a silica gel column withpetrolether/acetone (7:1). 0.7 g (28%) colourless oil; IR(NaCl) cm–1: 1 690 (C=O), 1 640 and 1 275 (N=O);1H-NMR δ 1.35 (s, 6H, C-CH3), 4.68 (s, 2H, CH2-ONO2), 7.34–7.52 (m, 5H, aromat. H). Anal.C11H13NO4S(C, H, N).

4.3.8. Methyl O-(3-nitrooxypivaloyl)-salicylate(SE 136)1.52 g (10.0 mmol) methyl salicylate were treated with

1.63 g (10.0 mmol)3c and 1.78 g (11.0 mmol) of CDI asdescribed forSE 135. The crude product was chromato-graphed on a silica gel column with petroether/ethylacetate (5:1). 2.37 g (80%) white powder; m.p. =47–50 °C; IR (KBr) cm–1: 1 750 (C=O), 1 730 (C=O);1 625 and 1 280 (N=O);1H-NMR δ 1.39 (s, 6H, C-CH3),3.80 (s, 3H, O-CH3), 4.75 (s, 2H, CH2-ONO2), 7.20 (dd,J = 8.1/1.0 Hz, 1H, aromat. H-3), 7.42 (ddd,J = 8.1/1.0Hz, 1H, aromat. H-5), 7.68 (ddd,J = 8.1/1.5 Hz, 1H,aromat. H-4), 7.93 (dd,J = 8.1/1.5 Hz, 1H, aromat. H-6).Anal. C13H15NO7 (C, H, N).

4.3.9. Methyl S-(3-nitrooxypivaloyl)-thiosalicylate(SE145)

2.58 g (15.3 mmol) 5a were treated with 2.50 g(15.3 mmol) 3c and 2.73 g (16.86 mmol) CDI as de-scribed forSE 135. The crude product was chromato-graphed on a silica gel column with petrolether/ethylacetate (5:1). 3.19 g (67%) white crystals; m.p. =38–39 °C (crystallized from petrolether in the cold); IR(KBr) cm–1: 1 730 (C=O), 1 690 (C=O), 1 630 and 1 275(N=O); 1H-NMR δ 1.35 (s, 6H, C-CH3), 3.79 (s, 3H,O-CH3), 4.66 (s, 2H, CH2-ONO2), 7.50–7.70 (m, 3H,aromat. H-3,4,5), 7.88 (dd,J = 7.6/2.0 Hz, 1H, aromat.H-6). Anal. C13H15NO6S (C, H, N).

4.3.10. Ethyl S-(3-nitrooxypivaloyl)-thiosalicylate(SE152)

Synthesized from 2.0 g (10.81 mmol) ethyl thiosalicy-late (6b), 1.76 g (10.81 mmol) 3c and 1.93 g(11.89 mmol) CDI as described forSE 135. The crudeproduct was chromatographed on a silica gel column withpetroleum ether/ethylacetate (4:1). 2.42 g (68%) yellow-ish oil; IR (NaCl) cm–1: 1 720 (C=O), 1 700 (C=O),1 635 and 1 280 (N=O);1H-NMR δ 1.28 (t,J = 7.1 Hz,2H, CH2-CH3), 1.35 (s, 6H, C-CH3), 4.26 (q,J = 7.1 Hz,2H, CH2-CH3), 4.65 (s, 2H, CH2-ONO2), 7.50–7.65 (m,3H, aromat. H-3,4,5), 7.87 (dd,J = 7.1/2.0 Hz, 1H,aromat. H-6). Anal. C14H17NO6S (C, H, N).

4.3.11. O-(3-Nitrooxypivaloyl)-salicylic acid dimethyl-amide(SE 157)

Synthesized from 1.65 g (10.0 mmol) salicylic aciddimethylamide (7b) [17], 1.63 g (10.0 mmol)3c and1.78 g (11.0 mmol) CDI as described forSE 135. Thecrude product was chromatographed on a silica gelcolumn with ethylacetate. 3.0 g (96%) colourless oil; IR(NaCl) cm–1: 1 750 (C=O), 1 650 (C=O), 1 640 and 1 280(N=O); 1H-NMR δ 1.32 (s, 6H, C-CH3); 2.76 (s, 3H,N-CH3); 2.94 (s, 3H, N-CH3); 4.68 (s, 2H, CH2-ONO2);7.19 (d,J = 8.1 Hz, 1H, aromat. H-3); 7.33–7.36 (m, 2H,

900

aromat. H-4,5); 7.48 (m, 1H, aromat. H-6). Anal.C14H18N2O6 (C, H, N).

4.3.12. S-(3-Nitrooxypivaloyl)-thiosalicylic acid dimethyl-amide(SE 158)

Synthesized from 1.81 g (10.0 mmol)7a [17], 1.63 g(10.0 mmol)3c and 1.78 g (11.0 mmol) CDI as describedfor SE 135. The crude product was chromatographed ona silica gel column with ethylacetate. 1.79 g (55%) thickbrown oil; IR (NaCl) cm–1: 1 685 (2× C=O), 1 630 and1 280 (N=O);1H-NMR δ 1.32 (s, 6H, C-CH3); 2.64 (s,3H, N-CH3); 2.95 (s, 3H, N-CH3); 4.65 (s, 2H, CH2-ONO2); 7.34–7.60 (m, 4H, aromat. H). Anal.C14H18N2O5S (C, H, N).

4.3.13.N, O-Di-(3-nitrooxypivaloyl)-salicylamide(SE 161)Synthesized from 1.0 g (7.29 mmol) salicylamide,

2.38 g (14.58 mmol)3c and 2.48 g (15.31 mmol) CDI asdescribed forSE 135. The crude product was chromato-graphed on a silica gel column with petroleum ether/acetone (1:1). 1.12 g (36%) white powder; m.p. =73–75 °C; IR (KBr) cm–1: 1 760 (C=O), 1 675 (2×C=O), 1 630 and 1 275 (N=O);1H-NMR δ 1.27 (s, 6H,C-CH3); 1.33 (s, 6H, C-CH3); 4.67 (s, 2H, CH2-ONO2);4.68 (s, 2H, CH2-ONO2); 7.21 (d,J = 7.6 Hz, 1H, aromat.H-3); 7.36 (dd,J = 7.6/7.1 Hz, 1H, aromat. H-5); 7.48(dd, J = 7.1/1.5 Hz, 1H, aromat. H-6); 7.56 (ddd,J =7.6/7.6/1.5 Hz, 1H, aromat. H-4); 10.93 (s, 1H, NH).Anal. C17H21N3O10 (C, H, N).

4.3.14. Methyl S-(4-nitrooxymethylbenzoyl)-thiosali-cylate(SE 175)

Synthesized from 0.50 g (2.97 mmol)6a, 0.55 g(2.97 mmol)3f and 0.53 g (3.27 mmol) CDI as describedfor SE 135. The crude product was chromatographed ona silica gel column with petroleum ether/acetone (1:1).0.53 g (51%) white powder; m.p. = 61–64 °C; IR (KBr)cm–1: 1 715 (C=O), 1 660 (C=O), 1 620 and 1 270(N=O); 1H-NMR δ 3.76 (s, 3H, O-CH3); 5.70 (s, 2H,CH2-ONO2); 7.60–7.75 (m, 5H, aromat. H-3,4,5,3≠,5≠);7.90–8.08 (m, 3H, aromat. H-6,2≠,6≠). Anal.C16H13NO6S (C, H, N).

4.3.15. Stability

The HPLC equipment used was a Consta Metric 3200(LCD-Analytical) pump with a Spherisorb 5 ODS 2 pre-(20 × 4 mm) and main-column (250× 4 mm), a SpektroMonitor 3200 (LCD-Analytical) UV detector and an HP4496 Series II (Hewlett Packard) integrator. Eluent:acetonitrile/phosphoric acid 0.1% (50:50); samples: 3 mgsubstance in 10 mL 50 mM phosphate buffer pH 7.4/acetonitrile (1:5; flow: 1 mL/min; column temperature:20–25 °C; detection wavelength: 190 mm. Samples wereheated at 37 °C for 24 h in a common dry oven.

References

[1] Lehmann J., Kahlich R., Meyer Zum Gottesberge C., Fricke U.,Arch. Pharm. Pharm. Med. Chem. 330 (1997) 247–252.

[2] Vallance P., Br. J. Clin. Pharmacol. 45 (1998) 433–439.

[3] Bennett B.M., McDonald B.J., James S.T., J. Pharmacol. Exp. Ther.261 (1992) 716–723.

[4] Ahlner J., Andersson R.G.G., Torfgård K., Axelsson K.L.,Pharmacol. Rev. 43 (1991) 351–423.

[5] Feelisch M., Noack E.A., Eur. J. Pharmacol. 139 (1987) 19–30.

[6] Chung S., Fung H.L., Biochem. Pharmacol. 42 (1991) 1433–1439.

[7] Hütter J., Noack E., (Schwarz Pharma AG) (1990) E P 89116700 9;ref C A 113, 212672x.

[8] Ishihara S., Saito F., Yoshioka T., Koike H., Miyake S., Mizuno H.,(1993) PCT Int. Appl. WO 9303, 163.

[9] Nallet J.P., Dreux J., Berdeaux A., Richard V., Martorana P., BohnH., (1991) PCT Int. Appl. WO 9303, 037.

[10] Gais H.J., Angew. Chem. 89 (1977) 251–252.

[11] Gordon A.J., Tetrahedron 23 (1967) 863–870.

[12] Kojda G., Kottenberg K., Hacker A., Noack E., Pharm. Acta Helv.73 (1998) 27–35.

[13] Hellwinkel D., Bohnet S., Chem. Ber. 120 (1987) 1151–1174.

[14] Baldwin D., Duckworth P.A., Erickson G.R., Lindoy L.F., McPartlinM., Aust. J. Chem. 40 (1987) 1861–1872.

[15] McCallum K.S., Emmons W.D., J. Org. Chem. 21 (1956) 367–368.

[16] Ustarshchikov B.F., Podgornova V.A., Dormidontova N.V., Far-berov M.J., Dokl. Akad. Nauk. SSSR 157 (1964) 143–144.

[17] Schindlbauer H., Monatsh. Chem. 99 (1968) 1799–1807.

901

Original article

Synthesis of oxypropanolamine derivativesof 3,4-dihydro-2H-1,4-benzoxazine,â-adrenergic affinity, inotropic,

chronotropic and coronary vasodilating activities

Kriton Iakovoua*, Michalis Kazanisa, Andreas Vavayannisa,Giancarlo Brunib, Maria Raffaella Romeob, Paola Massarellib,

Shuji Teramotoc, Hiroyuki Fujikic, Toyoki Moric

aDepartment of Pharmacy, Division of Pharmaceutical Chemistry,University of Athens, Panepistimiopolis Zografou, GR-157 71 Athens, Greece

bIstituto di Farmacologia, Universita di Siena, 53100 Siena, Italyc2nd Tokushima Institute of New Drug Research, Otsuka Pharmaceutical Co. Ltd.,

463-10 Kagasuno, Kawauchi-cho, Tokushima 771-01, Japan

(Received 18 January 1999; accepted 26 April 1999)

Abstract – A series of oxypropanolamine derivatives of 3,4-dihydro-2H-1,4-benzoxazine were synthesized and evaluated for inotropic,chronotropic and coronary vasodilating activities in the canine heart, affinity toâ1-adrenergic receptor in turkey erythrocytes and affinity tothe â2-adrenergic receptor in the rat lung. Among these compounds, 4-acetyl-6-(3-tert-butylamino-2-hydroxy)propoxy-3,4-dihydro-2H-1,4-benzoxazine showed 2.1-fold more potent affinity to theâ1 receptor than propranolol and 7-(3-tert-butylamino-2-hydroxy)propoxy-N-butyryl-3,4-dihydro-2H-1,4-benzoxazine showed 2.5-fold more potent affinity to theâ2 receptor and furthermore 4 386-fold more potent selectivityto the â2 receptor than propranolol. In addition, 4-acetyl-6-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy]propoxy-3,4-dihydro-2H-1,4-benzoxazine showed 1.1-fold more potent affinity to theâ1 receptor than propranolol and also 1 147-fold more potent selectivity to theâ1receptor. With a few exceptions, negative inotropic and chronotropic actions of these compounds were dependent on the size of the4-substituent obeying the order: unsubstituted< acetyl< propanoyl< butanoyl, while the benzoyl substituent conferred even stronger negativeactions in the 6-oxypropanolamine derivatives. Neither negative inotropic and chronotropic actions related with affinity toâ1-adrenoceptor norcoronary vasodilator action related with affinity toâ2-adrenoceptor were observed. 4-acetyl-7-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy]propoxy-3,4-dihydro-2H-1,4-benzoxazine exerted potent positive inotropic, chronotropic and coronary vasodilating actions. Theinotropic and chronotropic actions of the latter compound may be attributed to the release of intrinsic catecholamines, as concluded by theabsence ofâ1-adrenoceptor affinity and disappearance of activity in the presence of aâ-blocker. © 1999 Éditions scientifiques et médicalesElsevier SAS

1,4-benzoxazine / oxypropanolamines /â-adrenoceptor affinity / cardiovascular effects

1. Introduction

A number of 1,4-benzoxazine derivatives [1–4] havebeen synthesized so far and various pharmacologicalactivities have been reported with this class of molecules.Ethanolamine and oxypropanolamine derivatives of 1,4-benzothiazine active on the adrenergic system are alreadyknown [5, 6], as well as of the isoster nuclei 1,4-benzoxazine [7–10] and 3,4-dihydro-2(1H)-quinolinone

[11, 12]. Among the carbostyril derivatives, 5-(3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2(1H)-qui-nolinone (carteolol) [13, 14] is aâ-blocker more potentthan propranolol and is apparently devoid of the sideeffects which are usually associated withâ-blockingtherapy as it retains an intrinsic sympathomimetic acti-vity. Oxypropanolamine derivatives of 1,4-benzo-dioxin [15, 16], which are analogues of 1,4-benzoxazine,have been recently synthesized. The 4-acyl derivatives of7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazineswhich are included in the present work may be consid-*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 903−917 903© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

ered as dicyclic analogues of practolol, a prototypecardioselective blocking agent, as shown infigure1 [9,17, 18].

It is well known that the catecholamineâ-stimulants(arylethanolamines and aryloxypropanolamines) havebeen used as cardiotonic agents in the management ofheart failure. Unfortunately these compounds have somedisadvantages, such as powerful vasodilating effects,which cause a reflex rise in heart rate, lack of oralabsorption and short half-life [19]. On the other hand,some non-oxypropanolamine [20] or oxypropanolaminederivatives of 2(1H)-quinolinone [21, 22], have demon-strated remarkable positive inotropic activities, the latterin some cases without havingâ-agonistic action. Further-more, a series of non-oxypropanolamine derivatives of3,4-dihydro-1,4-benzoxazine [23] have exhibited potent,long-acting positive inotropic and peripheral vasodilatingactivities.

In order to extend the investigation on inotropic,chronotropic and vasodilating activities of non-catecholderivatives with possible affinity to theâ-adrenergicreceptors, we prepared molecules structurally similar tothe above. The synthesized compounds are novel 4-acylsubstituted 3,4-dihydro-2H-1,4-benzoxazines bearing thetypical oxypropanolamine chain at positions 6- and 7- ofthe aromatic ring.

2. Chemistry

The target compounds24–49 (tables I and II ) wereprepared in a standard three-step procedure, shown infigure 2, which involves: a) synthesis of the 6-and 7-hydroxy-4-acyl-3,4-dihydro-2H-1,4-benzoxazines[24], b) reaction thereof with epichlorhydrin in thepresence of potassium carbonate [25] and c) treatment ofthe resulting 1-aryloxy-2,3-epoxypropanes with the ap-propriate amines [26].

The derivatives 2–7 (table III) were synthesizedthrough the reaction of 6- or 7-hydroxy-3,4-dihydro-2H-

1,4-benzoxazine in aqueous medium with the appropriateanhydrides. The reaction of 6- or 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine with benzoic or chloroacetic anhy-dride was not successful in water and therefore ethylacetate was used instead. The intermediates12 and 13(table III) were synthesized via the corresponding4-chloroacetyl compounds10 and 11 (table III) whichwere condensed with diethylamine in ethanol. The gen-eral synthetic procedure for the hydroxy derivatives2–13is shown infigure 2.

The 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine wassynthesized by a method which is presented infigure 3.Initially 7-methoxy-3,4-dihydro-2H-1,4-benzoxazin-3-one [27–29] was converted into1 through reduction withlithium aluminium hydride in anhydrous tetrahydrofu-ran [30]. The intermediate1 was demethylated by meansof concentrated hydrobromic acid to give 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine hydrobromide, which wasalkalized by concentrated ammonium hydroxide to affordthe free base [31].

The synthetic route of 6-hydroxy-3,4-dihydro-2H-1,4-benzoxazine, as shown infigure 4, was achieved asfollows: 2,5-dimethoxyaniline was refluxed with 2-bromo-ethanol in the presence of calcium carbonate in water toyield 2,5-dimethoxy-N-(2-hydroxyethyl)aniline [32]which was isolated in pure form by distillation in vacuo.The latter was refluxed with concentrated hydrobromicacid leading to 6-hydroxy-3,4-dihydro-2H-1,4-benzo-xazine hydrobromide [32], which in turn, was alkalizedby concentrated ammonium hydroxide to provide the freebase.

The 4-unsubstituted products44–47 (table II) wereprepared, as shown infigure 2, by hydrolysis of thecorresponding 4-acetyl derivatives24–27with potassiumhydroxide in aqueous methanol [33].

IR absorption and NMR spectra were in conformitywith the structures expected. However, we observed thatthe resonance of the aromatic proton 5 either appears asa broad peak or disappears completely, depending on thetemperature. This is possibly due to stereochemical

Figure 1. Structure of the synthesized 4-acyl derivatives of 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazines compared topractolol.

904

effects caused by the equilibrium between two extremeconformations of the morpholine ring. A further study ofthis effect is still under investigation and will be pub-lished elsewhere.

3. Pharmacology

The biological profiles of the compounds listed intable IV on â1 and â2 adrenoceptors were respectively

Table I. 4-Acyl-6- and 7-(3-isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines and 4-acetyl-6-and 7-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy-propoxy]-3,4-dihydro-2H-1,4-benzoxazines.

Compound Position R1 R2 M.p. (°C) Yield (%) Formula

24 7 CH3 iPr 176–179a 24 C16H24N2O4. 0.5 C4H4O4

25 7 CH3 tBu 161–164a 25 C17H26N2O4. 0.5 C4H4O4

26 6 CH3 iPr 134a 20 C16H24N2O4. 0.5 C4H4O4

27 6 CH3 tBu 169–171a 25 C17H26N2O4. 0.5 C4H4O4

28 7 CH2CH3 iPr 120a 26 C17H26N2O4. 0.5 C4H4O4

29 7 CH2CH3 tBu 203a 25 C18H28N2O4. 0.5 C4H4O4

30 6 CH2CH3 iPr 129–131a 27 C17H26N2O4. 0.5 C4H4O4

31 6 CH2CH3 tBu 140–142a 23 C18H28N2O4. C4H4O4

32 7 CH2CH2CH3 iPr 175–178a 25 C18H28N2O4. 0.5 C4H4O4

33 7 CH2CH2CH3 tBu 171–172a 24 C19H30N2O4. 0.5 C4H4O4. 0.5 H2O34 6 CH2CH2CH3 iPr 117–120a 26 C18H28N2O4. 0.5 C4H4O4

35 6 CH2CH2CH3 tBu 142–145a 23 C19H30N2O4. 0.5 C4H4O4

36 7 Ph iPr 96–98 57 C21H26N2O4

37 7 Ph tBu 83–85 54 C22H28N2O4

38 6 Ph iPr 176–180a 17 C21H26N2O4. 0.5 C4H4O4

39 6 Ph tBu 112–115b 21 C22H28N2O4. C4H4O4. 0.5 H2O40 7 CH2N(C2H5)2 iPr 54–56 15 C20H33N3O4. 0.5 H2O41 7 CH2N(C2H5)2 tBu 85–90 13 C21H35N3O4. H2O42 6 CH2N(C2H5)2 iPr 68–70 16 C20H33N3O4. 0.5 H2O43 6 CH2N(C2H5)2 tBu 70–73 14 C21H35N3O4. 0.5 H2O48 7 CH3 3,4-dimethoxybenzyl 140–142b 31 C22H28N2O6. C4H4O4. 2H2O49 6 CH3 3,4-dimethoxybenzyl 138–140b 33 C22H28N2O6. C4H4O4. 2H2O

a: fumarate,b: maleate.

Table II. 6- and 7-(3-Isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines.

Compound Position R2 M.p. (°C ) Yield (%) Formula

44 7 iPr 130–134a 19 C14H22N2O3. C4H4O4. 0.5 H2O45 7 tBu 76–80b 17 C15H24N2O3. 2 C4H4O4

46 6 iPr 128–130c 20 C14H22N2O3. 2 C2H2O4

47 6 tBu 138–140a 22 C15H24N2O3. C4H4O4

a: fumarate,b: maleate,c: oxalate.

905

Figure 2. General synthetic procedure of the target 6- and 7-(alkylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines.

906

assessed on turkey erythrocytes and on rat lung. Theinotropic (changes in contractile force, CF), chronotropic(changes in sinus rate, SR) and coronary vasodilator(changes in coronary blood flow, CBF) effects of thesynthesized products were evaluated in the isolatedblood-perfused preparations of canine heart.

4. Results and discussion

4.1.â1-Adrenenoceptor binding

The affinities of the 4-acetyl substituted 6-oxypro-panolamines obeyed the order:27 > 49 > 26. The

Table III. 4-Acyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines.

Compound Position R1 M.p (°C ) Yield (%) Solvents of crystalliza-tion

2 7 CH3 152–154 72 a3 6 CH3 185–188 65 a4 7 CH2CH3 104–106 79 a5 6 CH2CH3 146–148 84 a6 7 CH2CH2CH3 123–125 80 a7 6 CH2CH2CH3 102–103 79 a8 7 Ph 190–193 65 a9 6 Ph 155 61 a10 7 CH2CI 120–122 82 a11 6 CH2CI 128–130 57 a12 7 CH2N(C2H5)2 110–113 57 b13 6 CH2N(C2H5)2 124–125 63 b

a: benzene-pentane, b: cyclohexane.

Figure 3. Preparation of 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine.

Figure 4. Preparation of 6-hydroxy-3,4-dihydro-2H-1,4-benzoxazine.

907

observed affinity order of the side-chain amino substitu-ents was tert-Bu > 3,4-dimethoxybenzyl> iPr. Thehighest affinities observed with derivatives27 and 49,were respectively 2.1- and 1.1-fold higher than that ofpropranolol. For the 6-iPr derivatives the elongation fromthe acetyl to the propanoyl chain led to a significant dropin affinity (30). Additional lengthening of the chain by amethylene group led to a loss in affinity (34). In the caseof the corresponding 6-tert-Bu-derivatives, the presenceof the propanoyl group (31) reduced the affinity, whereasthe replacement thereof with butanoyl (35) led to in-creased activity. In the group of the 4-unsubstitutedderivatives,44 and47 showed affinities ca. 10-fold lowerthan that of propranolol, while45 and46 had no affinityto the â1 adrenoceptor. The 4-acetyl substituted 7-oxypropanolamine compounds (24 and25) showed a lowdegree of affinity. For the 7-oxypropanolamine deriva-tives, the elongation from the acetyl to the propanoylchain led to little increase in affinity (28 and 29). Theobserved affinity order for the amino substituents follows

the order tert-Bu > iPr. Additional lengthening of thechain by a methylene group increased the affinity (32).The replacement of the iPr group of32 with tert-Bu ledto a drastic decrease in affinity (33). Among the4-benzoyl and 4-diethylaminoacetyl derivatives (36, 37,39, and40–43) only compound36showed an EC50 in the10–7 range.

4.2.â2-Adrenenoceptor binding

The high affinities for theâ2-adrenenoceptor wereobserved with the 4-acetyl substituted 6-oxypro-panolamine derivatives. The affinities of26 and27 were1.1-fold lower and 2.5-fold higher than that of propra-nolol, respectively. On the contrary, the 7-isomers24 and25 had low affinities to theâ2-adrenenoceptor. As to the4-acetyl substituted 6-oxypropanolamine derivatives, thechange of the amino substituent from iPr or tBu into3,4-dimethoxybenzyl dramatically decreased the affinity(49), whereas it led to a loss of affinity of the correspond-

Table IV. Inhibition of [3H]DHA binding on â1 andâ2 adrenoreceptors.

Compound â1 â2 â1/â2

Ki (± s.e.) Ki (± s.e.) selectivity ratio(M) (M)

Propranolol 1.6× 10–9 ± 0.132 2.5× 10–9 ± 0.171 1.52024a 6.0 × 10–6 ± 0.116 6.4× 10–7 ± 0.201 0.10625a 2.9 × 10–6 ± 0.327 3.6× 10–5 ± 0.319 12.29026a 7.4 × 10–8 ± 0.684 2.8× 10–9 ± 0.267 0.35027a 8.01 × 10–10 ± 0.782 1.0× 10–9 ± 0.326 1.28028a 2.7 × 10–6 ± 0.296 3.3× 10–8 ± 0.715 0.01229a 1.3 × 10–6 ± 0.131 1.6× 10–8 ± 0.220 0.01330a 1.2 × 10–6 ± 0.214 inactive31a 1.1 × 10–5 ± 0.121 1.6× 10–7 ± 0.359 0.01432a 2.0 × 10–9 ± 0.188 1.3× 10–6 ± 0.264 55533a 2.9 × 10–6 ± 0.282 1.0× 10–9 ± 0.136 0.000334a inactive 5.6× 10–6 ± 0.45435a 8.1 × 10–9 ± 0.849 5.5× 10–6 ± 0.481 68136 1.3 × 10–7 ± 0.121 8.2× 10–6 ± 2.720 6137 5.6 × 10–7 ± 0.628 2.3× 10–7 ± 0.234 4.1339b inactive 4.5× 10–6 ± 0.45940 inactive 1.2× 10–7 ± 0.16941 6.1 × 10–7 ± 0.606 inactive42 7.1 × 10–5 ± 0.684 inactive43 3.1 × 10–6 ± 0.306 2.5× 10–5 ± 0.651 7.9544a 1.9 × 10–8 ± 0.186 5.3× 10–6 ± 0.386 28045b inactive 1.2× 10–6 ± 0.17246c inactive 3.4× 10–7 ± 0.23847a 2.7 × 10–8 ± 0.296 4.5× 10–8 ± 0.185 1.7048b inactive inactive49b 1.4 × 10–9 ± 0.106 2.5× 10–6 ± 0.306 1 744

a: fumarate,b: maleate,c: oxalate.

908

ing 7-isomer (48). For the 6-derivatives the elongationfrom the acetyl to the propanoyl or butanoyl chain led toa substantial reduction, up to deletion, of affinity (30, 31,34, and35). On the contrary, the lengthening of the acylchain of the 7-isomers, led to considerable increase inaffinity, except for compound32. The 4-unsubstitutedderivatives 44–47 showed low affinity to the â2-adrenoceptor. The 4-benzoyl (36, 37, and 39) and4-diethylaminoacetyl derivatives (40–43) showed practi-cally no affinity.

4.3.â1/â2 selectivity

Compound32 showed considerableâ1-adrenergic af-finity with a high â1/â2 selectivity ratio (365 times morethan that of propranolol). Compound35 was 5-fold lessâ1-affinitive than propranolol, but 448-fold moreâ1-selective. Compound44 was 11-fold lessâ1-affinitivethan propranolol, but 185-fold moreâ1-selective. Com-pound 49 was 1.1-fold moreâ1-affinitive than propra-nolol and also 1 147-fold moreâ1-selective. Compounds28 and29 exhibited a moderateâ2-affinity, but consider-able â2-selectivity (about 120-fold more than that ofpropranolol). It is noteworthy that compound33 was2.5-fold more â2-affinitive and furthermore 4 386-foldmoreâ2-selective than propranolol.

The results discussed in the preceding paragraphs 4.1to 4.3 and presented intable IV, have shown that it isimpossible to draw any straightforward conclusions re-garding structure-activity relationships. For example, al-though compound24has the closest structure to practolol(figure 1), it did not show the expected activity, which infact was much lower than its 6-substituted isomer26. It ispossible that other parameters, such as the difference insteric configuration or the absence of amidic hydrogen,affect the biological activity of these cyclic analogues ofpractolol. Another example lies in the group of N-acyl7-substituted compounds, where each additional methyl-ene in the acyl chain may have a drastic effect on receptoraffinity and selectivity. Although this fact is obvious inthe cases of compounds25 vs.29 and28 vs.32 it cannotserve as a general rule.

4.4. Inotropic, chronotropic and coronary vasodilatingactivities

As shown infigures 5–7, negative inotropic and chro-notropic actions of the tested compounds were dependenton the size of the 4-substituent. In the series of the6-oxypropanolamine derivatives the actions followed theorder: unsubstituted (46 and47) < acetyl (26 and27) <propanoyl (30 and 31) < butanoyl (34 and 35). Thebenzoyl compounds (38 and 39) exerted stronger nega-

tive actions. Similar structure-activity relationships wereobserved in the series of 7-oxypropanolamine deriva-tives, i.e. unsubstituted (44 and45) < acetyl (24 and25)< propanoyl (28 and29) < butanoyl (32 and33), but nofurther negative action was observed in the benzoylsubsutituted compounds36 and37. The negative inotro-pic and chronotropic actions of these compounds weredifficult to explain by theâ1-adrenoceptor antagonisticactions because of inconsistency with theâ1-adrenoceptor affinity as described earlier.

In general, coronary vasodilator actions of the6-oxypropanolamine derivatives were more potent thanthose of their 7-substituted counterparts (figure 7). It wasalso difficult to explain the coronary vasodilator action ofthese compounds in relation to theirâ2-adrenoceptoragonistic actions.

The 4-diethylaminoacetyl substituted 6-oxypro-panolamine derivatives with iPr (42) and tert-Bu (43)functions, as well as the 4-acetyl substituted compoundbearing a 3,4-dimethoxybenzyl group (49), exerted weakpositive inotropic and chronotropic actions. The samewas true for the isomer of42 bearing the side chain atposition 7 (40). The 4-unsubstituted compounds44 and45 showed moderate positive actions (figure 8).

The 3,4-dimethoxybenzyl derivative48 exerted potentpositive inotropic, chronotropic and coronary vasodila-tory actions (figure 8). It should be noted that the inotro-

Figure 5. Structure-inotropic activity relationships of 6- and7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine deriva-tives in canine isolated blood-perfused heart preparations.Effects of test compounds on right ventricular papillary musclecontractile force (CF) were expressed by the % changes frombasal CF. Test compounds were dissolved in DMSO and theaction of the solvent itself was subtracted for compensation.

909

pic and chronotropic actions of this compound wereinhibited by pretreatment with theâ-blocker carteolol(figure 9).

5. Experimental protocols

5.1. ChemistryMelting points were determined on a Büchi micro

melting point apparatus without correction.1H-NMR and13C-NMR spectra were recorded on a 200 MHz BrukerAC 200 spectrometer in CDCl3 or DMSO-d6 usingtetramethylsilane as internal standard. All target com-pounds were analysed for C, H and some additionallyanalysed for N. Elemental analyses indicated by thesymbols of the elements or functions were within± 0.4%of the theoretical values.

5.1.1. 7-Methoxy-3,4-dihydro-2H-1,4-benzoxazine1A solution of 7-methoxy-3,4-dihydro-2H-1,4-

benzoxazin-3-one (0.028 mol) in dry tetrahydrofuran(95 mL) was added dropwise to a suspension of lithiumaluminum hydride (0.066 mol) in dry tetrahydrofuran

Figure 6. Structure-chrononotropic activity relationships of 6-and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine de-rivatives in canine isolated blood-perfused heart preparations.Effects of test compounds on right atrial sinus rate (SR) wereexpressed by changes from basal SR. Test compounds weredissolved in DMSO and the action of the solvent itself wassubtracted for compensation.

Figure 7. Structure-coronary vasodilator activity relationshipsof 6- and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzo-xazine derivatives in canine isolated blood-perfused heartpreparations. Effects of test compounds on coronary blood flow(CBF) through anterior septal arteries were expressed by themL/min changes from basal CBF. Test compounds were dis-solved in DMSO and the action of the solvent itself wassubtracted for compensation.

Figure 8. Effects of 6- and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine derivatives on inotropic, chronotropic, andcoronary vasodilator actions in canine isolated blood-perfusedheart preparations. Effects of test compounds on the rightventricular papillary muscle contractile force (CF) were ex-pressed by the % changes from basal CF, on the right atrialsinus rate (SR) were expressed by changes from basal SR andcoronary blood flow (CBF) through anterior septal arteries wereexpressed by the mL/min changes from basal CBF. Testcompounds were dissolved in DMSO and the action of thesolvent itself was subtracted for compensation.

910

(75 mL). The mixture was stirred and refluxed for 3 h andcooled. Aq. sodium hydroxide 5% (40 mL) was addeddropwise under cooling and continuous stirring. Themixture was stirred at room temperature for 1 h and theliquid part separated through a filter. The solution wasdried over anhydrous sodium sulfate and evaporated invacuo to give an oil. Yield 85%. 1H-NMR (CDCl3, 200MHz) δ 3.00 (brs, 1H, NH), 3.35 (brs, 2H, CH2NH), 3.7(s, 3H, CH3O), 4.22 (t, 2H, OCH2, J = 4.3Hz), 6.38–6.58(m, 3H, arom).

5.1.2. 7-Hydroxy-3,4-dihydro-2H-1,4-benzoxazineA solution of 1 (0.04 mol) in 25 mL concentrated

hydrobromic acid 62% (0.528 mol) was stirred and re-fluxed for 2 h. The reaction mixture was then basifiedwith concentrated ammonium hydroxide 28% (32 mL)and evaporated in vacuo. Ethyl acetate (250 mL) wasadded to the residue and the resulting mixture was stirredfor 1 h and filtered. The filtrate was dried over anhydroussodium sulfate and evaporated in vacuo to give 0.031 molof 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine. Yield:85%, m.p.: 95–97 °C (benzene-hexane).1H-NMR(DMSO-d6, 200 MHz) δ 2.07 (s, 1H, NH), 3.17 (t, 2H,CH2NH, J = 4.3 Hz), 4.06 (t, 2H, OCH2, J = 4.3 Hz),6.11–6.16 (m, 2H, arom), 6.41 (d, 1H, arom,J = 8.8 Hz),8.62 (s, 1H, OH).

5.1.3. 2,5-Dimethoxy-N-(2≠-hydroxyethyl)anilineA mixture of 2,5-dimethoxy-aniline (0.19 mol), cal-

cium carbonate (0.0135 mol), water (150 mL) and2-bromoethanol (0.24 mol) was refluxed for 4 h. Thereaction mixture was cooled, then extracted with ethylacetate (300 mL) and evaporated in vacuo. The oilyresidue was fractionated in vacuo (160 °C, 2 mm Hg).Yield: 41%, m.p.: 43–47 °C.1H-NMR (CDCl3, 200MHz) δ 2.15 (s, 1H, NH), 3.25 (t, 2H, NHCH2, J = 5.5Hz), 3.71–3.84 (m, 9H, 2CH3O, CH2OH and OH), 6.15(dd, 1H, arom,J = 8.5 Hz, 2.9 Hz), 6.24 (d, 1H, arom,J= 2.8 Hz), 6.65 (d, 1H, arom,J = 8.6 Hz).

5.1.4. 6-Hydroxy-3,4-dihydro-2H-1,4-benzoxazineA solution of 2,5-dimethoxy-N-(2≠-hydroxyethyl)-

aniline (0.02 mol) in 20 mL of concentrated hydrobromicacid 62% (0.44 mol) was stirred and refluxed for 2 h. Thereaction mixture was basified with concentrated ammo-nium hydroxide 28% (22 mL) and evaporated in vacuo.Ethyl acetate (120 mL) was added to the residue and theresultant mixture was stirred for 1 h and filtered. Thefiltrate was dried over anhydrous sodium sulfate andevaporated in vacuo to give 0.015 mol of 6-hydroxy-3,4-dihydro-2H-1,4-benzoxazine. Yield: 76%, m.p.:108–109 °C (benzene).1H-NMR (DMSO-d6, 200 MHz)δ 3.19 (t, 2H, CH2NH, J = 3.9 Hz), 3.97 (t, 2H, OCH2, J

Figure 9. Effects ofâ-blocker carteolol on the positive inotropic and chronotropic response induced by compound48 (4-acetyl-7-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy]propoxy-3,4-dihydro-2H-1,4-benzoxa-zine) in canine isolated blood-perfused prepara-tions. The atrial and papillary muscle preparations were preloaded with 2 g and 1 g respectively.

911

= 4.2 Hz), 5.62 (brs, 1H, NH), 5.83 (dd, 1H, arom,J = 8.4Hz, 2.8 Hz), 5.97 (d, 1H, arom,J = 2.2 Hz), 6.38 (d, 1H,arom,J = 8.4 Hz), 8.52 (s, 1H, OH).

5.1.5. 4-Acyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines2–7

To a suspension of 6- or 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine (0.02 mol) in water (12 mL), acetic orpropanoic or butanoic anhydride (0.025 mol) was added.The reaction mixture was heated in a water bath for15 min. After cooling, ethyl acetate (130 mL) was addedand the resulting mixture was stirred, filtered and ex-tracted with ammonium hydroxide 14% (24 mL). Thealkaline solution was discarded and the organic layer waswashed with water (100 mL). The organic layer wasseparated, shaken with hydrochloric acid 5% (35 mL),water (35 mL), dried over anhydrous sodium sulfate andevaporated to dryness under reduced pressure.1H-NMR(DMSO-d6, 200 MHz)2: δ 2.15 (s, 3H, NCOCH3), 3.75(t, 2H, CH2N, J = 4.4 Hz), 4.16 (brs, 2H, OCH2),6.24–6.30 (m, 2H, arom), 7.03 (brs, 1H, arom), 7.34 (s,1H, OH). Anal. (C10H11NO3) C, H, N. 3: δ 2.21 (s, 3H,NCOCH3), 3.77 (t, 2H, CH2N, J = 4.4 Hz), 4.14 (t, 2H,OCH2, J = 4.5 Hz), 6.43 (dd, 1H, arom,J = 8.6 Hz, 2.2Hz), 6.66 (d, 1H, arom,J = 8.7 Hz), 8.97 (brs, 1H, OH).Anal. (C10H11NO3) C, H, N. 1H-NMR (CDCl3, 200MHz) 4: δ 1.13 (t, 3H, CH3, J = 6.2 Hz), 2.56 (q, 2H,NCOCH2, J = 6.8 Hz), 3.89 (s, 2H, CH2N), 4.24 (t, 2H,OCH2, J = 4.8 Hz), 6.35–6.40 (m, 2H, arom), 6.93 (brs,1H, arom), 7.34 (s, 1H, OH). Anal. (C11H13NO3) C, H, N.1H-NMR (DMSO-d6, 200 MHz)5: δ 1.03 (t, 3H, CH3, J= 6.2 Hz), 2.50–2.57 (q, 2H, NCOCH2, J = 7.0 Hz), 3.79(t, 2H, CH2N, J = 4.1 Hz), 4.13 (t, 2H, OCH2, J = 4.0 Hz),6.46 (dd, 1H, arom,J = 8.6 Hz, 2.1 Hz), 6.67 (d, 1H,arom,J = 8.6 Hz), 7.32 (brs, 1H, arom), 8.95 (brs, 1H,OH). Anal. (C11H13NO3) C, H, N.1H-NMR (CDCl3, 200MHz) 6: δ 0.91 (t, 3H, CH3, J = 7.2 Hz), 1.6–1.8 (m, 2H,COCH2CH2), 2.51 (t, 2H, COCH2CH2, J = 7.6 Hz), 3.9(brs, 2H, CH2N), 4.24 (t, 2H, OCH2, J = 4.7 Hz), 5.3 (brs,1H, OH), 6.35–6.45 (m, 2H, arom), 6.9 (brs, 1H, arom).Anal. (C12H15NO3) C, H, N. 7: δ 0.94 (t, 3H, CH3, J =7.3 Hz), 1.60–1.76 (m, 2H, COCH2CH2), 2.54 (t, 2H,COCH2CH2, J = 7.4 Hz), 3.68 (brs, 1H, OH), 3.87 (brs,2H, CH2N), 4.21 (t, 2H, OCH2, J = 4.7 Hz), 6.57 (dd, 1H,arom,J = 8.9 Hz, 2.7 Hz), 6.75 (d, 1H, arom,J = 8.9 Hz),7.34 (s, 1H, arom). Anal. (C12H15NO3) C, H, N.

5.1.6. 4-Benzoyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines8 and9

A mixture of 6- or 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine (0.02 mol) and benzoic anhydride

(0.02 mol) in ethyl acetate (50 mL) was refluxed for 1 h.The mixture was then cooled, filtered, extracted withammonium hydroxide 14% (60 mL). The alkaline solu-tion was discarded and the organic phase was washedwith water, shaken with hydrochloric acid 10% (25 mL)and finally dried over anhydrous sodium sulfate andconcentrated in vacuo.1H-NMR (DMSO-d6, 200 MHz)8: δ 3.80 (brs, 2H, CH2N), 4.24 (brs, 2H, OCH2),6.10–6.30 (m, 2H, arom), 7.37–7.52 (m, 6H, arom), 9.37(s, 1H, OH), IR (Nujol): νOH 3 389 cm–1, νC=O

1 710 cm–1. Anal. (C15H13NO3) C, H, N. 1H-NMR(DMSO-d6, 200 MHz) 9: δ 3.77 (t, 2H, CH2N, J = 4.4Hz), 4.18 (t, 2H, OCH2, J = 4.3 Hz), 6.43 (dd, 1H, arom,J = 8.4 Hz, 2.0 Hz), 6.68 (d, 1H, arom,J = 8.4 Hz), 6.83(brs, 1H, arom), 7.40–7.55 (m, 5H, arom), 8.9 (s, 1H,OH), IR (Nujol): νOH: 3 406 cm–1, νC=O: 1 720 cm–1.Anal. (C15H13NO3) C, H, N.

5.1.7. 4-Chloroacetyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines10 and11

Compounds10 and11 were synthesized through reac-tion of 6- or 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazinewith chloroacetic anhydride in ethyl acetate according tothe previous method.1H-NMR (DMSO-d6, 200 MHz)10: δ 3.79 (t, 2H, CH2N, J = 4.5 Hz), 4.22 (t, OCH2, J =4.1 Hz), 4.55 (s, 2H, CH2CI), 6.25–6.33 (m, 2H, arom),7.75 (brs, 1H, arom), 9.45 (brs, 1H, OH). IR (Nujol):νOH: 3 327 cm–1, νC=O: 1 688 cm–1. Anal.(C10H10ClNO3) C, H, N. 1H-NMR (DMSO-d6, 200MHz) 11: δ 3.82 (t, 2H, CH2N, J = 4.4 Hz), 4.19 (t, 2H,OCH2), 4.60 (s, 2H, CH2CI), 6.47 (dd, 1H, arom), 6.68(d, 1H, arom), 7.35 (s, 1H, arom), 9.03 (s, 1H, OH). IR(Nujol): νOH: 3 315 cm–1, νC=O: 1 679 cm–1

. Anal.(C10H10ClNO3) C, H, N.

5.1.8. 4-Diethylaminoacetyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines12 and13

A mixture of 10 or 11 (0.012 mol) and diethylamine(0.034 mol) in absolute ethanol (100 mL) was refluxedfor 3 h. The mixture was then cooled, filtered andevaporated in vacuo. After addition of ethyl acetate(70 mL) the mixture was stirred, filtered and the filtratewas concentrated in vacuo. Water (20 mL) was added tothe obtained viscous oil and the resulting precipitate wascollected by filtration.1H-NMR (DMSO-d6, 200 MHz)12: δ 0.94 (t, 6H, N(CH2CH3)2, J = 6.7 Hz), 2.48–2.58(m, 4H, N(CH2CH3)2), 3.44 (s, 2H, COCH2N), 3.84 (brs,2H, CH2N), 4.19 (brs, 2H, OCH2), 6.22–6.30 (m, 2H,arom), 7.75 (brs, 1H, arom), 9.34 (brs, 1H, OH). Anal.(C14H20N2O3) C, H, N. 13: δ 0.95 (t, 6H, N(CH2CH3)2,J = 7.0 Hz), 2.48–2.58 (m, 4H, N(CH2CH3)2), 3.41 (s,2H, COCH2N), 3.86 (t, 2H, CH2N, J = 4.2 Hz), 4.15 (t,

912

2H, OCH2, J = 4.2 Hz), 6.42 (dd, 1H, arom,J = 8.7 Hz,2.4 Hz), 6.65 (d, 1H, aromJ = 8.7 Hz), 7.46 (brs, 1H,arom), 8.92 (s, 1H, OH). Anal. (C14H20N2O3) C, H, N.

5.1.9. 4-Acyl-6- and 7-(2,3-epoxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines14–23

A mixture of 2–9, 12 or 13 (0.013 mol), epichlorohy-drin (0.11 mol) and potassium carbonate (0.026 mol) wasstirred at 90–95 °C for 3 h. After cooling, ethyl acetate(50 mL) or chloroform was added and the mixture wasstirred, filtered and extracted with sodium hydroxide 5%(40 mL). The organic phase was dried over anhydroussodium sulfate and evaporated in vacuo to give a viscousoil. All epoxides were used in the next step withoutpurification, except for compounds15and20which werepurified as follows: the crude oily residue was worked upwith hot cyclohexane and filtered. After cooling thefiltrate, compound15 separated as an oily layer whichwas isolated after decantation and evaporation of theremaining solvent (yield 59%), whereas compound20was isolated as a white crystalline solid (m.p. 75–77 °C,yield 65%).1H-NMR (CDCl3, 200 MHz) 15: δ 2.32 (s,3H, COCH3), 2.72 (dd, 1H, CH2-oxiranic,J = 4.7 Hz, 2.6Hz), 2.88 (t, 1H, CH2-oxiranic, J = 4.5 Hz), 3.29–3.33(m, 1H, CH-oxiranic), 3.80–3.89, 4.15–4.25 (2m, 6H,CH2O, CH2CH2N, OCH2CH2), 6.66 (dd, 1H, arom,J =8.9 Hz, 2.5 Hz), 6.8 (d, 1H, arom,J = 9.0 Hz), 6.95 (brs,1H, arom). Anal. (C13H15NO4) C, H, N. 20: δ 2.7 (dd,1H, CH2-oxiranic, J = 4.8 Hz, 2.6 Hz), 2.87 (t, 1H,CH2-oxiranic, J = 4.5 Hz), 3.26–3.35 (m, 1H, CH-oxiranic), 3.84 (dd, 1H, CH2O, J = 10.9 Hz, 5.6 Hz),3.95–3.97 (m, 2H, CH2CH2N), 4.13 (dd, 1H, CH2O, J =11.0 Hz, 3.0 Hz), 4.3–4.32 (m, 1H, OCH2CH2), 6.26 (d,1H, arom,J = 6.9 Hz), 6.44 (d, 1H, arom,J = 2.8 Hz),6.91 (brs, 1H, arom), 7.34–7.48 (m, 5H, arom). Anal.(C18H17NO4) C, H, N.

5.1.10. 4-Acyl-6- and 7-(3-isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-ben-zoxazines24–43

A mixture of14–23(0.008 mol) and isopropylamine ortert-butylamine (0.08 mol) in isopropyl alcohol (90 mL)was refluxed for 3 h. The resulting solution was thor-oughly evaporated in vacuo, the viscous oily residue wasworked up with hot cyclohexane, filtered immediatelyand the solvent was evaporated. Compounds40–43 wereinsoluble in hot cyclohexane and were converted directlyinto salts. With the exception of36 and 37 which wereisolated as crystalline solids, the oily amines were con-verted into the fumarates (dry acetone-diethyl ether),except for39which was converted into the maleate (ethylacetate). Because the salts of the amines40–43could not

be crystallized, the bases were purified as follows: theoily salts were diluted in water, made alkaline withsodium hydroxide, extracted with chloroform and theorganic layer was dried over anhydrous sodium sulfateand evaporated in vacuo.1H-NMR (CDCl3, 200 MHz)24: δ 1.06 (d, 6H, CH(CH3)2, J = 6.3 Hz), 2.25 (s, 3H,COCH3), 2.39 (brs, 2H, OH and NH), 2.62–2.89 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.88–3.99 (m, 5H,OCH2CH(OH)CH2, CH2CH2N), 4.24 (t, 3H, OCH2CH2,J = 4.7 Hz), 6.43–6.49 (m, 2H, arom), 6.93 (brs, 1H,arom).1H-NMR (DMSO-d6, 200 MHz)24 (fumarate):δ1.15 (dd, 6H, CH(CH3)2, J = 6.2 Hz, 1.9 Hz), 2.18 (s, 3H,COCH3), 2.75–3.19 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.35–4.50 (m, OCH2CH(OH)CH2,CH2CH2N, OCH2CH2, NH, OH, HOOCCH=CHCOOH),6.37 (s, 1H, HOOCCH=CHCOOH), 6.42–6.55 (m, 2H,arom), 7.20, 7.90 (2brs, 1H, arom). Anal. C18H26N2O6

(C, H, N). 1H-NMR (CDCl3, 200 MHz). 25: δ 1.08 (s,9H, C(CH3)3), 2.25–2.83 (m, 7H, COCH3, OH, NH andOCH2CH(OH)CH2), 3.90 (s, 5H, OCH2CH(OH)CH2 andCH2CH2N), 4.24 (t, 3H, CH(CH3)2 and OCH2CH(OH)CH2) 3.88–3.99 (m, 5H, OCH2CH(OH)CH2,CH2CH2N), 4.24 (t, 2H, OCH2CH2, J = 4.5 Hz),6.44–6.49 (m, 2H, arom), 6.97 (brs, 1H, arom).1H-NMR(DMSO-d6, 200 MHz) 25 (fumarate): δ 1.2 (s, 9H,C(CH3)3), 2.2 (s, 3H, COCH3), 2.7–3.1 (m, 2H,OCH2CH(OH)CH2), 3.20–4.40 (m, OCH2CH(OH)CH2,CH2CH2N, OCH2CH2, NH, OH, HOOCCH=CHCOOH),6.37 (s, 1H, HOOCCH=CHCOOH), 6.42–6.55 (m, 2H,arom), 7.20, 7.90 (2brs, 1H, arom).13C-NMR (DMSO-d6) δ 22.26, 25.47, 30.27, 43.81, 52.84, 53.63, 59.3,65.62, 69.87, 101.76, 106.21, 121.01, 124.27, 135.29,145, 167.94, 169. Anal. C19H28N2O6 (C, H, N).1H-NMR(CDCl3, 200 MHz)26: δ 1.08 (d, 6H, CH(CH3)2, J = 6.3Hz), 2.29 (s, 3H, COCH3), 2.64–2.91 (m, 3H, CH(CH3)2and OCH2CH(OH)CH2), 3.14 (brs, 2H, NH and OH),3.86–4.23 (m, 7H, OCH2CH(OH)CH2, CH2CH2N andOCH2CH2), 6.6–6.8 (m, 2H, arom).1H-NMR (DMSO-d6, 200 MHz)26 (fumarate):δ 1.1 (d, 6H, CH(CH3)2, J= 6.3 Hz), 2.24 (s, 3H, COCH3), 2.61–3.09 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.20–4.20 (m,OCH2CH (OH)CH2, CH2CH2N, OCH2CH2, NH, OH,HOOCCH=CHCOOH), 6.37 (s, 1H, HOOCCH=CHCOOH), 6.64 (dd, 1H, arom,J = 6.9 Hz), 6.79 (d, 1H,arom, J = 7.1 Hz). Anal. C20H28N2O8 (C, H, N).1H-NMR (CDCl3, 200 MHz)27: δ 1.09 (s, 9H, C(CH3)3),2.31 (s, 3H, COCH3), 2.57–2.84 (m, 2H,OCH2CH(OH)CH2), 3.89 (brs, 5H, OCH2CH(OH)CH2

and CH2CH2N), 4.22 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.65(dd, 1H, arom,J = 8.9 Hz, 2.6 Hz), 6.79 (d, 1H, arom,J= 9.0 Hz). Anal. C19H28N2O6 (C, H).28: δ 1.09–1.26 (m,9H, CH(CH3)2 and COCH2CH3), 2.50–2.92 (m, 5H,

913

COCH2CH3, CH(CH3)2 and OCH2CH(OH)CH2),3.65–4.21 (m, 9H, OH, NH, OCH2CH(OH)CH2,CH2CH2N and OCH2CH2), 6.35–6.44 (m, 2H, arom),6.95 (brs, 1H, arom). Anal. C19H28N2O6 (C, H, N).29: δ1–1.25 (m, 12H, C(CH3)3 and COCH2CH3), 2.47–2.87(m, 6H, COCH2CH3, OCH2CH(OH)CH2, OH and NH),3.65–3.98 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N),4.17–4.27 (m, 2H, OCH2CH2), 6.4–6.5 (m, 2H, arom),6.95 (brs, 1H, arom). Anal. C20H28N2O6 (C, H, N).30: δ1.07 (d, 6H, CH(CH3)2, J = 6.3 Hz), 1.17 (t, 3H,COCH2CH3, J = 6.2), 2.52–2.88 (m, 7H, COCH2CH3,CH(CH3)2, OCH2CH(OH)CH2, OH and NH), 3.84–4.05(m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H,OCH2CH2, J = 4.0 Hz), 6.64 (dd, 1H, arom,J = 7.5 Hz,2.0 Hz), 6.77 (d, 1H, arom,J = 8.7 Hz). Anal.C19H28N2O6 (C, H). 31: δ 1.06 (s, 9H, C(CH3)3), 1.15 (t,3H, COCH2CH3, J = 6.2), 2.4–2.85 (m, 6H, COCH2CH3,OCH2CH(OH)CH2, OH and NH), 3.86 (brs, 5H,OCH2CH(OH)CH2 and CH2CH2N), 4.19 (t, 2H,OCH2CH2, J = 4.0 Hz), 6.64 (dd, 1H, arom,J = 7.5 Hz,2 Hz), 6.77 (d, 1H, arom,J = 8.7 Hz). Anal. C22H30N2O8

(C, H). 32: δ 0.91 (t, 3H, COCH2CH2CH3, J = 7.2 Hz),1.07 (d, 6H, CH(CH3)2, J = 6.2 Hz), 1.6–1.8 (m, 2H,COCH2CH2CH3), 1.83 (brs, 2H, OH and NH), 2.51 (t,2H, COCH2CH2CH3, J = 7.6 Hz), 2.63–2.9 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.89–4.07 (m, 5H,OCH2CH(OH)CH2 and CH2CH2N), 4.24 (t, 2H,OCH2CH2, J = 4.7 Hz), 6.45–6.5 (m, 2H, arom), 6.97(brs, 1H, arom). Anal. C20H30N2O6 (C, H). 33: δ0.87–1.13 (m, 12H, COCH2CH2CH3 and C(CH3)3),1.6–1.8 (m, 2H, COCH2CH2CH3), 1.86 (brs, 2H, OH andNH), 2.45–2.89 (m, 4H, COCH2CH2CH3 andOCH2CH(OH)CH2), 3.9 (brs, 5H, OCH2CH(OH)CH2

and CH2CH2N), 4.25 (brs, 2H, OCH2CH2), 6.4–6.5 (m,2H, arom), 6.97 (brs, 1H, arom). Anal. C21H32N2O6 (C,H). 34: δ 0.95 (t, 3H, COCH2CH2CH3, J = 7.2 Hz), 1.08(d, 6H, CH(CH3)2, J = 6.2 Hz), 1.6–2 (m, 4H,COCH2CH2CH3, OH and NH), 2.5–2.9 (m, 5H,COCH2CH2CH3, CH(CH3)2, and OCH2CH(OH)CH2),3.85–4.04 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N),4.22 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.64 (dd, 1H, arom,J = 7.5 Hz, 2.0 Hz), 6.8 (d, 1H, arom,J = 8.7 Hz). Anal.C20H30N2O6 (C, H). 35: δ 0.95 (t, 3H, COCH2CH2CH3,J = 7.2 Hz), 1.09 (s, 9H, C(CH3)3), 1.6–1.8 (m, 2H,COCH2CH2CH3), 2 (brs, 2H, OH and NH), 2.5–2.88 (m,4H, COCH2CH2CH3 and OCH2CH(OH)CH2), 3.87 (s,5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H,OCH2CH2, J = 4.7 Hz), 6.67 (dd, 1H, arom,J = 7.5 Hz,2.0 Hz), 6.78 (d, 1H, arom,J = 8.7 Hz). Anal.C21H32N2O6 (C, H).36: δ 1.05 (d, 6H, CH(CH3)2, J = 6.2Hz), 2.1 (brs, 2H, NH and OH), 2.6–2.86 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.85–3.99 (m, 5H,

OCH2CH(OH)CH2 and CH2CH2N), 4.32 (t, 2H,OCH2CH2, J = 4.4 Hz), 6.25 (d, 1H, arom,J = 7.5 Hz),6.44 (d, 1H, arom,J = 2.7 Hz), 7.29–7.48 (m, 5H, arom).Anal. C21H26N2O4 (C, H, N). 37: δ 1.20 (s, 9H,C(CH3)3), 2.60–2.95 (m, 4H, OCH2CH(OH)CH2, OHand NH), 3.87–4.00 (m, 5H, OCH2CH(OH)CH2 andCH2CH2N), 4.30 (d, 2H, OCH2CH2, J = 3.7 Hz), 6.25 (d,1H, arom,J = 6.9 Hz), 6.44 (d, 1H, arom,J = 2.8 Hz),7.34–7.48 (m, 5H, arom). Anal. C22H28N2O4 (C, H, N).38: δ 1.10 (d, 6H, CH(CH3)2, J = 6.2 Hz), 2.55–3.00 (m,3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.55–4.47 (m,7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2),6.55–6.84 (m, 1H, arom), 7.30–7.57 (m, 6H, arom), 8.03(d, 1H, arom,J = 8.2 Hz). Anal. C23H28N2O6 (C, H).1H-NMR (DMSO-d6, 200 MHz)39 (maleate):δ 1.27 (d,9H, C(CH3)3), J = 4.5 Hz), 2.71–3.00 (m, 2H, OCH2CH(OH)CH2), 3.65–4.29 (m, 7H, OCH2CH(OH)CH2,CH2CH2N and OCH2CH2), 5.83–6.00 (m, 2H, CH=CH-maleic), 6.64–6.90 (m, 1H, arom), 7.49–7.68 (m, 6H,arom), 8.02 (d, 1H, arom,J = 8.2 Hz), 8.28 (brs, 1H,COOH). Anal. C26H32N2O8 (C, H). 1H-NMR (CDCl3,200 MHz) 40: δ 0.9–1.25 (m, 12H, CH(CH3)2, andN(CH2CH3)2), 1.70–2.10 (m, 6H, N(CH2CH3)2, OH andNH), 2.50–2.85 (m, 3H,CH(CH3)2 and OCH2CH(OH)CH2), 3.15 (d, 2H, NCOCH2N, J = 5.7 Hz), 3.40(brs, H2O), 3.71–3.91 (m, 5H, OCH2CH(OH)CH2 andCH2CH2N), 4.19 (brs, 2H, OCH2CH2), 6.40 (d, 2H,arom,J = 2.5 Hz), 6.63 (d, 1H, arom,J = 9.5 Hz). Anal.C20H33N3O4 (C, H). 41: δ 0.85–1.30 (m, 15H, C(CH3)3and N(CH2CH3)2), 1.70–2.10 (m, 6H, N(CH2CH3)2, OHand NH), 2.50–2.85 (m, 2H, OCH2CH(OH)CH2), 3.13(brs, 2H, NCOCH2N), 3.40 (brs, H2O), 3.7–4.2 (m, 7H,OCH2CH(OH)CH2, CH2CH2N and OCH2CH2),6.34–6.69 (m, 3H, arom). Anal. C21H35N3O4 (C, H). 42:δ 0.9–1.27 (m, 12H, CH(CH3)2, and N(CH2CH3)2),1.7–2.1 (m, 6H, N(CH2CH3)2), OH and NH), 2.5–2.85(m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.20 (d, 2H,NCOCH2N) 3.40 (brs, H2O), 3.8–4.2 (m, 7H,OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.13 (dd,1H, arom,J = 8.0 Hz, 2.8 Hz), 6.3 (d, 1H, arom,J = 2.8Hz), 6.65 (d, 1H, arom,J = 8.5 Hz).13C-NMR (CDCl3)δ 22.58, 29.66, 49.08, 50.04, 55.88, 64.19, 65.83, 67.71,68.27, 70.98, 100.01, 102.08, 116.25, 135.96, 138.44,153.63, 163.58. Anal. C20H33N3O4 (C, H). 1H-NMR(CDCl3, 200 MHz) 43: δ 0.85–1.27 (m, 15H, C(CH3)3and N(CH2CH3)2), 1.65–2.06 (m, 6H, N(CH2CH3)2), OHand NH), 2.45–2.8 (m, 2H, OCH2CH(OH)CH2), 3.20(brs, 2H, NCOCH2N) 3.40 (brs, H2O), 3.7–4.2 (m, 7H,OCH2CH(OH)CH2, CH2CH2N and OCH2CH2),), 6.13(dd, 1H, arom,J = 8.0 Hz, 2.8 Hz), 6.3 (d, 1H, arom,J =2.8 Hz), 6.65 (d, 1H, arom,J = 8.5 Hz). Anal.C21H35N3O4 (C, H).

914

5.1.11. 6- and 7-(3-Isopropylamino- and 3-tert-butyl-amino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxa-zines44–47

A mixture of 24–27(0.0031 mol), potassium hydrox-ide (0.0372 mol), water (2.5 mL) and methanol (6 mL)was stirred at 65–70 °C for 3 h. The methanol wasremoved in vacuo and the residue was extracted succes-sively with ethyl acetate (50 mL) and water (20 mL). Theaqueous layer was discarded, the organic phase was driedover anhydrous sodium sulfate and evaporated in vacuoto give a viscous dark-coloured oil. The amines weretaken up from the crude product with warm cyclohexaneand the light-coloured oils which remained after evapo-ration, were converted into fumarates or maleates. Com-pound 46 was converted into the oxalate from ethylacetate.1H-NMR (CDCl3, 200 MHz)44: δ 1.05 (d, 6H,CH(CH3)2, J = 6.3 Hz), 2.34 (brs, 3H, OH and 2NH),2.60–2.86 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2),3.34 (t, 2H, CH2CH2N, J = 4.4 Hz), 3.83–3.98 (m, 3H,OCH2CH(OH)CH2), 4.21 (t, 2H, OCH2CH2, J = 4.3 Hz),6.32–6.40 (m, 2H, arom), 6.51 (d, 1H, arom,J = 8.2 Hz).Anal. C18H26N2O7 (C, H). 45: δ 1.09 (s, 9H, C(CH3)3),2.2 (brs, 3H, OH and 2NH), 2.60–2.85 (m, 2H,OCH2CH(OH)CH2), 3.35 (t, 2H, CH2CH2N, J = 4.4 Hz),3.80–3.95 (m, 3H, OCH2CH(OH)CH2), 4.20 (t, 2H,OCH2CH2, J = 4.3 Hz), 6.30–6.40 (m, 2H, arom), 6.52(d, 1H, arom,J = 8.2 Hz). Anal. C23H32N2O11 (C, H, N).46: δ 1.12 (s, 6H, CH(CH3)2), 2.55–2.87 (m, 6H, OH,2NH, CH(CH3)2 and OCH2CH(OH)CH2), 3.70–4.04 (m,5H, CH2CH2N and OCH2CH(OH)CH2), 4.22 (t, 2H,OCH2CH2, J = 4.3 Hz), 6.65 (dd, 1H, arom,J = 8.4 Hz,2.8 Hz), 6.80 (d, 1H, arom,J = 8.5 Hz). Anal.C18H26N2O11 (C, H). 47: δ 1.10 (s, 9H, C(CH3)3), 2.35(brs, 3H, OH and 2NH), 2.57–2.85 (m, 2H,OCH2CH(OH)CH2), 3.38 (t, 2H, CH2CH2N, J = 4.3 Hz),3.70–3.87 (m, 3H, OCH2CH(OH)CH2), 4.17 (t, 2H,OCH2CH2, J = 4.3 Hz), 6.15–6.23 (m, 2H, arom), 6.66(dd, 1H, arom,J = 8.4 Hz, 2.8 Hz). Anal. C19H28N2O7 (C,H).

5.1.12. 4-Acetyl-6- and 7-[3-(3,4-dimethoxybenzyl)amino-2-hydroxypropoxy]-3,4-dihydro-2H-1,4-benzoxa-zines48 and49

A mixture of 14 or 15 (0.0025 mol) and 3,4-dimethoxybenzylamine (0.025 mol) in isopropyl alcohol(30 mL) was refluxed for 3 h. The solvent was evaporatedin vacuo and the residue was purified by crystallizationfrom diethyl ether.1H-NMR (CDCl3, 200 MHz) 48: δ1.60 (brs, 2H, OH and NH), 2.25 (s, 3H, NCOCH3),2.6–2.85 (m, 2H, OCH2CH(OH)CH2), 3.70–3.92 (m,13H, CH2CH2N, CH2NHCH2CH(OH), 2CH3O andOCH2CH(OH)CH2), 4.20–4.30 (m, 2H, OCH2CH2),

6.40–6.50 (m, 2H, arom), 6.72–6.87 (m, 3H, arom), 7.75(s, 1H, arom). Anal. C26H32N2O10 (C, H). 49: δ 2 (brs,2H, OH and NH), 2.30 (s, 3H, NCOCH3), 2.67–2.92 (m,2H, OCH2CH(OH)CH2), 3.70–3.95 (m, 13H, CH2CH2N,CH2NHCH2CH(OH), 2CH3O and OCH2CH(OH)CH2),4.23 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.60–6.88 (m, 5H,arom), 7.35 (s, 1H, arom). Anal. C26H32N2O10 (C, H).

5.2. Pharmacological methods

5.2.1.â1-Adrenoceptor binding assayPellets containingâ1 type adrenergic receptors were

obtained from turkey erythrocyte membranes as de-scribed in the literature [34]. [3H]Dihydroalprenolol([3H]DHA) (NEN), having a specific activity of 99.9Ci/mmol and radiochemical purity> 98.5%, was used asligand.

â1-adrenergic receptor binding assays were determinedas follows: 100µL of membrane (≈ 431µg/mL of proteindiluted 1:8 v/v) were incubated for 15 min at 37 °C with100µL of 6 nM [3H]DHA and 100µL of various con-centrations of the test compounds (dissolved in salinewith DMSO 2.5%) and 12 mM Tris-HCI, pH = 7.5 (totalvol. 1 mL). The incubations were stopped by adding4 mL of cold buffer (12 mM Tris-HCI) followed by rapidfiltration through glass fibre filter disks (Whatman GF/B).The samples were subsequently washed 3 times with4.5 mL of the same buffer and placed into scintillationvials, 10 mL of Filter-Count (Packard) liquid scintillationcocktail was then added to each vial and counting wascarried out by scintillation spectrometer (Packard TRI-CARB 300C). Non-specific binding was defined as non-displaceable binding in the presence of 100µL of 10 µMpropranolol. Blank experiments were carried out to de-termine the effect of the solvent (2.5%) on the binding.

The concentrations of the test compounds that inhib-ited [3H]DHA binding by 50% (IC50) were determined bylog-probit analysis with six concentrations of the displac-ers, each performed in triplicate. The IC50 values ob-tained were used to calculate apparent inhibition con-stants (Ki) by the method of Cheng and Prussoff [35],from the following equation: Ki = IC50/(1 + S/KD) whereS represents the concentration of the ligand used and KD

its receptor dissociation constant (KD values, obtained byScatchard analysis [36], for [3H]DHA is 3.6 × 10–9 M).

5.2.2.â2-Adrenoceptor binding assayPreparation of lung homogenate: male Sprague-

Dawley rats were sacrificed by decapitation. The rightlung was removed free of the major bronchi. Lungs werehomogenized with a Brinkman Polytron (setting 5 for15 s) in 50 volumes of buffer, 75 mM Tris-HCI (pH 7.65),25 mM MgCl2 and then centrifuged at 30 000g for

915

10 min twice. The resulting pellets were resuspended in100 volumes of buffer, 75 mM Tris-HCI (pH 7.65),25 mM MgCl2, then were frozen at –80 °C before beingassayed [37, 38]. [3H]Dihydroalprenolol was used asligand. 300µL of lung membrane were incubated for30 min at 37 °C with 50µL of 6 nM [3H]DHA, 50 µL ofketanserine 10–7 M 5HT antagonist, 50µL of practolol10–6 M asâ2 antagonist and 50µL of various concentra-tions of the test compounds (dissolved in saline withDMSO 5%, or H2O) and 75 mM Tris-HCI (pH 7.65),25 mM MgCl2 (total volume 0.5 mL). The samples weresubsequently washed with 4.5 mL of the same buffer andplaced into scintillation vials. 10 mL of Filter-Count(Packard) liquid scintillation cocktail was then added toeach vial and counting was carried out by scintillationspectrometer (Packard Tricarb 300C). Non-specific bind-ing was defined as non-displaceable binding in thepresence of 50µL of 10 µM ICI 118551.

The concentration of the test compounds that inhibited[3H]DHA binding by 50% (IC50) were determined bylog-probit analysis with four concentrations of the dis-placers, each performed in triplicate. Non-specific bind-ing was measured in the presence of 50µL of 10 µMunlabelled ICI 118551, and specific binding as the differ-ence between total and non-specific binding. Blank ex-periments were carried out to determine the effect of thesolvent (5%) on the binding.

5.2.3. Inotropic, chronotropic and coronary vasodilatingactivity assays

The inotropic and chronotropic effects of the testcompounds were examined with the use of isolatedblood-perfused dog heart preparations. The hearts wereexcised from mongrel dogs of either sex weighing8–14 kg. The isolated blood-perfused papillary muscleand sinoatrial node preparations were prepared accordingto the methods of Endoh and Hashimoto [39] and Kubotaand Hashimoto [40] respectively. The preparations werecross-circulated through the cannulated arteries withblood from a donor anaesthetized with sodium pentobar-bital and receiving heparin. The perfusion pressure waskept constant at 100 mm Hg. The papillary muscle wasstimulated at a frequency of 2 Hz and the tensiondeveloped by the muscle was measured with a forcedisplacement transducer (Shinkoh, UL-20-240). Sinusrate was measured with the use of a cardiotachometer(Data Graph, T-149) triggered by the developed tensionof the right atrium. Blood flow through the cannulatedarteries was measured with an electromagnetic flowmeter (Nihon Kohden, MF-27). Signals of these param-eters were recorded on a thermal pen recorder (NEC-

Sanei, Recti-Horiz 8K). The compounds were injectedintra-arterially with microsyringes.

Acknowledgements

The authors thank Sig. Michele Guerrini for technicalassistance. Financial support from MURST (60%) andLAVIFARM S.A. is gratefully acknowledged.

References

[1] Kajino M., Shibouta Y., Nishikawa K., Meguro K., Chem. Pharm.Bull. 39 (1991) 2896–2905.

[2] Matsumoto Y., Tsuzuki R., Matsuhisa A., Takayama K., Yoden T.,Uchida W. et al., Chem. Pharm. Bull. 44 (1996) 103–114.

[3] Kuroita T., Sakamori M., Kawakita T., Chem. Pharm. Bull. 44(1996) 756–764.

[4] Matsuoka H., Ohi N., Mihara M., Suzuki H., Miyamoto K.,Maruyama N. et al., J. Med. Chem. 40 (1997) 105–111.

[5] Yoshitomi Pharmaceutical Industries, Ltd., Japan Kokai Tokyo81 51, 465; Chem. Abstr. 95 (1981) 115577.

[6] Cecchetti V., Fravolini A., Fringuelli R., Schiaffella F., MascellaniG., Pagella P., Rugarli P., Il Farmaco Ed. Sci 42 (1987) 61–75.

[7] Schromm K., Mentrup A., Renth E., Fuegner A., U.S. Patent, US 4,460, 581; Chem. Abstr. 101 (1984) 191939.

[8] Jaeggi K., Ostermayer F., Schroeter H., Ger. Offen. 2, 700, 193;Chem. Abstr. 87 (1977) 152206.

[9] Erez M., Shtacher G., Weinstock M., J. Med. Chem. 21 (1978)982–984.

[10] Angelopoulos C., (1997) Ph. D., University of Athens, Departmentof Pharmacy.

[11] Nakagawa K., Nishi T., Oshiro Y., Japan Kokai 76 68, 575; Chem.Abstr. 86 (1977) 89632.

[12] Nakagawa K., Murakami N., Yoshizaki S., Tominaga M., Mori H.,Yabuuchi Y., Shihtani S., J. Med. Chem. 17 (1974) 529–533.

[13] Winkler W., Arzn. Forsch. 33 (1983) 279–280.

[14] Yabuuchi Y., Kinoshita D., Jpn. J. Pharmacol. 24 (1974) 853–886.

[15] Khouili M., Guillaumet G., Coudert G., Pujol M.D., Il Farmaco. 51(1996) 175–184.

[16] Khouili M., Guillaumet G., Coudert G., Pujol M.D., Il Farmaco. 51(1996) 185–188.

[17] Crowther A.F., Howe R., Smith L.H., J. Med. Chem. 14 (1971)511–513.

[18] Hoefle M.L., Hastings S.G., Meyer R.F., Corey R.M., Holmes A.,Stratton C.D., J. Med. Chem. 18 (1975) 148–152.

[19] Keh G.S., Raper C., Dowd H., Clin. Exp. Pharmacol. Physiol 5(1978) 393–398.

[20] Tominaga M., Yo E., Ogawa H., Yamashita S., Yabuuchi Y.,Nakagawa K., Chem. Pharm. Bull. 32 (1984) 2100–2110.

[21] Tominaga M., Ogawa H., Yo E., Yamashita S., Yabuuchi Y.,Nakagawa K., Chem. Pharm. Bull. 35 (1988) 3699–3704.

[22] Fujioka T., Teramoto S., Mori T., Hosokawa T., Sumida T., Tomi-naga M., Yabuuchi Y., J. Med. Chem. 35 (1992) 3607–3612.

[23] Combs D.W., Rampulla M.S., Bell S.C., Klaubert D.H., Tobia A.J.,Falotico R., Haertlein B., Lakas-Weiss C., Moore J.B., J. Med.Chem. 33 (1990) 380–386.

916

[24] Furniss B.S., Hannaford A.J., Rogers V., Smith P.W.G., Tatchell A.R.(Eds.), Vogel’s textbook of practical Organic Chemistry fourthedition, Longman, Harlow, 1981, p. 753.

[25] Liu L.L., Chen H.C., Cao S.L., Li R.T., Synth. Commun. 24 (1994)833–838.

[26] Shtacher G., Erez M., Cohen S., J. Med. Chem. 16 (1973) 516–519.

[27] Loudon J.D., Ogg J., J. Chem. Soc. (1955) 739–744.

[28] Booher R.N., Kornfeld E.C., Smalstig E.B., Clemens J.A., J. Med.Chem. 30 (1987) 580–583.

[29] Ramage G.R., Hill J., British Patent, 1, 024, 626 (1966); Chem.Abstr. 64 (1966) 17614.

[30] Hill J., Ramage G.R., J. Chem. Soc. (1964) 3709–3713.

[31] Grollier J.F., Ger. Offen. DE 3, 545, 371 (1986); Chem. Abstr, 105(1986) 158608a.

[32] Bugant A., Estradier F., Ger. offen 1, 940, 085 (1970); Chem. Abstr.,73 (1970) 36576p.

[33] Coudert G., Guillaumet G., Loubinoux B., Synthesis (1979) 543.

[34] Minneman K.P., Weilland G.A., Molinoff P.B., Mol. Pharmacol. 16(1979) 34–46.

[35] Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973)3099–3108.

[36] Scatchard G., Ann. NY Acad. Sci. 51 (1949) 660–672.

[37] Minneman K.P., Hegstrand L.R., Molinoff P.B., Mol. Pharmacol. 16(1979) 21–33.

[38] Aarons R.D., Molinoff P.B., J. Pharmacol. Exp. Ther. 221 (1982)439–443.

[39] Endoh M., Hashimoto K., Am. J. Physiol. 218 (1970) 1459–1463.

[40] Kubota K., Hashimoto K., Naunyn-Schmiederg’s Arch. Pharmacol.278 (1973) 135–150.

917

Original article

â1- and â2-Adrenoceptor antagonist activity of a seriesof para-substituted N-isopropylphenoxypropanolamines

Simon N. Louis*, Tracy L. Nero, Dimitri Iakovidis, Felicia M. Colagrande, Graham P. Jackman,William J. Louis

Clinical Pharmacology and Therapeutics Unit, The University of Melbourne, Department of Medicine, Austin and Repatriation MedicalCentre Heidelberg 3084, Victoria, Australia

(Received 8 July 1998; revised 14 April 1999; accepted 20 May 1999)

Abstract – To further explore the structure-activity relationships ofâ-adrenoceptor (â-AR) antagonists, a series of 25para-substitutedN-isopropylphenoxy-propanolamines were synthesised, nine of which are new compounds. All have been examined for their ability toantagoniseâ1-ARs in rat atria andâ2-ARs in rat trachea. Substitution in thepara-position of the phenyl ring is thought to conferâ3-specificityand the selectivity of these compounds for theâ1-AR ranges from 1.5–234. None of the compounds tested were selective for theâ2-AR. Ofthe 25 compounds studied, 22 had reasonable (pA2 > 7) potencies for the ratâ1-AR. Only compound1 displayed reasonable (pA2 > 7) potencyfor the ratâ2-AR. Twenty two compounds were used as the training set for comparative molecular field analysis (CoMFA) of antagonistpotency (pA2) at the ratâ1- andâ2-ARs. The inclusion of a number of additional physical characteristics improved the QSAR analysis overmodels derived solely using the CoMFA electrostatic and steric fields. The final models predicted theâ1- andâ2-AR potency of the compoundsin the training set with high accuracy (r2 = 0.93 and 0.86 respectively). The finalâ1-AR model predicted theâ1-potencies of two out of thethree test compounds, not included in the training set, with residual pA2 values< –0.14, whereas the test compounds were not as well predictedby our finalâ2-AR model (residual pA2 values< –0.38). © 1999 Éditions scientifiques et médicales Elsevier SAS

comparative molecular field analysis / rat â-adrenoceptors / â-adrenoceptor antagonists /N-isopropylphenoxypropanolamines /quantitative structure activity relationships

1. Introduction

â-Adrenoceptors (â-ARs) are members of the largefamily of G-protein coupled receptors [1–4]. It has beenestablished so far that there are at least threeâ-ARsubtypes, designated theâ1-, â2- and â3-ARs [5–7].N-Isopropylphenoxypropanolamine (figure 1, R = H) isconsidered to be a non-selectiveâ-AR antagonist. Manystructure-activity studies in the past have focused uponthe N-isopropylphenoxypropanolamine core structurewith substitutions in theortho and/ormetapositions [8,9]. However, the effects ofpara-substituents upon theâ-blocking activity of this core structure have received

less attention [10–12]. Previous work by us [10, 13] andothers [11, 12, 14, 15] has established that highâ1-ARselectivity and potency can be achieved bypara-substitution of the phenyl ring and/or appropriate substi-tution of the phenoxypropanolamine amino group (e.g.-ethyl-3-(4-hydroxyphenyl)urea [13]).

*Correspondence and reprintsAbbreviations: Aryloxypropanolamine, AOPA;â-adrenoceptors,â-ARs; comparative molecular field analysis, CoMFA; ESP phenylring charge, ESP; flexibility of the para-substituent, flexibility;length of the para-substituent, length; optimal number of compo-nents, ONC; partial least squares, PLS; standard error, SE.

Figure 1. General structure ofpara-substitutedN-isopropyl-phenoxypropanolamines. The torsion anglesτ1 and τ2 definethe conformation of thepara-substituent and the oxypropano-lamine side-chains respectively in relation to the phenyl ring,for example whenτ1 = 0 thepara-substituent is planar to thephenyl ring.

Eur. J. Med. Chem. 34 (1999) 919−937 919© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

To further examine the effect ofpara-substitution uponboth â1- and â2-AR selectivity and potency, we havesynthesised a series ofpara-substitutedN-isopropyl-phenoxypropanolamines. Nine of these compounds arenew (6, 7, 9, 10, 15, 16, 18, 20 and22), that is to say thata chemical abstracts structure database search indicatedthat these compounds have not been previously reported.Eleven previously reported compounds (1, 2, 3, 4, 5, 8,12, 14, 19, 23 and 25) and five commercially availablecompounds (11 (metoprolol),13 (H 87/07), 17 (betax-olol), 21 (RO 31-1118) and24 (cicloprolol)) have alsobeen synthesised and included in the study in order toobtain comparative pharmacological data in the rat. All ofthese compounds have been examined in our laboratoryfor their ability to antagoniseâ1-ARs in rat atria andâ2-ARs in rat trachea, which are well established assources of theâ1- andâ2-AR subtypes respectively [16].The synthesis and structure-activity relationships, usingcomparative molecular field analysis (CoMFA), of thesepara-substitutedN-isopropylphenoxypropanolamines arepresented here.

2. Chemistry

All compounds were prepared as their racemic mix-tures by the general procedure shown infigure 2. It has

been established for simple phenoxypropanolamines thatthe S-isomer is the active isomer, with littleâ-AR activityresiding with the R-isomer [17–19]. Resolution of theracemate into the individual isomers or stereospecificsynthesis was therefore not carried out. The correspond-ing phenolsA reacted under aqueous alkaline conditionswith epichlorohydrin to produce the epoxidesB. Afterisolation, the crude epoxidesB were allowed to react withisopropylamine overnight to furnish the desired com-poundsC.

Compounds1, 2, 3, 4, 5, 8, 11 (metoprolol),12, 13 (H87/07),14, 17 (betaxolol),22, 21 (RO 31-1118),23, 24(cicloprolol) and25, have been previously reported andwere prepared following literature procedures [20–33].All intermediates and final compounds had virtuallyidentical physical and chemical data with those reported.

The precursor phenols of compounds6, 10, 15and16were prepared by reacting bromoethane, n-bromo-propane, 4-cyclohexylethyl bromide and 4-fluorophenyl-ethyl bromide with 4-benzyloxyphenol, thus producingthe benzyl ethersD which were subsequently hydroge-nated giving the desired phenolsE (4-ethoxyphenol),F(4-n-propoxyphenol),G (4-cyclohexylethoxyphenol) andH (4-fluorophenethoxyphenol) as shown infigure 3.

The precursor phenols of compounds18, 20 and 22were prepared by the sequence of reactions shown in

Figure 2. General procedure for preparingN-isopropylphenoxypropanolamines. Reagents: (i) aqueous KOH, ETOH; (ii) isopropy-lamine, ETOH.

Figure 3. Reagents: (i) K2CO3/NaI, anhydrous acetone; (ii) H2, 10% Pd/C.

920

figure 4. The corresponding alcohols 2-methylpropane-1-ol, 2-hydroxymethyltetrahydrofuran and n-pentane-1-olreacted with chloroacetic acid to furnish the correspond-ing alkoxy acetic acidsI . These were reduced to theircorresponding alkoxy-ethanolsJ whose tosylatesK re-acted with 4-benzyloxyphenol, to produce the benzylethersL which were subsequently hydrogenated givingthe desired phenolsM , N andO.

The precursor phenol of compound7 was prepared bythe reactions shown infigure 5[34, 35]. Phenyl-n-butanoate reacted with aluminium chloride to furnish amixture of regioisomers P (2- and 4-hydroxypropio-phenone), of which 4-hydroxypropiophenone, after iso-

lation, was reduced with sodium amalgam to produce thedesired phenolQ.

The precursor phenol of compound9 was preparedaccording to the reactions shown infigure 6. Ethyl-4-hydroxyphenyl propionate was benzylated to produce theether R, which was subsequently reduced to give theprotected alcoholS. Hydrogenation ofS produced thedesired phenolT.

The above mentioned phenols were then used toproduce compounds6, 7, 9, 10, 15, 16, 18, 20 and 22following the general procedure shown infigure 2. Thegeneral physical data for these compounds are given intable I.

Figure 4. Reagents: (i) ClCH2COOH, NaH, DMF; (ii) LiAlH4, THF; (iii) pTsCl, pyridine; (iv) NaH, DMF; (v) H2, 10% Pd/C.

Figure 5. Reagents: (i) anhydrous AlCl3; (ii) Zn-Hg, MeOH, HCl.

921

3. Pharmacology

All compounds were evaluated for their ability toantagoniseâ1- andâ2-ARs in rat atria and tracheal ringsrespectively. Cumulative concentration-response curveswere obtained in each preparation as described by VanRossum [36] and curves were fitted by computer analysisaccording to the method of Zabrowsky et al. [37] usingthe sigmoidal fit function of the Origin graphics pack-age [38]. The antagonist potencies, or pA2 values, werecalculated using equation 1 [39] and represent the mean±

SE from 4–9 individual experiments. Theâ-AR antago-nist potencies and the subtype selectivity are given intable II.

pA2 = –logS @antagonist#dose ratio− 1

D (1)

where the dose ratio=@ (-)-isoprenaline# atEC50 in the presence of antagonist

@ (-)-isoprenaline# atEC50 in the absence of antagonist

Figure 6. Reagents: (i) BzBr, K2CO3, NaI, anhydrous acetone; (ii) LiALH4, THF; (iii) H2, 10% Pd/C.

Table I. Physical data for compounds6, 7, 9, 10, 15, 16, 18, 20 and22.

Compound Yield M.p. (°C) Formulab

(from phenol) HCl salta

6 3.66 g 87.5–88.5 C14H23NO3.HCl(63.4 %)

7 4.35 g 70.5–73.0 C16H27NO2.HCl(72.1 %)

9 2.41 g 68.5–71.5 C15H25NO3.HCl(39.7 %)

10 3.56 g 116.5–119.0 C15H25NO3.HCl(58.6 %)

15 4.91 g 131.0–133.0 C20H33NO3.HCl(66.3 %)

16 11.71 g 136.0–138.0 C20H26FNO3.HCl(61.4 %)

18 4.00 g 89.5–91.3 C18H31NO4.HCl(55.3 %)

20 3.59 g 165.5–167.0 C11H31NO5.HCl(48.4 %)

22 3.90g 77.0–79.0 C19H33NO4.HCl(52.1 %)

aEthanol-ether was the solvent system used for recrystallisation.bEmpirical formula; elemental analyses were carried out for the ninecompounds shown and were within 0.4% per element.

922

Table II. Antagonist activity, subtype selectivity, flexibility of para-substituent,π, ESP phenyl ring charge at 90° and 0° and length of thepara-substituent for the compounds examined in this study.

Compound R Group â1-ARa pA2 â2-ARb pA2 â1-ARselectivityc

Flexi-bilityd

πe ESP phenylring chargeat 90o (e)

ESP phenylring chargeat 0o (e)

Lengthf

(Å )

Training Compounds1 8.09± 0.14 7.47± 0.14 4.2 0 0.00 –0.83 –0.83 0.00

2 6.60± 0.18 6.43± 0.17 1.5 0 –0.86 –0.67 –0.67 1.41

3 7.13± 0.04 6.47± 0.11 4.6 0 0.47 –0.71 –0.71 1.53

4 6.58± 0.19 6.34± 0.09 1.7 1 –0.04 –0.68 –0.72 2.30

5 7.41± 0.13 6.56± 0.18 7.1 2 1.37 –0.76 –0.72 3.84

6 7.08± 0.27 5.94± 0.16 14 2 0.36 –0.71 –0.74 3.61

7 7.49± 0.26 6.23± 0.23 18 3 1.82 –0.72 –0.68 5.00

8 7.38± 0.19 5.39± 0.16 98 3 –0.09 –0.75 –0.69 5.06

9 7.13± 0.10 6.58± 0.34 3.5 3 0.06 –0.72 –0.69 4.86

10 7.16± 0.14 5.72± 0.38 28 3 0.84 –0.68 –0.72 4.79

11 7.60± 0.17 6.43± 0.13 15 3 0.11 –0.72 –0.69 4.86

12 7.98± 0.21 6.77± 0.28 16 4 0.57 –0.74 –0.70 6.14

13 7.37± 0.09 5.97± 0.19 25 4 –0.13 –0.68 –0.72 5.91

14 6.67± 0.09 5.36± 0.18 20 2 2.06 –0.68 –0.69 5.08

15 7.65± 0.12 5.93± 0.23 52 4 2.57 –0.66 –0.70 6.45

16 8.01± 0.12 5.64± 0.12 234 4 2.13 –0.70 –0.73 8.72

17 8.06± 0.15 6.38± 0.13 48 5 0.96 –0.73 –0.70 7.39

923

4. Results and discussion

4.1. Isolated tissue preparations

Functional potencies of compounds1–25 for inhibiting(-)-isoprenaline-induced: (i)â1-AR mediated chronotro-pic effects in spontaneously beating rat atria; and (ii)â2-AR mediated relaxation of rat tracheal chain previ-ously contracted with 1µM carbachol are listed intable II. The unsubstituted reference compound1 had ahigh potency at both theâ1-AR (pA2 = 8.09) andâ2-AR(pA2 = 7.47). Para-substitution reduced the potency ofthe compounds for bothâ1- and â2-ARs (compounds2–11, 13–15, 18, 19, 20, 23 and24; table II), except forcompounds12, 16, 17, 21, 22 and 25 (â1-AR pA2s =7.98–8.06) which had similarâ1-AR potencies to theunsubstituted compound1. Compound12 (â2-AR pA2 =

6.77) was the most potent of thepara-substituted com-pounds at inhibiting ratâ2-ARs (c.f. reference compound1, pA2 = 7.47).

Overall, the compounds had higher potencies for theâ1-AR than the â2-AR, with the â1/â2 selectivitiesranging from 1.5–234. Compounds1, 2, 3, 4, 9 and23,however, were at the lower end of the selectivity scaleand are considered to be non-selective in the rat.

The animal species used in this study to determine boththe â1- andâ2-AR functional potency of the compoundswas the rat. Previously publishedâ1-AR antagonistfunctional potency, or activity data has been obtainedfrom both the rat and guinea-pig [30, 40], whereas theguinea-pig is the commonly used animal species forpublishedâ2-AR antagonist activity [30, 40]. By usingthe rat as the source for bothâ1- and â2-ARs, any

Table II. Continued

Compound R Group â1-ARa pA2 â2-ARb pA2 â1-ARselectivityc

Flexi-bilityd

πe ESP phenylring chargeat 90o (e)

ESP phenylring chargeat 0o (e)

Lengthf

(Å )

Training Compounds18 7.75± 0.22 5.69± 0.20 115 6 1.13 –0.73 –0.70 8.54

19 7.61± 0.15 5.40± 0.18 162 5 1.76 –0.72 –0.69 9.51

20 7.74± 0.09 5.49± 0.27 178 6 0.32 –0.69 –0.73 9.36

21 7.98± 0.21 5.68± 0.10 200 7 2.05 –0.72 –0.73 12.4

22 8.04± 0.23 5.80± 0.24 174 8 1.66 –0.72 –0.72 10.8

Test Compounds23 7.22± 0.08 6.47± 0.17 5.6 1 0.92 –0.77 –0.75 2.50

24 7.73± 0.07 5.44± 0.10 195 6 0.72 –0.73 –0.70 8.59

25 8.04± 0.30 5.81± 0.25 170 5 1.31 –0.74 –0.70 8.36

aâ1-AR antagonist pA2 value± SEM determined in isolated spontaneously beating rat atria.bâ2-AR antagonist pA2 value± SEM determinedin rat tracheal chain preparation.cSelectivity = antilog (pA2 â1-AR – pA2 â2-AR). dFlexibility was defined as the number of fully rotatablenon-H bonds in thepara-substituent.eπ was logP of the compound in question minus logP of compound1, logP was calculated using PrologP.fLength was defined as the distance between thepara-carbon atom of the phenyl ring and the most distant non-H atom.

924

interspecies differences between the receptor subtypeswas removed.

4.2. Alignment of molecules

The alignment of the compounds is the most importantfeature of CoMFA analysis [41]. It has been proposed thatAOPA type compounds interact withâ-ARs via a threepoint pharmacophore consisting of theâ-hydroxyl group,the amino group and an electron rich moiety which isusually a phenyl ring [42, 43] and more recent site-directed mutagenesis studies support this hypoth-esis [44–47]. If we assume the compounds all act via thesame mechanisms, then these common points of interac-tion must superimpose and hence we have assumed thatthe AOPA core structure remains fixed for all compoundsin our conformational analysis.

Previous studies within our laboratory [10] examinedtwo conformations of each molecule ie.τ1 = 90° or 0°. Inthe present study we have examined the effect of thesevalues ofτ1 on our CoMFA analysis.Figure 7adisplays

the superimposition of the compounds whenτ1 = 90° andfigure 7bwhenτ1 = 0°.

4.3. CoMFA analysis for antagonist potency at the ratâ1-AR

The CoMFA (SYBYL version 6.40) results for antago-nist potency at ratâ1-ARs for the twenty-two compoundsin the training set are given intable III. A range ofcolumn filtering values were used (1.0–16.0 kcal/mol).Column filtering omits from the analysis columns, latticepoints whose variance is less than the specified value(SYBYL). Partial least squares analysis (PLS) of the dataidentified that whens1 was set to 90° (figure 1) and thedefault CoMFA settings were used, which included bothsteric and electrostatic fields, the highest cross validatedr2 (q2) value was obtained when column filtering was setto 4.0 kcal/mol (q2 = 0.24; optimum number of compo-nents (ONC) = 4;table III). When s1 = 0°, q2 wasmaximised when column filtering was set at 5.0 kcal/mol(q2 = 0.38 and ONC = 2;table III).

Figure 7. (a)Structural alignment of the training set of compounds (compounds1–22) whenτ1 = 90°. (b) Structural alignment of thetraining set of compounds whenτ1 = 0°.

925

In addition to the PLS analysis of the CoMFA data,several other physical parameters of the compounds weredetermined and included in further CoMFA analysis.These parameters included the flexibility, length andπ ofthe para-substituent and the ESP phenyl ring charge(table II). The flexibility of thepara-substituent (flexibil-ity) was determined by assigning an integer value for thenumber of torsion angles which affected the conformationof the substituent. The term ‘ESP phenyl ring charge’(ESP) was defined in this study as the sum of the sixphenyl ring carbon atom esp charges. The phenyl ringcharges of all the compounds studied (table II) werecalculated using the AM1 Hamiltonian within MOPAC(version 6.0), specifying the key words ‘ESP’, ‘PRE-CISE’ and ‘NOMM’. The length of thepara-substituent(length) was defined as the distance between thepara-

carbon of the phenyl ring and the most distant non-Hatom, hence thepara-substituent length of compound1 iszero (table II). The π value of thepara-substituent wasdefined as the logP of the compound minus the logP ofcompound1, logP having been calculated using PallasPrologP (version 1.1). When these additional parameterswere included with CoMFA they generally improved q2.For instance, whenτ1 = 90° a combination of the CoMFAfields and ESP gave the highest q2 (q2 = 0.769; ONC =11; SE = 0.332;table III), and whenτ1 = 0° a combina-tion of the CoMFA fields, ESP and flexibility maximisedq2 (q2 = 0.657; ONC = 2; SE = 0.29;table III). Theformer model was selected for further analysis as itpossessed the highest q2 of any of theâ1-AR modelsexamined (table III).

Table III. CoMFA results for ratâ1-ARs (n = 22).

Analysis variables τ1 Column filtering q2 ONC SE

CoMFA 90o 4 0.244 4 0.461CoMFA, flexibilitya 90o 4 0.527 3 0.354CoMFA, ESPb,c 90o 13 0.747 10 0.332CoMFA, lengthd 90o 4 0.322 6 0.465CoMFA, π 90o 4 0.186 5 0.493CoMFA, flexibility, ESP 90o 9 0.686 4 0.331CoMFA, flexibility, length 90o 4 0.375 4 0.419CoMFA, flexibility, π 90o 4 0.398 4 0.411CoMFA, ESP, length 90o 13 0.672 10 0.358CoMFA, ESP,π 90o 13 0.638 10 0.388CoMFA, length,π 90o 4 0.247 5 0.474CoMFA, flexibility, ESP, length 90o 9 0.655 4 0.311CoMFA, flexibility, ESP,π 90o 10 0.632 5 0.332CoMFA, flexibility, length,π 90o 4 0.259 4 0.457CoMFA, ESP, length,π 90o 5 0.636 7 0.352CoMFA, flexibility, ESP, length,π 90o 9 0.610 5 0.341CoMFA 0o 5 0.384 2 0.394CoMFA, flexibilitya 0o 5 0.592 2 0.320CoMFA, ESPb 0o 5 0.604 4 0.334CoMFA, lengthc 0o 5 0.524 2 0.346CoMFA, π 0o 5 0.553 2 0.335CoMFA, flexibility, ESP 0o 5 0.657 2 0.294CoMFA, flexibility, length 0o 5 0.571 2 0.328CoMFA, flexibility, π 0o 5 0.515 3 0.359CoMFA, ESP, length 0o 5 0.618 2 0.310CoMFA, ESP,π 0o 5 0.601 4 0.335CoMFA, length,π 0o 5 0.407 3 0.397CoMFA, flexibility, ESP, length 0o 5 0.635 2 0.303CoMFA, flexibility, ESP,π 0o 2 0.599 6 0.357CoMFA, flexibility, length,π 0o 5 0.482 3 0.371CoMFA, ESP, length,π 0o 5 0.563 4 0.351CoMFA, flexibility, ESP, length,π 0o 5 0.564 6 0.373

aFlexibility of the para-substituent.bBold indicates the options used for the final model.cESP phenyl ring charge.dLength of thepara-substituent.

926

The final model, without cross validation, was obtainedusing the default CoMFA options, 10 ONC and 13 kcal/mol column filtering,τ1 = 90o, including ESP. This modelhad an r2 value of 0.95, a standard error (SE) of 0.15 andan F (11, 10) of 21.47. The regression equation for ESP,taken from the PLS output file, was:

â1-pA2 = 0.995–8.49 ESPThe relative contributions of the components were:

steric, 0.88; electrostatic, 0.00; and ESP, 0.12.Figure 8adisplays the relationship between calculated and mea-

sured pA2 values for the non-cross validated analysis andresidual values are given intable IV.

4.4. CoMFA analysis for antagonist potency at the ratâ2-AR

Similarly, the default CoMFA settings and a range ofcolumn filtering values (1.0–16.0 kcal/mol) were usedwhen analysing theâ2-AR data. Whenτ1 = 90o thehighest q2 value was obtained when column filtering was

Figure 8. (a)Plot of the actualâ1-AR pA2 verses the predictedâ1-AR pA2 compounds intable II (n = 22) using the final model, noncross validated, and an optimum number of components of 4.(b) Plot of the actualâ2-AR pA2 verses the predictedâ2-AR pA2compounds intable II (n = 22) using the final model, non cross validated, and an optimum number of components of 4.

927

set to 4 kcal/mol (q2 = 0.50; ONC = 5;table V). Whenτ1

= 0°, q2 was maximised when column filtering was set to5 kcal/mol (q2 = 0.45; ONC = 2; table V). When theadditional parameters were included in the analysis thehighest q2 occurred whenτ1 = 90o, column filtering wasset at 6 kcal/mol and both flexibility and ESP wereconsidered in addition to the steric and electrostatic fieldsgenerated within CoMFA (q2 = 0.60; ONC = 4; SE =0.38).

The final model, without cross-validation, for theâ2-AR was obtained using 4 ONC, column filtering set at6 kcal/mol and included flexibility and ESP. This modelhad an r2 value of 0.83, SE of 0.25 and an F (4, 17) of20.71. The regression equation for flexibility and ESPwas:

â2-pA2 = 2.05 + 0.07 flexibility–6.46 ESP.

The relative contributions of the components were:steric, 0.73; electrostatic, 0.00; flexibility, 0.11; and ESP,0.161. Figure 8b displays the relationship between thecalculated and measured pA2 values for the cross vali-dated analysis and residual values are given intable VI.

It is important to note that by using the column filteringvalues to maximise the q2 for the â1- andâ2-AR models(column filtering = 13 and 6 kcal/mol, respectively) iteffectively removes the electrostatic field parameter fromthe CoMFA analysis. For both models this occurs at

column filtering values of 5 kcal/mol whenτ1 = 90° and6 kcal/mol whenτ1 = 0°.

4.5. Comparison of contour maps

Figures 9Aand9B show compound16 binding withinthe steric contour maps for the finalâ1- and â2−ARmodels. The experimental data suggests thatpara-substitution decreases antagonist potency at bothâ1- andâ2-ARs. However, the contour maps, of theâ1- andâ2-AR pharmacophores display important differences,the most striking of which appears to be a bulk preferringpocket in theâ1-AR which can be accessed by longflexible para-substituents. Compounds that access thispocket have theirâ1-AR potency restored and the longerthe substituent the greater the restoration of their activity,compare the activity of compounds4, 13 and22. Com-pounds with shorter bulky substituents, like compound14cannot access this pocket and have lowâ1-AR potencies.The â2-AR, on the other hand, seems far more stericallyrestricted than theâ1-AR in the para-position. Forinstance, the addition of the short bulky substituent ofcompound14 reduces theâ1-AR potency by 1.42 logunits when compared to the base compound, whereas thissubstituent reducesâ2-AR potency by 2.11 log units.

Using compound16 as an example the differencesbetween the two steric maps can be clearly illustrated. Inthe â1-AR model (figure 9A), the para-substituent ofcompound16does not impinge on the sterically restrictedregion (ie. yellow region), however, the terminal fluoro-phenyl ring is embedded in the bulk preferring pocket(i.e. green region) and hence theâ1-potency is almostrestored to that of compound1 (table II). By contrast, thepara-substituent of compound16 is clearly embedded inthe sterically restricted region of theâ2-AR model(figure 9B) and hence this compound has a much lowerâ2-pA2 than compound1 (table II).

4.6. Predictive abilities of the CoMFA models

The potencies of three test compounds (table VII), notin the training set, were predicted at ratâ1- andâ2-ARsusing the final models and the predict property function inthe QSAR package of SYBYL (version 6.40). For theâ1-AR the predicted potencies for two of the compoundswere in close agreement with those obtained experimen-tally (residuals< 0.15). For theâ2-AR, however, all threecompounds were relatively poorly predicted (residuals–0.25 to –0.38) suggesting that our CoMFA model of theâ2-AR is not sufficiently resolved.

Table IV. CoMFA Experimental, calculated and residual activitiesof the training set of compounds forâ1-ARs.

Compound Experimental Calculated Residual

1 8.09 8.04 0.052 6.60 6.68 –0.083 7.13 7.02 0.114 6.58 6.73 –0.155 7.41 7.48 –0.076 7.08 6.96 0.127 7.49 7.31 0.188 7.38 7.58 –0.209 7.13 7.12 0.0110 7.16 7.19 –0.0311 7.60 7.56 0.0412 7.98 7.95 0.0313 7.37 7.43 –0.0614 6.67 6.68 –0.0115 7.65 7.61 0.0416 8.01 8.03 –0.0217 8.06 8.06 0.0018 7.75 7.75 0.0019 7.61 7.80 –0.1920 7.74 7.52 0.2221 7.98 8.03 0.0522 8.04 8.01 0.03

928

5. Discussion

The aim of this study was to determine the optimalstructural requirements ofpara-substitutedN-isopropyl-phenoxypropanolamines to maximise antagonist potencyand selectivity for â1-ARs. The study also providesinformation on the extent to which the CoMFA derivedmodels of the ratâ1- andâ2-ARs differed for a series ofpara-substituted â-AR antagonists. Althoughpara-substitution generally reduced antagonist potency at bothâ1- and â2-ARs, differences did exist between the twomodels. For both models, steric factors were the mostimportant (0.88 forâ1- and 0.73 forâ2-ARs). Theâ1-ARalso possesses a pocket which prefers bulky substituents.Long flexible compounds that accessed this region hadtheirâ1-AR potency restored which is consistent with ourprevious work [10] that suggested a hydrophobic binding

site existed that was accessible to longpara-substituentscontaining a ring system. Our data, however, suggeststhat although the pocket accommodated large, hydropho-bic substituents, eg. compounds15, 16and21, it will alsoaccept large, less hydrophobic substituents such as com-pounds17, 24and25and long flexible carbon chains likecompound22. The important characteristic is that thesubstituent is able to access this pocket, for example,shorter hydrophobic compounds that cannot reach thepocket, like compound14, encounter steric hindrance andhave lowâ1-potency.

In addition, a number of other physical characteristicsof the compounds seemed to influence their potency atâ1- and â2-ARs. For bothâ1- and â2-ARs, ESP wasnegatively correlated with potency suggesting that themore negative the phenyl ring charge the higher theâ1- or

Table V. CoMFA results for ratâ2-ARs (n = 22).

Analysis variables τ1 Column filtering q2 ONC SE

CoMFA 90o 4 0.501 5 0.434CoMFA, flexibilitya 90o 4 0.574 4 0.389CoMFA, ESPb 90o 6 0.586 4 0.383CoMFA, lengthc 90o 4 0.499 6 0.449CoMFA, π 90o 4 0.364 4 0.475CoMFA, flexibility, ESPd 90o 6 0.602 4 0.376CoMFA, flexibility, length 90o 4 0.498 4 0.422CoMFA, flexibility, π 90o 4 0.387 5 0.481CoMFA, ESP, length 90o 4 0.575 5 0.400CoMFA, ESP,π 90o 5 0.454 3 0.428CoMFA, length,π 90o 4 0.284 3 0.490CoMFA, flexibility, ESP, length 90o 5 0.587 4 0.383CoMFA, flexibility, ESP,π 90o 7 0.430 5 0.464CoMFA, flexibility, length,π 90o 4 0.372 3 0.459CoMFA, ESP, length,π 90o 12 0.473 2 0.409CoMFA, flexibility, ESP, length,π 90o 12 0.455 2 0.416CoMFA 0o 5 0.445 2 0.420CoMFA, flexibilitya 0o 5 0.462 4 0.437CoMFA, ESPb 0o 2 0.453 5 0.454CoMFA, lengthc 0o 5 0.463 4 0.437CoMFA, π 0o 5 0.351 5 0.495CoMFA, flexibility, ESP 0o 2 0.454 5 0.454CoMFA, flexibility, length 0o 5 0.479 4 0.430CoMFA, flexibility, π 0o 7 0.538 9 0.482CoMFA, ESP, length 0o 2 0.459 5 0.452CoMFA, ESP,π 0o 1 0.386 6 0.497CoMFA, length,π 0o 7 0.494 9 0.505CoMFA, flexibility, ESP, length 0o 2 0.491 5 0.438CoMFA, flexibility, ESP,π 0o 1 0.387 6 0.497CoMFA, flexibility, length,π 0o 1 0.440 10 0.555CoMFA, ESP, length,π 0o 1 0.451 6 0.470CoMFA, flexibility, ESP, length,π 0o 1 0.475 6 0.460

aFlexibility of the para-substituent.bESP phenyl ring charge.cLength of thepara-substituent.dBold indicates the options used for the finalmodel.

929

â2-AR potency which supports the hypothesis of previousworkers that an electron rich moiety is an essentialpharmacophoric element for binding toâ-ARs [42, 43].In addition, flexibility was positively correlated withâ2-AR potency and this may be related to the ability ofmore flexible para-substituents to avoid sterically re-stricted regions in theâ2-AR.

Unfortunately, our synthetic program was designed toexamineâ1-AR potency and hence only compound1 hadhigh potency (pA2 > 7.0) at the ratâ2-AR and higheraffinity compounds are required before we can confi-dently predict the structure of theâ2-AR pharmacophore.We can say, however, that theâ2-AR appears to be verysterically restricted around thepara-position of the phe-nyl ring of AOPA antagonists.

On the whole, the potency of the training set ofcompounds were predicted better by theâ1-AR model (r2= 0.95; residuals< ± 0.22) than theâ2-AR model (r2 =0.83; residuals< ± 0.80). In particular, the potency ofcompound8 was over estimated for theâ2-AR. This mayhave been due to the existence of a negative chargepreferring region in the region of the positively chargedamine in thepara-substituent. This, however, was notidentified by CoMFA perhaps due to the small number ofcompounds examined that possess this characteristic. It isof interest that the OH-group of compound9, which hasan electronegative nature, has significantly higherâ2-ARpotency than compound8 and slightly higher activitythan the uncharged substituent of compound7 (table II),therefore, this region requires further study.

Figure 9. (A) CoMFA contour map of the final model of theâ1-AR with compound16 embedded. Green areas show favoured areaswhere increasing bulk is associated with higher potency, yellow areas where increasing bulk is associated with lower potency and blueareas where positive charge is associated with higher potency.(B) CoMFA contour map of the final model of theâ2-AR with compound16 embedded.

930

The two CoMFA models expand the results of earlierstudies [10–15] that highâ1-selectivity and potency canbe achieved bypara-substitution of the phenyl ring. Forpara-substitutedN-isopropylphenoxy-propanolamines itappears at least two factors contribute toâ1-potency.These are the ability of thepara-substituent to access abulk preferring pocket and the phenyl ring charge. Thelatter is negatively correlated with potency.

6. Experimental section

6.1. Chemistry

6.1.1. GeneralMelting points were determined using a manual Gal-

lenkamp electrothermal apparatus (range 0–200 °C) inglass capillary tubes and are uncorrected. IR spectra wererecorded on a Perkin-Elmer FT IR 1600. NMR spectrawere recorded on a Varian Associates EM 360 spectrom-eter and are expressed inδ using TMS (tetramethylsilane) as reference. Mass spectra were recorded on aFinnigan 4000 series GC/MS Mass spectrometer or aThermo Instruments GCQ Mass spectrometer usingmethane gas as the ionising medium for CI (chemicalionisation spectra). All spectra were consistent with theassigned structures. Where analyses are indicated only by

the symbols of the elements, results obtained were within± 0.4% of the theoretical values. Descriptions of thesynthetic procedures for preparing compounds6, 7, 9, 10,15, 16, 18, 20 and 22 are outlined below. No attemptswere made to maximise yields.

6.1.2. MaterialsPhenol, 4-methylphenol, 4-ethylphenol, 4-methoxy-

phenol, 4-benzyloxyphenol, 4-fluorophenol, p-cresol,4-aminophenol, 4-fluorophenethyl bromide, 2-cyclo-hexyl-1-bromoethane, 2-(4-benzyloxyphenyl)-1-ethanol,benzyl bromide, 4-toluene sulfonylchloride, phenyl-n-butanoate, ethylchloroformate, bromoethane, iodomethane,2-hydroxymethyl tetrahydrofuran, 2-methyl-propane-1-ol, n-pentanol, isobutanol, cyclopentanol, cyclohexanol,chloroacetic acid, ethyl-4-hydroxyphenyl propionate,lithium aluminium hydride and dimethyl formamide(DMF) were purchased from Aldrich Chemicals. Sodiumhydride and isopropylamine were obtained from FlukaChemicals. Anhydrous aluminium chloride was obtainedfrom Merck. Inorganic reagents were supplied by AjaxChemicals. Analytical grade solvents, including carbondisulphide, were purchased from Rhone Poulenc Austra-lia. (-)-Isoprenaline and carbachol were obtained fromSigma. Silica plates (5× 10 cm, Silica F254) werepurchased from Merck. Silica for column chromatogra-phy (100 Å, 50 mm) was supplied by Amicon Inc., MA01923, USA.

6.1.3. General procedure 1 for the synthesis of 1-phenoxy-2,3-epoxypropanes(B)

The phenolA (20 mmol), KOH (1.23 g, 22 mmol) andepichlorohydrin (5.55 g, 60 mmol) in EtOH (60 mL)were stirred overnight at room temperature. The reactionmixture was evaporated to dryness and the residue waspartitioned between ether (or EtOAc) and water. Theorganic layer was washed with aqueous NaOH (5%),followed by a water wash and then dried over Na2SO4.After the solvent was evaporated, the residue was freedfrom excess epichlorohydrin under vacuum and used

Table VI. CoMFA Actual, calculated and residual activities of thetraining set of compounds forâ2-ARs.

Compound Actual Calculated Residual

1 7.47 7.41 0.062 6.43 6.34 0.093 6.47 6.53 –0.064 6.34 6.27 0.075 6.56 6.50 0.066 5.94 6.17 –0.237 6.23 6.03 0.208 5.39 6.19 –0.809 6.58 6.32 0.2610 5.72 5.76 –0.0411 6.43 6.21 0.2212 6.77 6.49 0.2813 5.97 6.11 –0.1414 5.36 5.39 –0.0315 5.93 6.07 –0.1416 5.64 5.49 0.1517 6.38 6.47 –0.0918 5.69 5.69 0.0019 5.40 5.33 0.0720 5.49 5.40 0.0921 5.68 5.67 0.0122 5.80 5.84 –0.04

Table VII. Comparison of predicted and experimentally determi-ned potencies at ratâ1- andâ2-ARs for the test compounds.

Compound Receptor Actual Predicted Residual

23 â1- 7.22 7.52 –0.3024 â1- 7.73 7.88 0.1525 â1- 8.04 7.94 0.10

23 â2- 6.47 6.85 –0.3824 â2- 5.44 5.80 –0.3625 â2- 5.81 6.06 –0.25

931

without further purification. All epoxides were homog-enous by TLC (silica, eluent dichloromethane).

6.1.4. General procedure 2 for preparing N-isopropyl-phenoxypropanolamines(C)

The crude 1-phenoxy-2,3-epoxypropaneB (20 mmol)and isopropylamine (100 mmol, 5 molar excess) in EtOH(60 mL) were stirred overnight at room temperature. Thereaction mixture was evaporated to dryness and theresidue was dissolved in ethanol and then treated withexcess ethereal HCl. The crude precipitate was recrystal-lised from a suitable solvent or purified by columnchromatography (silica, eluent: CH2Cl2/MeOH/NH4OH28%, 90:9:1).

6.1.5. 4-Ethoxyphenol(E)Ethyl bromide (6.0 g, 55 mmol), 4-benzyloxyphenol

(10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol)and NaI (0.5 g) were stirred and refluxed in anhydrousacetone (250 mL) for 48 h. The solvent was evaporatedand the residue was partitioned between dichloromethaneand water. The organic layer was washed with water anddried over anhydrous MgSO4. The solvent was evapo-rated and the residual 4-ethoxyphenylbenzyletherD (R =H) was recrystallised from EtOH. Yield = 8.2 g, 81.5%;m.p. = 64–66 °C; MS m/e 229 (M + 1);1H-NMR(CDCl3) δ 1.58 (3H, t), 4.15 (2H, q), 5.17 (2H, s), 7.06(5H, s), 7.58 (4H, s).

The ether D (R = H) (7.2 g, 36 mmol) in EtOH(200 mL) was hydrogenated over 10% Pd/C (0.1 g) atroom temperature and atmospheric pressure. When hy-drogen absorption ceased, the reaction mixture wasfiltered and evaporated to dryness. The productE wasrecrystallised from EtOH. Yield = 4.26 g, 87.0%; m.p. =64.5–65.5 °C; MS m/e 139 (M + 1);1H-NMR (CDCl3) δ1.61 (3H, t), 4.19 (2H, q), 5.63 (1H, s), 6.93 (4H, s).

6.1.6. 1-(4-Ethoxyphenoxy)-3-isopropylamino-2-pro-panol (6)

Prepared fromE according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 3.66 g, 63.4%; m.p. = 87.5–88.5 °C; MS m/e254 (M + 1); 1H-NMR (CD3OD) δ 1.33 (6H, d), 1.62(3H, t), 2.16–3.23 (6H, bm), 4.23 (2H, q), 7.10 (4H, s).Anal. (C14H23NO3.HCl) C, H, Cl, N.

6.1.7. 4-n-Butylphenol(Q)Phenyl-n-butanoate (82.0 g, 500 mmol) was added into

a stirred suspension of aluminium chloride (75 g) incarbon disulphide (80 mL) to furnish, after work-up, amixture of 2- and 4-hydroxyphenylpropiophenones (bu-tyrophenones) (64.5 g, 78.6%) which were separated byrepeated fractional distillation. The desired 4-hydroxy-

phenyl propiophenoneP was obtained in overall 33.4%yield (27.4 g) and had the following physical data: b.p. =148–150 °C (11.5 mm Hg); MS m/e 165 (M + 1);1H-NMR (CD3OD) δ 1.14 (3H, t), 2.39–2.82 (2H, m),3.62–3.98 (2H, m), 6.97 (5H, dd, the phenolic hydrogenis hiding under the aromatic protons).

Reduction ofP (27.0 g, 166 mmol) was achieved by itsportion wise addition into freshly prepared sodium amal-gam (68.0 g of powdered Zn and 5.03 g of HgCl2 in650 mL of methanol and 320 mL of concentrated HCl),evaporation of methanol, isolation and chromatographyon silica (eluent CH2Cl2:hexane, 1:1).Q was produced in62.5% yield (15.56 g); MS m/e 151 (M + 1);1H-NMR(CD3OD) δ 1.08 (3H, t), 1.25–2.05 (4H, m), 2.65 (2H, t),3.87 (5H, m), 6.89 (5H, dd, the phenolic hydrogen ishiding under the aromatic protons).

6.1.8. 1-(4-n-Butylphenoxy)-3-isopropylamino-2-pro-panol (7)

Prepared fromQ according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 4.35 g, 72.1%; m.p. = 70.5–73.0 °C; MS m/e266 (M + 1);1H-NMR (CD3OD) δ 1.13 (3H, t), 1.75 (6H,d), 2.51–2.93 (3H, m), 3.23–3.83 (5H, bm), 4.06–4.31(2H, m), 4.63–5.12 (2H, m), 6.77–7.28 (4H, dd). Anal.(C16H27NO2.HCl) C, H, Cl, N.

6.1.9. 4-(n-3≠-Hydroxypropyl)phenol(T)Ethyl-4-hydroxyphenylpropionate (54.5 g, 250 mmol)

was refluxed with benzyl bromide (44.5 g, 30.9 mL,260 mmol), anhydrous potassium carbonate (38.7 g,270 mmol) and sodium iodide (2.0 g) in dried acetone for48 h. Filtration of the cooled mixture and evaporation ofthe solvent produced a solid residue which was taken upin ethyl acetate, washed with diluted sodium hydroxidesolution and water, dried, filtered and evaporated. Theproduct, ethyl-4-benzyloxyphenylpropionate,R, was pu-rified by chromatography (silica, eluent CH2Cl2) andisolated as an oil. Yield = 68.6 g, 89.1%; MS m/e 309 (M+ 1); 1H-NMR (CDCl3) δ 1.16 (3H, t), 2.76 (2H, t), 3.82(2H, t), 4.64 (2H, q), 5.06 (2H, s), 6.88 (4H, s), 7.26 (5H,s).

Ethyl-4-benzyloxyphenylpropionate (61.6 g, 200 mmol)in anhydrous THF (200 mL) was added dropwise into asuspension of lithium aluminium hydride (8.0 g,210 mmol) in THF (500 mL). The resultant mixture wasrefluxed for 2 h, cooled, the excess reagent decomposedby the dropwise addition of water and filtered. The filtratewas evaporated to dryness and the residue was purified bychromatography (silica, eluent CH2Cl2). The isolatedproduct, 1-(4≠-benzyloxyphenyl)-3-propanol,S, was re-crystallised from ethanol-ether. Yield = 32.1 g, 66.2%;

932

m.p. = 59.5–61.5 °C; MS m/e 243 (M + 1);1H-NMR(CDCl3) δ 0.72 (1H, bs), 0.87–1.23 (2H, m), 2.81 (2H, t),3.87 (2H, t), 5.13 (2H, s), 7.11 (4H, dd), 7.52 (5H, s).

A solution of 1-(4≠-benzyloxyphenyl)-3-propanol(24.2 g, 100 mmol) in ethanol (600 mL) was hydroge-nated over 10% Pd/C at ambient temperature and pres-sure. Filtration of the catalyst and evaporation of thesolvent furnished the desired phenolT which was puri-fied by chromatography (silica, eluent CH2Cl2) andisolated as an oil. Yield = 12.81 g, 84.3%; MS m/e 153(M + 1); 1H-NMR (CDCl3/CD3OD 1:1)δ 1.77–2.23 (2H,m), 2.77 (2H, m), 3.81 (2H, m), 7.02 (4H, dd).

6.1.10. 1-(4-(n-3≠-Hydroxypropyl)-phenoxy)-3-isopropyl-amino-2-propanol(9)

Prepared fromT according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 2.41 g, 39.7%; m.p. = 68.5–71.5 °C; MS m/e268 (M + 1); 1H-NMR (CD3OD) δ 1.62 (6H, d),1.83–2.29 (2H, m), 2.61–3.12 (2H, m), 3.15–3.94 (6H,m), 4.07–4.36 (2H, m), 6.90–7.48 (4H, dd). Anal.(C15H25NO3.HCl) C, H, Cl, N.

6.1.11. 4-n-Propoxyphenol(F)n-Propyl bromide (6.8 g, 55 mmol), 4-benzyloxy-

phenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g,60 mmol) and NaI (0.5 g) were stirred and heated toreflux in anhydrous acetone (250 mL) for 48 h. Thesolvent was evaporated and the residue was partitionedbetween dichloromethane and water. The organic layerwas washed with water and dried over anhydrousMgSO4. The solvent was evaporated and the residual4-n-propoxyphenylbenzyletherD (R≠ = Me) was recrys-tallised from EtOH. Yield = 11.07 g, 83.2%; m.p. =67–69 °C; MS m/e 243 (M + 1);1H-NMR (CDCl3) δ1.22 (3H, t), 1.70–2.23 (2H, m), 4.03 (2H, t), 5.14 (2H, s),7.03 (5H, s), 7.52 (4H, s).

The etherD (R≠ = Me) (10.1 g, 42 mmol) in EtOH(200 mL) was hydrogenated over 10% Pd/C (0.1 g) atroom temperature and atmospheric pressure. When hy-drogen absorption ceased, the reaction mixture wasfiltered and evaporated to dryness. The productF wasrecrystallised from EtOH. Yield = 5.12 g, 80.2%; m.p. =67–69 °C; MS m/e 153 (M + 1);1H-NMR (CDCl3) δ1.19 (3H, t), 1.77–2.26 (2H, m), 4.06 (2H, t), 6.73 (1H, s),6.99 (4H, s).

6.1.12. 1-(4-n-Propoxyphenoxy)-3-isopropylamino-2-propanol(10)

Prepared fromF according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 3.56 g, 58.6%; m.p. = 116.5–119.0 °C; MSm/e 268 (M + 1);1H-NMR (CD3OD) δ 1.22 (3H, t), 1.72

(6H, d), 1.79–2.18 (2H, m), 3.27–3.78 (4H, m), 3.87–4.29(4H, m), 4.55–6.07 (1H, m), 5.48–5.74 (1H, m), 6.92(4H, s). Anal. (C15H25NO3.HCl) C, H, Cl, N.

6.1.13. 4-Cyclohexylethoxyphenol(G)4-Cyclohexylethyl bromide (10.5 g, 55 mmol), 4-ben-

zyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3

(8.29 g, 60 mmol) and NaI (0.5 g) were stirred andrefluxed in anhydrous acetone (250 mL) for 48 h. Thesolvent was evaporated and the residue was partitionedbetween dichloromethane and water. The organic layerwas washed with water and dried over anhydrousMgSO4. The solvent was evaporated and the residual4-(4≠-cyclohexylethoxy)phenylbenzyletherD (R≠ = cy-lohexyl) was recrystallised from EtOH. Yield = 13.6 g,81.4%; m.p. = 68–70 °C; MS m/e 311 (M + 1);1H-NMR(CDCl3) δ 0.89–1.96 (13H, m), 3.65–4.02 (2H, m), 4.95(2H, s), 6.72 (4H, s), 7.24 (5H, s).

The etherD (R≠ = cyclohexyl) (12.0 g, 36 mmol) inEtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g)at room temperature and atmospheric pressure. Whenhydrogen absorption ceased, the reaction mixture wasfiltered and evaporated to dryness. The productG wasrecrystallised from EtOH. Yield = 7.78 g, 85.6%; m.p. =91–93 °C; MS m/e 221 (M + 1);1H-NMR (CDCl3) δ0.77–2.01 (13H, m), 3.65–4.21 (2H, m), 4.47 (1H, s),6.73 (4H, s).

6.1.14. 1-(4-Cyclohexylethoxyphenoxy)-3-isopropylamino-2-propanol(15)

Prepared fromG according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 4.91 g, 66.3%; m.p. = 131.0–133.0 °C; MSm/e 336 (M + 1);1H-NMR (CD3OD) δ 0.91 (6H, d),0.64–1.53 (11H, bm), 2.80 (4H, m), 3.87 (5H, m), 6.43(4H, s). Anal. (C20H33NO3.HCl) C, H, Cl, N.

6.1.15. 4-Fluorophenethoxyphenol(H)4-Fluorophenethyl bromide (11.16 g, 55 mmol),

4-benzyloxyphenol (10.11 g, 50 mmol), anhydrousK2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirredand refluxed in anhydrous acetone (250 mL) for 48 h. Thesolvent was evaporated and the residue was partitionedbetween dichloromethane and water. The organic layerwas washed with water and dried over anhydrousMgSO4. After evaporation of the solvent, the residual4-(4≠-fluorophenethoxy)phenylbenzyletherD (R≠ =4-fluorophenyl) was recrystallised from EtOH. Yield =15.2 g, 87.9%; m.p. = 113–114 °C; MS m/e 323 (M + 1);1H-NMR (CDCl3) δ 3.05 (2H, m), 4.12 (2H, m), 5.11(2H, s), 6.88 (5H, s), 7.10 (4H, dd), 7.55 (5H, dd).

The etherD (R≠ = 4-fluorophenyl) (13.84 g, 40 mmol)in EtOH (250 mL) was hydrogenated over 10% Pd/C

933

(0.1 g) at room temperature and atmospheric pressure.When hydrogen absorption ceased, the reaction mixturewas filtered and evaporated to dryness. The productHwas recrystallised from toluene. Yield = 7.12 g, 76.1%;m.p. = 166–168 °C; MS m/e 233 (M + 1);1H-NMR(CDCl3) δ 2.45 (2H, m), 3.56 (2H, m), 4.67 (1H, s), 6.53(4H, dd), 6.93 (5H, dd, the phenolic hydrogen is hidingunder the aromatic protons).

6.1.16. 1-(4-(4≠-Fluorophenethoxy)-phenoxy)-3-iso-propylamino-2-propanol(16)

Prepared fromH according to the general procedure,chromatographed and isolated as the hydrochloride. Yield= 11.71 g, 61.4%; m.p. = 136.0–138.0 °C; MS m/e 348(M + 1); 1H-NMR (CD3OD) δ 1.16 (6H, d), 2.71–3.32(6H, m), 3.71–4.22 (4H, m), 6.46 (4H, s), 6.73 (4H, dd).Anal. (C20H26FNO3.HCl) C, H, Cl, F, N.

6.1.17. 4-(2≠-Methylpropoxy)ethoxyphenol(M )2-Methylpropoxyethyl bromide (9.96 g, 55 mmol),

4-benzyloxyphenol (10.11 g, 50 mmol), anhydrousK2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirredand refluxed in anhydrous acetone (250 mL) for 48 h. Thesolvent was evaporated and the residue was partitionedbetween dichloromethane and water. The organic layerwas washed with water and dried over anhydrousMgSO4. The solvent was evaporated and the residual4-(2-methylpropoxyethoxy)phenylbenzyletherL (R≠≠ =2-methylpropyl) was purified by chromatography (silica,eluent CH2Cl2) and isolated as an oil, homogenous byTLC. Yield = 13.8 g, 83.6%; MS m/e 301 (M + 1);1H-NMR (CDCl3) δ 1.16 (6H, d), 1.87–2.35 (1H, m),3.33–3.60 (2H, m), 3.77–4.35 (4H, m), 5.12 (2H, s), 6.94(4H, dd), 7.52 (5H, s).

The etherL (R≠≠ = 2-methylpropyl) (10.8 g, 36 mmol)in EtOH (200 mL) was hydrogenated over 10% Pd/C(0.1 g) at room temperature and atmospheric pressure.When hydrogen absorption ceased, the reaction mixturewas filtered and evaporated to dryness. The productMwas purified by chromatography (silica, eluent CH2Cl2)and isolated as an oil, homogenous by TLC. Yield =6.85 g, 86.5%; MS m/e 221 (M + 1);1H-NMR (CDCl3)δ 1.18 (6H, d), 1.97–2.42 (1H, m), 3.38–3.64 (2H, m),3.80–4.39 (4H, m), 6.96 (1H, s), 7.02 (4H, dd).

6.1.18. 1-4≠-(2≠≠-Methylpropoxyethoxyphenoxy)-3-iso-propylamino-2-propanol(18)

Prepared fromM according to the general procedures1 and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 4.00 g, 55.3%; m.p. = 89.5–91.3 °C; MS m/e326 (M + 1); 1H-NMR (CD3OD) δ 1.13 (6H, d), 1.75(6H, d), 3.37–3.62 (6H, bm), 3.80–4.37 (7H, bm), 6.94(4H, s). Anal. (C18H31NO4.HCl) C, H, Cl, N.

6.1.19. 4-(2≠-Tetrahydrofurylmethoxy)ethoxyphenol(N)2-Bromomethyl tetrahydrofuran (11.5 g, 55 mmol),

4-benzyloxyphenol (10.11 g, 50 mmol), anhydrousK2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirredand heated to reflux in anhydrous acetone (250 mL) for48 h. The solvent was evaporated and the residue waspartitioned between dichloromethane and water. Theorganic layer was washed with water and dried overanhydrous MgSO4. The solvent was evaporated and theresidual 4-(2≠-tetrahydrofurylmethoxy)ethoxyphenyl-benzyl etherL (R≠≠ = 2≠-tetrahydrofurylmethyl) waspurified by chromatography (silica, eluent CH2Cl2) andisolated as an oil, homogenous by TLC. Yield = 14.6 g,80.9%; MS m/e 329 (M + 1);1H-NMR (CDCl3) δ1.73–2.29 (4H, m), 3.48–4.39 (9H, m), 5.12 (2H, s), 7.03(4H, s), 7.55 (5H, s).

The etherL (R≠≠ = 2≠-tetrahydrofurylmethyl) (11.8 g,36 mmol) in EtOH (200 mL) was hydrogenated over 10%Pd/C (0.1 g) at room temperature and atmospheric pres-sure. When hydrogen absorption ceased, the reactionmixture was filtered and evaporated to dryness. TheproductN was purified by chromatography (silica, eluentCH2Cl2) and isolated as an oil, homogenous by TLC.Yield = 6.85 g, 80.1%; MS m/e 238 (M + 1);1H-NMR(CDCl3) δ 1.73–2.32 (4H, m), 3.57–4.52 (9H, m), 6.81(1H, s), 6.91 (4H, s).

6.1.20. 1-4≠-(2≠≠-Tetrahydrofurylmethoxy)ethoxyphe-noxy)-3-isopropylamino-2-propanol(20)

Prepared fromN according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 3.59 g, 48.4%; m.p. = 165.5–167.0 °C; MSm/e 336 (M + 1);1H-NMR (CD3OD) δ 0.93 (6H, d),1.73–2.32 (4H, m), 2.38–3.23 (6H, bm), 3.57–4.52 (9H,m), 6.93 (4H, s). Anal. (C19H31NO5.HCl) C, H, Cl, N.

6.1.21. 4-n-Pentyloxyethoxyphenol(O)n-Pentyloxyethyl bromide (10.72 g, 55 mmol),

4-benzyloxyphenol (10.11 g, 50 mmol), anhydrousK2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirredand refluxed in anhydrous acetone (250 mL) for 48 h. Thesolvent was evaporated and the residue was partitionedbetween dichloromethane and water. The organic layerwas washed with water and dried over anhydrousMgSO4. The solvent was evaporated and the residual4-(4≠-n-pentyloxyethoxy)phenylbenzyl etherL (R≠≠ =n-pentyl) was purified by chromatography (silica, eluentCH2Cl2) and isolated as an oil, homogenous by TLC.Yield = 13.7 g, 79.4%; MS m/e 315 (M + 1);1H-NMR(CDCl3) δ 1.03 (3H, t), 1.19–2.02 (8H, m), 3.53–4.22(4H, m), 5.16 (2H, s), 6.87 (4H, s), 7.24 (5H, s).

934

The etherL (R≠≠ = n-pentyl) (11.3 g, 36 mmol) inEtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g)at room temperature and atmospheric pressure. Whenhydrogen absorption ceased, the reaction mixture wasfiltered and evaporated to dryness. The productO waspurified by chromatography (silica, eluent CH2Cl2) andisolated as an oil, homogenous by TLC. Yield = 7.52 g,93.25%; MS m/e 225 (M + 1);1H-NMR (CDCl3) δ 1.07(3H, t), 1.22–2.07 (8H, m), 3.57–4.29 (4H, m), 6.87 (4H,s).

6.1.22. 1-(4-n-Pentyloxyethoxyphenoxy)-3-isopropyl-amino-2-propanol(22)

Prepared fromO according to the general procedures 1and 2, chromatographed and isolated as the hydrochlo-ride. Yield = 3.90 g, 52.1%; m.p. = 77.0–79.0 °C; MS m/e340 (M + 1);1H-NMR (CD3OD) δ 1.13 (3H, t), 1.75 (6H,d), 1.36–2.10 (8H, bm), 3.26–4.04 (6H, m), 3.86–4.28(4H, dt), 7.03 (4H, s). Anal. (C19H33NO4.HCl) C, H, Cl,N.

6.2. Pharmacological methods

6.2.1. Isolated tissue preparationsStudies were carried out on Sprague-Dawley rat iso-

lated atria and tracheal rings according to our methoddescribed previously [16]. All tissues were allowed toequilibrate for 45 min with Krebs Ringer physiologicalsalt solution; the composition of which in mmol L–1 wasNaCl, 120; KCl, 5.6; MgSO4, 1.2; CaCl2, 2.5; KH2PO4,1.4; NaHCO3, 25; glucose 11.2 and EGTA, 0.0025.

Cumulative concentration-response curves were ob-tained for the non-selectiveâ-AR agonist (-)-isoprenalinein each preparation [16]. (-)-Isoprenaline was dissolvedand diluted in 1 mg mL–1 ascorbic acid to preventoxidation. For the measurement of antagonist activity theappropriate agent was added to the organ bath at least30 min after the first control concentration-responsecurve was completed and allowed to equilibrate for10 min before the next concentration-response curveestablished. The shift in this curve to the right wascalculated as a pA2 value [39]. At least three concentra-tions of each antagonist were examined to verify the pA2.

6.2.1.1. Rat isolated spontaneously beating atriaRat hearts were removed from adult animals

(200–250 g) and placed in Krebs Ringer salt solution (pH7.4) aerated with 5% CO2 in O2. The atria were dissectedfree of the ventricles and overlying tissue and placed in a20 mL bath maintained at 37 °C and connected to anisotonic transducer. A tension of 1 g was applied andchronotropic activity was amplified and recorded on aGrass Polygraph.

6.2.1.2. Rat isolated tracheal chainsTrachea were excised from adult rats (200–250 g),

dissected free of overlying tissue and cut transversely intosegments about 2 mm wide. Five segments were mountedin a 20 mL bath maintained at 37 °C at a tension of 1 g.Relaxation of the segments by (-)-isoprenaline was re-corded by an isotonic transducer connected to a GrassPolygraph after tone had been established by administra-tion of 1 µM carbachol (45 min prior to concentration-response curves).

6.3. Computational methods

6.3.1. Molecular modellingModelling studies were performed using the SYBYL

software package (Tripos Inc., version 6.40) on a SiliconGraphics Indigo 2 UNIX Workstation. The structure ofthe active S-isomer of compound1 was constructed fromstandard bond lengths and bond angles using the sketchroutine in SYBYL. The geometry of compound1 wasfully optimised in vacuo as previous studies [19] haveidentified that the propanolamine side-chain of aryloxy-propanolamine (AOPA) type compounds can adopt com-mon conformations in solid, theoretical gas and solutionstates. Optimisation was conducted using the AM1Hamiltonian in MOPAC (version 6.0) [48], the keywords‘PRECISE’, ‘ESP’ and ‘NOMM’ were specified. Thekeyword ESP calculates the electrostatic potential derivedatomic charges, these charges have been used since theyare reported to be reliable and highly correlated with abinitio ESP charges [49, 50].

6.3.2. Molecular alignmentCompounds2–25 were constructed using the mini-

mised structure of compound1 as a starting point. Theappropriatepara-substituent was added using the addmultiple atom function in SYBYL. The structure wasthen optimised using the AM1 Hamiltonian in MOPAC,with the keywords ‘PRECISE’, ‘ESP’ and ‘NOMM’specified, however,τ1 was fixed at 0° or 90° (figure 1)and the bond, dihedral and valence angles of the AOPAcore structure were fixed at the optimised values forcompound1. Therefore, two series of compounds werecreated, one whereτ1 = 0° and the other whereτ1 = 90°.After optimisation the compounds were superimposedover the entireN-isopropylphenoxypropanolamine struc-ture (compound1) using the linear least squares routinewithin SYBYL, as shown infigure 7.

In addition to the optimisation and superimposition ofeach compound a number of other physical parameterswere examined:

935

6.3.3. Phenyl ring charge calculationsFor each compound intable II (1–25), the AM1 ESP

charges calculated for the six phenyl ring carbon atomswere added together to obtain the parameter ‘phenyl ringcharge’. Various oxypropanolamine andpara-substituentconformations (ie. differentτ1, τ2 values,figure 1) wereexamined for their effect upon phenyl ring charge. Bothτ1 andτ2 were varied independently in 90° steps througha full 360°, giving a total of sixteen conformers examinedfor each compound, and the phenyl ring charge wascalculated for each conformer. The phenyl ring charge didnot vary greatly between conformers and therefore thevalue obtained whenτ1 = 90° andτ2 = 0° was used in thisstudy. All sixteen phenyl ring charge values calculated fora particular compound were within± 0.04e of the valueobtained whenτ1 = 90° andτ2 = 0°.

6.3.4. LogP calculationsThe logP values for the twenty five compounds were

calculated using the Pallas PrologP (version 1.1) programat Swinburne University of Technology, Hawthorn, Vic-toria, Australia. The π values [51] for the para-substituents were determined by subtracting the logP ofcompound1 from the compound in question and aregiven in table II.

6.3.5. Length of the para-substituentThe length was defined as the distance between the

para-carbon of the phenyl ring and the most distantnon-H atom in thepara-substituent, hence thepara-substituent length of compound1 is zero (table II). Thechain length of thepara-substituent (R-group,table II)was calculated for the relaxed, fully extended low energyconformation of each molecule.

6.3.6. Flexibility of the para-substituentA measure of the flexibility was incorporated into the

QSAR analysis by assigning an integer for the number oftorsion angles which affect the conformation of thepara-substituent and is given intable II. For example, theflexibility of compound4 is assigned the integer 1, asrotation of terminal methyl, hydroxyl and amine groupsdo not affect the conformation of thepara-substituent.

6.3.7. Development of the CoMFA modelCoMFA [41] was performed using the QSAR option in

SYBYL. Both steric and electrostatic fields were consid-ered. The probe atom had the properties of an sp3 carbonatom and a charge of +1.0. Cut-off values were SYBYLdefault values and the grid step size was 2.0 Å. For eachgroup of compounds (ie.τ1 = 0° or 90°) the effects ofchanging column filtering values (≈ 1.0–16.0 kcal/mol)were examined. The CoMFA QSAR equations were

generated using the PLS algorithm using the ‘leave-one-out’ cross validation procedure, and the number ofcomponents with the lowest standard error of predictionvalue was selected as the ONC. The additional physicalparameters were also included in the cross validated PLSanalysis to determine whether they improved q2. The finalmodels, non cross-validated, for theâ1- andâ2-ARs weregenerated using CoMFA, the combination of additionalparameters that maximised q2 and the ONC as deter-mined above.

Acknowledgements

The authors would like to thank Ms. Leanne Styan forher technical assistance and Dr Margaret Wong for herhelpful comments. This work was supported by a grantfrom the National Health and Medical Research Councilof Australia.

References

[1] Greenidge P.A., Pharm. Acta Helv. 68 (1994) 201–210.

[2] Lafontan M., Belan M., J. Lipid Res. 34 (1993) 1057–1091.

[3] Ostrowski J., Kjelsberg M.A., Caron M.G., Lefkowitz R.J., Ann.Rev. Pharmacol. Toxicol. 32 (1992) 167–183.

[4] Strosberg A.D., Prot. Sci. 2 (1993b) 1198–1209.

[5] Emorine L.J., Marullo S., Briend-Sutren M.M., Patey G., Tate K.,Delavier-Klutchko C., Strosberg A.D., Science 245 (1989)1118–1121.

[6] Frielle T.J., Collins S., Daniel K.W., Caron M.G., Lefkowitz R.J.,Kobilka B.K., Proc. Natl. Acad. Sci. USA 84 (1987) 7920–7924.

[7] Kobilka B.K., Dixon R.A.F., Frielle T., Dohlman H.G., BolanowskiM.A., Sigal I.S. et al., Proc. Natl. Acad. Sci. USA 84 (1987) 46–50.

[8] Donne-Op Den Kelder G.M., Bultsma T., Timmerman H., Rade-maker B., J. Med. Chem. 31 (1988) 1069–1079.

[9] Linschoten M.R., Bultsma T., Ijzerman A.P., Timmerman H., J. Med.Chem. 29 (1986) 278–286.

[10] Ball J.B., Nero T.L., Iakovidis D., Tung L., Jackman G., Louis W.J.,J. Med. Chem. 35 (1992) 4676–4682.

[11] Baldwin J.J., Christy M.E., Denny G.H., Habecker C.N., FreedmanM.B., Lyle P.A. et al., J. Med. Chem. 29 (1986) 1065–1080.

[12] Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T.,J. Med. Chem. 27 (1984) 503–509.

[13] Berthold R., Louis W.J., US Patent, US 4, 425, 362, (1984).

[14] Hoefle M.L., Hastings S.G., Meyer R.F., Corey R.M., Holmes A.,Stratton C.D., J. Med. Chem. 18 (1975) 148–152.

[15] Large M.S., Smith L.H., J. Med. Chem. 25 (1982) 1286–1292.

[16] Louis S.N.S., Nero T.L., Iakovidis D., Jackman G.P., Louis W.J.,Eur. J. Pharmacol. 367 (1999) 431–435.

[17] Beckett A.H., Fortschr. Arzneim. Forsch. 1 (1959) 455–530.

[18] Easson L.H., Stedman E., Biochem. J. 27 (1933) 1257–1266.

[19] Nero T.L., Louis W.J., Iakovidis D., J. Mol. Struct. (Theochem.) 285(1993) 251–272.

[20] Brandstrom A.E., Carlsson P.A.E., Carlsson S.A., Corrodi H.R., EkL., Ablad B.A.H., US Patent, US 3928601, (1975).

936

[21] Carlsson P.A.E., Brandstrom A.E., Lamm B.R., Ablad B.A.H.,Carlsson S.A.I., Corrodi H.R., Ek L., DE Patent, DE 2020864,(1970).

[22] Crowther A.F., Gilman D.J., McLoughlin B.J., Smith L.H., TurnerR.W., Wood T.M., J. Med. Chem. 12 (1969) 638–642.

[23] Erez M., Shtacher G., Weinstock M., J. Med. Chem. 21 (1978)982–984.

[24] Ferrari G., Ferrini R., Casagrande C., Boll. Chim. Farm. 107 (1968)234–248.

[25] Jain P.C., Belleau B., Rastogi S.N., Anand N., Tripathi O.N., SrimalR.C., Indian J. Chem. 12 (1974) 1021–1027.

[26] Leigh T., GB Patent 1185046 (1970).

[27] Loevgren K., Hedberg A., Nilsson J., Lars G., J. Med. Chem. 24(1981) 451–454.

[28] Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T.,Fryer A.D., Melarange R.A., Shivdasani C., J. Med. Chem. 26(1983) 1570–1576.

[29] Malguzzi R., Giordano C., EP Patent, EP 50885 (1982).

[30] Manoury P.M., Binet J.L., Rousseau J., Lefevre-Borg F.M., CaveroI.G., J. Med. Chem. 30 (1987) 1003–1011.

[31] Rosell G., Mauleon D., Pujol M.D., Granados R., Arch. Farmacol.Toxicol. (1986) 65–70.

[32] Wick A.E., Binet J., US Patent, US 4515814 (1985).

[33] Zaagsma J., Nauta W.T., J. Med. Chem. 17 (1974) 507–513.

[34] Vogel A.I., Practical Organic Chemistry 3rd ed. Longmann, London,1962 pp. 676–962.

[35] Isagulyants V.I., Tisukova V.N., Favorskaya N.A., Oganesyan P.O.,Chem. Abstracts 62 (1962) 14545a.

[36] Van Rossum J.M., Hurkmans J.A.T.M., Wolters C.J.J., Arch. Int.Pharmacodyn. Ther. 143 (1963) 299–330.

[37] Zaborowsky B.R., McMahan W.C., Griffin W.A., Norris F.H.,Ruffolo Jr. R.R., J. Pharmacol. Methods 4 (1980) 165–178.

[38] MicroCal Origin, MicroCal Software Inc., USA.

[39] Mackay D.J., Pharm. Pharmacol. 30 (1978) 312–313.

[40] Baldwin J.J., Denny G.H., Hirschmann R., Freedman M.B., Ponti-cello G.S., Gross D.M., Sweet C.S., J. Med. Chem. 26 (1983)950–957.

[41] Cramer R.D., Paterson D.E., Bunce J.D., J. Am. Chem. Soc. 110(1988) 5959–5967.

[42] Triggle D.J., in: Triggle D.J., Triggle C.R. (Eds.), Chemical phar-macology of the synapse, Academic Press, London, 1976 p. 233.

[43] Belleau B., Ann. NY Acad. Sci. 139 (1967) 580–605.

[44] Dixon R.A.F., Sigal I.S., Strader C.D., Cold Spring Harbor Sympo-sium Quant. Biology. 3 (1988) 487–497.

[45] Lewell X.Q., Drug Design Discovery 9 (1992) 29–48.

[46] Strader C.D., Candalore M.R., Hill W.S., Dixon R.A.F., Sigal I.S., J.Biol. Chem. 264 (1989) 13572–13578.

[47] Strader C.D., Sigal I.S., Register R.B., Candelore M.R., Rands E.,Dixon R.A.F., Proc. Natl. Acad. Sci. USA 84 (1987) 4384–4388.

[48] Dewar M.J.S., Zoebisch E.G., Healy E.F., Stewart J.J.P., J. Am.Chem. Soc. 107 (1985) 3902–3909.

[49] Besler B.H., Merz Jr. K.M., Kollman P.A., J. Computat. Chem. 11(1990) 431–439.

[50] Ferenczy G.G., Reynolds C.A., Richards W.G., J. Computat. Chem.11 (1990) 159–169.

[51] Hansch C., Leo A., Unger S.H., Kim K.H., Nikaitani D., Lien E.J.,J. Med. Chem. 16 (1973) 1207–1216.

937

Original article

Protease inhibitors – Part 3. Synthesis of non-basic thrombin inhibitorsincorporating pyridinium-sulfanilylguanidine moieties at the P1 site

Andrea Scozzafava, Fabrizio Briganti, Claudiu T. Supuran*

Università degli Studi, Laboratorio di Chimica Inorganica e Bioinorganica, Via Gino Capponi 7, I-50121 Florence, Italy

(Received 15 March 1999; revised 17 May 1999; accepted 20 May 1999)

Abstract – Using benzamidine and sulfaguanidine as lead molecules, three series of derivatives have been prepared by reaction ofsulfaguanidine with pyrylium salts, with the pyridinium derivatives of glycine and with the pyridinium derivatives ofâ-alanine, respectively.The new compounds were assayed as inhibitors of two serine proteases, thrombin and trypsin. The study showed that in contrast to the leads,possessing KI’s around 100–300 nM against thrombin, and 1 200–1 500 nM against trypsin, respectively, the new derivatives showedinhibition constants in the range of 15–50 nM against thrombin, whereas their affinity for trypsin remained relatively low. Derivatives ofâ-alanine were more active than the corresponding Gly derivatives, which in turn were more inhibitory than the pyridinium derivatives ofsulfaguanidine possessing the same substitution pattern at the pyridinium ring. Thus, the present study proposes two novel approaches for thepreparation of high affinity, specific thrombin inhibitors: a novel S1 anchoring moiety in the already large family of arginine/amidine-basedinhibitors, i.e., the SO2N=C(NH2)2 group, and novel non-peptidomimetic scaffolds obtained by incorporating alkyl-/aryl-substituted-pyridinium moieties in the hydrophobic binding site(s). The first one is important for obtaining bioavailable thrombin inhibitors, devoid ofthe high basicity of the commonly used arginine/amidine-based inhibitors, whereas the second one may lead to improved water solubility ofsuch compounds due to facilitated salt formation as well as increased stability at hydrolysis (in vivo). © 1999 Éditions scientifiques etmédicales Elsevier SAS

thrombin / trypsin / sulfaguanidine / pyridinium salts / pyridinium amino acid / non-basic thrombin inhibitor

1. Introduction

Thrombin (EC 3.4.21.5) has become an importanttarget for drug design in recent years, in the search forlow molecular-weight potent and selective inhibitors withapplications as diagnostic and therapeutic agents for theincreasingly common thrombotic diseases [1–8]. Al-though a large number of potent active site-directedthrombin inhibitors, such as peptide aldehydes [9, 10],boronates [11], benzamidine- [2, 3, 12, 13] or arginine/guanidine-derived [14] inhibitors are reported, none ofthem meet all the criteria needed for an ideal antithrom-botic drug [2, 15]. Thus, the largest majority of thepresently available low-molecular weight inhibitors, suchas argatroban (MQPA)1 [16], inogatran2 [8], NAPAP3 [17], 4-TAPAP 4 or its 3-amidino-isomer, 3-TAPAP5 [2, 17] (figure 1), are poorly bioavailable, either due totheir high basicity, connected with the presence of

guanidino-/amidino moieties in their molecule, or are notabsorbable orally, or are rapidly eliminated from thecirculation, mainly due to their peptidic nature. Althoughrecently some non-basic S1 anchoring groups have beenincorporated in the molecules of some thrombin inhibi-tors [3, 7, 18], the presence of guanidino-/benzamidinomoieties in such compounds is critical, since it is bymeans of the interaction of these highly polar groups withAsp 189, the central amino acid residue from the speci-ficity pocket, that the enzyme-inhibitor adduct is initiallyformed (obviously, a lot of other secondary interactionsare responsible for the formation of high affinity adductsbetween thrombin and its inhibitors) [3–5, 12–14]. Inorder to exploit the intrinsically high affinity ofguanidino-/benzamidino-containing inhibitors for thethrombin active site, but also to avoid undesired proper-ties connected with their too high basicity, we proposehere a novel approach for designing tight-binding suchinhibitors, by using sulfanilylguanidino moieties as an-*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 939−952 939© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

choring groups to the S1 specificity pocket. Obviously,the presence of the SO2 group in the neighbourhood ofthe guanidino moiety strongly reduces the basicity of thelatter, presumably without precluding to the binding ofinhibitors within the enzyme active site.

In this paper we report the preparation and serineprotease inhibitory properties (against human thrombinand human trypsin) of three series of compounds ob-tained by reaction of sulfaguanidine with pyrylium salts,with the pyridinium derivatives of glycine (prepared fromGly and pyrylium salts) and with the pyridinium deriva-tives of â-alanine (obtained fromâ-Ala and pyryliumsalts), respectively. From the point of view of theirthrombin inhibitory properties, as well as that of theirspecificity for thrombin over trypsin, some of our com-pounds showed inhibition constants of the same order ofmagnitude as those of the clinically used compounds,argatroban (MQPA)1 [16], and inogatran2 [8], in the15–50 nM range against thrombin, whereas maintaining amuch lower trypsin affinity (inhibition constants around

1 200–1 500 nM) as compared to the above-mentionedclinically used derivatives.

2. Chemistry

Compounds prepared by reaction of di-, tri- or tetra-substituted pyrylium salts with sulfaguanidine, of typesA1–A16, as well as the corresponding Gly derivatives oftypesB1–B16 andâ-Ala derivativesC1–C16are shownin table I.

Non-exceptional synthetic procedures have been usedfor the reactions of pyrylium salts with nucleophiles (forthe preparation of compoundsA, B, C (1–16) as well asthe pyridinium amino acid intermediates10 and11) [19,20], whereas for attaching the pyridinium-amino acylmoieties, the condensation reactions in the presence ofcarbodiimide derivatives has been used, as outlined infigure 2[21, 22].

Sulfanilylguanidine7 was reacted with di-, tri- ortetrasubstituted pyrylium salts6 leading to the pyridinium

Figure 1. Structures of serine protease inhibitors1–4.

940

Table I. Pyridinium-sulfanilylguanidines (A1–A16), pyridinium-methylcarboxamido- (B1–B16) and pyridinium-ethylcarboxamido-sulfanilylguanidines (C1–C16) prepared in the present study, with their inhibition data against human thrombin and human trypsin.

Compound n R1 R2 R3 R4 KIa

Thrombin Trypsin(nM)

A1 – Me H Me Me 83± 5 1 290± 80A2 – i-Pr H Me Me 76± 6 1 180± 90A3 – i-Pr H Me i-Pr 102± 5 1 440± 105A4 – Me H Ph Me 41± 2 1 120± 75A5 – Et H Ph Et 37± 3 1 100± 62A6 – n-Pr H Ph n-Pr 54± 7 1 165± 63A7 – i-Pr H Ph i-Pr 48± 5 1 200± 85A8 – Me H Ph Ph 32± 2 1 265± 102A9 – Et H Ph Ph 30± 4 1 240± 77A10 – n-Pr H Ph Ph 36± 5 1 260± 80A11 – i-Pr H Ph Ph 34± 2 1 210± 104A12 – n-Bu H Ph Ph 60± 5 1 340± 120A13 – t-Bu H Ph Ph 33± 3 1 170± 96A14 – Ph H Ph Ph 54± 4 1 950± 130A15 – Ph H H Ph 58± 6 1 950± 140A16 – Me Me Me Me 79± 6 1 300± 90B1 1 Me H Me Me 75± 5 1 210± 91B2 1 i-Pr H Me Me 62± 4 1 120± 60B3 1 i-Pr H Me i-Pr 80± 7 1 320± 95B4 1 Me H Ph Me 34± 3 1 100± 72B5 1 Et H Ph Et 30± 3 1 020± 60B6 1 n-Pr H Ph n-Pr 50± 4 1 150± 45B7 1 i-Pr H Ph i-Pr 43± 5 1 175± 75B8 1 Me H Ph Ph 21± 2 1 250± 88B9 1 Et H Ph Ph 17± 1 1 200± 105B10 1 n-Pr H Ph Ph 23± 2 1 210± 65B11 1 i-Pr H Ph Ph 22± 2 1 175± 90B12 1 n-Bu H Ph Ph 50± 5 1 210± 60B13 1 t-Bu H Ph Ph 25± 3 1 100± 55B14 1 Ph H Ph Ph 50± 4 1 900± 60B15 1 Ph H H Ph 57± 5 1 905± 65B16 1 Me Me Me Me 73± 6 1 265± 90C1 2 Me H Me Me 69± 4 1 140± 87C2 2 i-Pr H Me Me 47± 4 1 100± 100C3 2 i-Pr H Me i-Pr 71± 5 1 290± 97C4 2 Me H Ph Me 29± 2 1 055± 73C5 2 Et H Ph Et 26± 2 1 010± 55C6 2 n-Pr H Ph n-Pr 47± 5 1 105± 102C7 2 i-Pr H Ph i-Pr 41± 4 1 100± 79C8 2 Me H Ph Ph 18± 2 1 210± 75C9 2 Et H Ph Ph 15± 1 1 165± 50C10 2 n-Pr H Ph Ph 21± 2 1 200± 60C11 2 i-Pr H Ph Ph 20± 1 1 155± 70C12 2 n-Bu H Ph Ph 47± 5 1 200± 85C13 2 t-Bu H Ph Ph 19± 2 1 130± 50C14 2 Ph H Ph Ph 42± 5 1 380± 45C15 2 Ph H H Ph 50± 3 1 520± 50C16 2 Me Me Me Me 70± 5 1 210± 40

aKI values were obtained from Dixon plots using a linear regression program [26], from at least three different assays.

941

derivativesA1–A16. Alternatively, reaction of pyryliumsalts with Gly orâ-Ala afforded the pyridinium aminoacid derivatives10 and11, which were coupled with7 inthe presence of EDCI or diisopropylcarbodiimide ascondensing agents, leading to compoundsB1–B16, andC1–C16, respectively.

3. Pharmacology

Inhibition data against two serine proteases, humanthrombin and human trypsin are shown intable I. Thechromogenic substrate Chromozym TH (Ts-Gly-Pro-Arg-p-nitroanilide) was used in the assay, with the spectro-photometric method of Lottenberg et al. [23]. Inhibitiondata with the standard serine protease inhibitors1–3 arealso provided for comparison intable II.

pKa values for the guanidino/amidino and sulfonamidomoieties for some of the thrombin inhibitors reportedhere, as well as standard compounds, are shown intable III.

4. Results and discussion

The lead molecule for obtaining novel types of throm-bin inhibitors considered by us was benzamidine12, oneof the simplest of such compounds, which possesses aninhibition constant KI = 300 nM against human thrombin;moreover, the X-ray crystallographic structure for the

complex of benzamidine with this enzyme has recentlybeen reported (PDB entry: 1DWB) [24]. From the X-raydata it was observed that the amidino moiety of theinhibitor is anchored to the S1 specificity pocket of theenzyme, interacting electrostatically and by means ofhydrogen bonds with Asp 189. Several other van derWaals contacts between the inhibitor molecule and theenzyme were also evidenced [24]. Obviously, benzami-dine is a weak thrombin inhibitor, since the bindingenergy is only gained due to the strong electrostaticinteraction of the carboxylate of Asp 189 and the posi-tively charged amidino moiety. On the other hand, asalready mentioned in the introductory section, the ami-dino moiety possesses too high a basicity for allowing theformation of bioavailable enzyme inhibitors, and it thus

Figure 2. Synthesis of derivativesA1–A16, B1–B16 andC1–C16.

Table II. Inhibition data of two serine proteases with standardinhibitors 1–3 and12, and sulfaguanidine7.

Compound KI (nM)a

Thrombin Trypsin1 Argatrobanb 19 ± 2 –2 Inogatran 15± 1 540± 113 NAPAP 6.5± 0.05 690± 247 Sulfaguanidine 95± 4 1 350± 12012 Benzamidine 300± 5 450± 6

aKI values were obtained from Dixon plots using a linear regres-sion program [26], from at least three different assays.bFromref. [5].

942

appeared of great interest to elaborate on non-basicvariants of this interesting serine protease anchoringgroup. The sulfonylguanidino moiety appeared as anattractive candidate for such a purpose, since the presenceof the SO2 moiety in the neighbourhood of the strongbase, guanidine, should drastically weaken its basicity.Such modified anchoring groups should not presumablyinterfere with the binding of the inhibitor to the enzyme,since the hydrogen-bonding donor/acceptor properties, aswell as the possibility to interact electrostatically with theenzyme for the compounds incorporating them, shouldnot differ too much from those of the classical amidino-/guanidino-based inhibitors of types1–5 or 12. Thesulfonyl-guanidines possess a large number of possibletautomeric forms, and this factor might also be a criticalone for the binding of such a compound to thrombin.Thus, in previous work [25] we have shown that arylsul-fonyl guanidines, including sulfaguanidine7, possessmoderate but specific thrombin inhibitory properties.Moreover, by means of AM1 and MOPAC calculations itwas demonstrated that the tautomer of type13A ofbenzenesulfonylguanidine is much more stable than thetautomer 13B (figure 3), a situation that seems to beimportant for binding to the enzyme [25]. Thus, wepresume that the same is true for the pyridinium-basedcompounds reported here, i.e., that the symmetricaltautomers of type13A are more stable than the corre-sponding non-symmetrical tautomers of type13B. It isobvious from the above data that the symmetrical natureof the favoured tautomer should enable stronger interac-tions with the carboxylate moiety of Asp 189 andpresumably, the formation of high affinity E-I adducts.

Thus, three series of pyridinium containing sulfanilyl-guanidinesA1–A16, B1–B16 and C1–C16 were pre-pared in order to test the above-mentioned hypothesis(table I). These compounds were obtained by reactions of

pyrylium salts with sulfaguanidine, or alternatively, bycondensation of sulfaguanidine with the pyridinium de-rivatives of glycine orâ-alanine, obtained from the twomentioned amino acids and pyrylium salts, by the origi-nal procedure of Balaban’s group [27–32].

The following should be noted regarding the serineprotease inhibition data oftables I and II with the newcompounds and standard inhibitors: (i) the pyridiniumderivatives A, B, C (1–16) reported here generallybehave as stronger thrombin inhibitors as compared to thelead molecule from which they were derived, i.e., benza-midine 12 and sulfaguanidine7. At the same time, theiraffinity for trypsin is relatively low, which constitutes apositive feature for the putative clinical use of suchcompounds; (ii) in the three subseries of investigatedcompounds, thrombin inhibitory properties increasedfrom the pyridinium derivatives of sulfaguanidineA(1–16) to the corresponding pyridinium-Gly-derivativesB (1–16), with the pyridinium-â-Ala derivativesC(1-16)behaving as the most active inhibitors in the whole seriesof reported compounds (obviously, this discussion takesinto account the same substitution pattern at the pyri-dinium ring for compounds in the three investigatedsubseries); (iii) the nature of R1–R4 groups substitutingthe pyridinium ring was critical for the biological activityof the obtained compounds, similarly to the situationevidenced for the carbonic anhydrase sulfonamide inhibi-tors reported previously [19, 20]. Thus, tri- or tetraalkyl-pyridinium- as well as 2,6-di- or 2,4,6-triphenyl-pyridinium moieties were generally less effective than2-alkyl-4,6-diphenyl-pyridinium groups in inducingstrong thrombin inhibitory properties to the compoundsincorporating them. Practically, the most active deriva-tives in all three subseries were those containing 2-alkyl-4,6-diphenyl-pyridinium moieties, such as 2-methyl-;2-ethyl-; 2-iso-propyl- or 2-tert-butyl-4,6-diphenyl-pyridinium groups. Replacing the 2-alkyl group men-tioned above with a bulky phenyl one (such as incompoundsA14, B14 or C14) or with a longer aliphatic

Table III. pKa data of serine protease inhibitors1–3, 7, A9, B9andC9.

Compound pKaa

Guanidino/amidino SO2NHmoiety moiety

1 Argatrobanb 12.5 –2 Inogatranb 12.3 –3 NAPAP 12.6 –7 Sulfaguanidine 8.4 7.0A9 8.1 7.0B9 8.3 7.1C9 8.5 7.1

apKa values were determined in 30% Et-OH/water (v/v) as descri-bed in the Experimental section.bFrom ref. [8].

Figure 3. Tautomeric forms of compound13.

943

chain (n-butyl, such as inA12, B12 or C12) led to adrastic reduction of the thrombin inhibitory effects of thecorresponding compounds. On the other hand, com-pounds possessing 2,6-dialkyl-4-phenyl-pyridinium moi-eties in their molecule (such asA, B, C (4 and 5))possessed a behaviour intermediate between the stronginhibitors of the typeA, B, C (8, 9, 11 and13) and therelatively weak inhibitors of typeA, B, C (1–3 and14–16). Anyhow, the best substitution for inducing strongthrombin inhibitory properties was that incorporating the2-ethyl-4,6-diphenylpyridinium moiety in the moleculesof the new derivatives. Some of the compounds contain-ing this substitution pattern, such asB9 andC9 (but alsothe structurally-related compoundsB8, B10, B11, C8,C10 andC11) showed thrombin inhibitory properties ofthe same order of magnitude as the clinically usedderivatives argatroban1 and inogatran2, although theyare less effective as compared to the very potent inhibitorNAPAP (table II). A special mention should be maderegarding the fact that the new compounds reported herepossess a much lower affinity for trypsin as compared tothe standard inhibitors1–3, which constitutes a highlydesirable feature for a compound to be developed forclinical use.

pKa values for the amidino/guanidino as well assulfonamido moieties of some of the newly synthesizedserine protease inhibitors and standard compounds suchas inogatran, argatroban and NAPAP (table III) prove thatthe approach proposed here for reducing the basicity ofsuch an enzyme inhibitor is a successful one. Thus, unlikethe highly basic guanidines/amidines of type1–3 (pKa’saround 12.3–12.6), sulfaguanidine7 and its derivativesreported here (such as compoundsA9, B9 or C9) havepKa values of the guanidino moiety around 8.1–8.5, beingat least 104 times less basic than the previously men-tioned derivatives. Furthermore, due to the presence ofthe sulfonyl moiety in their molecules, these compoundsalso possess a weakly acidic character, with anotherionization step around the pKa value of 7, due to the lossof the SO2NH proton. These features should positivelyinfluence the pharmacological profile of a thrombininhibitor of the type described here.

The strong thrombin inhibitory properties of somecompounds reported in this study might be explained bytaking into account the X-ray crystallographic structureof the enzyme as well as those of some of its adducts withguanidine/amidine-based inhibitors [4, 5, 33, 34]. Thus, itwas shown that effective binding is achieved when aproline, a pipecolic acid or a similarly non-hydrophilicmoiety is present in the P2 position, which allowsfavourable interactions with the enzyme S2 cavity, com-prising among others, amino acid residues Trp 60D and

Tyr 60A as well as when hydrophobic (generally aro-matic: Ph, Ts; naphthyl) groups are present at P3, whichallow strong interactions with the S3 site, comprisingresidues Leu 99; Trp 215 and Ile 174 among others [4, 5,33, 34]. Some moieties present in the compounds pre-pared by us might possess just the required structuralelements for the formation of high affinity adducts withthrombin. For example, for the strongest inhibitor re-ported in this paper,C9 (KI = 15 nM against thrombin),the CH2CH2CO moiety might interact with the S2 cavity,whereas the two phenyls substituting the pyridiniummoiety probably bind within the aryl binding site (S3).Obviously, the sulfonylguanidino moiety of all theseinhibitors probably fills the S1 specificity pocket, inter-acting with Asp 189, as discussed earlier. But anotheraspect might be important for explaining the relativelyhigh affinity of this entire class of inhibitors for thrombin.Thus, around the entrance of the specificity pocket of thisenzyme, ten negatively-charged amino acid residues areclustered [4, 5, 33, 34]. Some of these (such as Asp 189and Glu 192) are directly involved in the substrate/inhibitor recognition process, whereas some others mightbe crucial for driving or stabilizing the inhibitor withinthe active site [4]. In the case of cationic inhibitors, suchas the compounds reported in the present paper, thepresence of such a cluster of ten negatively-chargedresidues at the entrance of the active site might beextremely favourable for obtaining strong E-I adducts,due to the possibility of salt bridge formation between thecationic moiety of the inhibitor and the anionic groups ofthe enzyme. Associated with precise geometric require-ments for binding to the S2 and S3 sites (mentionedabove) our approach led to high affinity, less basicthrombin inhibitors.

5. Conclusion

Three series of cationic sulfaguanidine derivativeshave been prepared by reaction of sulfaguanidine withdi-, tri- or tetrasubstituted pyrylium salts bearing alkyl,aryl or a combination of the two moieties in theirmolecule, and with the corresponding Gly-pyridiniumand â-Ala pyridinium derivatives, respectively. Qualita-tive SAR proved that the best activity for inhibitingthrombin was obtained for compounds bearing 2-alkyl-4,6-diaryl- pyridinium moieties, and that theâ-Ala de-rivatives were more active than the corresponding Glyderivatives, which in turn were more active than thecorresponding pyridinium-sulfaguanidines. The obtainedcompounds generally possessed a low affinity for trypsin,which might be considered a positive feature for theputative pharmacological development of such a throm-

944

bin inhibitor. Thus, our study proposes two novel ap-proaches for the preparation of high affinity, specificthrombin inhibitors: 1) a novel S1 anchoring moiety ofthe arginine/amidine type, i.e., the SO2N=C(NH2)2group; and 2) novel non-peptidomimetic scaffolds ob-tained by incorporating alkyl-/aryl-substituted-pyridi-nium moieties in the hydrophobic binding site(s). Thefirst approach is important for obtaining bioavailablethrombin inhibitors devoid of the high basicity of thecommonly used arginine/amidine-based inhibitors, withsome of the new derivatives proving to be 104 times lessbasic than the standard compounds in clinical use. Thesecond one may lead to improved water solubility of suchderivatives due to facilitated salt formation as well asincreased in vivo stability at hydrolysis.

6. Experimental

6.1. Chemistry

Melting points: heating plate microscope (not cor-rected); IR spectra: KBr pellets, 400–4 000 cm–1 Perkin-Elmer 16PC FTIR spectrometer;1H-NMR spectra:Varian 300CXP apparatus (chemical shifts are expressedas δ values relative to Me4Si as standard); Elementalanalysis (± 0.4% of the theoretical values, calculated forthe proposed formulas, data not shown): Carlo ErbaInstrument CHNS Elemental Analyzer, Model 1106. Allreactions were monitored by thin-layer chromatography(TLC) using 0.25 mm precoated silica gel plates (E.Merck). Preparative HPLC was performed on aDynamax-60A column (25× 250 mm), with a BeckmanEM-1760 instrument. The detection wavelength was254 nm. Sulfaguanidine, triethylamine, carbodiimides,and amino acids used in the syntheses were commerciallyavailable compounds (from Sigma, Acros or Aldrich).Pyrylium salts were prepared as described in the litera-ture [30–32]. Acetonitrile, acetone, dioxane, ethyl acetate(E. Merck, Darmstadt, Germany) or other solvents usedin the synthesis were doubly distilled and kept onmolecular sieves in order to maintain them in anhydrousconditions. Inogatran was from Astra Hassle (Molndal,Sweden). Benzamidine, NAPAP, human thrombin, hu-man trypsin and Chromozym TH were from SigmaChem. Co. (St Louis, MO, USA).

6.1.1. General procedure for the preparation ofcompoundsA (1–16)

Method A: an amount of 0.21 g (1 mM) of sulfaguani-dine 7 and the stoichiometric amount of pyrylium salt6and 140µL of triethylamine (1 mM) were dissolved/suspended in 20 mL of absolute methanol. The mixture

was refluxed for 30 min, then 0.45 mL of glacial aceticacid were added and refluxation was continued foranother 2 h. The cold mixture was treated with100–200 mL of diethyl ether for the precipitation of thepyridinium saltsA1–A16 which were recrystallized fromwater with 2–5% perchloric acid.

Method B: an amount of 0.60 g (2.9 mM) of sulfa-guanidine7 and 2.9 mM of pyrylium salt6 were sus-pended in 5 mL of anhydrous methanol and poured into astirred mixture of 14.5 mM of triethylamine and 5.8 mMof acetic anhydride. After 5 min of stirring, another10 mL of methanol were added to the reaction mixture,which was heated to reflux for 15 min. Then 14.5 mM ofacetic acid was added and heating was continued for2–5 h. The role of the acetic anhydride is to react with thewater formed during the condensation reaction betweenthe pyrylium salt and the aromatic amine, in order to shiftthe equilibrium towards the formation of the pyridiniumsalts of typeA1–A16. In the case of sulfaguanidine, thisprocedure is the only one which gave acceptable yields inpyridinium salts possessing 2-methyl groups. The pre-cipitated pyridinium salts obtained were then purified bytreatment with concentrated ammonia solution (whichalso converts the eventually unreacted pyrylium salt tothe corresponding pyridine which is soluble in acidicmedium), reprecipitation with perchloric acid and recrys-tallization from water with 2–5% HClO4.

6.1.1.1. 1-N-(4-Guanidinosulfonyl-phenyl)-2,4,6-trimethyl-pyridinium perchlorateA1

White crystals, m.p. 273–275 °C (yield of 34%); IR(KBr), cm–1: 625, 740, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 040, 3 245, 3 335;1H-NMR (TFA), δ ppm: 2.56(s, 6H, 2,6-(Me)2); 2.81 (s, 3H, 4-Me); 7.35–7.85 (m,AA≠BB≠, 4H, ArH from 1,4-phenylene); 8.10 (s, 2H,ArH, 3,5-H from pyridinium). Anal. C15H19N4O2S

+

ClO4– (C, H, N, S).

6.1.1.2. 1-N-(4-Guanidinosulfonyl-phenyl)-2-iso-propyl-4,6-dimethylpyridinium perchlorateA2

Pale yellow crystals, m.p. 255–256 °C (yield of 51%);IR (KBr), cm–1: 625, 680, 1 100, 1 175, 1 290, 1 345,1 580, 1 675, 3 020, 3 235;1H-NMR (TFA), δ ppm: 1.50(d, 6H, 2Me fromi-Pr); 2.70 (s, 3H, 6-Me); 2.83 (s, 3H,4-Me); 3.48 (heptet, 1H, CH fromi-Pr); 7.25–8.45 (m,AA≠BB≠, 4H, ArH from 1,4-phenylene); 7.98 (s, 2H,ArH, 3,5-H from pyridinium). Anal. C17H23N4O2S

+

ClO4– (C, H, N, S).

6.1.1.3. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-iso-propyl-4-methylpyridinium perchlorateA3

Tan crystals, m.p. 201–202 °C (yield of 76%); IR(KBr), cm–1: 625, 685, 820, 1 100, 1 175, 1 290, 1 345,

945

1 580, 1 675, 3 030, 3 250;1H-NMR (TFA), δ ppm: 1.51(d, 12H, 4Me from 2i-Pr); 2.80 (s, 3H, 4-Me); 3.42(heptet, 2H, 2CH from 2i-Pr); 7.31–8.51 (m, AA≠BB≠,4H, ArH from 1,4-phenylene); 8.05 (s, 2H, ArH, 3,5-Hfrom pyridinium). Anal. C19H27N4O2S

+ ClO4– (C, H, N,

S).

6.1.1.4. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-dimethyl-4-phenylpyridinium perchlorateA4

White crystals, m.p. 280–281 °C (yield of 50%); IR(KBr), cm–1: 625, 690, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 030, 3 260, 3 330;1H-NMR (TFA), δ ppm: 2.58(s, 6H, 2,6-(Me)2); 8.10–9.12 (m, 11H, ArH from 1,4-phenylene, pyridinium and 4-Ph). Anal. C20H21N4O2S

+

ClO4– (C, H, N, S).

6.1.1.5. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-diethyl-4-phenylpyridinium perchlorateA5

Yellow crystals, m.p. 263–265 °C (yield of 37%); IR(KBr), cm–1: 625, 765, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 040, 3 270, 3 360;1H-NMR (TFA), δ ppm: 1.43(t, 6H, 2 Me from ethyl); 2.82 (q, 4H, 2 CH2 from Et);7.68–8.87 (m, 11H, ArH from 1,4-phenylene, pyridiniumand 4-Ph). Anal. C22H25N4O2S

+ ClO4– (C, H, N, S).

6.1.1.6. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-n-propyl-4-phenylpyridinium perchlorateA6

Yellowish crystals, m.p. 215–216 °C (yield of 61%);IR (KBr), cm–1: 625, 775, 1 100, 1 175, 1 290, 1 345,1 580, 1 675, 3 060, 3 220, 3 315;1H-NMR (TFA), δppm: 1.01 (t, 6H, 2 Me from propyl); 1.70 (sextet, 4H,2CH2 (â) from n-Pr); 2.80 (t, 4H, 2 CH2 (α) from n-Pr);7.55–8.78 (m, 11H, ArH from 1,4-phenylene, pyridiniumand 4-Ph). Anal. C24H29N4O2S

+ ClO4– (C, H, N, S).

6.1.1.7. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-iso-propyl-4-phenylpyridinium perchlorateA7

White crystals, m.p. 198–199 °C (yield of 24%); IR(KBr), cm–1: 625, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 060, 3 270, 3 315;1H-NMR (TFA), δ ppm: 1.45(d, 12H, 4 Me fromi-Pr); 2.95 (heptet, 2H, 2 CH fromi-Pr); 7.92–8.97 (m, 11H, ArH from 1,4-phenylene,pyridinium and 4-Ph). Anal. C24H29N4O2S

+ ClO4– (C, H,

N, S).

6.1.1.8. 1-N-(4-Guanidinosulfonyl-phenyl)-2-methyl-4,6-diphenylpyridinium perchlorateA8

White crystals, m.p. 262–263 °C (yield of 30%); IR(KBr), cm–1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 040, 3 245, 3 350;1H-NMR (TFA), δ ppm: 2.72(s, 3H, 2-Me); 7.55–8.73 (m, 16H, ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal.C25H23N4O2S

+ ClO4– (C, H, N, S).

6.1.1.9. 1-N-(4-Guanidinosulfonyl-phenyl)-2-ethyl-4,6-diphenylpyridinium perchlorateA9

White-yellow crystals, m.p. 233–234 °C (yield of39%); IR (KBr), cm–1: 625, 700, 770, 1 100, 1 175,1 290, 1 345, 1 580, 1 675, 3 040, 3 250, 3 350;1H-NMR(TFA), δ ppm: 1.50 (t, 3H, Me from ethyl); 2.97 (q, 2H,CH2); 7.40–8.57 (m, 16H, ArH from 1,4-phenylene,pyridinium and 4,6-Ph2). Anal. C26H25N4O2S

+ ClO4– (C,

H, N, S).

6.1.1.10. 1-N-(4-Guanidinosulfonyl-phenyl)-2-n-propyl-4,6-diphenylpyridinium perchlorateA10

White crystals, m.p. 243–244 °C (yield of 36%); IR(KBr), cm–1: 625, 700, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 030, 3 270, 3 350;1H-NMR (TFA), δ ppm: 1.05(t, 3H, Me from propyl); 1.93 (sextet, 2H,â-CH2 fromn-Pr); 2.93 (t, 2H,α-CH2 from n-Pr); 7.38–8.53 (m, 16H,ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal.C27H27N4O2S

+ ClO4– (C, H, N, S).

6.1.1.11. 1-N-(4-Guanidinosulfonyl-phenyl)-2-iso-propyl-4,6-diphenylpyridinium perchlorateA11

White crystals, m.p. 183–184 °C (yield of 25%); IR(KBr), cm–1: 625, 700, 770, 1 100, 1 175, 1 290, 1 345,1 580, 1 675, 3 040, 3 250, 3 360;1H-NMR (TFA), δppm: 1.52 (d, 6H, 2 Me fromi-propyl); 2.52–3.25 (m, 1H,CH from i-Pr); 7.33–8.60 (m, 16H, ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal.C27H27N4O2S

+ ClO4– (C, H, N, S).

6.1.1.12. 1-N-(4-Guanidinosulfonyl-phenyl)-2-n-butyl-4,6-diphenylpyridinium perchlorateA12

White crystals, m.p. 244–245 °C (yield of 72%); IR(KBr), cm–1: 625, 710, 770, 1 100, 1 175, 1 290, 1 345,1 580, 1 675, 3 060, 3 260, 3 345;1H-NMR (TFA), δppm: 0.90 (t, 3H, Me from butyl); 1.10–2.15 (m, 4H,CH3-CH2-CH2-CH2 from n-Bu); 2.97 (t, 2H,α-CH2 fromn-Bu); 7.25–8.52 (m, 16H, ArH from 1,4-phenylene,pyridinium and 4,6-Ph2). Anal. C28H29N4O2S

+ ClO4– (C,

H, N, S).

6.1.1.13. 1-N-(4-Guanidinosulfonyl-phenylmethyl)-2-tert-butyl-4,6-diphenylpyridinium perchlorateA13

White crystals, m.p. 197–198 °C (yield of 62%); IR(KBr), cm–1: 625, 765, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 060, 3 270;1H-NMR (TFA), δ ppm: 1.90 (s, 9H,t-Bu); 6.83–8.83 (m, 16H, ArH from 1,4-phenylene,4,6-Ph2 and 3,5-H from pyridinium). Anal.C28H29N4O2S

+ ClO4– (C, H, N, S).

946

6.1.1.14. 1-N-(4-Guanidinosulfonyl-phenyl)-2,4,6-tri-phenylpyridinium perchlorateA14

Yellow crystals, m.p. 234–235 °C (yield of 80%); IR(KBr), cm–1: 625, 700, 770, 1 100, 1 175, 1 290, 1 345,1 580, 1 675, 3 030, 3 260, 3 350;1H-NMR (TFA), δppm: 7.47–8.63 (m, 21H, ArH from 1,4-phenylene, pyri-dinium and 2,4,6-Ph3). Anal. C30H25N4O2S

+ ClO4– (C,

H, N, S).

6.1.1.15. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-phenylpyridinium perchlorateA15

Yellow-orange crystals, m.p. 250–252 °C (yield of36%); IR (KBr), cm–1: 625, 705, 765, 1 100, 1 175,1 290, 1 345, 1 580, 1 675, 3 050, 3 260;1H-NMR(TFA), δ ppm: 6.71–8.40 (m, 17H, ArH from 1,4-phenylene, 2,6-Ph2 and 3,4,5-H from pyridinium). Anal.C24H20N4O2S

+ ClO4– (C, H, N, S).

6.1.1.16. 1-N-(4-Guanidinosulfonyl-phenyl)-2,3,4,6-tetramethylpyridinium perchlorateA16

White crystals, m.p. 256–257 °C (yield of 28%); IR(KBr), cm–1: 625, 750, 1 100, 1 175, 1 290, 1 345, 1 580,1 675, 3 040, 3 245, 3 330;1H-NMR (TFA), δ ppm: 2.45(s, 3H, 3-Me); 2.50 (s, 3H, 4-Me); 2.55 (s, 3H, 6-Me);2.75 (s, 3H, 2-Me); 8.03–9.17 (m, 5H, ArH from 1,4-phenylene and pyridinium 5-H). Anal. C16H21N4O2S

+

ClO4– (C, H, N, S).

6.1.2. General procedure for the preparation ofderivatives10 and11

An amount of 10 mM of amino acid (Gly orâ-Ala)was suspended/dissolved in 50 mL of anhydrous acetoni-trile and the stoichiometric amount (10 mM) of pyryliumsalt 6 and triethyl amine (10 mM, 1.47 mL) were added.The reaction mixture was heated at reflux for 4 h, then2.5 mL of glacial acetic acid were added and refluxationwas continued for another 2 h. The obtained reactionmixture was treated as described above (Method A), inorder to obtain the pure intermediates10 and11 (recrys-tallized from water with 2–5% perchloric acid).

6.1.3.General procedure for the preparation ofcompoundsB andC (1–16)

An amount of 1 mM of pyridinium-amino acid deriva-tive 10 or 11 was dissolved/suspended in 25 mL ofanhydrous acetonitrile or acetone, and then treated with210 mg (1 mM) of sulfaguanidine7 and 190 mg (1 mM)of EDCI. HCl or di-isopropyl-carbodiimide. The reactionmixture was magnetically stirred at room temperature for15 min, then 30µL (2 mM) of triethylamine were addedand stirring was continued for 16 h at 4 °C. The solventwas evaporated in vacuo and the residue taken up in ethylacetate (5 mL), poured into a 5% solution of sodium

bicarbonate (5 mL) and extracted with ethyl acetate. Thecombined organic layers were dried over sodium sulfateand filtered, and the solvent removed in vacuo. Prepara-tive HPLC (Dynamax-60A column (25× 250 mm); 90%acetonitrile/8% methanol/2% water; flow rate of 30 mL/min) afforded the pure compoundsB and C (1–16) ascolourless solids.

6.1.3.1. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,4,6-trimethylpyridinium perchlorateB1

White-tan crystals, m.p. 280–282 °C (yield of 80%);IR (KBr), cm–1: 625, 680, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 030, 3 250;1H-NMR(TFA), δ ppm: 2.70 (s, 3H, 4-Me); 2.85 (s, 6H, 2,6-(Me)2); 4.12 (s, 2H, Gly CH2); 7.13–8.41 (m, AA≠BB≠,4H, ArH from 1,4-phenylene); 8.00 (s, 2H, ArH, 3,5-Hfrom pyridinium). Anal. C17H21N4O3S

+ ClO4– (C, H, N,

S).

6.1.3.2. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-iso-propyl-4,6-dimethylpyridinium perchlo-rate B2

Light orange crystals, m.p. 205–207 °C (yield of 64%);IR (KBr), cm–1: 625, 680, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 020, 3 235;1H-NMR(TFA), δ ppm: 1.50 (d, 6H, 2Me fromi-Pr); 2.80 (s, 3H,6-Me); 2.90 (s, 3H, 4-Me); 3.48 (heptet, 1H, CH fromi-Pr); 4.12 (s, 2H, Gly CH2); 7.25–8.43 (m, AA≠BB≠, 4H,ArH from 1,4-phenylene); 7.98 (s, 2H, ArH, 3,5-H frompyridinium). Anal. C19H25N4O3S

+ ClO4– (C, H, N, S).

6.1.3.3. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-di-iso-propyl-4-methylpyridiniumperchlo-rate B3

Tan crystals, m.p. 202–203 °C (yield of 75%); IR(KBr), cm–1: 625, 820, 1 100, 1 175, 1 290, 1 345, 1 535,1 580, 1 640, 1 675, 3 030, 3 250;1H-NMR (TFA), δppm: 1.51 (d, 12H, 4Me from 2i-Pr); 2.83 (s, 3H, 4-Me);3.42 (heptet, 2H, 2CH from 2i-Pr); 4.12 (s, 2H, CH2);7.31–8.51 (m, AA≠BB≠, 4H, ArH from 1,4-phenylene);8.03 (s, 2H, ArH, 3,5-H from pyridinium). Anal.C21H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.4. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-dimethyl-4-phenylpyridinium perchlorateB4

Orange-red crystals, m.p. 240–241 °C (yield of 69%);IR (KBr), cm–1: 625, 765, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 050, 3 265;1H-NMR(TFA), δ ppm: 3.00 (s, 6H, 2,6-(Me)2); 4.12 (s, 2H, CH2);7.21–8.51 (m, 11H, ArH from 1,4-phenylene, 4-Ph and3,5-H from pyridinium). Anal. C22H23N4O3S

+ ClO4– (C,

H, N, S).

947

6.1.3.5. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-diethyl-4-phenylpyridinium perchlorateB5

Tan crystals, m.p. 223–224 °C (yield of 53%); IR(KBr), cm–1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 535,1 580, 1 640, 1 675, 3 060, 3 230;1H-NMR (TFA), δppm: 1.55 (t, 6H, 2 Me from Et); 3.30 (q, 4H, 2 CH2 fromEt); 4.12 (s, 2H, N+ -CH2); 7.08–8.63 (m, 11H, ArH from1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal.C24H27N4O3S

+ ClO4– (C, H, N, S).

6.1.3.6. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-di-n-propyl-4-phenylpyridinium perchlo-rate B6

Tan crystals, m.p. 218–219 °C (yield of 55%); IR(KBr), cm–1: 625, 775, 1 100, 1 175, 1 290, 1 345, 1 535,1 580, 1 640, 1 675, 3 060, 3 240;1H-NMR (TFA), δppm: 1.15 (t, 6H, 2 Me from Pr); 1.90 (sextet, 4H, 2 CH2

from Pr); 3.18 (t, 4H, 2 CH2 from Pr); 4.12 (s, 2H,N+-CH2); 7.10–8.50 (m, 11H, ArH from 1,4-phenylene,4-Ph and 3,5-H from pyridinium). Anal. C26H31N4O3S

+

ClO4– (C, H, N, S).

6.1.3.7. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-di-iso-propyl-4-phenylpyridiniumperchlo-rate B7

Tan crystals, m.p. 210–213 °C (yield of 79%); IR(KBr), cm–1: 625, 775, 1 100, 1 175, 1 290, 1 345, 1 535,1 580, 1 640, 1 675, 3 060, 3 240;1H-NMR (TFA), δppm: 1.55 (d, 12H, 4 Me fromi-Pr); 3.53 (heptet, 2H, 2CH from i-Pr); 4.13 (s, 2H, N+-CH2); 7.23–8.65 (m, 11H,ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyri-dinium). Anal. C26H31N4O3S

+ ClO4– (C, H, N, S).

6.1.3.8. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-methyl-4,6-diphenylpyridinium perchlorateB8

Yellow crystals, m.p. 254–255 °C (yield of 51%); IR(KBr), cm–1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 535,1 580, 1 640, 1 675, 3 050, 3 250;1H-NMR (TFA), δppm: 3.00 (s, 3H, 2-Me); 4.12 (s, 2H, CH2); 7.08–8.58(m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-Hfrom pyridinium). Anal. C27H25N4O3S

+ ClO4– (C, H, N,

S).

6.1.3.9. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-ethyl-4,6-diphenylpyridinium perchlorateB9

White crystals, m.p. 221–223 °C (yield of 84%); IR(KBr), cm–1: 625, 705, 770, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 050, 3 250;1H-NMR(TFA), δ ppm: 1.60 (t, 3H, Me from Et); 3.27 (q, 2H, CH2

from Et); 4.12 (s, 2H, N+-CH2); 7.08–8.60 (m, 16H, ArH

from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium).Anal. C28H27N4O3S

+ ClO4– (C, H, N, S).

6.1.3.10. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-n-propyl-4,6-diphenylpyridinium perchlo-rate B10

White-yellowish crystals, m.p. 200–201 °C (yield of53%); IR (KBr), cm–1: 625, 685, 770, 1 100, 1 175,1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 080, 3 250;1H-NMR (TFA), δ ppm: 1.18 (t, 3H, Me from Pr); 2.10(sextet, 2H, CH2 from n-Pr); 3.20 (t, 2H, CH2 from n-Pr);4.12 (s, 2H, N+-CH2); 7.08–8.63 (m, 16H, ArH from1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium).Anal. C29H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.11. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-iso-propyl-4,6-diphenylpyridinium perchlo-rate B11

Tan crystals, m.p. 187–188 °C (yield of 62%); IR(KBr), cm–1: 625, 710, 770, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 070, 3 250;1H-NMR(TFA), δ ppm: 1.55 (d, 6H, 2 Me fromi-Pr); 3.55 (heptet,1H, CH from i-Pr); 4.10 (s, 2H, N+-CH2); 7.08–8.63 (m,16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H frompyridinium). Anal. C29H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.12. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-n-butyl-4,6-diphenylpyridinium perchlorateB12

Tan crystals, m.p. 198–199 °C (yield of 43%); IR(KBr), cm–1: 625, 690, 770, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 080, 3 250;1H-NMR(TFA), δ ppm: 0.93 (t, 3H, Me fromn-Bu); 1.55 (sextet,2H, CH2 from n-Bu); 2.05 (quintet, 2H, CH2 from n-Bu);3.17 (t, 2H, CH2 from n-Bu); 4.12 (s, 2H, N+-CH2);7.08–8.58 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and3,5-H from pyridinium). Anal. C30H31N4O3S

+ ClO4– (C,

H, N, S).

6.1.3.13. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2-tert-butyl-4,6-diphenylpyridinium perchlo-rate B13

White crystals, m.p. 201–203 °C (yield of 54%); IR(KBr), cm–1: 625, 705, 765, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 060, 3 270;1H-NMR(TFA), δ ppm: 1.90 (s, 9H,t-Bu); 4.22 (s, 2H, CH2);6.83–8.83 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and3,5-H from pyridinium). Anal. C30H31N4O3S

+ ClO4– (C,

H, N, S).

948

6.1.3.14. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,4,6-triphenylpyridinium perchlorateB14

Orange crystals, m.p. 234–235 °C (yield of 70%); IR(KBr), cm–1: 625, 705, 770, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 050, 3 270;1H-NMR(TFA), δ ppm: 4.09 (s, 2H, CH2); 6.70–8.56 (m, 21H,ArH from 1,4-phenylene, 2,4,6-Ph3 and 3,5-H frompyridinium). Anal. C32H27N4O3S

+ ClO4– (C, H, N, S).

6.1.3.15. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,6-diphenylpyridinium perchlorateB15

Yellow-orange crystals, m.p. 204–206 °C (yield of40%); IR (KBr), cm–1: 625, 705, 765, 1 100, 1 175,1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 260;1H-NMR (TFA), δ ppm: 4.13 (s, 2H, CH2); 6.71–8.40 (m,17H, ArH from 1,4-phenylene, 2,6-Ph2 and 3,4,5-H frompyridinium). Anal. C26H22N4O3S

+ ClO4– (C, H, N, S).

6.1.3.16. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylmethyl]-2,3,4,6-tetramethylpyridinium perchlorateB16

White-tan crystals, m.p. 253–255 °C (yield of 65%);IR (KBr), cm–1: 625, 800, 1 100, 1 175, 1 290, 1 345,1 535, 1 580, 1 640, 1 675, 3 030, 3 305;1H-NMR(TFA), δ ppm: 2.60 (s, 3H, 4-Me); 2.77 (s, 3H, 3-Me);2.87 (s, 6H, 2,6-(Me)2); 4.12 (s, 2H, CH2); 7.21–8.50 (m,AA≠BB≠, 4H, ArH from 1,4-phenylene); 7.90 (s, 1H,ArH, 5-H from pyridinium). Anal. C18H23N4O3S

+ ClO4–

(C, H, N, S).

6.1.3.17. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,4,6-trimethylpyridinium perchlorateC1

White crystals, m.p. 266–267 °C (yield of 84%); IR(KBr), cm–1: 625, 680, 1 100, 1 175, 1 285, 1 345, 1 540,1 580, 1 645, 1 675, 3 060, 3 250, 3 330;1H-NMR(TFA), δ ppm: 2.66 (s, 3H, 4-Me); 2.88 (s, 6H, 2,6-(Me)2); 3.12 (t, 2H, CH2); 4.05 (t, 2H, CH2); 7.47–8.38(m, 6H, ArH from 1,4-phenylene and 3,5-H from pyri-dinium). Anal. C18H23N4O3S

+ ClO4– (C, H, N, S).

6.1.3.18. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-iso-propyl-4,6-dimethylpyridinium perchlo-rate C2

White crystals, m.p. 244–245 °C (yield of 83%); IR(KBr), cm–1: 625, 685, 1 100, 1 175, 1 285, 1 345, 1 540,1 580, 1 645, 1 675, 3 040, 3 255, 3 380;1H-NMR(TFA), δ ppm: 1.47 (d, 6H, 2Me fromi-Pr); 2.68 (s, 3H,4-Me); 2.90 (s, 3H, 6-Me); 3.10–3.75 (m, 3H, CH fromi-Pr + CH2); 4.03 (t, 2H, CH2); 7.33–8.35 (m, 6H, ArHfrom 1,4-phenylene and 3,5-H from pyridinium). Anal.C20H26N4O3S

+ ClO4– (C, H, N, S).

6.1.3.19. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-di-iso-propyl-4-methylpyridinium perchlo-rate C3

White crystals, m.p. 250–251 °C (yield of 76%); IR(KBr), cm–1: 625, 685, 1 100, 1 175, 1 285, 1 345, 1 540,1 580, 1 645, 1 675, 3 040, 3 235, 3 410;1H-NMR(TFA), δ ppm: 1.48 (d, 12H, 4Me from 2i-Pr); 2.70 (s,3H, 4-Me); 3.15–3.79 (m, 4H, 2CH from 2i-Pr + CH2);4.02 (t, 2H, CH2); 7.33–8.27 (m, 6H, ArH from 1,4-phenylene and 3,5-H from pyridinium). Anal.C22H31N4O3S

+ ClO4– (C, H, N, S).

6.1.3.20. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-dimethyl-4-phenylpyridinium perchlorateC4

White crystals, m.p. 221–223 °C (yield of 72%); IR(KBr), cm–1: 625, 690, 780, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 280;1H-NMR(TFA), δ ppm: 3.08 (s, 6H, 2,6-(Me)2); 3.15 (t, 2H, CH2);4.03 (t, 2H, CH2); 7.55–8.37 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal.C23H25N4O3S

+ ClO4– (C, H, N, S).

6.1.3.21. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-diethyl-4-phenylpyridinium perchlorateC5

White crystals, m.p. 227–229 °C (yield of 80%); IR(KBr), cm–1: 625, 700, 780, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 060, 3 240, 3 335;1H-NMR(TFA), δ ppm: 1.67 (t, 6H, 2 Me from Et); 3.15–3.80 (m,6H, 2 CH2 from Et + CH2 from ethylene bridge); 4.07 (t,2H, CH2 from ethylene bridge); 7.57–8.50 (m, 11H, ArHfrom 1,4-phenylene, 4-Ph and 3,5-H from pyridinium).Anal. C25H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.22. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-di-n-propyl-4-phenylpyridinium perchlo-rate C6

White crystals, m.p. 212–214 °C (yield of 63%); IR(KBr), cm–1: 625, 685, 775, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 255, 3 335;1H-NMR(TFA), δ ppm: 1.23 (t, 6H, 2 Me from Pr); 2.03 (q, 4H,2 CH2 from Pr); 3.07–3.75 (m, 6H, 2 CH2 from Pr + CH2

from ethylene bridge); 4.05 (t, 2H, CH2 from ethylenebridge); 7.55–8.43 (m, 11H, ArH from 1,4-phenylene,4-Ph and 3,5-H from pyridinium). Anal. C27H33N4O3S

+

ClO4– (C, H, N, S).

6.1.3.23. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-di-iso-propyl-4-phenylpyridinium perchlo-rate C7

White crystals, m.p. 241–243 °C (yield of 69%); IR(KBr), cm–1: 625, 685, 765, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 060, 3 270, 3 350;1H-NMR

949

(TFA), δ ppm: 1.60 (d, 12H, 4 Me fromi-Pr); 3.10–3.83(m, 4H, 2 CH fromi-Pr + CH2 from ethylene bridge);4.13 (t, 2H, CH2 from ethylene bridge); 7.47–8.43 (m,11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H frompyridinium). Anal. C27H33N4O3S

+ ClO4– (C, H, N, S).

6.1.3.24. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-methyl-4,6-diphenylpyridinium perchlorateC8

White crystals, m.p. 233–234 °C (yield of 77%); IR(KBr), cm–1: 625, 675, 775, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 245, 3 435;1H-NMR(TFA), δ ppm: 3.03–3.39 (m, 5H, 2-Me + CH2 fromethylene bridge); 4.06 (t, 2H, CH2 from ethylene bridge);7.05–8.45 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and3,5-H from pyridinium). Anal. C28H27N4O3S

+ ClO4– (C,

H, N, S).

6.1.3.25. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-ethyl-4,6-diphenylpyridinium perchlorateC9

White crystals, m.p. 229–230 °C (yield of 54%); IR(KBr), cm–1: 625, 685, 750, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 220, 3 390;1H-NMR(TFA), δ ppm: 1.72 (t, 3H, Me from Et); 2.90–3.78 (m,4H, CH2 from Et + CH2 from ethylene bridge); 4.08 (t,2H, CH2 from ethylene bridge); 6.88–8.47 (m, 16H, ArHfrom 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium).Anal. C29H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.26. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-n-propyl-4,6-diphenylpyridinium perchlorateC10

White crystals, m.p. 235–236 °C (yield of 59%); IR(KBr), cm–1: 625, 705, 775, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 080, 3 255, 3 340;1H-NMR(TFA), δ ppm: 1.32 (t, 3H, Me from Pr); 2.17 (sextet, 2H,CH2 from n-Pr); 2.82–3.66 (m, 4H, CH2 from n-Pr + CH2

from ethylene bridge); 4.09 (t, 2H, CH2 from ethylenebridge); 6.83–8.43 (m, 16H, ArH from 1,4-phenylene,4,6-Ph2 and 3,5-H from pyridinium). Anal.C30H31N4O3S

+ ClO4– (C, H, N, S).

6.1.3.27. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-iso-propyl-4,6-diphenylpyridinium perchlo-rate C11

White crystals, m.p. 234–235 °C (yield of 73%); IR(KBr), cm–1: 625, 700, 765, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 070, 3 250, 3 350;1H-NMR(TFA), δ ppm: 1.70 (d, 6H, 2 Me fromi-Pr); 3.15 (t, 2H,CH2 from ethylenic bridge); 3.50–4.03 (m, 1H, CH fromi-Pr); 4.11 (t, 2H, CH2 from ethylenic bridge); 6.95–8.53

(m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-Hfrom pyridinium). Anal. C30H31N4O3S

+ ClO4– (C, H, N,

S).

6.1.3.28. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-n-butyl-4,6-diphenylpyridinium perchlorateC12

White crystals, m.p. 207–208 °C (yield of 78%); IR(KBr), cm–1: 625, 685, 7650, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 080, 3 255, 3 330;1H-NMR(TFA), δ ppm: 1.15 (t, 3H, Me fromn-Bu); 1.38–2.45 (m,4H, 2 CH2 from n-Bu); 3.00–3.68 (m, 4H, CH2 fromn-Bu + CH2 from ethylenic bridge); 4.10 (t, 2H, CH2from ethylenic bridge); 7.02–8.43 (m, 16H, ArH from1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium).Anal. C31H33N4O3S

+ ClO4– (C, H, N, S).

6.1.3.29. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2-tert-butyl-4,6-diphenylpyridinium perchlo-rate C13

White crystals, m.p. 220–222 °C (yield of 69%); IR(KBr), cm–1: 625, 700, 765, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 060, 3 250, 3 370;1H-NMR(TFA), δ ppm: 1.92 (s, 9H,t-Bu); 3.14 (t, 2H, CH2); 4.10(t, 2H, CH2 from ethylene bridge); 6.90–8.77 (m, 16H,ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyri-dinium). Anal. C31H33N4O3S

+ ClO4– (C, H, N, S).

6.1.3.30. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,4,6-triphenylpyridinium perchlorateC14

Yellow crystals, m.p. 213–214 °C (yield of 82%); IR(KBr), cm–1: 625, 680, 770, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 260, 3 335;1H-NMR(TFA), δ ppm: 3.12 (t, 2H, CH2 from ethylene bridge);4.05 (t, 2H, CH2 from ethylene bridge); 6.57–8.40 (m,21H, ArH from 1,4-phenylene, 2,4,6-Ph3 and 3,5-H frompyridinium). Anal. C33H29N4O3S

+ ClO4– (C, H, N, S).

6.1.3.31. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,6-diphenylpyridinium perchlorateC15

Yellow crystals, m.p. 212–214 °C (yield of 16%); IR(KBr), cm–1: 625, 700, 760, 1 100, 1 175, 1 285, 1 345,1 540, 1 580, 1 645, 1 675, 3 050, 3 240, 3 325;1H-NMR(TFA), δ ppm: 3.07 (t, 2H, CH2); 4.13 (t, 2H, CH2 fromethylene bridge); 6.55–8.50 (m, 17H, ArH from 1,4-phenylene, 2,6-Ph2 and 3,4,5-H from pyridinium). Anal.C27H24N4O3S

+ ClO4– (C, H, N, S).

6.1.3.32. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbo-nylethyl]-2,3,4,6-tetramethylpyridinium perchlorateC16

White crystals, m.p. 210–211 °C (yield of 51%); IR(KBr), cm–1: 625, 680, 1 100, 1 175, 1 285, 1 345, 1 540,1 580, 1 645, 1 675, 3 030, 3 245, 3 325;1H-NMR

950

(TFA), δ ppm: 2.52 (s, 3H, 3-Me); 2.62 (s, 3H, 4-Me);2.83 (s, 3H, 6-Me); 2.92 (s, 3H, 2-Me); 3.13 (t, 2H, CH2);4.07 (t, 2H, CH2); 7.61–8.55 (m, 5H, ArH from 1,4-phenylene + 5-H from pyridinium). Anal. C19H25N4O3S

+

ClO4– (C, H, N, S).

6.2. Pharmacology

6.2.1. Enzyme assays: KI determinationsHuman thrombin and human trypsin were purchased

from Sigma Chemical Co. (St. Louis, MO, USA); theirconcentrations were determined from the absorbance at280 nm and the extinction coefficients furnished by thesupplier. The activity of such preparations was in therange of 2 500–3 000 NIH units/mg. The potency ofstandard and newly obtained inhibitors was determinedfrom the inhibition of the enzymatic (amidolytic) activityof these serine proteases, at 21 °C, usingTs-Gly-Pro-Arg-pNA (Chromozym TH) from Sigma as substrate, by themethod of Lottenberg et al. [23]. The substrate wasreconstituted as a 4 mM stock in ultrapure water andbrought to pH 4 with hydrochloric acid. Substrate con-centrations were determined from absorbance at theisosbestic wavelength for the peptide-p-nitroanilide/p-nitroaniline mixtures. Extinction coefficients of8 270 L.mol–1.cm–1 in the used buffer (0.01 M Hepes,0.01 M Tris, 0.1 M NaCl, 0.1% polyethylene glycol 6000;pH 7.80) were employed. The rate ofp-nitroanilidehydrolysis was determined from the change in absor-bance at 405 nm using an extinction coefficient forp-nitroaniline of 9 920 L.mol–1.cm–1 for the above-mentioned reaction buffer. Measurements were madeusing a Cary 3 spectrophotometer interfaced with a PC.Initial velocities were thus estimated using the directlinear plot-based procedure as reported by Lottenberg etal. [23]. KI’s were then determined according to Dixon,using a linear regression program [26]. The KI valuesdetermined are the means of at least three determinations.

6.2.2. pKa determinationThe half neutralization point was measured by titrating

the organic acids/bases with 0.05 N NaOH and 0.05 NHCl in EtOH/water (30%, v/v), using a glass electrode, asdescribed by Bell and Roblin [35] for the structurally-related antibacterial sulfonamides.

Acknowledgements

This research was financed in part by the EU grantERB CIPDCT 940051. Thanks are addressed to Dr M.A.Ilies for expert technical assistance.

References

[1] Shafer J.A., Curr. Opin. Chem. Biol. 2 (1998) 458–465.

[2] Stürzebecher J., Meier J., J. Enzyme Inhib. 9 (1995) 1–2.

[3] Stürzebecher J., Prasa D., Hauptmann J., Vieweg H., Wikstrom P., J.Med. Chem. 40 (1997) 3091–3099.

[4] Pavone V., De Simone G., Nastri F., Galdiero S., Staiano N.,Lombardi A., Pedone C., Biol. Chem. 379 (1998) 987–1006.

[5] Engh R.A., Brandstetter H., Sucher G., Eichinger A., Baumann U.,Bode W. et al., Structure 4 (1996) 1353–1362.

[6] Babine R.E., Bender S.L., Chem. Rev. 97 (1997) 1359–1472.

[7] Salemme F.R., Spurlino J., Bone R., Structure 5 (1997) 319–324.

[8] Eriksson U.G., Renberg L., Bredberg U., Teger-Nilsson A.C.,Regardh C.G., Biopharm. Drug Dispos. 19 (1998) 55–64.

[9] Bajusz S., Szell E., Bagdy D., Barbas E., Horvath G., Dioszegi M.et al., J. Med. Chem. 33 (1990) 1729–1735.

[10] Krishnan R., Zhang E., Hakansson K., Arni R.K., Tulinsky A.,Lim-Wilby M.S. et al., Biochemistry 37 (1998) 12094–13103.

[11] Claeson G., Philipp M., Agner E., Scully M.F., Metternich R.,Kakkar V.V., Desoyza T., Niu L.H., Biochem. J. 290 (1993)309–312.

[12] Stürzebecher J., Prasa D., Bretschneider E., Bode W., Bauer M.,Brandstetter H., Wikstrom P., Vieweg H., New developments in thefield of benzamidine-derived thrombin inhibitors, in: Muller-Berghaus G., Madlener K., Blomback M., ten Cate J.W. (Eds.), DIC– Pathogenesis, Diagnosis, and Therapy of Disseminated Intravas-cular Fibrin Formation, Excerpta Medica, Amsterdam, London, NewYork, Tokyo, 1993, pp. 183–190.

[13] Lumma W.C., Witherup K.M., Tucker T.J., Brady S.F., Sisko J.T.,Naylor-Olsen A.M. et al., J. Med. Chem. 41 (1998) 1011–1013.

[14] Semple J.E., Rowley D.C., Brunck T.K., Ha-Uong T., Minami N.K.,Owens T.D. et al., J. Med. Chem. 39 (1996) 4531–4536.

[15] Sixma J.J., de Groot P.G., Thromb. Res. 68 (1992) 507–512.

[16] Okamoto S., Kinjo K., Hijikata A., Kikumoto R., Tamao Y., OhkuboK., Tonomura S., J. Med. Chem. 23 (1980) 827–830.

[17] Stürzebecher J., Markwardt F., Voigt B., Wagner G., Walsmann P.,Thromb. Res. 29 (1983) 635–642.

[18] Groutas W.C., Kuang R., Venkataraman R., Epp J.B., Ruan S.,Prakash O., Biochemistry 36 (1997) 4739–4750.

[19] Supuran C.T., Manole G., Dinculescu A., Schiketanz A., GheorghiuM.D., Puscas I., Balaban A.T., J. Pharm. Sci. 81 (1992) 716–719.

[20] Supuran C.T., Scozzafava A., Ilies M.A., Iorga B., Cristea T.,Chiraleu F., Banciu M.D., Eur. J. Med. Chem. 33 (1998) 577–594.

[21] Anderson G.W., Zimmerman J.E., Callahan F.M., J. Am. Chem. Soc.85 (1963) 3039.

[22] Sheehan J.C., Ledis S.L., J. Am. Chem. Soc. 95 (1973) 875–879.

[23] Lottenberg R., Christensen U., Jackson C.M., Coleman P.L., Meth.Enzymol. 80 (1981) 341–361.

[24] Banner D.W., Hadvary P., J. Biol. Chem. 266 (1991) 20085–20093.

[25] Supuran C.T., Scozzafava A., Briganti F., Clare B.W., J. Med. Chem.in press.

[26] Hemker H.C., Willems G.M., Beguin S.A., Thromb. Haemostas. 56(1986) 9–17.

[27] Balaban A.T., Toma C., Tetrahedron Suppl. 7 (1966) 1–7.

[28] Dinculescu A., Balaban A.T., Rev. Roum. Chim. 25 (1980)1505–1528.

[29] Supuran C.T., Banciu M.D., Balaban A.T., Rev. Roum. Chim. 38(1993) 199–205.

[30] Balaban A.T., Dinculescu A., Dorofeenko G.N., Fischer G.W.,Koblik A.V., Mezheritskii V.V., Schroth W., Pyrylium salts: synthe-

951

ses, reactions and physical properties, in: Katritzky A.R. (Ed.),Advances in Heterocyclic Chemistry, Academic Press, New York,1982, pp. 8–360.

[31] Supuran C.T., Pop E., Dinculescu A., Heterocycles 37 (1994)667–671.

[32] Neidlein R., Witerzens P., Monats. Chem. 106 (1975) 643–648.

[33] Bode W., Turk D., Karshikov A., Protein Sci. 1 (1992) 426–471.

[34] Stubbs M.T., Bode W., Thromb. Res. 69 (1993) 1–58.

[35] Bell P.H., Roblin R.O., J. Am. Chem. Soc. 64 (1942) 2905–2917.

952

Original article

Synthesis and antispasmodic activity of analogues of natural pterosins

Helen Sheridana*, Neil Frankishb,Ronan Farrella

aDepartment of Pharmacognosy, School of Pharmacy, Trinity College Dublin 2, IrelandbDepartment of Pharmacology, School of Pharmacy, Trinity College Dublin 2, Ireland

(Received 22 February 1999; accepted 31 May 1999)

Abstract – The synthesis of an extensive range of analogues of natural pterosins using modified Heck coupling is reported. The smoothmuscle relaxant activity of these compounds has been examimed. Several compounds with significant smooth muscle relaxant activity havebeen identified. © 1999 Éditions scientifiques et médicales Elsevier SAS

pterosins / indanones / indanes / smooth muscle relaxant activity

1. Introduction

As part of an ongoing study into the synthesis andsmooth muscle relaxant activity of sesquiterpene in-danones related to the pterosin family of fern metabo-lites [1], we have developed an improved synthetic routeto pterosin Z (1) [2], a metabolite ofPityrogrammaandPteris species [3, 4]. More recently we have shown that(1) exhibits potent smooth muscle relaxant activity [5].We now report on the synthesis and activity of a numberof indane and indanone analogues of pterosins in which avariety of three carbon side chains replace the hydroxy-ethyl side chain of the natural pterosins.

2. Chemistry

Palladium catalysed Heck coupling [6] was used in thisstudy for attaching a three carbon side chain to a range ofindanone nucleii to yield a series of pterosin analogues.Methylacrylate (2) was identified as a suitable threecarbon unit for coupling with bromoindanes (3–8), keyintermediates used in earlier pterosin synthetic studies [1,2]. The regioselectivity of the site of arylation on thedouble bond of methylacrylate has been reported to occurexclusively at the terminal alkene position giving rise to

a single structural isomer [6, 7]. Coupling of2 with 3–8was catalysed by palladium acetate in the presence oftriphenylphosphine and triethylamine and was carried outat 100 °C in sealed ampoules (Expt. l). The lowest yieldsin the coupling reaction were consistently observed forthe products10, 13 and14 while the highest yields wereobserved for the isopterosin type structure12. Theproducts of the initial coupling reactions9–14 (table I)were characterised spectroscopically and in all cases weresingle geometric isomers. The1H-NMR spectrum of10,typical for the series, shows the presence of two sets ofdoublets downfield at 6.2 and 7.8 ppm, characteristic ofthe methylpropenoate side chain. The coupling constantof the doubletsJ (16 Hz) is characteristic oftranscoupledprotons and is observed for all compounds in the series.It is therefore confirmed that in the Heck couplingbetween methylacrylate and bromoindanes only theEisomer is formed.

Compounds9–14were subjected to a range of chemi-cal reactions in which the methyl propenoate side chainwas modified (figure 1) to give rise to a series of pterosinanalogues15–39for which the change in smooth musclerelaxant activity could be measured relative to the changein structure. The structure of these compounds wasestablished by spectroscopic means and the smoothmuscle relaxant activity of9–39 was measured usingestablished methods.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 953−966 953© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

3. Pharmacological results and discussion

Addition of calcium (2.5 mM) to guinea-pig ileumbathed in high potassium (45 mM), calcium-free Kreb’ssolution caused a contracture of the tissue which wassustained for a period greater than 40 min. The com-pounds (added cumulatively) caused a dose-dependantinhibition of calcium (2.5 mM) contractures of guinea-pig ileum (for examples seefigure 2). The more potentcompounds (12, 16, 20, 28and 29), with EC50 values

ranging from 1.3× 10–6 to 1.1× 10–5 M, were more than3 orders of magnitude less potent than the positivecontrol, the calcium antagonist, nifedipine (figure 2)which had an EC50 value of 9.3 ± 0.5 × 10–9 M.Comparison of the activities of the compounds was madeat a single concentration of 10–5 M, with inhibitoryactivity ranging from 20–75% (figures 3–5). In compari-son, nifedipine, at a concentration of 10–8 M, inhibitedcalcium contractures by 48.1± 2.3%. The most potentsmooth muscle relaxant (figure 2) is the diol 39. The

Table I. Yield and phsical data for compounds9–14.

Compound Yield % Melting point M+

9 60 86–87 °C 244.1100

10 45 93–94 °C 272.1407

11 61 124–125 °C 244.1100

12 73 oil 272.1413

13 45 47–48 °C 230.1307

14 41 oil 258.1620

954

activity of this compound is statistically greater (P <0.05) than that of its close analogue29. These twocompounds (29 and 39) are structurally quite similar,having the three carbon side chain at the C-4 position andthe aromatic methyl groups at the 3,5-positions. The nextlevel of activity is shown by compounds12, 20 and28.These compounds are statistically (P < 0.05) less activethan29 and39 but more active than16 (P < 0.05).

No clear structure activity relationship is apparent fromthe results of the 30 compounds analysed in this study,although it does appear that an unsaturated three carbonside chain at the C-4 position coupled with aromaticmethyls at the 3,5- positions with oxygenation at C-1(isopterosin form) enhances activity relative to the posi-tioning of the carbon side chain at the C-6 positioncombined with the aromatic methyl substitution at the5,7- positions (pterosin form). The smooth muscle relax-ant activity (figure 2) of the most active compound39 inthis study (EC50 4.9± 0.6× 10–6 M) is significantly lower(P < 0.05) than the activity we have recently reported forpterosin Z (1) (EC50 1.3 ± 0.1 × 10–6 M) [5] and the

fungal indane40 (EC50 2.9± 1.6× 10–6 M) [1]. However,the structural modifications observed in this group ofpterosin analogues has led to an increase in activityrelative to the first reported naturally occurring smoothmuscle relaxants, onitin (EC50 1 × 10–4 M), onotisin(EC50 2 × 10–3 M) and otninoside (EC50 7 × 10–4M) [8,9].

Pterosins have been shown to inhibit contractile re-sponses of guinea-pig ileum by both histamine andacetylcholine [8], together with barium and potas-sium [9], suggesting a mechanism of action involvinginterference with calcium handling in the smooth musclecell. This conclusion is supported by the results of thisstudy, showing that the compounds inhibit calcium con-tractures of potassium-depolarised smooth muscle. How-ever it has not been determined at this time whether suchinterference with calcium handling involves inhibition ofextracellular calcium influx through membrane channelsor interference with the calcium/calmodulin cascade ofreactions within the cell.

Figure 1. Example of reactions carried out to modify the methylpropenoate side chain of pterosins.

955

4. Experimental protocols

Melting points were determined on a Me-Opta hotstage and are uncorrected. Infra red spectra were recordedon a Nicolet 205 FT-IR. Ultraviolet spectra were recorded

on a Varian Carey 3E UV-visible spectrophotometer.Mass spectra were determined at 70 eV on an AEI MS 30instrument.1H-NMR spectra were recorded on a BrukerMSL 300 instrument at 300 MHz.13C-NMR were re-corded at 75.47 MHz. Deuteriochloroform was used as

Figure 1. (Continued).

956

solvent with SiMe4 as internal standard. TLC’s were runon commercially pre-coated plates (Merck, Kieselgel

60F254). Merck Kieselgel 60 (9385) was used for columnchromatography.

Figure 1. (Continued).

957

4.1. Synthesis of bromoindanones3–8. As outlined inreference [1]

4.1.1. General procedure for the coupling on methyl-acrylate (2) with bromoindanones(3–8) to synthesiseproducts9–14

A solution of indanone (2 mmol), palladium acetate(0.11 mmol), triphenyl phosphine (0.21 mmol) and me-thylacrylate (2) (5 mmol) in triethylamine (10 mL) wassealed in a glass ampoule (25 mL). The ampoule wasshaken until the mixture was homogeneous and was thenheated under pressure 15 lbs/in2 at 100 °C for 48 h. Aftercooling, the ampoule was broken and the reaction mix-

ture was filtered and stirred on ice/HCl and extracted intoEtOAc. The organic layer was washed, dried (Na2SO4)and evaporated under vacuum. The residue was purifiedby column chromatography on silica gel (eluant: pet.ether:ethyl acetate, 9:1).

4.1.1.1. E-2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-propenoic acid methyl ester(9)

Prepared by coupling 6-bromo-5,7-dimethyl-1H-indan-1-one [1] with methacrylate. White solid (59.5%),m.p. 86–87 °C (EtOH). Found M+ 244.1109 (C15H16O3

requires 244.1099).υmax(KBr) 2 361, 1 707, 1 686 cm–1;δH (CDCl3): 2.41 (3H, s, CH3), 2.62 (2H, tJ 6.0 Hz,

Figure 2. Effect of compounds 3× 10–8–3 × 10–5 M and Nifedipine, 1× 10–10–1 × 10–7 M (added cumulatively) on inhibition ofcalcium (2.5 mM) contractions of guinea-pig ileum suspended in high-potassium, calcium-free modified Kreb’s solution. Values areexpressed as means± SEM, n = 6.

958

CH2), 2.65 (3H, s, CH3), 3.08 (2H, tJ 6.0 Hz, CH2), 3.81(3H, s, OCH3), 6.23 (1H, dJ 16.4 Hz), 6.96 (1H, s, Ar-H),7.84 (1H, dJ 16.4 Hz);δC 18.1(CH3), 21.0 (CH3), 25.94(CH2), 36.72 (CH2), 51.72 (OCH3), 122.1 (CH), 128.3(CH), 129.0 (ArC), 133.1 (ArC), 139.6 (ArC), 140.2(ArCH), 144.0 (ArC), 155.3 (ArC), 167.2 (CO), 207.0(CO); m/z 244 (M+, 100), 213 (59), 212 (66), 185 (58),184 (40), 171 (43), 157 (32), 142 (66), 141 (44), 128 (30),115 (25), 77 (19).

4.1.1.2. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1-one-6-propenoic acid methyl ester(10)

Prepared by coupling 6-bromo-2,3-dihydro-2,2,5,7-tetramethyl-1H-inden-1-one [1] with methacrylate. Whitecrystals (46%), m.p. 93–94 °C (EtOH). Found M+

272.1416 (C17H20O3 requires 272.1413);υmax (KBr)2 951, 1 712, 1 698 cm–1; δH (CDCl3) 1.19 (6H, s,2×CH3), 2.45 (3H, s, CH3), 2.59 (3H, s, CH3), 2.99 (2H,s, CH2), 3.81 (3H, s, OCH3), 6.23 (1H, dJ 16.5 Hz), 7.01(1H, s, Ar-H), 7.86 (1H, dJ 16.4 Hz);δC 18.2 (CH3), 21.0(CH3), 25.3 (2×CH3), 43.1 (CH2), 45.4 (C), 51.7(OCH3),122.0 (CH), 128.9 (ArC), 131.3(ArC), 132.5 (ArCH),140.4 (ArCH), 141.2 (ArC), 144.0 (ArC), 152.3 (ArC),167.2 (CO), 211.3 (CO); m/z 272 (M+, 100), 257 (68) 211(21), 213 (32), 195 (28), 170 (17), 128 (14).

4.1.1.3. E-2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-propenoic acid methyl ester(11)

Prepared by coupling 5-bromo-2,3-dihydro-4,6-dimethyl-1H-inden-1-one [1] with methacrylate. Whitecrystals (61%), m.p. 124–125 °C (EtOH). Found M+

244.1096 (C15H16O3 requires 244.1099);υmax (KBr)2 955, 1 721, 1 704, 1 647cm–1; δH (CDCl3): 2.15 (3H, s,CH3), 2.18 (3H, s, CH3), 2.50 (2H, tJ 5.5 Hz, CH2), 2.83(2H, t J 5.5 Hz, CH2), 3.69 (3H, s, OCH3), 5.91 (1H, dJ16.4 Hz), 7.31 (1H, s, ArH), 7.64 (1H, dJ 16.4 Hz);δC

15.7 (CH3), 20.6 (CH3), 24.5 (CH2), 35.9 (CH2), 59.8(OCH3), 121.7 (CH), 124.8 (CH), 133.0 (ArC), 135.6(ArC), 135.8 (ArC), 140.0 (ArC), 142.4 (ArCH), 151.9(ArC), 166.0 (CO), 206.3 (CO); m/z 244 (M+, 40), 229(48), 213 (49), 212 (42), 184 (44), 171 (61), 143 (39), 142(100), 141 (37), 128 (48), 115 (37).

4.1.1.4. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1-one-5-propenoic acid methyl ester(12)

Prepared by coupling 5-bromo-2,3-dihydro-2,2,4,6-tetramethyl-1H-indan-1-one [1] with methacrylate. Paleoil (72%). Found M+ 272.1408 (C17H20O3 requires272.1413)υmax (Film) 2 958, 1 714, 1 716, 1 644 cm–1;δH (CDCl3) 1.11 (6H, s, 2×CH3), 2.18 (3H, s, CH3), 2.22(3H, s, CH3), 2.77 (2H, s, CH2), 3.72 (3H, s, OCH3), 5.96

Figure 3. Effect of several compounds (1× 10–5 M) and Nifedipine 1× 10–8 M, on inhibition of calcium (2.5 mM) contractions ofguinea-pig ileum suspended in high-potassium, calcium free modified Kreb’s solution. Values are expressed as means± SEM, n = 6.

959

(1H, d J 16.5 Hz), 7.32 (1H, s, ArH), 7.70 (1H, dJ 16.5Hz); δC 15.6 (CH3), 20.6 (CH3), 24.9 (2×CH3), 41.7(CH2), 45.1 (C), 51.4 (OCH3), 122.5 (CH), 124.7 (CH),133.3 (ArC), 134.1 (ArC), 135.8 (ArC), 140.3 (ArC),142.6 (ArCH), 149.0 (ArC), 166.1 (CO), 210.9 (CO); m/z272 (M+, 45), 258 (19), 257 (100), 241 (28), 240 (23),225 (19), 213 (13), 197 (17), 153 (10), 128 (14), 77 (5).

4.1.1.5. E-2,3-Dihydro-2,4-dimethyl-1H-inden-3-prope-noic acid methyl ester(13)

Prepared by coupling 3-bromo-2,3-dihydro-2,4-dimethyl-1H-inden [1] with methacrylate. White solid(45%), m.p. 47–48 °C (EtOH). Found M+ 230.1300(C15H18O2 requires 230.1307);υmax KBr 2 951, 1 721,1 633, 1 161 cm–1; δH (CDCl3) 2.09 (2H, mJ 7.5 Hz,CH2), 2.29 (3H, s, CH3), 2.34 (3H, s, CH3), 2.86 (2H, tJ7.5 Hz, CH2), 2.93 (2H,J 7.5 Hz, CH2), 3.85 (3H, s,OCH3), 6.06 (1H, dJ 16.2 Hz, CH), 6.99 (1H, s, ArH),7.90 (1H, dJ 16.2 Hz);δC 17.4 (CH3), 21.1 (CH3), 24.6(CH2), 31.8 (CH2), 33.1 (CH2), 51.5 (OCH3), 123.0(CH), 123.9 (CH), 131.6 (ArC), 132.1 (ArC), 134.8

(ArC), 141.4 (ArC), 144.2 (ArCH), 144.4 (ArC), 167.2(CO); m/z 230 (M+, 67), 215 (39), 199 (100), 170 (67),156 (19), 155 (57), 141 (15), 128 (23).

4.1.1.6. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-6-propenoic acid methyl ester(14)

Prepared by coupling 6-bromo-2,3-dihydro-2,2,5,7-dimethyl-1H-inden [1] with methacrylate. Pale oil (41%).Found M+ 258.1610 (C17H22O2 requires 258.1602).υmax

(Film) 2 952, 1 723, 1 634 cm–1; δH (CDCl3): 1.16 (6H,s, 2×CH3), 2.21 (3H, s, CH3), 2.32 (3H, s, CH3), 2.65(2H, s, CH2), 2.73 (2H, s, CH2), 3.82 (3H, s, OCH3), 6.04(1H, d J 16.2 Hz, CH), 6.87 (1H, s, ArH), 7.90 (1H, tJ16.2 Hz, CH);δC 17.3 (CH3), 20.3 (CH3), 29.1 (2×CH3),39.2 (C), 45.8 (CH2), 47.9 (CH2), 51.5 (OCH3), 124.2(CH), 129.8 (CH), 131.6 (ArC), 134.8 (ArC), 135.4(ArC), 140.6 (ArC), 142.8 (ArCH), 143.7 (ArC), 167.3(CO); m/z 258 (M+, 96), 243 (39), 228 (18), 227 (100),212 (22), 211 (61), 199 (40), 197 (17), 171 (11), 157 (13),153 (20), 143 (19), 128 (17).

Figure 4. Effect of several compounds (1× 10–5 M) and Nifedipine 1× 10–8 M, on inhibition of calcium (2.5 mM) contractions ofguinea-pig ileum suspended in high-potassium, calcium free modified Kreb’s solution. Values are expressed as means± SEM, n = 6.

960

4.1.2. General procedure for the preparation ofpropenoic acids(15, 19, 23, 27, 31and35)

A solution of propenoic acid methyl ester (2.25 mmol)in HCl (5 M, 15mL) and THF (15 mL) was refluxed for6 h. After this time the THF was evaporated in vacuo. Theaqueous residue was then diluted with H2O (50 mL) andwas extracted with EtOAc (2× 30 mL). The organic layerwas washed with 1 M NaOH (2 × 20 mL) and theaqueous layer retained. Following acidification with HClthe aqueous layer was extracted with ethyl acetate,washed with water, dried over anhydrous sodium sul-phate and evaporated off under vacuum to give a cruderesidue which was purified by recrystallisation to yieldthe propenoic acids.

4.1.2.1. E-2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-propenoic acid(15)

Prepared from9. White crystalline powder, m.p.219–220 °C. Found M+ 230.0953 (C14H16O3 requires230.0943).υmax (KBr) 3 500–3 200b, 1 725, 1 696 cm–1;

δH (DMSO-D6), 2.41 (3H, s, CH3), 2.51 (3H, s, CH3),2.59 (2H, tJ 6 Hz, CH2), 3.06 (2H, tJ 6 Hz, CH2), 6.27(1H, d J 16 Hz, CH), 7.10 (1H, s, CH), 7.72 (1H, dJ 16Hz, CH); δC 17.7 (CH3), 20.5 (CH3), 25.5 (CH2), 36.5(CH2), 123.6 (CH), 129.0 (C), 132.0 (C), 138.2 (CH),139.1 (C), 140.2 (CH), 143.6 (C), 155.4 (C), 167.4 (CO),206.5 (CO). m/z 230(M+, 96), 215 (28), 212 (41), 186(18), 185 (82), 184 (29), 171 (37), 157 (68), 156 (19), 143(37), 128 (41).

4.1.2.2. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1-one-6-propenoic acid(19)

Prepared from10. Pale yellow needles (EtOH), m.p.173–174 °C, yield 76%. Found M+ 258.1245 (C16H18O3

requires 258.1256).υmax (KBr) 3 400–3 200b, 1 702,1 679 cm–1; δH (CDCl3) 1.19 (6H, s, 2×CH3), 2.45 (3H,s, CH3), 2.59 (3H, s, CH3), 3.01 (2H, s, CH2), 6.23 (1H,d J 16 Hz, CH), 7.01 (1H, s, CH), 7.93 (1H, dJ 16 Hz,CH); δC 18.2 (CH3), 21.0 (CH3), 25.3 (2×CH3), 43.1(CH2), 45.4 (C), 121.9 (CH), 128.6 (C), 131.3 (C), 132.6

Figure 5. Effect of several compounds (1× 10–5 M) and Nifedipine 1× 10–8 M, on inhibition of calcium (2.5 mM) contractions ofguinea-pig ileum suspended in high-potassium, calcium free modified Kreb’s solution. Values are expressed as means± SEM, n = 6.

961

(CH), 139.1 (C), 140.7 (C), 141.8 (CH), 144.3 (C), 152.6(C), 170.7 (CO), 211.6 (CO); MS(EI) m/z 258 (M+, 82),243 (91), 240 (12), 213 (38), 195 (31), 170 (15), 128 (16),115 (10), 73 (13).

4.1.2.3. E-2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-propenoic acid(23)

Prepared from 11. White powder (EtOH), m.p.225–226 °C yield 95%. Found M+ 230.0934 (C14H14O3

requires 230.0943).υmax (KBr) 3 500–3 000b, 1 722,1 711 cm–1; δH (DMSO d6) 2.28 (3H, s, CH3), 2.32 (3H,s, CH3), 2.63 (2H, tJ 6 Hz, CH2), 2.98 (2H, tJ 6 Hz,CH2), 6.09 (1H, dJ 16 Hz, CH), 7.37 (1H, s, CH), 7.71(1H, d J 16 Hz, CH), 12.73 (1H, s, COOH);δC 15.6(CH3), 20.6 (CH3), 24.5 (CH2), 36.0 (CH2), 121.3 (CH),126.4 (CH), 133.8 (C), 135.6 (C), 135.7 (C), 140.2 (C),141.8 (CH), 152.4 (C), 166.9 (CO), 206.3 (CO); m/z 230(M+, 100), 215 (61), 212 (44), 184 (32), 171 (29), 143(62), 142 (100), 141 (35), 128 (45), 91 (73), 77 (23).

4.1.2.4. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1-one-5-propenoic acid(27)

Prepared from 12. White solid (EtOH), m.p.181–182 °C yield 86%. Found M+ 258.1227 (C16H18O3

requires 258.1256).υmax (KBr) 3 400–3 200b, 1 714,1 702 cm–1; δH (CDCl3) 1.08 (6H, s, 2×CH3), 2.16 (3H,s, CH3), 2.21 (3H, s, CH3), 2.75 (2H, s, CH2), 5.94 (1H,d J 16 Hz, CH), 7.30 (1H, s, CH), 7.70 (1H, dJ 16 Hz,CH); δC (CDCl3) 15.7 (CH3), 20.7 (CH3), 25.0 (2×CH3),41.9 (CH2), 45.4 (C), 122.7 (CH), 125.3 (CH), 133.6 (C),134.1 (C), 136.1 (C), 140.7 (C), 143.5 (CH), 149.4 (C),168.5 (CO), 211.9 (CO); m/z 258(M+, 100), 244 (28), 243(100), 240 (28), 153 (20), 128 (35), 115 (23), 77 (12).

4.1.2.5. E-2,3-Dihydro-5,7-dimethyl-1H-indene-6-prope-noic acid(31)

Prepared from13. Pale oil, yield 64%. Found M+

216.1387 (C14H16O2 requires 216.1146);υmax (Film)3 200–2 900, 1 693, 1 428 cm–1; δH (CDCl3) 2.05 (2H,m, CH2), 2.25 (3H, s, CH3), 2.31 (3H, s, CH3), 2.83 (2H,t J 7.5 Hz, CH2), 2.86 (2H, tJ 7.5 Hz, CH2), 6.01 (1H, dJ 16 Hz, CH), 6.95 (1H, s, CH), 7.89 (1H, dJ 16 Hz, CH);δC 17.4 (CH3), 21.1 (CH3), 24.5 (CH2), 31.7 (CH2), 33.1(CH2), 116.4 (CH), 123.8 (CH), 131.5 (C), 132.2 (C),134.9 (C), 141.4 (C), 144.4 (CH), 145.2 (C), 169.0 (CO);m/z 216 (M+, 100) 198 (29), 183 (42), 143 (44), 91 (10),77 (17).

4.1.2.6. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5-propenoic acid(35)

Prepared from 14. White solid (EtOH), m.p.138–139 °C, yield 83%. Found M+ 244.1459 (C16H18O2

requires 244.1436).υmax (KBr) 3 000b, 1 689 cm–1; δH

(CDCl3) 1.17 (6H, s, 2×CH3), 2.26 (3H, s, CH3), 2.35(3H, s, CH3), 2.68 (2H, s, CH2), 2.73 (2H, s, CH2), 6.07(1H, d J 16 Hz, CH), 6.92 (1H, s, CH), 8.01 (1H, dJ 16Hz, CH);δC 19.1 (CH3), 20.3 (CH3), 29.1 (2×CH3), 39.4(C), 45.7 (CH2), 48.9 (CH2) 124.4 (CH), 127.4 (C), 129.9(CH), 132.6 (C), 136.0 (C), 140.8 (C), 143.2 (C), 144.9(CH), 168.5 (CO); m/z 244 (M+, 100), 229 (51), 226 (22),199 (23), 185 (22), 183 (42), 171 (24), 157 (20), 143 (23),128 (25), 91 (15), 77 (15).

4.1.3. General procedure for the preparation ofpropanoic acid methyl esters(16, 20, 24, 28, 32and36)

A solution of the corresponding propenoic acid methylester (2 mmol) in EtOH was stirred with H2 underWilkinson’s catalyst [RhCl(PPh3)3] at room temperaturefor 24 h. On completion, the mixture was filtered and theEtOH was removed under vacuum. The residue waspurified by column chromatography (eluant pet. ether-:ethyl acetate, 4:1) to yield the following:

4.1.3.1. 2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-pro-panoic acid methyl ester(16)

Prepared by reduction of15. White solid (n-Hexane),m.p. 66–67 °C yield 86%. Found M+ 246.1262(C15H18O3 requires 246.1256);υmax (KBr) 1 738, 1 703,1 438 cm–1; δH (CDCl3) 2.33 (3H, s, CH3), 2.45 (2H, tJ8 Hz, CH2), 2.51 (3H, s, CH3), 2.61 (2H, tJ 6 Hz, CH2),2.95 (2H, tJ 8 Hz, CH2), 2.97 (2H, tJ 6 Hz, CH2), 3.66(3H, s, OCH3), 6.89 (1H, s, CH);δC 17.7 (CH3), 19.2(CH3), 23.8 (CH2), 24.0 (CH2), 33.0 (CH2), 36.7 (CH2),51.6 (OCH3), 131.8 (CH), 132.6 (C), 133.1 (C), 136.4(C), 142.5 (C), 154.7 (C), 173.0 (CO), 210 (CO); m/z 246(M+, 88), 214 (74), 186 (19), 173 (100), 159 (43), 144(40), 129 (35), 115 (24), 91 (17).

4.1.3.2. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5-pro-panoic acid methyl ester(20)

Prepared by reduction of19. White solid (n-Hexane),m.p. 54–55 °C yield 92%. Found M+ 274.1573(C17H22O3 requires 274.1569);υmax (KBr) 1 731, 1 693,1 441 cm–1; δH (CDCl3) 1.14 (6H, s, 2×CH3), 2.30 (3H,s, CH3), 2.42 (2H, tJ 9 Hz, CH2), 2.49 (3H, s, CH3), 2.83(2H, s, CH2), 2.90 (2H, t J 9 Hz, CH2), 3.62 (3H, s,OCH3), 6.87 (1H, s, CH):δC 17.6 (CH3), 19.2 (CH3),23.9 (CH2), 25.4 (2×CH3), 33.0 (CH2), 41.0 (CH2), 45.2(C), 51.5 (OCH3), 130.7 (C), 132.0 (CH), 133.1 (C),137.1 (C), 142.5 (C), 151.5 (C), 172.82 (CO), 211.6

962

(CO); m/z 274(M+, 100), 260 (17), 259 (97), 242 (27),201 (56), 200 (29), 199 (23), 187 (14), 185 (21).

4.1.3.3. 2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-pro-panoic acid methyl ester(24)

Prepared by reduction of23. White solid (n-Hexane),m.p. 97–98 °C, yield 86%. Found M+ 245.1255(C15H18O3 requires 246.1256;υmax (KBr) 1 738,1 708 cm–1. δH (CDCl3) 2.26 (3H, s, CH3), 2.32 (3H, s,CH3), 2.40 (2H, tJ 9 Hz, CH2), 2.58 (2H, tJ 6 Hz, CH2),2.92 (2H, tJ 6 Hz, CH2), 3.00 (2H, tJ 9 Hz, CH2), 3.67(3H, s, OCH3), 7.35 (1H, s, CH);δC 14.3 (CH3), 19.9(CH3), 24.9 (CH2), 25.2 (CH2), 32.7 (CH2), 36.2 (CH2),51.7 (OCH3), 122.3 (CH), 133.6 (C), 134.8 (C), 135.8(C), 144.2 (C), 152.6 (C), 172.8 (CO), 207.1 (CO); m/z245 (M+, 65), 214 (59), 186 (22), 173 (100), 159 (29),144 (25), 128 (22), 115 (17), 91 (16).

4.1.3.4. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1-one-5-propanoic acid methyl ester(28)

Prepared by reduction of27. Colourless oil yield 91%.Found M+ 274.1570 (C17H22O3 requires 274.1569).υmax

(KBr) 1 739, 1 711 cm–1; δH (CDCl3) 1.15 (6H, s,2×CH3), 2.23 (3H, s, CH3), 2.32 (3H, s, CH3), 2.40 (2H,t J 9 Hz, CH2), 2.80 (2H, s, CH2), 3.00 (2H, tJ 9 Hz,CH2), 3.66 (3H, s, OCH3), 7.45 (1H, s, CH);δC 14.4(CH3), 19.9 (CH3), 25.2 (2×CH3), 25.3 (CH2), 32.6(CH2), 42.1 (CH2), 45.3 (C), 51.7 (OCH3), 123.1 (CH),133.0 (C), 133.5 (C), 136.0 (C), 144.5 (C), 149.6 (C),172.9 (CO), 211.5 (CO); m/z 274 (M+, 76), 260 (26), 259(100), 227 (17), 201 (26), 187 (20), 128 (10).

4.1.3.5. 2,3-Dihydro-4,6-dimethyl-1H-indene-propanoicacid methyl ester(32)

Prepared by reduction of31. White solid (Pet. ether:e-ther), m.p. 42–43 °C, yield 83%. Found M+ 232.1464(C15H20O2 requires 232.1463).υmax (KBr) 2 952, 1 740,1 194cm–1. δH (CDCl3) 2.09 (2H, quintetJ 7 Hz, CH2),2.29 (3H, s, CH3), 2.37 (3H, s, CH3), 2.49 (2H, tJ 9 Hz,CH2), 2.91 (4H, 2×t J 7 Hz, 2×CH2), 3.02 (2H, tJ 9 Hz,CH2), 3.76 (3H, s, OCH3), 6.97 (1H, s, CH);δC 15.8(CH3), 19.9 (CH3), 24.7 (CH2), 24.9 (CH2), 31.9 (CH2),32.9 (CH2), 33.6 (CH2), 51.5 (OCH3), 123.8 (CH), 131.8(C), 134.0 (C), 134.4 (C), 141.3 (C), 141.7 (C), 173.5(CO); m/z 232 (M+, 69), 217 (3), 160 (36), 159 (100), 146(25), 149 (59), 128 (13).

4.1.3.6. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1-one-5-propanoic acid methyl ester(36)

Prepared by reduction of35. Colourless oil, yield 87%.Found M+ 260.1754 (C17H24O2 requires 260.1776).υmax

(neat) 2 952, 1 742, 1 195cm–1; δH (CDCl3) 1.24 (6H, s,2×CH3), 2.23 (3H, s, CH3), 2.35 (3H, s, CH3), 2.52 (2H,

t J 8 Hz, CH2), 2.71 (2H, s, CH2), 2.80 (2H, s, CH2), 2.95(2H, t J 8 Hz, CH2), 3.76 (3H, s, OCH3), 6.86 (1H, s,CH); δC 18.6 (CH3), 18.7 (CH3), 25.6 (CH2), 29.3(2×CH3), 33.7 (CH2), 39.2 (C), 46.3 (CH2), 46.6 (CH2),51.5 (OCH3), 129.5 (CH), 131.7 (C), 131.9 (C), 133.4(C), 139.8 (C), 142.0 (C), 173.4 (CO); m/z 260 (M+, 77),245 (3), 228 (29), 187 (76), 173 (100), 141 (11).

4.1.4. General procedure for the preparation ofpropanoic acids(17, 21, 25, 29, 33 and37)

A solution of the corresponding propanoic acid methylester (1.45mmol) in 4 M HCl (10 mL) and THF (20 mL)was refluxed for 6 h. On completion, the THF wasremoved under vacuum and the remaining mixture waspoured onto iced water (50 mL). The aqueous mixturewas extracted with EtOAc (2× 50 mL). The organic layerwas then washed with NaOH (2× 20 mL) and theaqueous layer was acidified with HCl and extracted intoEtOAc. The EtOAc was dried over anhydrous Na2SO4,filtered and evaporated under vacuum. The residue waspurified by recrystallisation to yield the following;

4.1.4.1 2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-pro-panoic acid(17)

Prepared by HCl reflux from16. White solid (Pet.ether:ether), m.p. 154 °C, yield 88%. Found M+ 232.1075(C14H16O3 requires 232.1099).υmax (KBr) 3 300, 1 732,1 162 cm–1; δH (CDCl3) 2.38 (3H, s, CH3), 2.55 (2H, tJ6 Hz, CH2), 2.57 (3H, s, CH3), 2.67 (2H, tJ 6 Hz, CH2),3.01 (2H, tJ 8 Hz, CH2), 3.04 (2H, tJ 8 Hz, CH2), 6.94(1H, s, CH);δC 17.79 (CH3), 19.25 (CH3), 23.71 (CH2),24.07 (CH2), 33.07 (CH2), 36.80 (CH2), 131.99 (CH),132.76 (C), 133.06 (C), 136.72 (C), 142.66 (C), 154.85(C), 177.84 (CO), 207.83 (CO). m/z 232 (M+, 86), 214(11), 173 (100), 159 (59), 149 (55), 144 (15), 128 (17), 91(11), 71 (18).

4.1.4.2. 2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1-one-6-propanoic acid(21)

Prepared by HCl reflux from20. White needles(EtOH), m.p. 84–86 °C, yield 70%. Found M+ 260.1408(C16H20O3 requires 260.1412). υmax (KBr)3 400–2 800b, 1 737, 1 700 cm–1. δH (CDCl3); 1.20 (6H,s, 2×CH3), 2.37 (3H, s, CH3), 2.51 (2H, tJ 8 Hz, CH2),2.57 (3H, s, CH3), 2.89 (2H, s, CH2), 2.97 (2H, tJ 8 Hz,CH2), 6.95 (1H, s, CH);δC 17.86 (CH3), 19.35 (CH3),23.82 (CH2), 25.47 (2×CH3), 33.21 (CH2), 41.17 (CH2),45.47 (C), 130.84 (C), 132.12 (CH), 132.91 (C), 137.36(C), 142.83 (C), 151.81 (C), 178.42 (CO), 212.26 (CO).m/z 260 (M+, 74), 246 (20), 245 (100), 187 (19), 185(20), 157 (17), 143 (17), 128 (19), 91 (15), 77 (11).

963

4.1.4.3. 2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-pro-panoic acid(25)

Prepared by HCl reflux from24. White needles(EtOH), m.p. 209–210 °C, yield 93%. Found M+

232.1102 (C14H16O3 requires 232.1099).υmax (KBr)3 100–2 800b, 1 731, 1 178 cm–1; δH (CDCl3) 2.33 (3H,s, CH3), 2.40 (3H, s, CH3), 2.69 (2H, tJ 8 Hz, CH2), 2.68(2H, t J 6 Hz, CH3), 3.00 (2H, tJ 6 Hz, CH2), 3.09 (2H,t J 8 Hz, CH2), 7.44 (1H, s, CH);δC 14.17 (CH3), 19.76(CH3), 24.80 (CH2), 24.84 (CH2), 32.46 (CH2), 36.06(CH2), 122.31 (CH), 133.45 (C), 134.68 (C), 135.70 (C),143.90 (C), 152.90 (C), 177.25 (CO), 207.36 (CO); m/z232 (M+, 61), 173 (100), 172 (35), 160 (25), 159 (38),129 (20), 128 (22), 115 (20), 91 (18), 77 (13).

4.1.4.4. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1-one-5-propanoic acid(29)

White cubes (n-Hexane), m.p. 124–125 °C, yield 94%.Found M+ 260.1405 (C16H20O3 requires 260.1412).υmax

(KBr) 3 200, 1 709, 1 692 cm–1; δH (CDCl3) 1.19 (6H, s,2×CH3), 2.28 (3H, s, CH3), 2.37 (3H, s, CH3), 2.49 (2H,t J 8 Hz, CH2), 2.84 (2H, s, CH2), 3.06 (2H, tJ 8 Hz,CH2), 7.41 (1H, s, CH);δC 14.41 (CH3), 19.92 (CH3),25.18 (CH2), 25.26 (2×CH3), 32.79 (CH2), 42.26 (CH2),45.30 (C), 123.27 (CH), 133.10 (C), 133.59 (C), 136.09(C), 144.50 (C), 149.75 (C), 177.41 (CO), 211.78 (CO);m/z 260 (M+, 25), 245 (44), 149 (29), 139 (10), 125 (19),119 (16), 97 (43), 77 (100).

4.1.4.5. 2,3-Dihydro-4,6-dimethyl-1H-inden-5-propanoicacid (33)

Prepared by HCl reflux from32. White solid (EtOH),m.p. 112 °C, yield 89%. Found M+ 218.1302 (C14H18O2

requires 218.1099).υmax(KBr) 2 953, 1 699, 1 297 cm–1;δH (CDCl3) 2.10 (2H, quinJ 7 Hz, CH2), 2.31 (3H, s,CH3), 2.38 (3H, s CH3), 2.54 (2H, tJ 9 Hz, 8 Hz, CH2),2.91 (4H, m, 2×CH2), 3.05 (2H, tJ 9 Hz, CH2), 6.98 (1H,s, CH). δC 15.78 (CH3), 19.92 (CH3), 24.72 (2×CH2),31.96 (CH2), 32.95 (CH2) 33.64 (CH2), 123.84 (CH),131.82 (C), 134.00 (C), 134.17 (C), 141.40 (C), 141.79(C), 179.12 (CO); m/z 218 (M+, 59), 160 (28), 159 (100),145 (44), 128 (14), 115 (10).

4.1.4.6. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5-pro-panoic acid(37)

Prepared by HCl reflux from36. White cubes (EtOH),m.p. 63–64 °C, yield 86%. Found 246.1617 (C16H22O2

requires 246.1620).υmax(KBr) 3 200, 1 736, 1 170 cm–1;δH (CDCl3) 1.16 (6H, s, 2×CH3), 2.16 (3H, s, CH3), 2.28(3H, s, CH3), 2.48 (2H, tJ 9 Hz, CH2), 2.64 (2H, s, CH2),2.72 (2H, s, CH2), 2.89 (2H, tJ 9 Hz, CH2), 6.80 (1H, s,CH); δC 18.67 (2×CH3), 25.36 (CH2), 29.28 (2×CH3),33.62 (CH2), 39.21 (C), 46.28 (CH2), 46.53 (CH2),

129.46 (CH), 131.60 (C), 131.90 (C), 133.53 (C), 139.96(C), 142.02 (C), 178.99 (CO); m/z 246 (M+, 6), 228 (45),187 (57), 186 (38), 185 (16), 173 (83), 85 (17), 72 (100).

4.1.5. General procedure for the preparation of 3≠hydroxypropyl derivatives(18, 26, 30, 34 and38)

Freshly generated diborane was introduced under N2 toa solution of the appropriate propanoic acids (1.3 mmol)in dry THF (10 mL) at 0 °C. After 15 min EtOH wasadded and the solvents were removed under vacuum. Theresidue was purified by column chromatography (pet-ether, EtOAc; 9:1–7:3) to yield the corresponding alco-hols.

4.1.5.1. (±)-2,3-Dihydro-6-(3≠-hydroxypropyl)-5,7-di-methyl-1H-inden-1-ol(18)

Prepared by diborane reduction of17. White powder(n-hexane), m.p. 86 °C, yield 63%. Found M+ 220.1457(C14H20O2 requires 220.1463).υmax(KBr) 3 350b, 2 936,1 465cm–1; δH (CDCl3) 1.74 (2H, m, CH2), 2.08 (1H, m,CH), 2.30 (3H, s, CH3), 2.36 (3H, s, CH3), 2.39 (2H, m,CH), 2.65 (2H, tJ 8 Hz, CH2), 2.81 (1H, dqJ 3,9 Hz,CH), 3.08 (1H, quinJ 8 Hz, CH), 3.69 (2H, tJ 6 Hz,CH2), 5.28 (1H, ddJ 2,7 Hz, CH), 6.87 (1H, s, CH);δC

17.8 (CH3), 18.8 (CH3), 26.2 (CH2), 28.7 (CH2), 32.2(CH2), 35.0 (CH2), 62.7 (CH2), 75.4 (CH), 130.5 (CH),132.2 (C), 133.7 (C), 136.4 (C), 140.6 (C), 142.8 (C); m/z220 (M+, 50), 205 (26), 184 (26), 175 (30), 169 (36), 161(41), 159 (39), 157 (100), 145 (35), 141 (24).

4.1.5.2. 2,3-Dihydro-6-(3≠-hydroxypropyl)-2,2,5,7-tetra-methyl-1H-inden-1-one(22)

LiAlH 4 (15 mg, 0.39 mmol) was added to a solution of20 (200 mg, 0.73 mmol) in dry Et2O (10 mL) and themixture was stirred at 0 °C for 10min. The reactionmixture was poured onto ice/HCl and extracted withEtOAc. The organic layer was evaporated and the residuewas purified by column chromatography on silica geleluant (pet. ether:EtOAc, 7:3), followed by preparativeTLC using the same developer to give22 as a pale oilyield 45%. Found M+ 246.1016 (C16H22O2 requires246.1614). υmax (film) 3 400, 1 705, 1 465 cm–1; δH

(CDCl3) 1.21 (6H, s, CH3), 1.64 (2H, m, CH2), 2.37 (3H,s, CH3), 2.47 (2H, tJ 8 Hz, CH2), 2.57 (3H, s, CH3), 2.89(2H, s, CH2), 4.06 (2H, tJ 7, CH2), 6.95 (1H, s, CH);δC

17.8 (CH3), 19.3 (CH3), 24.0 (CH2), 25.5 (2×CH3), 32.3(CH2), 41.1 (CH2), 45.4 (C), 62.1 (CH2), 131.0 (C), 132.0(CH), 133.2 (C), 137.1 (C), 142.6 (C), 151.6 (C), 215.0(C=O). 246 (M+, 8), 245 (42), 201 (100), 200 (93).

964

4.1.5.3. (±)-2,3-Dihydro-5-(3≠-hydroxypropyl)-4,6-di-methyl-1H-inden-1-ol(26)

Prepared by diborane reduction of25. Waxy solid(EtOH), yield 47%. M+ 220.1431 (C14H20O2 requires220.1463).υmax (KBr) 3 400b, 2 954, 1 457cm–1; δH

(CDCl3) 1.68 (2H, m, CH2), 1.71 (1H, m, CH), 2.09 (1H,m, CH), 2.24 (3H, s, CH3), 2.35 (3H, s, CH3), 2.70 (2H,t J 8 Hz, CH2), 2.77 (1H, m, CH), 2.98 (1H, quinJ 8 Hz,CH), 3.70 (2H, tJ 9 Hz, CH2), 4.89 (1H, ddJ 4 Hz, 7 Hz,CH), 7.09 (1H, s, CH);δC 15.5 (CH3), 20.1 (CH3), 25.8(CH2), 29.4 (CH2), 31.9 (CH2), 32.2 (CH2), 62.7(CH2OH), 83.4 (CH), 124.2 (CH), 131.9 (C), 134.3 (C),138.6 (C), 139.9 (C), 140.9 (C); m/z 220 (M+, 67), 175(100), 169 (19), 160 (16), 131 (33), 115 (35), 91 (29).

4.1.5.4. (±)-2,3-Dihydro-5-(3≠-hydroxypropyl)-2,2,4,6-tetramethyl-1H-inden-1-ol(30)

Prepared by diborane reduction of29. White needles(EtOH), m.p. 130–131 °C yield 85%. M+ 248.1773(C16H24O2 requires 220.1776).υmax (KBr) 3 600b,1 464 cm–1. δH (acetone D6) 0.98 (3H, s, CH3), 1.13 (3H,s, CH3), 2.17 (3H, s, CH3), 2.24 (3H, s, CH2), 2.28 (3H,s, CH3), 2.48 (1H, dJ 15 Hz, CH), 2.66 (1H, dJ 14 Hz,CH), 2.70 (2H, tJ 8 Hz, CH2), 2.90 (1H, bs, OH), 3.62(2H, t J 7 Hz, CH2), 3.97 (1H, dJ 5 Hz, OH), 4.56 (1H,d J 5 Hz, CH), 6.94 (1H, s, CH);δC 15.4 (CH3), 20.1(CH3), 25.9 (CH2), 27.2 (CH2), 32.1 (CH2), 43.6 (C),44.2 (CH2), 62.6 (CH2), 83.6 (CH), 123.8 (C), 132.0 (C),134.4 (C), 138.3 (C), 138.7 (C), 141.4 (C); m/z 248 (M+,94), 247 (28), 233 (15), 203 (46), 189 (100), 186 (32),171 (21), 91 (14).

4.1.5.5. 2,3-Dihydro-5-(3≠-hydroxypropyl)-4,6-dimethyl-1H-indene(34)

Prepared by diborane reduction of33. White needles(EtOH), m.p. 51–52 °C yield 64%. Found M+ 204.1508(C14H20O requires 204.1514).υmax (KBr) 3 400b,1 459cm–1; δH (CDCl3) 1.80 (1H, quinJ 8 Hz, CH2), 2.08(2H, quinJ 8 Hz, CH2), 2.28 (3H, s, CH3), 2.35 (3H, s,CH3), 2.74 (2H, tJ 8 Hz, CH2), 2.90 (4H, 2×t J 8 Hz,2×CH2), 3.78 (2H, tJ 8 Hz, CH2), 6.96 (1H, s, CH);δC

15.8 (CH3), 20.0 (CH3), 24.7 (CH2), 25.7 (CH2), 32.0(CH2), 32.4 (CH2), 32.9 (CH2), 63.0 (CH2), 123.7 (CH),131.7 (C), 133.9 (C), 136.0 (C), 141.0 (C), 141.2 (C); m/z204 (M+, 61), 160 (39), 159 (100), 146 (19), 145 (28),129 (11).

4.1.5.6. 2,3-Dihydro-5-(3≠-hydroxypropyl)-2,2,4,6-tetra-methyl-1H-indene(38)

Prepared by diborane reduction of37. Pale oil, yield59%. Found M+ 232.1785 (C16H24O requires 232.1821).υmax (film) 3 400b, 1 464 cm–1. δH (CDCl3) 1.18 (6H, s,2×CH3), 1.76 (2H, m, CH2), 2.21 (3H, s, CH3), 2.32 (3H,

s, CH3), 2.60 (2H, tJ 7 Hz, CH2), 2.66 (2H, s, CH2), 2.72(2H, s, CH2), 3.76 (2H, tJ 6 Hz, CH2), 6.86 (1H, s, CH);δC 15.6 (CH3), 18.6 (CH3), 25.7 (CH2), 29.2 (2×CH3)32.5 (CH2), 39.1 (C), 46.3 (CH2), 47.1 (CH2), 62.8(CH2), 123.9 (CH), 129.3 (C), 131.8 (C), 132.4 (C), 139.6(C), 140.3 (C); m/z 232 (M+, 100), 187 (90), 173 (90),159 (58), 128 (40).

4.1.5.7. 2,3-Dihydro-5-(3≠-hydroxypropyl)-2,2,4,6-tetra-methyl-1H-inden-1-one(39)

A solution of NaOCl (5 mL, 12.5% w/v) was addeddropwise to a stirred solution of30 (250 mg, 1 mmol) inglacial acetic acid (3 mL) at 0 °C. The reaction mixturewas stirred for 30 min and was then poured onto ice andextracted with EtOAc. The organic layer was washedwith H2O, dried over Na2SO4 and evaporated undervacuo. The residue was purified by column chromatog-raphy on silica gel (eluant: pet. ether:EtOAc, 7:3) to yield39 as a waxy solid yield 43%. Found M+ 246.1599(C16H22O2 requires 246.1614).υmax (KBr) 3 400b,1 712 cm–1; δH (CDCl3) 1.23 (6H, s, 2×CH3), 1.70 (2H,m, CH2), 2.28 (3H, s, CH3), 2.38 (3H, s, CH3), 2.80 (2H,t J 8 Hz, CH2), 2.86 (2H, s, CH2), 3.77 (2H, tJ 7 Hz,CH2), 7.43 (1H, s, CH);δC 15.9 (CH3), 20.0 (CH3), 25.3(2×CH3), 25.8 (CH2), 31.7 (CH2), 43.6 (C), 44.2 (CH2),62.9 (CH2), 123.0 (CH), 132.7 (C), 134.5 (C), 138.3 (C),141.4 (C), 146.4 (C), 209.5 (C=O); m/z 246 (M+, 73), 231(40), 215 (100), 128 (22).

4.2. Pharmacological methods

Smooth muscle relaxant activity was assessed as de-scribed previously [1]. Guinea-pigs (250–400 g) of eithersex were killed by cervical dislocation and exsan-guination. The abdomen was opened by midline incisionand the ileum removed. The tissue was stored at 4 °C inKreb’s solution (composition (mM): NaCl 118, KCl 4.7,CaCl2 2.5, MgCl2 1.15, NaH2PO4 1.17, NaHCO3 25, glu-cose 14.4). Segments of ileum 2.5 cm in length weresuspended in a high potassium calcium-free modifiedKreb’s solution (composition (mM): NaCl 12.5, KCl 45,MgCl2 1.15, NaH2PO4 1.17, NaHCO3 25, glucose 11.1)at 37 °C, gassed with 95% O2/5% CO2 under a restingtension of 1.5 g, from Grass FT.03 transducers fitted withblack springs to record contractions isometrically. Con-tractions were displayed on a Grass 79D oscillograph.Sustained (> 40 min) contactures were elicited by addi-tion of CaCl solution sufficient to raise the calciumconcentration in the bath to 2.5 mM. When contractureshad reached a stable maximum, test compounds wereadministered at a single concentration of 1× 10–5 M or inthe case of sufficiently active compounds, they wereadded cumulatively (3× 10–8–3 × 10–5M). Compounds

965

were dissolved in 0.5% ethanol, addition of which(0.1 mL) induced an inhibition of calcium contractions ofapproximately 14%. However, this inhibition was tempo-rary, and contractions returned to control levels within5 min. In contrast, the inhibitory effect of all compoundswas sustained, and subsequently all responses weremeasured at least 10 min following addition of com-pound.

5. Statistics

Results are expressed as means± SE, mean of %inhibition of calcium contractures at a single dose of 1×10–5 M. EC50 values were determined from the linearportion of individual dose/inhibition curves by regressionanalysis, and expressed as a mean± SE, mean for eachcompound shown. Statistical analysis was performedusing Studentst-test for unpaired samples, with a prob-ability value (P) of less than 0.05 being taken assignificant.n denotes the number of preparations used inthat series of experiments.

Acknowledgements

We would like to acknowledge an EOLAS grant whichsupported this work.

References

[1] Sheridan H., Lemon S., Frankish N., McCardle P., Higgins T., JamesJ.P., Bhandari P., Eur. J. Med. Chem. 25 (1990) 603–608.

[2] Farrell R., Kelleher F., Sheridan H., J. Nat. Prod. 59 (1996) 446–447.

[3] Murakami T., Tanaka N., in: Herz W., Griesbeck H., Kirby G.,Tamm C. (Eds.), Progress in the Chemistry of Organic NaturalProducts Vol 54, Springer-Verlag, Vienna, New York, 1988, pp.51–56 and 176–195.

[4] Bardouille V.S., Mooto B., Hirotsu K., Clardy J., Phytochemistry 17(1978) 275–278.

[5] Sheridan H., Frankish N., Farrell R., Planta Medica 65 (1998)271–272.

[6] Cabri W., Candiani I., Acc. Chem. Res. 28 (1995) 2.

[7] Heck R., Palladium Reagents in Organic synthesis, Academic Press,London, 1985, p. 415.

[8] Ho S., Yang M., Wu T., Wang C., Planta Medica 51 (2) (1985)148–150.

[9] Yang M., Planta Medica 52 (1) (1986) 25–27.

966

Original article

Synthesis and pharmacological evaluation of carboxamide derivativesas selective serotoninergic 5-HT4 receptor agonists

Katsuhiko Itoh*, Koji Kanzaki, Tsuguo Ikebe, Takanobu Kuroita, Hideo Tomozane,Shuji Sonda, Noriko Sato, Keiichiro Haga, Takeshi Kawakita

Research Laboratories, Yoshitomi Pharmaceutical Industries Ltd., 955 Koiwai, Yoshitomi-cho, Chikujo-gun, Fukuoka 871-8550, Japan

(Received 29 September 1998; revised 8 December 1998; accepted 9 December 1998)

Abstract – A number of new carboxamide derivatives were synthesized. The affinity of these compounds for the serotoninergic 5-HT4receptor was evaluated by use of radioligand-binding techniques. The agonistic activity was evaluated as the contractile effect of the ascendingcolon isolated from guinea-pigs. Among these compounds, 4-amino-5-chloro-2-methoxy-N-[1-[2-[(methylsulfonyl)amino]ethly]-4-piperidinylmethyl]benzamide (24) showed a high affinity for the 5-HT4 receptor (Ki = 9.6 nM). Compound24 displayed a higher affinity for5-HT4 receptors than the other receptors, including, 5-HT3 and dopamine D2 receptors. In addition, compound24was confirmed to be a potent5-HT4 receptor agonist (ED50 = 7.0 nM). An interaction model between compound24 and 5-HT4 receptor was proposed. © 1999 Éditionsscientifiques et médicales Elsevier SAS

5-HT4 receptor / 5-HT4 receptor agonist / structure–activity relationship / carboxamide / receptor model construction

1. Introduction

Serotonin (5-HT) is a neurotransmitter responsible fora wide range of pharmacological reactions. Serotonergicreceptors are now classified into four broad subtypes suchas 5-HT1, 5-HT2, 5-HT3, and 5-HT4 receptors, and clonesfor additional subtypes termed 5-HT5, 5-HT6, and 5-HT7receptors have been identified [1]. Among diverse 5-HTreceptors, we have investigated 5-HT4 function. Activa-tion of the 5-HT4 receptor mediates widespread effects inthe central and peripheral nervous systems [2]. Therefore,selective 5-HT4 receptor ligands would be useful forelucidating the physiological function of this receptor.Agonists for the 5-HT4 receptor known to date are indolederivatives such as 5-methoxytryptamine, benzamide de-rivatives such as cisapride [3–5], benzimidazolone de-rivatives such as BIMU8 [6], and benzoate derivativessuch as ML 10302 [7](figure 1). However, nonselectiveaffinity or lability of these compounds prevents purepharmacological characterization of this receptor in vivo.For example, 5-methoxytryptamine activates all 5-HTreceptor subtypes except the 5-HT3 receptor [8].

Cisapride and BIMU8 show obvious affinity for the5-HT3 receptor [9, 10]. While, ML 10302 is a selective5-HT4 receptor agonist, it shows limited pharmacologicalactivities in vivo because of possible hydrolysis at itsbenzoate moiety. Recently, some ketone derivatives werereported to show a high affinity for the 5-HT4 receptorand to be selective 5-HT4 partial agonists [11]. In addi-tion, the original 5-TH4 receptor mapping by the activeanalogue approach by using several 5-HT4 receptorantagonists and inactive ligands was reported [12].

One of our purposes is to supply a useful tool forelucidating the physiological function of the 5-HT4 re-ceptor. In the course of our synthetic study on 5-HT3

receptor antagonists [13], it has been already clarified thata benzoxazine-8-carboxamide derivative1 (figure 2)shows a weak affinity for the 5-HT4 receptor. This resultlead us to design selective 5-HT4 receptor agonistshaving novel chemical structure and chemical stability.Thus, compound1, a leading compound for this study,could further be optimized at four points: optimization of(1) the structure of the cycloamine moiety, (2) thedistance between the amide nitrogen in the carboxamidemoiety and the basic nitrogen in the cycloamine moiety,*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 977−989 977© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

(3) the variety and number of the substituents on thearomatic moiety, and (4) the structure of the side chain onthe cycloamine moiety. Herein, we describe the design,synthesis, and pharmacological evaluation of carboxam-ide derivatives as selective 5-HT4 receptor agonists.Interaction between putative 5-HT4 receptor models andagonistic ligand is also discussed here.

2. Chemistry

The general synthetic procedure used in this study isillustrated infigure 3. By the mixed anhydrides method(method A) or the 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrochloride (WSC) method (method B),carboxylic acids (2a–2h) were coupled with appropriate

amines (4a, 4b and 6a–6f) to afford target benzamides(7–24), on which the synthetic data are listed intable I.Among the acidic starting materials, 3,4-dihydro-1,4-benzoxazine-8-carboxylic acid (2a) was prepared by ourmethod reported previously [14], and 2,3-dihydro-5-chlorobenzofuran-7-carboxlic acid (2b) [15] and5-chloro-2-methoxy-4-methylaminobenzoic acid (2h)[16] were prepared by known procedures. As for the othergroup of starting materials, pyrrolidine-3-methanol (3)was condensed with phthalimide by the Mitsunobu reac-tion [17] to afford a corresponding phthalimide deriva-tive, which gave 1-benzyl-3-pyrrolidinylmethylamine(4a) through a hydrazinolysis(figure 4). 1-Substituted-4-piperidinylmethylamines (6a–6d) were prepared fromtheir corresponding isonipecotamides (5a–5d) by lithiumaluminium hydride reduction(figure 5). 2-(1-Benzyl-4-piperidiyl)ethylamine (6f) and [1-(2-methansufonyl-amino)ethyl]-4-piperidylmethylamine (6g) were preparedusing our procedures [18].

3. Pharmacological data and discussion

3.1. Structure–activity relationships

The affinity of compounds7–24 for the 5-HT4 receptorwas determined as their ability to inhibit the binding of[3H]GR113808 to the receptor. Their affinities for 5-HT3

and dopamine D2 receptors were similarly evaluated byusing [3H]granisetron and [3H]spiperone as radioligand,respectively. Here, membrane preparations of the stria-

Figure 1. Chemical structures of 5-HT4 receptor agonists.

BIMU 8 ML 10302

Figure 2. Chemical structure of compound1.

978

tum of guinea-pigs, the cerebral cortex of rats, and thestriatum of rats were used for 5-HT4, 5-HT3, and D2

receptor binding assays, respectively. Either agonistic orantagonistic activity of these compounds was evaluatedas the contractile ability of the ascending colon ofguinea-pigs.

Pharmacological data on compounds1 and 7–24,4-amino-N-(1-benzyl-4-piperidinyl)-5-chloro-4-methoxy-

benzamide (clebopride), and 5-HT are listed intable II,where clebopride and 5-HT are reference compounds. Asconcerns the cycloamine part, we chose not the bicy-cloamine group (tropane) but the monocycloamine group(pyrrolidine and piperidine) to obtain the compoundshaving a higher affinity for 5-HT4 receptor than for 5-HT3

Figure 3. The general synthetic procedures used in this study. (a) method A: i-BuOCOCl, Et3N, AcOEt or method B: WSC, HOBt,DMF.

Figure 4. Condensation of3 by the Mitsunobu reaction. (a)PPh3, phthalimide, DEAD, THF; (b) NH2NH2H2O, EtOH.

Figure 5. Lithium aluminium hydride reduction of5a–5d.(a)LiA1H4, THF.

979

Table I Physicochemical properties for compounds7–24.

Compound Ar n X R Method Mp (°C) Formula

7 0 Bna A 94–95 C21H24CIN3O2

8 0 Bna A 232–233 C22H26CIN3O2C4H4O4f

9 1 Bna A 186–187 C22H26CIN3O2C2H4O4g

10 1 Bna A 97–98 C23H28CIN3O2

11 2 Bna A 122–124 C24H30CIN3O2C2H2O41/2EtOH1/2H2O

g

12 1 Bna B 176–178/DEC C22H25CIN2O2C2H2O4g

13 1 Bna B 161–163 C20H24CIN3O2C2H2O4g

14 1 Bna B 155–156 C21H26CIN3O2

15 2 Bna B 135–137 C22H28CIN3O2C2H2O4g

16 1 Bna B 169–170 C21H26N2O2C2H2O4g

17 1 Bna B 90–93 C21H25CIN2O2C2H2O4 1/4H2Og

18 1 Bna B 110–112/DEC C20H24CIN3O2C2H2O4 4/5H2Og

980

receptor [7]. As a result, monocycloamine derivatives (7and 8) in benzoxazine series showed a low affinity for5-HT3 receptor. Furthermore, piperidine derivatives (10and14) showed a higher affinity for 5-HT4 receptor thanpyrrolidine derivatives (9 and13).

Next, we investigated the distance requirements be-tween the carbonyl group in the amide bond and the basicnitrogen in the cyclic amine of pyrrolidine and piperidinederivatives. Here, two benzoxazinecarboxamide deriva-tives (7 and8) and clebopride, where the hydrogen atomon their carboxamide nitrogen is replaced by a 1-benzyl-3-pyrrolidinyl or 1-benzyl-4-piperidinyl group, showedaffinities for 5-HT3, 5-HT4 and D2 receptors. On the otherhand, compounds9, 10 and12–14, where the hydrogenatom on their carboxamide nitrogen is replaced by a1-benzyl-3-pyrrolidinylmethyl or 1-benzyl-4-piperi-dinylmethyl group, showed a higher affinity for the5-HT4 receptor than for the 5-HT3 receptor, and showeda low affinity for the D2 receptor. Compounds11 and15,where the hydrogen atom on their carboxamide nitrogenis replaced by a 2-(1-benzyl-4-piperidinyl)ethyl group,exhibited only low affinities for these kinds of receptors.The results indicate that the optimum distance for selec-tive affinity for the 5-HT4 receptor was one carbon [9].

Compound10 was confirmed to be a 5-HT4 receptorantagonist (pKB = 7.9). Compound12, a benzofurananalogue of benzoxazine derivative10, was also con-firmed to behave as a 5-HT4 receptor antagonist (pKB =8.2) although they showed a high affinity for the receptor.However, compounds13 and 14, both of which have aclassic 4-amino-5-chloro-2-methoxybenzamide skeleton,were confirmed to be a 5-HT4 receptor agonist. We nextfocused on the effect of each substituent (4-amino,5-chloro and 2-methoxy group) on the phenyl ring inbenzamide derivatives. A simple 2-methoxybenzamidederivative 16 showed a low affinity to the receptor.Although 5-chloro-2-methoxybenzamide derivative17exhibited a 5-HT4 receptor affinity similar to compound14, the former never behaved as a agonist for the receptor.On the other hand, such an affinity was practicallyundetectable for the 4-amino-3-chlorobenzamide deriva-tive 18, a 2-desmethoxy analogue of compound14. Themoderate affinity of compound16 and the low affinity ofcompound 18 indicate that the 2-methoxy group isessential for binding to the 5-HT4 receptor. Compound17showed a similar affinity to compound14. This resultsuggests that the 5-chloro group helped ligands to bind tothe receptor complementarily. The 4-Amino-2-

Table I (Continued)

Compound Ar n X R Method Mp (°C) Formula

19 1 Bna B 160–161 C21H27N3O2C2H2O4g

20 1 Bna B 148–149 C22H28CIN3O2C4H4O4f

21 1 Etb B 115–116 C16H24N3O2

22 1 Buc B 179–181 C18H28CIN3O2C2H2O4g

23 1 Hexd B 113–115 C20H32CIN3O2

24 1 MSAEe B 177–178/DEC C17H27CIN4O4S2C2H2O4g

aBn, benzyl;bEt, ethyl; cBu, butyl; dHex, hexyl;eMSAE, (methylsulfonylamino)ethyl;ffumalate;goxalate.

981

methoxybenzamide derivative19, a 5-deschloro analogueof compound14, showed a potent 5-HT4 receptor ago-nistic activity, but showed only a moderate affinity for thereceptor. These results indicate that the 4-amino groupplayed an important role in 5-HT4 receptor agonisticactivity. In addition, this consideration is supported by thefacts, which 2,3-dihydro-4-amino-5-chlorobenzofurancarboxamides showed 5-HT4 receptor agonistic activ-ity [19], while 4-desamino-2,3-dihydro-5-chlorobenzo-furan carboxamide12 showed low 5-HT4 receptor ago-nistic activity. Although 5-chloro-2-methoxy-4-methylaminobenzamide derivative20, an N-methylanalogue of compound14, also exhibited a moderateaffinity, it showed low 5-HT4 receptor agonistic activity.Consequently, all the three substituents on the phenyl ringare essential for both affinity and agonistic activity.

Finally, we examined the influence of the side chains ofN-substituted piperidinylmethyl group on the interactionwith the 5-HT4 receptor. Ethyl derivative21 showed amoderate affinity for the 5-HT4 receptor. Butyl derivative22was almost equal to compound14 in the affinity, hexylderivative23 was equal to compound14 in the affinity,but with reduced 5-HT4 receptor agonistic activity. Whenthe substituent was a polar methylsulfonylamino group,agonistic activity increased. Compound24 was a more

potent agonist at the 5-HT4 receptor than other com-pounds (14, 21–23). Therefore, compound24 was as-sayed for its binding ability for several kinds of receptors.The results were as follows (IC50 value (nM), ligand):α1 > 1 000, [3H]prazosin; 5-HT1A > 1 000, [3H]8-OH-DPAT; 5-HT2 > 1 000, [3H]ketanserin; MACh> 1 000,[3H]QNB. Thus Compound24 represented a selective5-HT4 receptor agonist.

3.2. Computational modelling study

3.2.1. 5-HT4 receptor model constructionA 5-HT4 receptor model was constructed in order to

obtain rational structure–activity relationships and toclarify the interaction between ligands and the receptor.The 5-HT4 receptor is a member of the G-protein coupledreceptor (GPCR). Generally, GPCRs are assumed toconsist of seven transmembrane helices and eight intraand extracellular loops and their models are constructedfrom coordinates of bacteriorhodopsin [20–27]. The cred-ibility of the model is evaluated by the consistency withstructure–activity relationships and mutational data. Inthis study, affinity for the guinea-pig 5-HT4 receptor wasmeasured. Therefore the receptor model should be con-structed from the sequence of guinea-pig. But the guinea-

Table II Pharmacological data of compound1, compounds7–24and two reference compounds.

Binding affinities, Ki (nM)a Contractile effects in guinea pigs ascending colon

Compound 5-HT4 5-HT3 D2 EC 50 (nM)d Maximal response (%)e

1 280 0.27 > 1000b NTc –7 160 10 12 NTc –8 530 21 67 NTc –9 19 95 > 1000b > 1000010 0.93 120 > 1000b > 1000011 530 > 1000b > 1000b NT –12 1.7 340 NT > 1000013 20 160 > 1000b 47 2114 6.7 290 > 1000b 20 1815 > 1000b 470 > 1000b NTc –16 66 > 1000b 880 NTc –17 8.1 > 1000b NTc > 1000018 > 1000b > 1000b NTc NTc –19 110 > 1000b NTc 150 2020 100 > 1000b NTc 900 8.921 28 > 1000b NTc 61 1622 8.8 > 1000b NTc 13 1523 6.7 > 1000b NTc 130 2024 9.6 > 1000b > 1000b 7.0 27clebopride 92 230 63 NTc –5-HT 130 35 21

aEach value is the mean from triplicate assays in a single experiment;bIC50 value; cnot tested;dEC50 values were determined by linearregression;e% of contraction to methacoline at 30µM.

982

pig sequence has not been reported yet. Although se-quences of receptors are not identical among animals,amino acids involved in ligand recognition are mostlikely to be conserved. Therefore, in the agonist recogni-tion site, the conserved amino acids between human andrat 5-HT4 receptors should be conserved in guinea-pig5-HT4 receptors. The 5-HT4 receptor model was con-structed with the residues in the human 5-HT4 receptorand the conserved residues were used for speculation ofreceptor-ligand interaction. The seven transmembraneregions of the 5-HT4 receptor was determined from ahydropathy plot [28]. To determine sequence alignmentof bacteriorhodopsin and the 5-HT4 receptor, over 70GPCRs such as dopamine, adrenergic, muscarine recep-tor, and rhodopsin were included in the study.

3.2.2. Docking of compound24 into the 5-HT4 receptormodel

Mutational studies for the 5-HT4 receptor have notbeen reported yet. Therefore, the structure–activity rela-tionships for this series of compounds and mutationaldata for the 5-HT1A and 5-HT2A receptors [29–33] wereused for the docking study of compound24 into thereceptor model. From the structure–activity relationships,the 2-methoxy, 5-chloro, and 4-amino groups on thephenyl ring, and the methylsulfonylamino group in thepiperidinomethyl moiety were defined to be essential forboth affinity and agonistic activity. These groups are

assumed to interact with certain residues in the 5-HT4

receptor. From some mutational data, Asp308, Thr312,Ser507, Ala510, and Phe619 are defined to be importantfor agonist binding and receptor activation. Asp308 issupposed to form a salt bridge with the ammoniumnitrogen of the ligand. Thr312 and Ser507 are supposedto form a hydrogen bond. Ala510 is supposed to performa hydrophobic interaction withN-alkyl substituents insome tryptamine derivatives. Phe619 is assumed to rec-ognise an aromatic ring of the ligand. From thesestructure–activity relationships and mutational data, aninteraction model between the 5-HT4 receptor and com-pound24 was proposed :(figure 6)the ammonium nitro-gen in the piperidine ring forms a salt bridge withAsp308. The 2-methoxy group forms an intramolecularhydrogen bond with the NH group in the benzamide andconstrains the conformation of the benzamide group. The5-chloro group performs a hydrophobic interaction withAla510. The 4-amino group forms a hydrogen bond withSer507. The methylsulfonylamino group forms a hydro-gen bond with Thr312. Furthermore, aromatic residuessuch as Phe515, Phe612, Trp616, Phe619, and Phe620 arepresumed to locate near the ligand and they are supposedto form an aromatic binding pocket or stabilize thepositive charge of the cationic amine. As for Trp409,there may be a hydrogen bond to the CO group in thebenzamide.

Figure 6. Stereoview of the interaction of compound24 with the 5-HT4 receptor.

983

4. Conclusion

We describe the synthesis and biological evaluationof a series of carboxamides as selective 5-HT4 receptoragonists. Based on our results, we propose that 4-amino-5-chloro-2-methoxy-N-[(4-piperidinyl)methyl]benzamidewith a polar group at the 1-position on piperidinemoieties are necessary for the selective agonistic activityfor 5-HT4 receptors. Compound24 represents a selective5-HT4 agonist and would be a useful tool for probing the5-HT4 receptor function in vivo. Detailed analyses ofstructure–activity relationships for the side chain moietieswill be reported in due course.

5. Experimental protocols

5.1. Chemistry

All melting points were measured in open capillariesand uncorrected. Proton nuclear magnetic resonance(1H-NMR) spectra were recorded on Jeol JNM-EX270spectrometers and chemical shifts are expressed in ppmwith tetramethylsilane (TMS) as an internal standard.Signal multiplicities are represented by s (singlet), d(doublet), t (triplet), q (quartet), br-s (broad singlet) andm (multiplet). Mass spectra (MS) were taken on JeolJMS-O1SG spectrometers. Elementary analysis was per-formed for C, H and N, and were within 0.4% of thecalculated values. Silica-gel plates (Merck F254) andsilica gel 60 (Merck, 70–230 mesh) were used foranalytical and preparative column chromatography, re-spectively.

5.1.1. General procedure for the preparation of7–11(method A)

5.1.1.1. N-(1-Benzyl-3-pyrrolidinyl)-6-chloro-3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide7

Isobutyl chloroformate (1.2 g, 9.7 mmol) was added toa mixture of2a (2.0 g, 8.8 mmol), triethylamine (2.0 g,19 mol), and ethyl acetate (40 mL) at –10 °C. The mix-ture was stirred below –5 °C for 30 min and a solution of4b (1.5 g, 8.8 mmol) in ethyl acetate (10 mL) was addedwith stirring at –10 °C. Stirring was continued at the sametemperature for 30 min. The resulting mixture was addedto water and extracted with ethyl acetate. The extract waswashed with brine, and dried over anhydrous magnesiumsulfate. After evaporation in vacuo, the residue waschromatographed on silica gel (CHCl3:MeOH = 10:1).Recrystallized from ethyl acetate/diisopropylether to give7 (0.50 g, 15%);1H-NMR (CDCl3) δ: 1.70–1.82 (2H, m),2.64 (2H, d,J = 11 Hz), 2.90 (3H, s), 3.35 (2H, t,J = 5.6

Hz), 3.55 (1H, d,J = 11 Hz), 3.70 (1H, d,J = 11 Hz), 4.35(2H, t, J = 5.6 Hz), 4.44–4.80 (1H, m), 6.62 (1H, d,J =3.0 Hz), 7.20–7.30 (5H, m), 7.39 (1H, d,J = 3.0 Hz), 7.95(1H, br-s); MS m/z: 385 (M+).

5.1.1.2. N-(1-Benzyl-4-piperidinyl)-6-chloro-3,4-dihy-dro-4-methyl-2H-1,4-benzoxazine-8-carboxamide fuma-late 8

Similarly to 7, 8 was prepared starting from2a (3.4 g,15 mmol), triethylamine (3.3 g, 33 mmol), isobutyl chlo-roformate (2.3 g, 17 mmol), ethyl acetate (70 mL), and6e(3.0 g, 15 mmol). The resulting oil was transformed intofumalate and recrystallized from ethanol to give8 (3.4 g,55%); 1H-NMR (DMSO-d6) δ: 1.54 (2H, dd,J = 12, 24Hz), 1.82 (2H, d,J = 12 Hz), 2.20–2.39 (2H, t,J = 12 Hz),2.80 (2H, d,J = 12 Hz), 2.87 (3H, s), 3.29 (2H, t,J = 4.0Hz), 3.50 (2H, s), 3.70–3.83 (1H, m), 4.29 (2H, t,J = 4.0Hz), 6.60 (2H, s), 6.75 (1H, d,J = 3.0 Hz), 6.82 (1H, d,J = 3.0 Hz), 7.22–7.38 (5H, m), 7.98 (1H, d,J = 7.2 Hz);MS m/z: 399 (M+).

5.1.1.3. N-[(1-Benzyl-3-pyrrolidinyl)methyl]-6-chloro-3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamideoxalate9

Similarly to 7, 9 was prepared starting from2a (1.8 g,7.9 mmol), triethylamine (1.6 g, 16 mmol), isobutyl chlo-roformate (1.1 g, 7.9 mmol), ethyl acetate (30 mL), and4a (1.5 g, 7.9 mmol). The resulting oil was transformedinto oxalate and recrystallized from methanol to give9(2.4 g, 62%);1H-NMR (DMSO-d6) δ: 1.60–2.22 (2H,m), 2.64–2.75 (2H, m), 2.80–2.92 (2H, m), 2.90 (3H, s),3.17 (2H, d,J = 5.9 Hz), 3.55 (2H, s), 3.78–3.95 (1H, m),4.21 (2H, t,J = 4.0 Hz), 4.29 (2H, t,J = 4.0 Hz), 6.74(1H, d, J = 3.0 Hz), 6.78 (1H, d,J = 3.0 Hz), 7.25–7.60(5H, m), 8.28 (1H, t,J = 6.0 Hz); MS m/z: 399 (M+).

5.1.1.4. N-[(1-Benzyl-4-piperidinyl)methyl]-6-chloro-3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide10

Similarly to7, 10was prepared starting from2a (1.2 g,5.3 mmol), triethylamine(0.80 g, 8.0 mmol), isobutylchloroformate (0.67 g, 5.3 mmol), ethyl acetate (25 mL),and6a (1.0 g, 5.3 mmol). The resulting solid was recrys-tallized from ethyl acetate to give10 (1.4 g, 64%);1H-NMR (CDCl3) δ: 1.23–1.32 (2H, m), 1.55 (1H, br-s),1.73 (2H, d,J = 11 Hz), 2.00–2.42 (2H, m), 2.80 (2H, d,J = 11 Hz), 2.90 (3H, s), 3.30 (2H, t,J = 5.6 Hz), 3.35(2H, d,J = 11 Hz), 3.50 (2H, s), 4.35 (2H, t,J = 5.6 Hz),6.64 (1H, d,J = 3.0 Hz), 7.20–7.30 (5H, m), 7.42 (1H, d,J = 3.0 Hz), 7.69 (1H, br-s); MS m/z: 413 (M+).

984

5.1.1.5. N-[2-(1-Benzyl-4-piperidinyl)ethyl]-6-chloro-3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamideoxalate11

Similarly to7, 11was prepared starting from2a (1.0 g,4.6 mmol), triethylamine (1.0 g, 10 mmol), isobutyl chlo-roformate (0.70 g, 5.1 mmol), ethyl acetate (50 mL), and6f (1.0 g, 4.6 mmol). The resulting oil was transformedinto oxalate and recrystallized from ethanol to give11(0.56 g, 24%);1H-NMR (DMSO-d6) δ: 1.32–1.60 (4H,m), 1.84 (2H, d,J = 12 Hz), 2.80 (2H, d,J = 12 Hz), 2.84(2H, t, J = 11 Hz), 2.87 (3H, s), 3.29 (5H, m), 4.17 (2H,s), 4.28 (2H, t,J = 4.6 Hz), 6.73 (1H, d,J = 2.7 Hz), 6.82(1H, d, J = 2.7 Hz), 7.41–7.49 (5H, m), 8.12 (1H, t,J =5.2 Hz); MS m/z: 427 (M+).

5.1.2. General procedure for the preparation of12–24(method B)

5.1.2.1. 4-Amino-N-[(1-benzyl-3-pyrrolidinyl)methyl]-5-chloro-2-methoxybenzamide oxalate13

A mixture of 2c (1.6 g, 7.9 mmol), 4a (1.5 g,7.9 mmol), 1-hydroxybenzotriazole (HOBt) (1.1 g,8.1 mmol), and dimethylformamide (50 mL) was stirredunder ice-cooling for 1 h and then 1-ethyl-3-[3-(dimethylamino)propyl] carbodiimide hydrochloride(WSC) (2.3 g, 8.4 mmol) was added at the same tempera-ture. Stirring was continued overnight at room tempera-ture. After evaporation, 5% aqueous sodium bicarbonatewas added to the residue and extracted with ethyl acetate.The extraction was washed with brine, and was dried overanhydrous magnesium sulfate. After evaporation invacuo, the residue was dissolved in ethanol, an alcoholicsolution of oxalic acid (2.5 equiv.) was added. Theprecipitates were collected and recrystallized from etha-nol to give 13 (1.9 g, 52%);1H-NMR (DMSO-d6) δ:1.60–2.22 (2H, m), 2.64–2.75 (2H, m), 2.80–2.92 (3H,m), 3.17 (2H, d,J = 5.9 Hz), 3.55 (2H, s), 3.80 (3H, s),4.23 (2H, br-s), 6.48 (1H, s), 7.25–7.55 (5H, m), 7.62(1H, s), 8.02 (1H, t,J = 6.0 Hz); MS m/z: 373 (M+).

5.1.2.2. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro-3,4-dihydro-1,4-benzofuran-7-carboxamide oxalate12

Similarly to 13, 12 was prepared starting from2b(0.50 g, 2.5 mmol),6a (0.49 g, 2.5mmol), HOBt (0.33 g,2.5 mmol), dimethylformamide (20 mL), and WSC(0.47g, 2.5 mmol). The resulting oil was transformed intooxalate and recrystallized from isopropanol to give12(0.89 g, 77%),1H-NMR (CDCl3) δ: 1.65–2.02 (6H, m),2.45–2.60 (2H, br-s), 3.16 (2H, t,J = 6.0 Hz), 3.30 (2H,br-s), 3.58 (2H, br-s), 4.18 (2H, br-s), 4.72 (2H, t,J = 6.0Hz), 7.30–7.42 (5H, m), 7.62 (1H, t,J = 6.0 Hz), 8.82(1H, d, J = 2.0 Hz); MS m/z: 384 (M+).

5.1.2.3. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-5-chloro-2-methoxybenzamide14

Similarly to 13, 14 was prepared starting from2c(1.0 g, 5.0 mmol),6a (1.0 g, 5.0 mmol), HOBt (0.86 g,6.4 mmol), dimethylformamide (30 mL), and WSC(1.1 g, 5.7 mmol). The resulting solid was recrystallizedfrom ethyl acetate to give14 (1.4 g, 72%);1H-NMR(CDCl3) δ: 1.35–1.55 (5H, m), 2.00 (2H, d,J = 12 Hz),2.90 (2H, d,J = 12 Hz), 3.30 (2H, t,J = 5.6 Hz), 3.50 (2H,s), 3.92 (3H, s), 4.38 (2H, br-s), 7.22–7.38 (5H, m), 7.70(1H, br-s), 8.10 (1H, s); MS m/z: 387 (M+).

5.1.2.4. 4-Amino-N-[2-(1-benzyl-4-piperidinyl)ethyl]-5-chloro-2-methoxybenzamide oxalate15

Similarly to 13, 15 was prepared starting from2c(0.93 g, 4.6 mmol),6f (1.0 g, 4.6 mmol), HOBt (0.65 g,0.48 mmol), dimethylformamide (30 mL), and WSC(0.92 g, 4.8 mmol). The resulting oil was transformedinto oxalate and recrystallized from ethanol to give15(0.56 g, 24%);1H-NMR (DMSO-d6) δ: 1.35–1.55 (5H,m), 1.85 (2H, d,J = 12 Hz), 2.75 (2H, t,J = 6.8 Hz),3.19–3.30 (4H, m), 3.88 (3H, s), 4.14 (2H, s), 5.91 (2H,br-s), 6.47 (1H, s), 7.41–7.47 (5H, m), 7.66 (1H, s), 7.89(1H, t, J = 5.9Hz); MS m/z: 401 (M+).

5.1.2.5. N-[(1-Benzyl-4-piperidinyl)methyl]-2-methoxy-benzamide oxalate16

Similarly to 13, 16 was prepared starting from2d(0.74 g, 4.9 mmol),6a (1.0 g, 4.9 mmol), HOBt (0.73 g,5.4 mmol), dimethylformamide (30 mL), and WSC(1.0 g, 5.4 mmol). The resulting oil was transformed intooxalate and recrystallized from ethanol to give16 (1.5 g,71%); 1H-NMR (DMSO-d6) δ: 1.44 (2H, dd,J = 12, 24Hz), 1.70–1.80 (1H, m), 1.82 (2H, d,J = 12 Hz), 2.78(2H, t, J = 12 Hz), 3.17–3.24 (4H, m), 3.85 (3H, s), 4.15(2H, s), 6.76 (1H, d,J = 8.0 Hz), 7.01 (1H, d,J = 7.2 Hz),7.43–7.50 (6H, m), 7.54 (1H, dd,J = 2.0, 7.2 Hz), 8.22(1H, t, J = 6.0 Hz); MS m/z: 338 (M+).

5.1.2.6. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro-2-methoxybenzamide oxalate17

Similarly to 13, 17 was prepared starting from2e(0.82 g, 4.4 mmol),6a (0.90 g, 4.4 mmol), HOBt (0.66 g,4.8 mmol), dimethylformamide (30 mL), and WSC(0.93 g, 4.8 mmol). The resulting oil was transformedinto oxalate and recrystallized from ethanol to give17(0.27 g, 16%);1H-NMR (CDCl3-CD3OD) δ: 1.69 (2H,dd,J = 12, 24 Hz), 1.93 (2H, d,J = 12 Hz), 2.00 (1H, m),2.78 (2H, br-s), 3.32 (2H, d,J = 6.6 Hz), 3.48 (2H, d,J= 12 Hz), 3.96(3H, s), 4.24 (2H, s), 6.97 (1H, d,J = 10Hz), 7.41 (1H, dd, J = 2.7, 10 Hz), 7.50–7.62 (5H, m),8.00 (1H, d,J = 2.7 Hz), 7.93 (1H, br-s); MS m/z: 372(M+).

985

5.1.2.7. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-5-chlorobenzamide oxalate18

Similarly to 13, 18 was prepared starting from2f(0.76 g, 4.4 mmol),6a (0.90 g, 4.4 mmol), HOBt (0.66 g,4.8 mmol), dimethylformamide (30 mL), and WSC(0.93 g, 4.8 mmol). The resulting oil was transformedinto oxalate and recrystallized from ethanol to give18(0.41 g, 21%);1H-NMR (CDCl3-CD3OD) δ: 1.74 (2H,dd,J = 12, 24 Hz), 1.90 (2H, d,J = 12 Hz), 1.98 (1H, m),2.74 (2H, br-s), 3.24 (2H, d,J = 6.6 Hz), 3.44 (2H, d,J= 12 Hz), 4.26 (2H, s), 6.76 (1H, d,J = 10 Hz), 7.32–7.48(5H, m), 7.54 (1H, dd,J = 2.7, 10 Hz), 7.79 (1H, d,J =2.7 Hz), 7.93 (1H, br-s); MS m/z: 357 (M+).

5.1.2.8. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-2-methoxybenzamide oxalate19

Similarly to 13, 19 was prepared starting from2g(0.82 g, 4.9 mmol),6a (1.0 g, 4.9 mmol), HOBt (0.73 g,5.4 mmol), dimethylformamide (30 mL), and WSC(1.0 g, 5.4 mmol). The resulting oil was transformed intooxalate and recrystallized from ethanol to give19 (0.49 g,23%); 1H-NMR (DMSO-d6) δ: 1.43 (2H, t,J = 12 Hz),1.70 (1H, br-s), 1.77 (2H, d,J = 12 Hz), 2.78 (2H, t,J =12 Hz), 3.18 (2H, t,J = 5.9 Hz), 3.25 (2H, d,J = 12 Hz),3.80 (3H, s), 4.16 (2H, br-s), 6.16 (1H, dd,J = 2.0, 8.0Hz), 6.22 (1H,J = 2.0 Hz), 7.46 (5H, m), 7.59 (1H, d,J= 8.0 Hz), 7.90 (1H, t,J = 6.0 Hz); MS m/z: 353 (M+).

5.1.2.9. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro-2-methoxy-4-methylaminobenzamide fumalate20

Similarly to 13, 20 was prepared starting from2h(1.2 g, 5.6 mmol),6a (1.1 g, 5.6 mmol), HOBt (0.84 g,6.2 mmol), dimethylformamide (30 mL), and WSC(1.2 g, 6.2 mmol). The resulting oil was transformed intofumalate and recrystallized from ethanol/acetone to give20 (0.31 g, 14%);1H-NMR (DMSO-d6)δ: 1.27 (2H, dd,J = 12, 21 Hz), 1.59 (1H, m), 1.65 (2H, d,J = 12 Hz),2.21 (2H, t,J = 11 Hz), 2.83 (3H, d,J = 4.7 Hz), 2.95 (2H,d, J = 11 Hz), 3.18 (2H, t,J = 6.0 Hz), 3.69 (2H, s), 3.93(3H, s), 6.50 (1H, q,J = 4.7 Hz), 6.22 (1H, s), 6.59 (2H,s), 7.31–7.70 (5H, m), 7.93 (1H, t,J = 6.0Hz); MS m/z:401 (M+).

5.1.2.10. 4-Amino-5-chloro-N-[(1-ethyl-4-piperidinyl)methyl]-2-methoxybenzamide21

Similarly to 13, 21 was prepared starting from2c(2.0 g, 9.9mmol),6b (1.4 g, 9.9 mmol), HOBt (1.7 g,13 mmol), dimethylformamide (30 mL), and WSC (2.3 g,1.2 mmol). The resulting solid was recrystallized fromethyl acetate to give21 (2.0 g, 62%);1H-NMR (CDCl3)δ: 1.14 (3H, t,J = 12 Hz), 2.20–3.05 (7H, m), 2.20 (2H,q, J = 12 Hz), 2.92 (2H, d,J = 12 Hz), 3.32 (2H, t,J = 6.6

Hz), 3.88 (3H, s), 4.38 (2H, br-s), 6.23 (1H, s), 7.68 (1H,br-s), 8.03 (1H, s); MS m/z: 325 (M+).

5.1.2.11. 4-Amino-N-[(1-butyl-4-piperidinyl)methyl]-5-chloro-2-methoxybenzamide oxalate22

Similarly to 13, 22 was prepared starting from2c(0.40 g, 2.0 mmol),6c (0.31 g, 2.0 mmol), HOBt (0.29 g,2.2 mmol), dimethylformamide (10 mL), and WSC(0.42 g, 2.2 mmol). The resulting oil was transformedinto oxalate and recrystallized from ethanol to give22(0.26 g, 29%);1H-NMR (CD3OD) δ: 1.23 (3H, t,J = 8.5Hz), 1.35–1.42 (2H, m), 1.62 (2h, br-s), 1.70–1.85(2H,m), 1.95 (1H, br-s), 1.97 (2H, d,J = 13Hz), 2.90 (2H,br-s), 3.07 (2H, t,J = 8.5 Hz), 3.31 (2H, br-s), 3.35 (2H,br-s), 3.91 (3H, s), 6.23 (1H, s), 7.78 (1H, s); MS m/z:353 (M+).

5.1.2.12. 4-Amino-5-chloro-N-[(1-hexyl-4-piperidinyl)methyl]-2-methoxybenzamide23

Similarly to 13, 23 was prepared starting from2c(1.7 g, 8.4 mmol),6d (1.5 g, 8.4 mmol), HOBt (1.5 g,11 mmol), dimethylformamide (30 mL), and WSC (2.1 g,11 mmol). The resulting solid was recrystallized fromethyl acetate/diisopropylether to give23 (2.1 g, 66%);1H-NMR (CDCl3) δ: 0.83 (3H, t,J = 12 Hz), 1.20–2.05(15H, m), 3.30 (2H, t,J = 12, 24 Hz), 3.92 (2H, d,J = 12Hz), 3.33 (2H, t,J = 6.6 Hz), 3.90 (3H, s), 4.38 (2H, br-s),6.28 (1H, s), 7.70 (1H, br-s), 8.10 (1H, s); MS m/z: 381(M+).

5.1.2.13. 4-Amino-5-chloro-N-[[2-[(methylsulfonyl)amino]ethyl]-4-piperidinyl]-2-methoxybenzamide oxa-late 24

Similarly to 13, 24 was prepared starting from2c(2.0 g, 9.9 mmol),6g (2.3 g, 9.9 mmol), HOBt (1.6 g,12 mmol), dimethylformamide (50 mL), and WSC (2.3 g,12 mmol). The resulting oil was transformed into oxalateand recrystallized from ethanol to give24 (1.3 g, 26%);1H-NMR (DMSO-d6) δ: 1.20–2.00 (5H, m), 2.60–3.80(13H, m), 2.96 (3H, s), 3.79 (3H, s), 4.38 (2H, br-s), 6.42(1H, s), 7.60 (1H, s), 7.94 (1H, t,J = 6.0 Hz); MS m/z:418 (M+).

5.1.3. 1-Benzyl-3-pyrrolidinylmethylamine4aA mixture of 3 (5.0 g, 23 mmol), triphenylphosphine

(6.6 g, 25 mmol), phthalimide (3.7 g, 25 mmol) and tet-rahydrofuran (30 mL) was stirred under ice-cooling andthen diethyl azodicarboxylate (3.6 mL, 23 mmol) wasadded at the same temperature. Stirring was continuedovernight at room temperature. After evaporation, theresidue was dissolved in ethyl alcohol and then hydrazinemonohydrate (3.3 mL, 68.4 mmol) was added. The mix-ture was refluxed for 2 h. After cooling, the reaction

986

mixture was filtered by suction though celite and to thefiltrate was added diethylether. The precipitate was fil-tered off and filtrate was extracted with 10% hydrochloricacid. The aqueous layer was basified by 10% NaOH andextracted with chloroform. The extraction was washedwith brine, and was dried over anhydrous magnesiumsulfate. After evaporation in vacuo to give4a (2.8 g,56%); 1H-NMR (CDCl3) δ: 1.10–1.29 (2H, m),1.38–1.52 (1H, m), 1.55 (2H, t,J = 7.2 Hz), 1.63 (2H, d,J = 6.6 Hz), 1.92 (2H, dt,J = 1.0, 12 Hz), 2.50 (2H, br-s),2.72 (2H, t,J = 7.2 Hz), 2.82 (2H, d,J = 12 Hz), 3.61 (2H,s), 7.12–7.23 (5H, m); MS m/z: 218 (M+).

5.1.4. General procedure for the preparation of6a–6d

5.1.4.1. 4-Aminomethyl-1-benzylpipeidine6aA solution of 5a (5.7 g, 26 mmol) in tetrahydrofuran

(50 mL) was added dropwise to a suspension of lithiumaluminium hydride (LiAlH4) (2.0 g, 52 mmol) in tetrahy-drofuran (50 mL) at 0 °C, and the mixture was heated at45 °C with stirring for 4 h. After cooling, water wasadded dropwise to the mixture at a temperature below0 °C to destroy the excess LiAlH4. After filtration, thefiltrate was dried over anhydrous magnesium sulfate, andafter evaporation in vacuo, the residue was distilled underreduced pressure to give6a (5.0 g, 94%), b.p. 125 °C/1.0 mmHg;1H-NMR (CDCl3) δ: 1.28 (2H, dd,J = 4.0, 12Hz), 1.30 (2H, br-s), 1.35–1.50 (1H, m), 1.72 (2H, d,J =6.6 Hz), 1.92 (2H, dt,J = 2.0, 6.6 Hz), 2.64 (2H, d,J =6.6 Hz), 2.87 (2H, d,J = 6.6Hz), 3.00 (2H, t,J = 6.6 Hz),7.18–7.35 (5H, m); MS m/z: 142 (M+).

The other piperidines (6b–6d) were prepared in asimilar manner.

5.2. Pharmacology

5.2.1. Radioligand binding assays

5.2.1.1. 5-HT4 receptorMale Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka,

Japan) were sacrificed by cervical dislocation and thestriatum was separated from each brain. The striatum washomogenized in 15 volumes of 50 mmol/L ice-coldHEPES buffer (pH 7.4) with Polytron PT-10 and thencentrifuged at 35 000× g for 20 min. The resulting pelletwas resuspended in the HEPES buffer and finally dilutedto the appropriate concentration for assay (6 mg wetweight per assay tube). This suspension was used as thetissue preparation. Assay tubes contained 50µL ofHEPES buffer or a solution of the test agents, 50µLsolution of [3H]GR113808 (Amersham International,UK) to give a final concentration of 0.1 nmol/L and900µL of tissue preparation. Each tube was incubated for

30 min at 37 °C and the reaction was terminated by rapidfiltration through a Whatmann GF/B filter (presoaked in0.01% v/v polyethyleneimine) followed by washing with1 × 4 mL of ice-cold HEPES buffer. Then the filter wasplaced in 3 mL of scintillator and the radioactivity wasdetermined by scintillation counting in a Beckman modelLS3801 scintillation counter. Non specific binding wasdefined in the presence of unlabelled GR113808 to give afinal concentration of 1µmol/L. The IC50 value wasdetermined by non-linear regression of the displacementcurve, and theKi value was calculated according to theformula (Ki = IC50/(1 + L/Kd)), where L is the concen-tration of radioligand andKd is the dissociation constantof the radioligand.

5.2.1.2. 5-HT3 receptor[3H]Granisetron binding assays were performed ac-

cording to the method of Nelson and Thomas [34]. MaleWistar rat (Japan SLC, Ltd., Shizuoka, Japan) cerebralcortex was homogenized in 20 volumes of 0.32 mol/Lsucrose and the centrifuged at 1 000× g for 10 min. Thesupernatant was centrifuged at 40 000× g for 15 min. Thepellet was suspended in 20 volumes of HEPES buffer(50 mmol/L, pH 7.4) and the suspension was incubated at37 °C for 10 min and centrifuged at 40 000× g for15 min. The pellet was washed and centrifuged (40 000×g for 15 min). The final pellet was resuspended in 30volumes of HEPES buffer and used as tissue homogenate.The binding assay consisted of 50µmol/L of [3H]Gran-isetron, 50µL of displacing drugs and 900µL of tissuehomogenate. Following a 30 min incubation at 25 °C, theassay mixture was rapidly filtered under reduced pressurethrough Whatman GF/B glass filters which had beenpresoaked in 0.1% polyethyleneimie. Filters were washedimmediately with 3× 3 mL of ice-cold Tris-HCl buffer(50 mM, pH 7.4). ICS 205930 (100 mmol/L) was usedfor the determination of nonspecific binding.

5.2.1.3. D2 receptor[3H]Spiperone binding assays were performed accord-

ing to the method of Crees et al. Male Wistar rat (JapanSLC, Ltd., Shizuoka, Japan) striatal membrane washomogenized in 100 volumes of ice-cold Tris-HCl buffer(50 mmol/L, pH 7.7) and centrifuged (500× g, 10 min,0 °C). The supernatant was centrifuged at 50 000× g for15 min. The pellet was suspended in 100 volumes ofice-cold Tris-HCl buffer (50 mmol/L, pH 7.7) and recen-trifuged (500× g, 10 min, 0 °C). The final pellet wasresuspended in 150 volumes (50 mmol/L, pH 7.7) con-taining 120 mmol/L NaCl, 5 mmol/L KCl, 2 mmol/LCaCl2, 1 mmol/L MgCl2, 1.1 mmol/L ascorbic acid and10 µmol/L pargyline, and incubated at 37 °C for 10 min.

987

A portion of this membrane suspension (900µmol/L) wasplaced in a tube, and 50µmol/L of either test compoundor vehicle solution was added, followed by 50µL of[3H]Spiperone (40 Ci/mmol) at a final concentration of0.2 nmol/L. The tubes were incubated at 37 °C for 20 minand filtered through Whatman GF/B glass filters, whichwere then washed three times with 3 mL of Tris-HClbuffer (50 mmol/L, pH 7.7). Sulpiride (100µmol/L) wasused for the determination of nonspecific binding. Theradioactivity trapped on the filters was measured byliquid scintillation spectrometry.

5.2.2. Contractile effects

5.2.2.1. 5-HT4 receptor agonismMale Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka,

Japan) were killed by cervical dislocation and the ascend-ing colon (a 10 cm segment starting 1 cm from thecaecum) was removed. The longitudinal muscle layer wasseparated from the underlying circular muscle and di-vided into four segments of about 2.5 cm. Four musclestrip preparations were individually mounted verticallyfor isotonic measurement into a tissue bath containing10 mL Tyrode solution. Only 5-HT was tested in theTyrode solution containing methysergide (1µmol/L) andgranisetron (1µmol/L) to inhibit responses mediated by5-HT2 and 5-HT1-like and 5-HT3 receptors, respectively.This solution was kept at 37 °C and gassed with 95% O2,5% CO2. The strips were subjected to a preload of 1 g andallowed to stabilize for 20 min. After stabilization, theresponse of the longitudinal muscle to 30µmol/L metha-choline was measured. Agonist concentration-effectcurves were constructed using sequential dosing, leaving15 min between doses. A 15 min dosing cycle wasrequired to prevent desensitization. The agonist was leftin contact with a preparation until the response hadreached a maximum, the preparation was washed. Fortyminutes was left between the determination ofconcentration-effect curves. GR113808 (10 nmol/L) wereincubated for 10 min before repeating agonistconcentration-effect curves. After each determination ofconcentration-effect curve, 30µmol/L of methacholinewas added to the tissue bath again. All responses wereexpressed as a percentage of the mean of the twocontractions induced by 30µmol/L methacholine. TheEC50 value, the concentration causing 50% of the maxi-mal response, was determined by linear regression analy-sis.

5.2.2.2. 5-HT4 receptor antagonism5-HT4 receptor antagonism was expressed in the form

of pKB value on the contractile response to 5-MeOT.

5.3. Construction of receptor model

5.3.1. SequenceThe sequence in this study, except for the 5-HT4

receptor, was from the Swiss-Prot Protein Sequence DataBank [35]. The coordinates of Bacteriorhodopsin (entry1BRD) [36] were from the Protein Data Bank [37] atBrookhaven National Laboratory. Modelling wasachieved with the molecular package SYBYL 6.2 [38].The interactive modelling and display were performed ona Silicon Graphics IRIS INDIGO/Elan 4 000 computer.Five main steps were used: stretch ofα-helical structurewhere putative transmembrane region is longer thanbacteriorhodopsin, amino acid substitution, local geom-etry optimization, docking of ligand, and side-chainrotation to minimize overlaps between helices. Somemanual adjustments were made to remove bad stericinteractions in geometry optimization. In Energy minimi-zation procedure, the backbone was aggregated until theRMS gradient was less than 0.05 (kcal/mol Å2). Adielectric constant of compound24 was calculated wherea cut off distance of 8 Å was used and no solventmolecules were included in the calculation.

Acknowledgements

We thank Mrs. F. Matsugaki for some of the biologicalresults. We also thank Dr. M. Terasawa and Dr. K. Adachifor helpful discussion.

References

[1] Alexander S.H., Peters J.A., TPiS Receptor and Ion ChanenelNomenclature Supplement, Elsevier, 1998, pp. 46–48.

[2] Eglen R.M., Wong E.F., Dumuis A., Bockaert J., Trends Pharmacol.Sci. 16 (1995) 391–398.

[3] Craig D.A., Clarke D.E., J. Pharmacol. Exp. Ther. 252 (1990)1378–1386.

[4] Elswood C.J., Bunce K.T., Humphrey P.P.A., Eur. J. Pharmacol. 196(1991) 149–155.

[5] Briejer M.R., Akkerman L.S., Meulemans A.L., Lefebvre R.A.,Schuurkes L.J., Naunyn-Schmiederg’s Arch. Pharmacol. 347 (1993)464–470.

[6] Rizzi C.A., Coccini T., Onori L., Manzo L., Tonini M., J. Pharmacol.Exp. Ther. 261 (1992) 412–419.

[7] Langlois M., Zhang L., Yang D., Brémont B., Shen S., Manara L.,Croci T., Bioorg. Med. Chem. Lett. 4 (1994) 1433–1436.

[8] Dumuis A., Bouhelal R., Sebben M., Cory R., Bockaert J., Mol.Pharmacol. 34 (1988) 880–887.

[9] Flynn D.L., Zabroski D.L., Becker D.P., Nosal R., Villamil C.L.,Gullikson G.W., Moummi C., Yang D., J. Med. Chem. 35 (1992)1487–1491.

[10] Turconi M., Schiantarelli P., Borsini F., Rizzi C.A., Ladinsky H.,Donetti A., Drugs Future 16 (1991) 1011–1026.

988

[11] Clark R.D., Jahangir A., Langston J.A., Weinhardt K.K., MillerA.B., Leung E., Eglen R.M., Bioorg. Med. Chem. Lett. 4 (1994)2477–2480.

[12] López-Rodríguez M.L., Morcillo M.J., Benhaú B., Rosado M.L., J.Comput.-Aided Mol. Des. 11 (1997) 589–599.

[13] Kuroita T., Sakamori M., Kawakita T., Chem. Pharm. Bull. 44(1996) 756–764.

[14] Kuroita T., Ikebe T., Murakami S., Takehara S., Kawakita T., Bioorg.Med. Chem. Lett. 5 (1995) 1245–1250.

[15] Tahara T., Hayano K., Murakami S., Fukuda T., Setoguti M., IkedaK., Marubayasi N., Chem. Pharm. Bull. 38 (1990) 1609–1615.

[16] Iwanami S., Takasima M., Hirata Y., Hasegawa O., Usuda S., J.Med. Chem. 24 (1981) 1224–1230.

[17] Mitsunobu O., Synthesis (1981) 1–28.

[18] Kawakita T., Kuroita T., Murozono T., Hakira H., WO 95/26953Chem. Abstr. 124 (1996) 14913.

[19] Fancelli D., Caccia C., Fornaretto M.G., McArthur R., Severino D.,Vaghi F., Varasi M., Bioorg. Med. Chem. Lett. 6 (1996) 236–266.

[20] Hibert M.F., Trumpp-Kallmeyer S., Bruinvels A., Hoflack J., Mol.Pharmacol. 40 (1991) 8–15.

[21] Ijzerman A.P., Van Gallen P.M., Jacobson K.A., Drug Design Dev. 9(1992) 49–67.

[22] Trumpp-Kallmeyer S., Hoflack J., Bruinvels A., Hibert M., J. Med.Chem. 35 (1992) 3448–3462.

[23] Yamamoto Y., Kamiya K., Terao S., J. Med. Chem. 36 (1993)820–825.

[24] Nordvall G., Hacksell U., J. Med. Chem. 36 (1993) 976–976.

[25] Glennon R.A., Dukat M., Westkaemper R.B., Ismaiel A.M., IzzarelliD.G., Parker E.A., Mol. Pharmacol. 49 (1996) 198–206.

[26] Teeter M.M., Froimowitz M.B., Durand C.J., J. Med. Chem. 37(1996) 2874–2888.

[27] Kotoyianni M., Deweese C., Penzotti J.E., Lybrand T.P., J. Med.Chem. 39 (1996) 4406–4420.

[28] Kyte J., Doolittle R.F., J. Mol. Biol. 157 (1982) 105–132.

[29] Begonia Y.H., Karschin A., Branchek T., Davidson N., Lester H.A.,FEBS. Lett. 312 (1992) 259–262.

[30] Johnson M.P., Loncharich R.J., Baez M., Nelson D.L., Mol. Phar-macol. 45 (1993) 277–286.

[31] Almaula N., Ebersole B.J., Zhang D., Weinstein H., Sealfon S.C., J.Biol. Chem. 271 (1996) 14672–14675.

[32] Almaula N., Ebersole B.J., Ballesteros J.A., Weinstein H., SealfonS.C., Mol. Pharmacol. 50 (1996) 34–42.

[33] Choudhary M.S., Craigo S., Roth B.L., Mol. Pharmacol. 43 (1993)755–761.

[34] Nelson D.R., Thomas D.R., Biochem. Pharmacol. 38 (1989)1693–1695.

[35] Bairoch A., Boeckmann B., Nucleic Acids Res. 20 (1992)2019–2022.

[36] Henderson R., Baldwin J.M., Ceska T.A., Zemlin F., BeckmannE.D., J. Mol. Biol. 213 (1990) 899–929.

[37] Bernstein F.C., Koetzle T.F., Williams G.B., Meyer E.F. Jr., BriceM.D., Rodgers J.R., Kennard O., Shimanouchi T., Tasumi M., J.Mol. Biol. 112 (1977) 535–542.

[38] Tripos E & S, 6548 Clayton Road, St. Louis, MO 62117. SYBYLmolecular modelling software.

989

Preliminary communication

Synthesis and antioxidant activity of new tetraarylpyrroles

Jacques Lehuédéa, Bernard Fauconneaub, Laurence Barrierb,Marina Ourakowa, Alain Pirioub, Jean-Michel Vierfonda*

aLaboratoire de Chimie Organique, Faculté de Pharmacie, BP 199, 34, rue du Jardin des Plantes, F-86005 Poitiers, FrancebCentre d’Études et de Recherche sur les Xénobiotiques, EA 1223, Faculté de Pharmacie,

BP 199, 34, rue du Jardin des Plantes, F-86005 Poitiers, France

(Received 26 January 1998; revised 18 January 1999; accepted 11 March 1999)

Abstract – The synthesis and in vitro antioxidant activity of 17 new tetraarylpyrroles are investigated by 2 tests highly documented in theliterature: capability to prevent Fe2+-induced lipid peroxidation on microsomes, which is a membrane preparation rich in polyunsaturated fattyacids, and direct scavenging effect on a stable free radical, 1,l-diphenyl-2-picryl-hydrazyl (DPPH). For the Fe2+-induced microsomal lipidperoxidation system, the results show that molecules which possess 2-pyrazinyl or 2-pyridyl in the 3- and 4-positions on the pyrrole ring arethe most efficient. Introduction of methoxy groups on the phenyl ring in the 2- and 5-positions increases the effects but the higher activity isobtained with 2-furyl or 2-thienyl. The only compounds which possess a direct scavenger effect on trapping the stable free radical DPPH arethose which have 2-pyridyl in the 3- and 4-positions and 2-furyl or 2-thienyl in the 2- and 5-positions. © 1999 Éditions scientifiques etmédicales Elsevier SAS

tetraarylpyrrole / antioxidant activity / radical scavenging effect / microsomes / DPPH

1. Introduction

Formation of reactive oxygen species and the ensuingoxidation of biological molecules is a well recognizedmechanism of tissue damage in many pathological situ-ations, such as, inflammation, stroke, acute myocardialinfarction and the subsequent reperfusion phase. Numer-ous natural or synthetic antioxidant compounds havebeen tested with success in various disease models aswell as in clinics [1].

In the literature, it is well known that pyrrolic com-pounds such as polypyrroles, poly- or hetero-arylpyrrolespresent an electronic delocalization, conferring to thesemolecules electric conductor and oxidizable proper-ties [2, 3]. Furthermore, pyrrolic structures such asbenzoylpyrrole-3-acetic acids were tested in vitro byexamining their effects on lipid peroxidation using rathepatic microsomes and as hydroxyl radical scaven-gers [4].

Recently we reported results on six pyrrolic com-pounds synthesized in our laboratory showing interestingantioxidant activities [5]. Consequently, it seems relevantto test large series of molecules bearing different substitu-ents on the pyrrole ring: phenyl rings (with or withoutmethoxy groups) and aromatic heterocycle rings in orderto have an electronic delocalization and consequently,potent antioxidant activities.

2. Chemistry

The synthesis of compounds2a–q (table I) was real-ized in two steps (figure 1): metalation of methylazinefrom lithium diisopropylamine and condensation with anaromatic nitrile to obtain an imine-enamine1 moisturesensitive (first step), and oxidation of imine-enaminel byPb(OAc)4 to give tetraarylpyrroles2 (second step) [6].The structure is given intable I.

2.1. Antioxidant activity

The antioxidant potency of compounds2a–q wasassessed by two tests commonly used in the literature*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 991−996 991© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

bearing on this topic: capability to prevent Fe2+-inducedlipid peroxidation on microsomes, which is a membranepreparation rich in polyunsaturated fatty acids, and direct

scavenging effect on a stable free radical, 1,l-diphenyl-2-picryl-hydrazyl (DPPH).

The results were compared to those observed withTrolox (water soluble vitamin E analogue), a referenceantioxidant.

3. Results and discussion

3.1. Effect on lipid peroxidation

Three distinctive steps can be distinguished in lipidperoxidation: initiation, propagation and termination re-actions. Lipid peroxidation may be induced, for example,by radical species which are sufficiently reactive toabstract a hydrogen atom from the unsaturated fatty acids.This is the starting point for the lipid radical chainpropagation reaction. Thus, many molecules of lipidsmay be oxidized to lipid hydroperoxides for every initia-tion event. The propagation cycle is broken by termina-tion reactions which result in the destruction of freeradicals. Termination reactions occur when two radicalspecies combine to form non-radical final products [7].

Table I. Structure and antioxidant effects of tetraarylpyrroles. IC50 values are expressed as means± SEM.

R Ar1 Ar2 Yields IC50 microsomes IC50 DPPH

(%) (µM) (µM)

2a H 2-pyrazinyl phenyl 90a 320± 16 > 1 0002b H 2-pyrazinyl 2-pyridyl 77a > 1 000 > 1 0002c H 2-pyrazinyl 4-pyridyl 52a > 1 000 > 1 0002d H 2-pyrazinyl 2-furyl 98a 19.0± 9.6 > 1 0002e H 2-pyrazinyl 2-thienyl 90a 8.0± 1.1 > 1 0002f H 2-quinoxalinyl phenyl 59a > 1 000 > 1 0002g H 2-pyridyl phenyl 50b 67.2± 6.7 > 1 0002h H 4-pyridyl phenyl 37b > 1 000 > 1 0002i H 2-pyridyl 2-methoxyphenyl 53c 28 ± 3 > 1 0002j H 2-pyridyl 3-methoxyphenyl 36c 63 ± 4 > 1 0002k H 2-pyridyl 4-methoxyphenyl 67c 12.3± 4.0 > 1 0002l H 2-pyridyl 3,4,5-methoxyphenyl 25c 26.2± 8.5 > 1 0002m H 4-pyridyl 4-methoxyphenyl 29c 100± 12 > 1 0002n H 2-pyridyl 2-furyl 17.5b 11.0± 1.9 200± 502o CH3 2-pyridyl 2-furyl 88d 61.1± 5.0 > 1 0002p H 2-pyridyl 2-thienyl 53b 2.7± 1.1 410± 1272q H 2-pyridyl 3-thienyl 32b 12.5± 2.2 > 1 000Trolox 5.0± 0.3 10.1± 0.5

a: based on isolated imine-enamine.b: based on aromatic nitrile.c: with TMEDA and heating, based on aromatic nitrile.d: methylation from2n.

Figure 1. Synthesis of tetraarylpyrroles.

992

A chain breaking antioxidant has the ability to donatehydrogen radicals and contributes to stop the chainreaction. It reduces the primary radical by a one electronreduction to a non-radical chemical species and, as aconsequence, is transformed into an oxidized antioxidantradical. However, this antioxidant must be more than justa hydrogen atom donor; it needs also to form a radical ofsuch low reactivity that no further reactions with lipidscan occur.

Peroxyl radicals are the major chain-propagating spe-cies in the process of lipid peroxidation in mem-branes [8]. Thus, a standard test for antioxidants is theaction of a substance in inhibiting peroxidation of mem-branes such as microsomes. The thiobarbituric acid(TBA) test is widely used in microsomal systems but it isessential to ensure that the apparent antioxidant effect ofan added compound is not due to the interference with theTBA test itself [9]. Consequently, relevant controls arenecessary to check the absence of such an interference.Peroxidation can be accelerated by adding iron salts tomembranes, e.g. Fe2+. In metal-ion dependent systems,an added antioxidant might act not only by scavengingperoxyl radicals but also by binding iron ions andstopping them from accelerating peroxidation. However,these two possibilities can be distinguished easily. Theefficient compounds tested in this study acted as chainbreaking antioxidants because at low concentrations, alag period into the peroxidation process was observed(data not shown), corresponding to the time taken for theantioxidant to be consumed, whereas metal-binding anti-oxidants give a constant inhibition throughout the reac-tion [9].

The effects of the compounds on lipid peroxidationinduced by FeCl2 are shown intable I. When Ar2 isconstant, i.e. phenyl ring, and if Ar1 is 2-pyridyl (2g), theactivity of the molecule is higher than that observed withAr1 = 2-pyrazinyl (2a). The activity disappears if Arl is4-pyridyl (2h) or 2-quinoxalinyl (2f). If the phenyl ring inthe Ar2 position is substituted, i.e. 4-methoxyphenyl, thehigher activity of 2-pyridyl compared with that of4-pyridyl in Arl is confirmed (2k versus2m). When theyare in the Arl position, the tested substituents can beclassified as follows: 2-pyridyl> 2-pyrazinyl> 4-pyridyl> 2-quinoxalinyl. Consequently, molecules with Ar1 =2-pyridyl or 2-pyrazinyl show the greatest interest and thesynthesis of molecules with different groups in Ar2

positions is performed to improve the antioxidant effect.When methoxy groups are added in the 2, 3, or 4

position on the phenyl ring in Ar2, the antioxidant activityis increased (2i, 2j, 2k versus2g), with a major activityfor 4-methoxyphenyl (2k). This activity may be corre-lated to the introduction of electron donor substituents

which stabilize the generated radical during oxidation.The electronic mesomeric donor effect is of the same typewhen methoxy groups are added in the 2 or 4 position onthe Ar2 phenyl ring (i.e.2i or 2k), but 2k has a betteractivity, probably due to the steric bulk observed with2-methoxyphenyl (2i). The substituent 3-methoxyphenylhas a lower electronic effect, and consequently is lessactive than2i and2k. But methoxy tri-substitution in the3, 4, 5 positions on the phenyl ring (2l) decreases 2 timesthe activity of2k.

The antioxidant tests performed with compounds bear-ing heterocyclic substituents in the Ar2 position make itpossible to classify them into two groups: those with 2-and 4-pyridyl which are inactive (2b and 2c) and thosewith pentagonal heterocycles (2d, 2e, 2n, 2p, 2q) whichseem to be of major importance because the antioxidantactivity is at least as great as that of the more activephenylmethoxylated compounds. The pentagonal sub-stituents have a donor mesomeric effect on a conjugatedsystem, whereas pyridyl has an attractor effect. Substitu-ent 3-thienyl (2q) is less active than 2-thienyl (2p).

On microsomes oxidized by Fe2+, 2p (Arl = 2-pyridyland Ar2 = 2-thienyl; IC50 = 2.7 ± 1.1 µM) is the mostefficient antioxidant. Its activity is higher than thatobtained with Trolox (5.0± 0.3µM), a reference antioxi-dant.

These results show the major importance of the pyrrolring in antioxidant effects since methylation of2n (i.e.2o) decreases about 6 times the activity of this compound.This may be explained by the difficulty in forming a freeradical with a pentasubstituted pyrrole.

3.2. Effect on DPPH

The DPPH test provided information about the reac-tivity of the tested compounds with a stable free radical.Because of its odd electron, the DPPH radical showed astrong absorption band at 515 nm in visible spectroscopy(a deep purple colour). As this electron is paired off in thepresence of a free radical scavenger, absorption vanishesand the resulting decoloration is stoichiometric withrespect to the number of electrons taken up. This bleach-ing of DPPH absorption, which occurs when the oddelectron of the radical is paired, is thus representative ofthe capacity of the compounds to scavenge free radicalsindependently of any enzymatic activity.

Only 2p and 2n have a radical scavenging activity,with a major activity for2n, whereas that of Trolox isconsiderably higher. It means that these compoundspossess a direct trapping effect (table I).

It is noteworthy that these molecules (i.e.2p and2n)also have a good activity with a microsome system.

993

Nevertheless,2e has no activity on DPPH, while it hasgood activity on Fe2+-induced microsomal lipid peroxi-dation. This result shows that this molecule doesn’tpossess a direct scavenger effect, at least on this radical.This effect is improved when a 2-pyridyl is substituted inthe Ar1 position (2e versus2p).

In conclusion, our study provides evidence that severaltetraarylpyrroles exhibit interesting antioxidant propertiesmainly expressed by their capacity to inhibit Fe2+-induced microsomal lipid peroxidation. These effectsmay be useful in the treatment of pathologies in whichfree radical oxidation plays a fundamental role.

The early trials on animals are encouraging, sincetoxicity tests have been performed on female Swiss mice(18–20 g, Iffa-Credo, l’Arbresle, France). At the presenttime, the only compound tested is2a and is responsiblefor no mortality when orally administered at the dose of1g/kg (batch of 10 animals).

4. Experimental protocols

Melting points were measured by using a Köflerapparatus and are uncorrected. The1H-NMR spectrawere recorded on Varian EM 360 and Bruker 200 A.C.spectrometers. IR spectra were realized on a Unicam SP1100 and on an ATI Mattson Genesis spectrometer.Elemental analyses were performed on a Perkin Elmer240 apparatus.

Acetonitrile was dried over Na2SO4. Tetrahydrofuran(THF) was dried and extemporaneously distillated oversodium. Diisopropylamine and tetramethylethylenedi-amine (TMEDA) were dried over BaO and distillated.Butyl lithium used was a Merck 1.6 M solution in hexane.

The synthesis of compounds2a–h was previouslydescribed [6]. The2n–q pyrroles were prepared accord-ing to the precedent method without isolation of theimine-enamine intermediate.

4.1. General procedure for the preparation ofmethoxylated phenylpyrroles (2i–m)

Imines were prepared according to the general proce-dure [6]. 13.8 mL (22 mmol) of a solution of n-butyl-lithium (1.6 M) in hexane was added, via syringe, to astirred solution of diisopropyl-amine (2.2 g; 22 mmol)solved in 30 mL of dry THF plus 10 mL of dry TMEDA,under 0 °C nitrogen atmosphere. After 20 min at 0 °C, thesolution was cooled to –40 °C and a solution of2-methylpyrazine (1.88 g; 20 mmol) in 5 mL of THF wasslowly added. Then, the solution was stirred for 45 min at–40 °C and a solution of methoxylated aromatic nitrile(10 mmol) in 5 mL of mixed solvent (THF 3/4 plus

TMEDA 1/4; v/v) was added at –40 °C. The mixture washeated for 2 h to refluxing THF afterwards. After coolingto 20 °C, 1 mL of water was added, with strong and quickstirring, the mixture was stored on anhydrous sodiumsulphate. After filtration and evaporation of solvents, theresidue (non-isolated imine-enamine) was dissolved inacetonitrile (100 mL) plus chloroform (50 mL) andcooled to –40 °C. The lead tetraacetate (2.22 g; 5 mmol)was introduced at –40 °C with stirring. After 15 min, themixture was heated to 20 °C and stirred for 1 h. Then,12 mL of aqueous saturated sodium carbonate was added.The precipitate was washed out with acetonitrile andchloroform. The organic phase was dried over anhydroussodium sulphate, evaporated and purified by chromatog-raphy on a silica gel column, elution with ethyl acetate.

4.1.1. 1H-2.5-di-(2-methoxyphenyl)-3.4-di-(2-pyridyl)pyrrole 2i

Yellow powder (2.29 g, 53%); m.p. 179 °C; IR (KBr)(cm–l): 3 480, 1 580, 783, 751;1H-NMR (CDCl3) δ 10.2(1H, s, NH), 8.4 (2H, d,J = 4–5 Hz), 7.5–6.6 (14H, m),3.7 (6H, s); Anal. C28H23N3O2 (C, H, N).

4.1.2. 1H-2.5-di-(3-methoxyphenyl)-3.4-di-(2-pyridyl)pyrrole 2j

Pale yellow powder (1.55 g, 36%); m.p. 165 °C; IR(KBr) (cm–1): 3 389, 1 608, 1 558, 871, 852, 784, 748;1H-NMR (CDCl3) δ 8.8 (1H, s, NH), 8.45 (2H, m),7.5–6.6 (14H, m), 3.6 (6H, s); Anal. C28H23N3O2 (C, H,N).

4.1.3. 1H-2.5-di-(4-methoxyphenyl)-3,4-di-(2-pyridyl)pyrrole 2k

Pale yellow powder (2.90 g, 67%); m.p. 128 °C; IR(KBr) (cm–l): 3 210, 1 590, 834, 792, 744;1H-NMR(CDCl3) δ 8.7 (1H, s, NH), 8.4 (2H, d,J = 4–5 Hz),7.5–6.6 (14H, m), 3.7 (6H, s); Anal. C28H23N3O2 (C, H,N).

4.1.4. 1H-2.5-di-(3,4,5-trimethoxyphenyl)-3,4-di-(2-pyridyl) pyrrole2l

Brown powder (1.39 g, 25%); m.p.> 260 °C; IR (KBr)(cm–l): 3 357, 1 584, 842, 793, 747, 699;1H-NMR(CDCl3) δ 9.0 (1H, s, NH), 8.5 (2H, d,J = 4–5 Hz),7.6–6.9 (6H, m), 6.6 (4H, s), 3.8 (6H, s), 3.65 (12H, s);Anal. C32H31N3O6 (C, H, N).

4.1.5. 1H-2.5-di-(4-methoxyphenyl)-3,4-di-(4-pyridyl)pyrrole 2m

White powder (1.25 g, 29%); m.p.> 260 °C; IR (KBr)(cm–1): 3 350, 1 550, 900, 645;1H-NMR (CDCl3) δ 8.75(1H, s, NH), 8.2 (4H, d,J = 4–5 Hz), 7.15–6.6 (12H, m),3.7 (6H, s); Anal. C28H23N3O2 (C, H, N).

994

4.1.6. 1H-2.5-di-(2-furyl) 3,4-di-(2-pyridyl)pyrrole2nPale yellow powder (0.62 g, 17.5%); m.p. 172 °C; IR

(KBr) (cm–l): 3 430, 1 589, 1 562, 803, 790, 733;1H-NMR (CDCl3) δ 9.4 (1H, s, NH), 8.6 (2H, m), 7.5–6.9(8H, m), 6.3 (4H, m); Anal. C22H15N3O2 (C, H, N).

4.1.7. 1H-1-methyl-2.5-di-(2-furyl)-3,4-di-(2-pyridyl)pyrrole 2o

Brown powder; m.p. 180 °C; IR (KBr) (cm–l): 3 134,1 587, 1 561, 909, 750;1H-NMR (CDCl3) δ 8.4 (2H, m),7.5–6.8 (8H, m), 6.3 (4H, m), 3.6 (3H, s); Anal.C23H17N3O2 (C, H, N).

4.1.8. 1H-2.5-di-(2-thienyl)-3.4-di-(2-pyridyl)pyrrole2pPale yellow powder (2.04 g, 53%); m.p. 230 °C; IR

(KBr) (cm–1): 3 250, 1 590, 840, 750, 690;1H-NMR(CDCl3) δ 8.7 (1H, s, NH), 8.5 (2H, m), 7.5–6.8 (12H,m); Anal C22H15N3S2 (C, H, N).

4.1.9. 1H-2.5-di-(3-thienyl)-3.4-di-(2-pyridyl)pyrrole2qPale yellow powder (1.23 g, 32%); m.p. 243 °C; IR

(KBr) (cm–1): 3 420, 1 590, 1 560, 865, 782, 749, 693,616; 1H-NMR (CDCl3) δ 9.6 (1H, s, NH), 8.4 (2H, m),7.5–6.8 (12H, m) Anal. C22H15N3S2 (C, H, N).

4.2. Procedures concerning antioxidant activities

4.2.1. ChemicalsAll chemicals were purchased from the Sigma Chemi-

cal Co. (St. Louis, MO, USA) except Trolox which wasobtained from Aldrich (St Quentin Fallavier, France).

The powders corresponding to the molecules2a–qwere kept under nitrogen. For the biological tests, thecompounds were dissolved extemporaneously in ethanoland the different dilutions were performed in ethanol.

4.2.2. Inhibitory effect on lipid peroxidationMale Sprague-Dawley rats (200–250 g, Iffa-Credo,

l’Arbresle, France) had been deprived of food overnight(16 h). The rats were anaesthetized by inhalation of ethylether. Livers were perfused, rapidly isolated and mincedthoroughly with scissors. The minced tissue was washedwith ice-cold 0.15 M KCl and homogenized with 5volumes of ice-cold 0.15 M KCl, using a Teflon-glasspotter homogenizer (3 000 rpm for 2 min). Microsomalfractions were isolated in Tris-HCl 0.05 M/KCl 0.15 M,pH 7.4, by removal of the nuclear fraction at 800g for15 min, removal of the mitochondrial fraction at 15 000gfor 15 min and sedimentation at 105 000g for 30 min.Pellets were washed twice in Tris-HCl 0.05 M/KCl 0.15M, pH 7.4 buffer by centrifugation, with subsequentsedimentation at 125 000g for 15 min. Microsome pel-lets were resuspended in the same buffered solution at

5 mg protein/mL and stored at –80 °C for a maximum of1 month. The protein content was determined by themethod of Bradford [10], using bovine albumin as astandard.

For the test, microsomal fractions were thawed justbefore use and were diluted with 0.05 M Tris-HCl, pH7.4, containing 0.15 M KCl. The final protein concentra-tion in the incubation mixture amounted to 0.75 mg/mL.Microsomes were pre-incubated with different concentra-tions of compounds in a shaking water-bath at 37 °C for10 min. Then, lipid peroxidation was initiated with 10µM FeCl2 and the samples were incubated at 37 °C for30 min. After action of thiobarbituric acid, the absor-bance was measured at 532 nm for thiobarbituric acidreactive substance (TBARS) determination [11], versuscontrol containing the same quantity of ethanol butwithout the compound studied. The measurements wereperformed in triplicate. The inhibition of lipid peroxida-tion was expressed as a percentage and the inhibitionconcentration 50% (IC50) was obtained from the inhibi-tion curve by graphical determination.

4.2.3. Radical scavenging effectFree radical scavenging capacity of the compounds

was determined using DPPH [12]. An ethanol DPPHsolution (0.1 mM) was mixed with different concentra-tions of compound and the absorbance change at 515 nmwas measured 10 min later with a spectrophotometer(Uvikon 940, Kontron) versus control containing thesame quantity of ethanol, but without compound studied.The measurements were performed in triplicate. Theinhibition of coloration was expressed as a percentageand the IC50 was obtained from the inhibition curve, bygraphical determination.

Acknowledgements

The authors wish to thank R. Pontcharraud and S.Mairesse-Lebrun for excellent technical assistance.

References

[1] Rice-Evans C.A., Diplock A.T., Free Rad. Biol. Med. 15 (1993)77–96.

[2] Niziurski-Mann R.E., Scordilis-Kelley C., Tae-Lane L., Carlin R.T.,J. Am. Chem. Soc. 115 (1993) 887–891.

[3] Demopoulos V.J., Rekka E., J. Pharm. Pharmacol. 46 (1994)740–744.

[4] Demopoulos V.J., Rekka E., Retsas S., Pharmazie 45 (1990)403–407.

[5] Fauconneau B., Lehuede J., Karge E., Vierfond J.M., Piriou A.,Klinger W., Exp. Toxicol. Pathol. 48 SII (1996) 116–120.

995

[6] Lehuédé J., Mettey Y., Vierfond J.M., Synth. Comm. 26 (1996)793–802.

[7] Gutteridge J.M.C., Clin. Chem. 41 (1995) 1819–1828.

[8] Halliwell B., Gutteridge J.M.C., Free Radicals in Biology andMedicine, second edition, Clarendon Press, Oxford, UK, 1989.

[9] Halliwell B., Free Rad. Res. Com. 9 (1990) 1–32.

[10] Bradford M.M., Anal. Biochem. 72 (1976) 248–254.

[11] Wilbur K.M., Bernheim F., Shapiro O.W., Arch. Biochem. Biophys.24 (1949) 305–313.

[12] Tamura A., Sato T., Fuhu T., Chem. Pharm. Bull. 38 (1990)255–257.

996

Preliminary communication

Synthesis and anti-inflammatory activity of N-substituted 2-oxo-2H-1-benzopyran-3-carboxamides and their 2-iminoanalogues

Igor E. Bylov*, Maksym V. Vasylyev, Yaroslav V. Bilokin‡

Department of Organic Chemistry, Ukrainian Academy of Pharmacy, 310002 Kharkov, Ukraine

(Received 12 November 1999; revised 6 May 1999; accepted 7 May 1999)

Abstract – A series ofN-arylsubstituted 2-imino-2H-1-benzopyran-3-carboxamides3a andb and 2-oxo-2H-1-benzopyran-3-carboxamides4a–h were synthesized and evaluated for their anti-inflammatory activity in carrageenan-induced rat paw oedema assays and in aceticacid-induced peritonitis tests in albino rats. The resulting products were found to be active anti-inflammatory agents and their effects werecomparable to that of piroxicam as the reference compound. In the consideration of the efficacy of the compounds in these assays,2-imino/oxo-2H-1-benzopyran-3-carboxamides3a andb and4a–h were further studied at graded doses for their acute toxicity (ALD50) inalbino mice and were essentially non-toxic at the highest dose tested. © 1999 Éditions scientifiques et médicales Elsevier SAS

coumarins / benzopyrans / amides / Knoevenagel condensation / anti-inflammatory activity

1. Introduction

Compounds comprising a coumarin (2-oxo-2H-1-benzopyran) backbone have a wide range of biologicalactivities. Thus, among the natural and synthetic cou-marin derivatives there are compounds possessing anti-microbial [1], antitumour [2], antiviral [3] and other [4]activities. Moreover, studies of the hydroalcohol extractof Justicia pectoralis(Eha) and its main constituents,coumarin (Cou) and umbelliferone (Umb), showed anal-gesic and anti-oedema activities on acetic acid-inducedwrithing in mice and on the carrageenan end dextran pawoedema in rats [5]. The Eha, Cou and Umb presented asignificant anti-oedema effect in the carrageenan modelbut only Cou decreased the rat paw volume in the dextranmodel. Anti-inflammatory activity of coumarins isolatedfrom Santolina oblongifoliaBoiss was also reported [6].The isolated coumarins identified as 7-methoxycoumarin(herniarin), 6,7-dihydroxycoumarin (aesculetin),6-methoxy-7-glucosidylcoumarin (scopolin), and6-hydroxy-7-methoxycoumarin (scopoletin) showedmarked activity as inhibitors of eicosanoid-release from

ionophore-stimulated mouse peritoneal macrophages. Itwas also revealed [7] that compounds containing a ben-zopyran moiety were potent and selective inhibitors ofcyclooxygenase (COX). Of the 3-substituted coumarinderivatives, our attention was called to theirN-substitutedamide derivatives, since marked anti-inflammatory activ-ity of structurally related N-substituted amides of4-hydroxy-2-quinolone-3-carboxylic acids has been re-ported [8]. Also taking into consideration the fact thatsuch anti-inflammatory drugs as mefenamic and meclofe-namic acids [4] constitute derivatives of aromatic aminoacids,N-substituted coumarin-3-carboxamide derivativescontaining aromatic amino acid residues were selected astargets for our anti-inflammatory studies.

2. Chemistry

The general synthetic strategy employed to prepareN-substituted 2-imino/oxo-2H-1-benzopyran-3-carbox-amide derivatives was based on Knoevenagel condensa-tion [9, 10] of active methylene compounds. As shown infigure 1, 2-imino-2H-1-benzopyran-3-N-R-carboxamides3a and b were prepared by condensingN-substitutedcyanoacetamides1a and b and salicylic aldehyde2 toform the expected imino compounds using piperidine as

*Correspondence and reprints‡Present address: Department of Organic Chemistry, The Wei-zmann Institute of Science, Rehovot 76100, Israel.

Eur. J. Med. Chem. 34 (1999) 997−1001 997© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

a catalyst in ethanol at room temperature [11, 12]. 2-Oxo-2H-1-benzopyran-3-N-R-carboxamides4a and b wereobtained by acidic hydrolysis of the corresponding iminoanalogues3a and b employing a mixture of ethanol/water/≈ 32% hydrochloric acid and refluxing for 1 h.

The synthesis of coumarin-3-(N-2-carboxyphenyl)carboxamides4c–h was carried out as shown infigure 2.The precursor 2-carboxymalonanilic acid ester5 [13] was

condensed with an equivalent amount of salicylic alde-hydes2a–f to produce the desired coumarin derivatives4c–h utilizing piperidine as a catalyst in ethanol at roomtemperature. An alternative approach for synthesis ofN-substituted coumarin-3-carboxamides of type4, basedon rearrangements of 2-imino-2H-1-benzopyran-3-carboxamides under the action of anthranilic acid asN-nucleophile, has been recently developed [14] in ourlaboratory.

Physicochemical data for the compounds3a andb and4a–h are given intable I.

3. Pharmacology

The compounds synthesized were evaluated for theiranti-inflammatory activity in carrageenan-induced ratpaw oedema assays [15] and in acetic acid-induced peri-tonitis tests in albino rats [16]. 2-Imino/oxo-2H-1-benzopyran-3-carboxamides3a and b and 4a–h werefurther studied for their acute toxicity [17]. Piroxicamwas used as a control compound.

4. Results and discussion

Pharmacological results on anti-inflammatory activi-ties and acute toxicity of the benzopyran derivatives3aand b and 4a–h and piroxicam are summarized intable II.

In carrageenan-induced rat paw oedema assays, thecompounds4g, 3b, 4b, and4h were found to be the mostactive anti-inflammatory agents at the graded dose10 mg/kg po and exhibited 54± 6.50%, 51± 5.05%, 48± 6.51%, and 47± 7.13% inhibition of inflammation,respectively. Their effects were comparable to that of

Figure 1. Synthesis ofN-arylsubstituted 2-imino/oxo-2H-1-benzopyran-3-carboxamides3a andb and4a andb.

Figure 2. Synthesis ofN-arylsubstituted 2-oxo-2H-1-benzo-pyran-3-carboxamides4c–h.

Table I. Physicochemical data of the synthesizedN-arylsubstituted 2-imino/oxo-2H-1-benzopyran-3-carboxamides3a andb and4a–h.

Compound R1 R Molecular Yield Recryst. M.p. (°C)

formula (%) solvent

3a H 2-CO2CH3 C18H14N2O4 74 i-PrOH 137–1383b H 4-CO2C2H5 C19H16N2O4 82 i-PrOH 222–2244a H 2-CO2CH3 C18H13NO5 87 i-PrOH 205–2084b H 4-CO2C2H5 C19H15NO5 85 i-PrOH 246–2474c H 2-CO2H C17H11NO5 67 AcOH 275–276a

4d 6-OCH3 2-CO2H C18H13NO6 59 BuOH 124–1254e 8-OCH3 2-CO2H C18H13NO6 67 BuOH 140–1424f 6-NO2 2-CO2H C17H10N2O7 63 BuOH 147–1504g 6-Cl 2-CO2H C17H10ClNO5 71 i-PrOH 232–2354h 8-allyl 2-CO2H C20H15NO5 54 MeCN 249–251

aLit. [21] m.p. for 4c: 279 °C.

998

piroxicam, the reference compound, which showed 57±6.61% inhibition at the same dose level.

At the same dose level (10 mg/kg po), in acetic acidperitonitis tests, all compounds exhibited moderate togood anti-inflammatory activity. The most active com-pounds4c, 3a, and4a showed 42± 7.20%, 40± 6.43%,and 39± 5.09% inhibition of inflammation, respectively.In this test, piroxicam revealed a protection of 29±7.24%. All tested compounds were essentially non-toxicat the highest dose graded.

5. Conclusion

The products synthesized were found to be activeanti-inflammatory agents and their effects were compa-rable to that of piroxicam as the reference compound. Themost active compounds3b, 4a, 4b and 4g have beenmarked for further detailed pharmacological studies to beevaluated for COX-2 and COX-1 inhibition in micro-somal and cellular assays.

6. Experimental protocols

6.1. Chemistry

Melting points (°C) were measured with a Büchimelting point apparatus and were uncorrected. Thin layer

chromatography (TLC) was performed on aluminiumsheets precoated with silica gel (Merck, Kieselgel 60F-254). 1H-NMR spectra were recorded on a VarianWXR-400 spectrometer in DMSO-d6 using TMS as aninternal standard (chemical shifts inδ ppm), but a studyon isomerization of benzopyran-2-imines in DMSO-d6

has to be mentioned [18]. O’Callaghan et al. [18] re-vealed that when unsubstituted 2-imino-2H-1-benzo-pyran-3-carboxamide was dissolved in DMSO-d6, NMRspectra showed that a mixture of 2-imino-2H-1-benzopyran-3-carboxamide and the isomeric 2-cyano-3-(2-hydroxyphenyl)prop-2-ene-1-carboxamide was pre-sent and other benzopyran-2-imines behaved similarlyand the degree of isomerization varied considerably,depending on the nature and position of the substituentspresented. In our case, isomerization ofN-arylsubstituted2-imino-2H-1-benzopyran-3-carboxamides3a andb didnot occur in dimethyl sulfoxide-d6 at room temperatureand only starting materials were present. Mass spectra(MS) were obtained with a Finnigan MAT-4615B spec-trometer at an ionization potential of 70 eV. Infraredspectra (IR) were recorded in KBr pellets on an IBM 486PC computer-controlled Specord M-80 spectrometer. El-emental analyses were performed at the MicroanalysisLaboratory, Kharkov State University, and the combus-tion analyses of all compounds synthesized indicated bythe symbols of the elements were within± 0.4% oftheoretical values. TheN-substituted cyanoacetamides1aand b, which are key intermediates for synthesis of thebenzopyran-2-imines3a andb, were prepared according

Table II. Anti-inflammatory activities and acute toxicity of the benzopyran derivatives3a andb and4a–h and piroxicam.

Acute toxicity Anti-inflammatory activity

Compound Approximate LD50 carrageenan-induced paw oedema acetic acid peritonitis(mg/kg) (in mice)

(Mean % inhibition± SE)po ip 10 mg/kg po

3a > 1 000 > 700 19± 2.11 40± 6.43b

3b > 1 000 > 700 51± 5.05 32± 6.894a > 1 000 > 700 44± 7.21 39± 5.094b > 1 000 > 700 48± 6.51 35± 1.024c > 1 000 > 700 38± 1.01 42± 7.204d > 1 000 > 700 30± 6.96 18± 4.634e > 1 000 > 700 41± 6.42 32± 7.01b

4f > 1 000 > 700 35± 7.11 29± 5.58b

4g > 1 000 > 700 54± 6.50 35± 4.88b

4h > 1 000 > 700 47± 7.13 31± 3.31b

Piroxicam 360a > 1 000a 57 ± 6.61 29± 7.24b

aData from ref. [22];bP < 0.05 Student’st test versus controls.

999

to reported methods [19, 20] from ethyl cyanoacetate andmethyl anthranilate or ethyl 4-aminobenzoate.

6.1.1. N-arylsubstituted 2-imino-2H-1-benzopyran-3-carboxamides3a andb

To a well-stirred solution ofN-substituted cyanoaceta-mides1aandb (4 mmol) in 15 mL of ethanol, was addedan equivalent amount of salicylic aldehyde2 and a fewdrops of piperidine as a catalyst. The reaction mixturewas stirred at room temperature for 2 h. The products,which precipitated in the course of the reactions werefiltered and recrystallized from the proper solvent. Yieldsand physicochemical data of the synthesizedN-substituted 2-imino-2H-1-benzopyran-3-carboxamides3aandb are listed intable I. 3a: 1H-NMR: δ 3.88 (s, 3H,CH3); 7.13–7.25 (m, 3H, ArH); 7.48–7.58 (m, 2H, ArH);7.70 (d, 1H,J = 7.9 Hz, ArH); 7.88 (d, 1H,J = 7.9 Hz,ArH); 8.42 (d, 1H,J = 8.41 Hz, ArH); 8.50 (s, 1H, 4-CH);8.79 (s, 1H, C=NH); 12.94 (s, 1H, NH). MS m/z 322(M+⋅). IR (KBr), cm–1: m 3 300 (NH), 3 207 (NH), 3 041(CH), 1 715 (C=O), 1 678 (C=O), 1 641 (C=N), 1 606(C=C). Anal. C18H14N2O4 (C, H, N). 3b: 1H-NMR: δ1.31 (t, 3H,J = 6.9 Hz, CH3); 4.30 (q, 2H,J = 6.9 Hz,CH2); 7.27–8.04 (m, 8H, ArH); 8.59 (s, 1H, 4-CH); 9.32(s, 1H, C=NH); 13.16 (s, 1H, NH). MS m/z 336 (M+⋅). IR(KBr), cm–1: m 3 315 (NH), 2 976 (CH), 1 704 (C=O),1 688 (C=O), 1 644 (C=N), 1 593 (C=C). Anal.C19H16N2O4 (C, H, N).

6.1.2. N-arylsubstituted 2-oxo-2H-1-benzopyran-3-carboxamides4a andb

A solution of the corresponding 2-iminobenzopyranderivatives3a andb (4 mmol) in 15–20 mL of a mixtureof ethanol/water/≈ 32% hydrochloric acid (30:1:1, v/v/v)was refluxed with vigorous stirring for 1 h. After coolingto room temperature the products precipitated were fil-tered and recrystallized from the appropriate solvent.Yields and physicochemical data of the synthesizedN-substituted 2-oxo-2H-1-benzopyran-3-carboxamides4aandb are listed intable I. 4a: 1H-NMR: δ 3.92 (s, 3H,CH3); 7.20–8.08 (m, 7H, ArH); 8.61 (d, 1H,J = 8.2 Hz,ArH); 9.04 (s, 1H, 4-CH); 12.16 (s, 1H, NH). MS m/z 323(M+⋅). IR (KBr), cm–1: m 3 248 (NH), 3 042 (CH), 1 735(C=O), 1 713 (C=O), 1 668 (C=O), 1 608 (C=C). Anal.C18H13NO5 (C, H, N). 4b: 1H-NMR: δ 1.38 (t, 3H,J =7.0 Hz, CH3); 4.34 (q, 2H,J = 7.0 Hz, CH2); 7.42–7.54(m, 2H, ArH); 7.72–8.06 (m, 6H, ArH); 9.00 (s, 1H,4-CH); 10.88 (s, 1H, NH). MS m/z 337 (M+⋅). IR (KBr),cm–1: m 3 250 (NH), 2 984 (CH) 1 704 (C=O), 1 671(C=O), 1 596 (C=C). Anal. C19H15NO5 (C, H, N).

6.1.3. N-arylsubstituted 2-oxo-2H-1-benzopyran-3-carboxamides4c–h

To a well-stirred solution of 2-carboxymalonanilic acidester5 [13] (4 mmol) in 10 mL of ethanol was added anequivalent amount of salicylic aldehydes2a–f and a fewdrops of piperidine as a catalyst. The reaction mixturewas stirred at room temperature for ca. 1 day and thenpoured into water. The products precipitated were filteredand recrystallized from the suitable solvent. Yields andphysicochemical data of the synthesizedN-substituted2-imino-2H-1-benzopyran-3-carboxamides4c–h arelisted in table I. 4c: 1H-NMR: δ 7.15 (dd 1H,J = 8.0, 8.0Hz, ArH); 7.39 (m, 2H, ArH); 7.56 (dd, 1H,J = 8.0, 8.0Hz, ArH); 7.73 (dd, 1H,J = 7.9, 7.9 Hz, ArH); 7.94 (d,1H, J = 8.2 Hz, ArH); 8.05 (d, 1H,J = 8.2 Hz, ArH); 8.65(d, 1H,J = 8.3 Hz, ArH); 8.85 (s, 1H, 4-CH); 13.52 (br s,1H, NH). MS m/z 309 (M+⋅). IR (KBr), cm–1: m 3 266(NH), 3 032 (CH), 1 731 (C=O), 1 696 (C=O), 1 673(C=O), 1 608 (C=C). Anal. C17H11NO5 (C, H, N). 4d:1H-NMR: δ 3.95 (s, 3H, OCH3); 7.14–7.32 (m, 4H, ArH);7.42 (d, 1H,J = 8.2 Hz, ArH); 8.01 (d, 1H,J = 8.0 Hz,ArH); 8.60 (d, 1H,J = 8.0 Hz, ArH); 8.89 (s, 1H, 4-CH);13.20 (s, 1H, NH). MS m/z 339 (M+⋅). IR (KBr), cm–1: m3 287 (NH), 2 952 (CH), 1 726 (C=O), 1 694 (C=O),1 675 (C=O), 1 614 (C=C). Anal. C18H13NO6 (C, H, N).4e: 1H-NMR: δ 3.95 (s, 3H, OCH3); 7.15–7.44 (m, 4H,ArH); 7.52 (d, 1H,J = 8.3 Hz, ArH); 8.00 (d, 1H,J = 7.9Hz, ArH); 8.64 (dd, 1H,J = 8.0, 8.0 Hz, ArH); 8.90 (s,1H, 4-CH); 13.12 (s, 1H, NH). MS m/z 339 (M+⋅). IR(KBr), cm–1: m 3 291 (NH), 2 947 (CH), 1 730 (C=O),1 689 (C=O), 1 677 (C=O), 1 604 (C=C). Anal.C18H13NO6 (C, H, N). 4f: 1H-NMR: δ 7.21 (dd, 1H,J =8.0, 8.0 Hz, ArH); 7.60 (dd, 1H,J = 8.0, 8.0 Hz, ArH);7.71 (d, 1H,J = 8.2 Hz, ArH); 8.02 (d, 1H,J = 7.9 Hz,ArH); 8.46 (d, 1H,J = 8.3 Hz, ArH); 8.72 (d, 1H,J = 7.9Hz, ArH); 8.95 (s, 1H, 5-CH); 9.10 (s, 1H, 4-CH); 13.21(s, 1H, NH). MS m/z 354 (M+⋅). IR (KBr), cm–1: m 3 295(NH), 3 085 (CH), 1 754 (C=O), 1 724 (C=O), 1 683(C=O), 1 618 (C=C). Anal. C17H10N2O7 (C, H, N). 4g:1H-NMR: δ 7.20 (t, 1H,J = 7.9 Hz, ArH); 7.60–7.74 (m,3H, ArH); 7.98–8.06 (m, 2H, ArH); 8.12 (s, 1H, 5-CH);9.03 (s, 1H, 4-CH); 12.47 (s, 1H, NH). MS m/z 345, 343(M+⋅). IR (KBr), cm–1: m 3 389 (OH + NH), 3 047 (CH),1 736 (C=O), 1 689 (C=O), 1 671 (C=O), 1 607 (C=C).Anal. C17H10ClNO5 (C, H, N). 4h: 1H-NMR: δ 3.63 (m,2H, CH2–CH=CH2); 5.14 (m, 2H, CH2–CH=CH2); 6.03(m, 1H, CH2–CH=CH2); 7.13 (dd, 1H,J = 8.0, 8.0 Hz,ArH); 7.32 (dd, 1H,J = 7.9, 7.9 Hz, ArH); 7.55 (m, 2H,ArH); 7.78 (d, 1H,J = 8.0 Hz, ArH); 8.03 (d, 1H,J = 7.9Hz, ArH); 8.76 (dd, 1H,J = 7.9, 7.9 Hz, ArH); 8.98 (s,1H, 4-CH); 12.53 (s, 1H, NH) MS m/z 349 (M+⋅). IR(KBr), cm–1: m 3 224 (NH), 2 942 (CH), 1 724 (C=O),

1000

1 684 (C=O), 1 657 (C=O), 1 600 (C=C). Anal.C20H15NO5 (C, H, N).

6.2. Pharmacology

6.2.1. Anti-inflammatory activity

6.2.1.1. Carrageenan-induced rat hind paw oedema testExperiments were carried out on groups of five

Sprague-Dawley rats (140–160 g). The tested compoundsand reference drug were administered orally (po) in 0.5%methylcellulose solution in water and 1 h later 0.1 mL of1% carrageenan solution was injected under the plantaraponeurosis of the right hind paw of the rat by the methodof Winter et al. [15]. The volume of the paw wasmeasured before and 3 h after carrageenan treatment by amercury plethysmometer. Anti-inflammatory activity wasgiven as percentage of inhibition of oedema in treatedgroups compared with controls and was calculated ac-cording to the formula:

% inhibition = 100× [1 – (Vt/Vc)]

where Vt is the mean increase in paw volume of the ratstreated with tested compounds and Vc is the meanincrease in paw volume of the control group of rats.

6.2.1.2. Acetic acid peritonitis assayThis test was performed according to the procedure of

Arrigoni-Martelli [16]. Groups of five rats were admin-istered intraperitoneally (ip) with 10 mL/kg of 0.5%acetic acid solution 1 h after oral administration of thetested compounds. After 30 min, the rats were killed withdiethyl ether and peritoneal exudate was collected andmeasured. The anti-exudate response was expressed asthe inhibition percentage in comparison to the vehicle-treated control:

% inhibition = 100× [1 – (Vt/Vc)]

where Vt is the mean volume of the peritoneal exudate intreated rats and Vc is the mean volume of the peritonealexudate in vehicle-treated rats.

6.2.1.3. Toxicity studiesAll tested compounds were investigated for their acute

toxicity and approximate lethal dose (ALD50). Albinomice (either sex) weighing 20–25 g were used for thestudy. ALD50 values were determined by observingmortality within 24 h after drug administration [17].

6.2.1.4. Statistical calculationData are expressed as mean± SE. The Student’st test

was applied to determine the significance of the differ-ence between the control and the treated groups. The

difference in results was considered to be significantwhenP < 0.05.

Acknowledgements

We are grateful to Dr S.N. Kovalenko (Department ofOrganic Chemistry) for initial suggestions and helpfuldiscussions. We thank Dr S.M. Drogovoz (Department ofPharmacology) for providing the Department’s facilitiesfor pharmacological tests. We are indebted to Dr I.V.Ukrainets (Department of Pharmaceutical Chemistry) forhis generous gift of the compound5.

References

[1] Maxwell A., Mol. Microbiol. 9 (1993) 681–686.

[2] Dexeus F.H., Logothetis C.J., Sella A., Fitz K., Amato R., ReubenJ.M., Dozier N., J. Clin. Oncol. 8 (1990) 325–329.

[3] Zembower D.E., Liao S.Y., Flavin M.T., Xu Z.Q., Stup T.L.,Buckheit R.W. et al., J. Med. Chem. 40 (1997) 1005–1017.

[4] Negwer M. (Ed.), Organic-Chemical Drugs and Their Synonyms,Academie-Verlag, Berlin, 1987.

[5] Lino C.S., Taveira M.L., Viana G.S.B., Matos F.J.A., Phytother. Res.11 (1997) 211–215.

[6] Silvan A.M., Abad M.J., Bermejo P., Sollhuber M., Villar A., J. Nat.Prod. 59 (1996) 1183–1185.

[7] Naka Y., Yamada I., Ochi H., Matsuura M., Nagamatsu Y., Oe T.,Bioorg. Med. Chem. Lett. 5 (1995) 959–962.

[8] Ukrainets I.V., Gorokhova O.V., Taran S.G., Turov A.V., Chem.Heterocycl. Compd. (NY) 30 (1994) 1211–1213.

[9] Tietze L.F., Beifuss U., in: Trost B.M., Fleming I. (Eds.), Compre-hensive Organic Synthesis Vol. 2, Pergamon Press, Oxford, 1991,pp. 341–394.

[10] Jones G., Org. React. (NY) 15 (1967) 204–599.

[11] Schiemenz G.P., Chem. Ber. 95 (1962) 483–486.

[12] Czerney P., Hartmann H., J. Prakt. Chem. 323 (1981) 691–693.

[13] Ukrainets I.V., Bezugly P.A., Treskach V.I., Taran S.G., GorokhovaO.V., Tetrahedron 50 (1994) 10331–10338.

[14] Bilokin Y.V., Kovalenko S.N., Bylov I.E., Chernykh V.P.,Heterocycl. Commun. 4 (1998) 257–260.

[15] Winter C.A., Risley E.A., Nuss G.W., Proc. Soc. Exp. Biol. Med.111 (1962) 544–547.

[16] Arrigoni-Martelli E., Boll. Chim. Farm. 107 (1968) 205–217.

[17] Swinyard E.A., Brown W.C., Goodman L.S., J. Pharmacol. Exp.Ther. 106 (1952) 319–330.

[18] O’Callaghan C.N., McMurry T.B.H., O’Brein J.E., J. Chem. Soc.Perkin Trans. 2 (1998) 425–429.

[19] Ukrainets I.V., Taran S.G., Bezugly P.A., Kovalenko S.N., TurovA.V., Marusenko N.A., Khim. Geterotsikl. Soedin. (1993)1223–1226.

[20] Huang C.K., Wu F.Y., Ai Y.X., Bioorg. Med. Chem. Lett. 5 (1995)2423–2428.

[21] Selim M.R., Aly F.M., Bendair A.H., Abu-Shanab F.A., J. IndianChem. Soc. 69 (1992) 688–690.

[22] Wiseman E.H., Chang Y.H., Lombardino J.G., Arzneim. -Forsch. 26(1976) 1300–1303.

1001

Short communication

Research on heterocyclic compounds, XLI.2-Phenylimidazo[1,2-b]pyridazine-3-acetic derivatives: synthesis

and anti-inflammatory activity

Antonia Sacchia, Sonia Laneria*, Francesca Arenaa, Enrico Abignentea,Marina Gallitellia, Michele D’amicob, Walter Filippellib, Francesco Rossib

aDipartimento di Chimica Farmaceutica e Tossicologica,Facoltà di Farmacia, Università di Napoli Federico II, Via Domenico Montesano 49, I-80131 Napoli, Italy

bIstituto di Farmacologia e Tossicologia, Facoltà di Medicina e Chirurgia,II Università di Napoli, Via Costantinopoli 16, I-80138 Napoli, Italy

(Received 21 January 1999; accepted 21 April 1999)

Abstract – The synthesis of a group of 2-phenylimidazo[1,2-b]pyridazine-3-acetic esters and acids is described. The structures of the newcompounds are supported by1H-NMR spectra. These compounds were tested in vivo for their anti-inflammatory, analgesic and ulcerogenicactivity. All new compounds showed remarkable anti-inflammatory action in the carrageenan rat paw oedema (one third of that forindomethacin) but no significant analgesic activity in the acetic acid writhing test together with negligible ulcerogenic action, and were alsofound to be lacking inhibitory activity on cyclooxygenase in vitro. © 1999 Éditions scientifiques et médicales Elsevier SAS

imidazo[1,2-b]pyridazines / anti-inflammatory activity / analgesic activity / ulcerogenic activity / cyclooxygenase inhibition

1. Introduction

In the context of our research on the structure-activityrelationships and mode of action of bicyclic imidazo-derivatives with anti-inflammatory and analgesic activity,we synthesized a series of 2-phenylimidazo[1,2-b]pyri-dazine-3-carboxylic acids1 (figure 1) which showed highanalgesic activity in the acetic acid writhing test in mice,low or no anti-inflammatory activity in the carrageenan-induced rat paw oedema and low ulcerogenic action onthe rat gastric mucosa [1, 2]. We then synthesized threeseries of analogues, namely 2-methylimidazo[1,2-b]-pyridazine-3-carboxylic acids2 [3], imidazo[1,2-b]-pyridazine-2-acetic acids3 [4] and imidazo[1,2-b]-pyridazine-2-carboxylic acids4 [5].

In comparison with the first series of acids1, com-pounds 2 and 3 showed a lower analgesic activity,whereas the same activity reached the lowest level in thelast series (acids4). However, the pharmacological pro-file was the same in all four groups of compounds, since

the analgesic activity was always coupled with low or noanti-inflammatory and ulcerogenic action.

In consideration of the fact that the highest analgesicactivity was shown by the first type of compounds, i.e.2-phenylimidazo[1,2-b]pyridazine-3-carboxylic acid1,and in order to extend the study of the structure-activityrelationships, we have synthesized a new series of2-phenyl derivatives, with only one structural change, i.e.the replacement of the carboxylic moiety in position 3 byan acetic group.*Correspondence and reprints

Figure 1. Imidazo[1,2-b]pyridazine acidic derivatives.

Eur. J. Med. Chem. 34 (1999) 1003−1008 1003© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

The required compounds were prepared using thesynthetic method depicted infigure 2. The reaction in dryethanol of 3-aminopyridazines5a–g with ethyl3-benzoyl-3-bromopropionate6, prepared ad hoc, af-forded the corresponding ethyl 2-phenylimidazo[1,2-b]pyridazine-3-acetates7a–g. These esters were con-verted into the respective acids8a–gby acid hydrolysis.Esters7a, 7b and 7c were prepared by reacting6 with3-aminopyridazine5a, 3-amino-6-chloropyridazine5b,and 3-amino-6-methoxypyridazine5c, respectively, pre-pared ad hoc following the methods described by Wer-muth et al. [6] and Steck et al. [7].

The couples of isomers7d–eand7f–g were obtainedwith the same method previously employed to obtain thecorresponding isomeric couples in the preceding series ofimidazo[1,2-b]pyridazines1–4[5]. The couple7f–g was

prepared using as starting material a mixture (nearly 1:1)of 3-amino-6-chloro-5-methylpyridazine (5f) and3-amino-6-chloro-4-methylpyridazine (5g): this mixturewas obtained from 3,6-dichloro-4-methylpyridazine [8]following the Mori procedure [9]. After reaction with6,the products7f and7g were separated by column chro-matography. The above mixture of the amines5f and5gafforded the corresponding dehalogenated mixture of3-amino-5-methylpyridazine (5d) and 3-amino-4-methyl-pyridazine (5e) by catalytic hydrogenation. The reactionof 6 with the mixture5d–5eafforded the esters7d and7e.

The correct structural assignments to these productswere performed by1H-NMR spectra (table I), and are inaccordance with the literature data, in particular with thechemical shifts found by Kobe et al. [10] for H-6, H-7,H-8 in the imidazo[1,2-b]pyridazine.

3. Pharmacology

The new esters7a–g and acids8a–g were tested invivo using the acetic acid writhing test in mice andcarrageenan induced rat paw oedema to study analgesicand anti-inflammatory activity. Higher doses were admin-istered to rats to study the irritative and ulcerogenicaction on the mucosa of the stomach and small intestine.Indomethacin was used in all tests as reference drug.These three tests should allow us not only to extend theSAR study to another series of imidazopyridazines, butalso to obtain some information about their mechanism ofaction: they should display a similar level of activity in allthree tests if they are inhibitors of prostaglandin biosyn-thesis. In order to unequivocally resolve this question,some new compounds (the more and less active ones)were also subjected to two different cyclooxygenaseactivity assays in vitro [11, 12].

4. Results and discussion

The anti-inflammatory activity displayed by the com-pounds under examination is reported intable II. Theesters7c, 7d and7g, and the acids8a, 8c, 8f and8g arethe most active compounds with comparable levels ofactivity, as can be seen on the basis of reported values ofED50 (approximately one third of that for indomethacin).The results obtained in the acetic acid writhing test arequite different (table III), in fact all compounds showweak or no activity. The gastrointestinal irritative andulcerogenic action (table IV) was almost completely ab-sent in all compounds.

There is no parallelism between the results of all threein vivo tests. This is certainly not the pharmacological

Figure 2. Synthetic method for 2-phenylimidazo[1,2-b]pyri-dazine-3-acetic derivatives.

1004

Table I. Yields, m.p. and1H-NMR spectral data of 2-phenylimidazo[1,2-b]pyridazine-3-acetic derivatives.

Compounds1 yield m.p. 1H-NMR in ppm2

% °C δ: H-6 H-7 H-8 J6, 7 J6, 8 J7, 8 Ph CH2 Ethyl SubstituentsHz (s)

7a 32 103–105 8.50 (dd) 7.20 (dd) 7.85 (dd) 4.5 2.0 10 7.70 (m) 7.60 (m) 4.10 4.00 (q) 1.00 (t)7b 20 108–110 7.00 (d) 7.90 (d) 10 7.70 (m) 7.40 (m) 4.10 4.05 (q) 1.20 (t)7c 40 104–105 6.65 (d) 8.70 (d) 10 8.60 (m) 8.40 (m) 4.10 4.05 (q) 1.10 (t) 6-OCH3: 3.90 (s)7d 40 110–112 8.25 (d) 7.95 (d) 2.4 7.50 (m) 7.40 (m) 4.10 4.08 (q) 1.20 (t) 7-CH3: 2.60 (s)7e 40 121–123 8.15 (d) 7.19 (d) 4.3 7.60 (m) 7.50 (m) 4.10 4.05 (q) 1.30 (t) 8-CH3: 2.80 (s)7f 30 113–115 7.70 (s) 7.65 (m) 7.50 (m) 4.20 4.10 (q) 1.30 (t) 7-CH3: 2.50 (s)7g 30 125–127 6.85 (s) 7.80 (m) 7.50 (m) 7.30 (m) 4.20 4.10 (q) 1.30 (t) 8-CH3: 2.70 (s)8a 55 > 200 8.49 (dd) 7.20 (dd) 7.80 (dd) 4.2 2.0 10 7.60 (m) 7.55 (m) 4.058b 55 > 200 7.21 (d) 7.95 (d) 10 7.90 (m) 7.68 (m) 4.108c 50 > 200 6.75 (d) 8.77 (d) 10 8.67 (m) 8.42 (m) 4.10 6-OCH3: 3.90 (s)8d 55 > 200 8.35 (d) 8.04 (d) 2.3 7.60 (m) 7.45 (m) 4.10 7-CH3: 2.65 (s)8e 60 > 200 8.25 (d) 7.20 (d) 4.3 7.65 (m) 7.50 (m) 4.10 8-CH3: 2.55 (s)8f 60 > 200 7.75 (s) 7.90 (m) 7.50 (m) 4.10 7-CH3: 2.55 (s)8g 60 > 200 6.80 (s) 7.82 (m) 7.45 (m) 4.20 8-CH3: 2.68

1All compounds were analysed for C, H, N (also for Cl when present): found values were within± 0.4% compared with theoretical values.2Solvents: CDCl3 for 7a–g, CD3OD for 8a–g.

Table II. Anti-inflammatory activity by the carrageenan rat paw oedema test.

Compound Dose mg/kgp.o.

% Oedema inhibition relative to control at the: ED50, mg/kg (fiducial limits)1st h 2nd h 3rd h 4th h 3rd h 4th h

7a 40 –17 –26 –29 –44 – –7b 40 –7 –14 –19 –21 – –7c 20 –62 –53 –34 –37 35.1 28.9

40 –17 –26 –47 –59 (29.4–41.8) (24.0–34.8)80 –52 –72 –80 –87

7d 20 –28 –14 –24 –42 36.5 –40 –63 –56 –53 –51 (32.1–41.7)80 –62 –78 –84 –90

7e 40 –17 –26 –47 –44 – –7f 40 0 –20 –43 –44 – –7g 20 –5 –17 –20 –37 – 33.2

40 –27 –34 –37 –51 (27.0–40.7)80 –57 –75 –82 –77

8a 10 –30 –31 –22 –32 – 21.420 –47 –42 –33 –47 (16.7–27.5)40 –53 –68 –73 –66

8b 40 –14 –31 –33 –24 – –8c 10 –30 –20 –22 –24 31.4 27.8

20 –58 –42 –41 –34 (24.5–40.1) (22.6–34.2)40 –83 –66 –55 –64

8d 40 –67 –51 –43 –25 – –8e 40 –33 –51 –43 –38 – –8f 10 –22 –33 –33 –22 17.8 19.8

20 –66 –65 –51 –55 (14.7–21.5) (17.0–23.1)40 –86 –81 –76 –75

8g 10 –22 –28 –22 –25 21.9 23.020 –58 –42 –41 –34 (18.9–25.4) (1.6–33.7)40 –86 –72 –76 –75

IMA 5 –3 –40 –38 –35 7.0 6.77.5 –16 –33 –49 –55 (4.5–10.8) (4.8–8.7)10 –39 –55 –67 –79

1005

profile expected from compounds acting as inhibitors ofthe prostaglandin biosynthesis. In order to investigate thisaspect of the question, our compounds were tested invitro for their cyclooxygenase-inhibiting activity.

The most active compounds7c, 7d, 8c, 8f and8g weretested for their cyclooxygenase-inhibiting activity bymeasuring the rate of conversion of [1-14C]arachidonicacid into PGE2 in the microsomal fraction of mucosapreparation of rabbit distal colon after incubation withtest compounds, following the method previously re-ported [13].

All compounds were found to be devoid of inhibitoryactivity, i.e. 0.7–9.0% relative to control, compared with90–92% of indomethacin at the same concentration(10 µM).

The second test was carried out by measuring the rateof conversion of exogenous arachidonic acid into PGE2

in the rat medullary and cortical kidney microsomes.Inhibition of microsomal PGE2 production was measuredby radioimmunoassay as reported previously [13]. Thistest confirmed the above results: no compound showedsignificant activity (< 12% relative to control), exceptindomethacin (90–94%) at the same concentration(10 µM).

Therefore in the present case the remarkable anti-inflammatory activity shown by these compounds wasfound to be independent of the cyclooxygenase inhibi-tion. It should be noted that we recently came to the sameconclusion for a group of imidazo[1,2-a]pyrimidine ana-logues [14].

From the point of view of our studies on structure-activity relationships of imidazo[1,2-b]pyridazine acidicderivatives1–4 (figure 1), the most significant finding isthat the 2-phenylimidazo[1,2-b]pyridazine-3-acetic de-rivatives (esters7 and acids8) showed weak analgesicactivity and proved to be completely lacking ulcerogenicaction, whereas showed a significant anti-inflammatoryactivity. The pharmacological profile resulting from theabove data for these 2-phenylimidazo[1,2-b]pyridazine-3-acetic derivatives is different from that previouslyobserved for1–4 series in which the analgesic activitywas clearly prevailing over the anti-inflammatory action,so further investigation will be necessary in order toexplain the activity of these imidazo-derivatives, which isprobably due to the multiple mechanism of action differ-ently influenced by the structural changes.

5. Experimental protocols

5.1. Chemistry

Thin layer chromatography by precoated silica gelplates (Merck 60 F254) was used to control the course ofreactions and purity of products: all compounds weredesignated as pure when they showed a single spot afterelution with chloroform/methanol mixture (95:5); detec-tion of components was made by UV light and/ortreatment with iodine vapors. Preparative separationswere performed in columns packed with silica gel fromFarmitalia Carlo Erba (RS,J mm 0.05:0.20). Meltingpoints were determined with a Kofler hot stage micro-scope and are uncorrected. Elemental analyses indicatedby the symbols of the elements were within± 0.4% of thetheoretical values. The1H-NMR spectra were recordedusing a Bruker AMX- 500 spectrometer equipped with aBruker X-32 computer; chemical shift values are reported

Table III. Analgesic activity by the acetic acid writhing test inmice.

Compound Dose mg/kg p.o. % Decrease of mean no. ofwrithes in 25 min after treat-ment relative to control

7a 40 –22.57b 40 –4.57c 40 –34.47d 40 –8.87e 40 –32.37f 40 –4.07g 40 –25.18a 40 –17.28b 40 –8.98c 40 –26.58d 40 –10.88e 40 –9.38f 40 –29.68g 40 –15.8IMA 5 –51.2

Table IV. Incidence of gastrointestinal lesions in rats.

Compound Dose mg/kgp.o.

Remarks at 6 h after treatment:% animal with:hyperaemia ulcers

7a 100 20 07b 100 10 07c 100 30 07f 100 30 07g 100 20 08a 100 30 08b 100 40 08c 100 20 08f 100 30 08g 100 30 0IMA 5 80 50

1006

in δ units (ppm) relative to tetramethylsilane used asinternal standard.

5.1.1. Ethyl 3-benzoyl-3-bromopropionate6A solution of 3-benzoylpropionic acid (19.6 g,

0.1 mol) in 100 mL of dry ethanol with concentratedH2SO4 (5 mL) was refluxed for 7 h. After cooling,ethanol was removed in vacuo, and the residue dissolvedin diethyl ether and extracted with NaHCO3 saturatedsolution. The organic extract was washed with water,dried on Na2SO4 and evaporated in vacuo to obtain ethyl3-benzoylpropionate as an oil in 75% yield.

A solution of ethyl 3-benzoylpropionate in 100 mL ofdiethyl ether was added slowly with an equimolar amountof bromine at 0 °C. The solution was stirred at roomtemperature for 1 h. The organic solution was washedthree times with NaHCO3 saturated solution, dried onNa2SO4 and evaporated in vacuo to obtain the requiredcompound6 as an oil in 65% yield.

5.1.2. Ethyl-2-phenylimidazo[1,2-b]pyridazine-3-aceta-tes7a and7c–g

General procedure: a mixture of the starting 3-amino-pyridazine and ethyl 3-benzoyl-3-bromopropionate6(molar ratio 1:1.5) in ligroine was refluxed for 3 h. Aftercooling, the mixture was filtered and the filtrate wasextracted with 10% aqueous HCl. The aqueous layer wasseparated and adjusted to pH 7–8 with NaHCO3 to obtainthe precipitation of the ester which was then recrystal-lized from n-hexane. This procedure allowed us to obtainthe esters7a and7c.

In the case of the isomeric mixtures7d–e and 7f–g,obtained starting from the mixtures of the isomericamines5d–e and 5f–g, respectively, the crude productprecipitated was subjected to column chromatographicseparation, eluting with n-hexane/diethyl ether mixtureswith increasing percentage of ether. The single productsobtained from the columns were then recrystallized fromn-hexane.

5.1.3. Ethyl 6-chloro-2-phenylimidazo[1,2-b]pyridazine-3-acetate,7b

A mixture of a 3-amino-6-chloropyridazine5b andethyl 3-benzoyl-3-bromopropionate6 (molar ratio 1:1.5)in anhydrous ethanol was refluxed for 10 h. After cooling,ethanol was removed in vacuo, and the residue treatedwith NaHCO3 saturated solution and extracted withCHCl3. The organic extract was washed with water, driedon Na2SO4 and evaporated in vacuo to obtain therequired product7b, which was recrystallized fromn-hexane.

5.1.4. 2-Phenylimidazo[1,2-b]pyridazine-3-acetic acids8a–g

General procedure: a mixture of 10 mmol of each ethylester and 40 mL of 10% aqueous HCl was refluxed for2–3 h. After cooling, the solution was adjusted to pH 4–5with NaHCO3 to obtain the precipitation of the acidwhich was recrystallized from ethanol.

5.2. Pharmacology

As regards the experiments carried out in vivo, testcompounds were administered orally by gavage in 1%methylcellulose suspension. In the oedema and writhingtest each compound was first tested at 40 mg/kg. If asignificant activity was observed, lower and/or higherdoses were administered in order both to study thedose-dependence of the pharmacological activity and tocalculate ED50 values, when possible. Gastric ulcero-genic action was studied in rats which were treated orallywith higher doses (100 mg/kg).

Indomethacin was included in all tests for comparisonpurposes (IMA intables II–IV).

5.2.1. Anti-inflammatory activityThe paw oedema inhibition test [15] was used on rats.

Groups of 5 rats of both sexes (body weight 150–200 g),pregnant females excluded, were given a dose of a testcompound. After 30 min, 0.2 mL of 1% carrageenansuspension in 0.9% NaCl solution was injected subcuta-neously into the plantar aponeurosis of the hind paw andthe paw volume was measured by a water plethysmom-eter Socrel and then measured again 1, 2, 3 and 4 h later.The mean increase of paw volume at each time intervalwas compared with that of a control group (5 rats treatedwith carrageenan, but not treated with test compounds) atthe same time intervals and percentage inhibition valueswere calculated. The experimental results are listed intable II.

5.2.2. Analgesic activityAcetic acid writhing test [16] was used on mice.

Groups of 5 mice of both sexes (body weight 20–25 g),pregnant females excluded, were given a dose of a testcompound. After 30 min the animals were injected intra-peritoneally with 0.25 mL/mouse of 0.5% acetic acidsolution and writhes were counted during the following25 min. The mean number of writhes for each experimen-tal group and percentage decrease compared with thecontrol group (5 mice not treated with test compounds)were calculated. The experimental results are listed intable III.

1007

5.2.3. Ulcerogenic actionGroups of 10 rats of both sexes (body weight

200–220 g), pregnant females excluded, fasted for 24 h,were given an oral dose of a test compound, except thecontrol group. All animals were sacrificed 6 h afterdosing and their stomachs and small intestine weremacroscopically examined to assess the incidence ofhyperaemia and ulcers. The experimental results arelisted in table IV.

Acknowledgements

The NMR spectra were performed at the “Centro diRicerca Interdipartimentale di Analisi Srtumentale,” Uni-versità di Napoli “Federico II,” Naples. The assistance ofstaff is much appreciated. This research has been finan-cially supported by the Italian Ministry of University andScientific and Technological Research (MURST, Rome).

References

[1] Abignente E., Arena F., Luraschi E., De Caprariis P., Marmo E.,Vitagliano S., Donnoli D., Res. Commun. Chem. Pathol. Pharmacol.67 (1990) 43–54.

[2] Abignente E., Arena F., Luraschi E., Saturnino C., Marmo E.,Berrino L., Donnoli D., Farmaco 45 (1990) 1075–1087.

[3] Abignente E., Arena F., Luraschi E., Saturnino C., Rossi F., BerrinoL., Cenicola M.L., Farmaco 47 (1992) 931–944.

[4] Luraschi E., Arena F., Sacchi A., Laneri S., Abignente E., D’AmicoM., Berrino L., Rossi F., Farmaco 50 (1995) 349–354.

[5] Luraschi E., Arena F., Sacchi A., Laneri S., Abignente E., AvalloneL., D’Amico M., Berrino L., Rossi F., Farmaco 52 (1997) 213–217.

[6] Wermuth C.G., Bourguignon J.J., Schlewer G., Gies J.P., Schoen-felder A., Melikian A. et al., J. Med. Chem. 30 (1987) 239–249.

[7] Steck E.A., Brundage R.P., Fletcher L.T., J. Am. Chem. Soc. 76(1954) 3225–3226.

[8] Steck E.A., Brundage R.P., Fletcher L.T., J. Am. Chem. Soc. 76(1954) 4454–4457.

[9] Mori K., Yakugaku Zasshi, 82 (1962) 303-309; Chem. Abs., 58(1963) 3427h.

[10] Kobe J., Stanovnik B., Tisler M., Tetrahedron 24 (1968) 239–243.

[11] Calderaro V., Parrillo C., Giovane A., Greco R., Matera M.G.,Berrino L., Rossi F., J. Pharmacol. Exp. Ther. 263 (1992) 579–587.

[12] Gans K.R., Galbraith W., Roman R.J., Haber S.B., Kerr J.S.,Schdmit W.K., Smith C., Hewes W.E., Ackerman N.R., J. Pharma-col. Exp. Ther. 254 (1990) 180–187.

[13] Abignente E., Sacchi A., Laneri S., Rossi F., D’Amico M., BerrinoL., Calderaro V., Parrillo C., Eur. J. Med. Chem. 29 (1994) 279–286.

[14] Laneri S., Sacchi A., Gallitelli M., Arena F., Luraschi E., AbignenteE., Filippelli W., Rossi F., Eur. J. Med. Chem. 33 (1998) 163–170.

[15] Winter C.A., Risley E.A., Nuss G.W., Proc. Soc. Exp. Biol. Med.111 (1962) 544–547.

[16] Davies J.E., Kellet D.N., Pennington J.C., Arch. Int. Pharmacodyn.Ther. 221 (1976) 274–282.

1008

Short communication

2-Substituted indazoles. Synthesis and antimicrobial activity

Tamás Lóránda*, Béla Kocsisb, Levente Emôdyb, Pal Sohárc

aDepartment of Medical Chemistry, University Medical School, H-7601 Pécs, POB 99, HungarybDepartment of Medical Microbiology and Immunology, University Medical School, H-7601 Pécs, POB 99, Hungary

cDepartment of General and Inorganic Chemistry, Loránd Eötvös University, H-1518 Budapest, POB 32, Hungary

(Received 24 November 1998; revised 4 May 1999; accepted 7 May 1999)

Abstract – 2-Isothiocarbamoyl substituted fused pyrazolines and their S-alkyl derivatives were prepared as potentially antimicrobial agents.Conventional methods were used to synthesize the novel derivatives starting from cyclic unsaturated ketones and thiosemicarbazide underacidic catalyst. These cyclizations yielded only one diastereoisomer of 3-H, 3a-Hcis. The alkylations were performed applying alkyl halides.The structures of the new compounds, including configurations and conformations, were elucidated by NMR spectroscopy, also making useof 2D-HSC, DEPT and DNOE measurements. TheS-alkyl derivatives were evaluated for activity against Gram-negative and Gram-positivebacteria and their in vitro toxicity was determined on HeLa cells. The structure-activity relationship was also studied. © 1999 Éditionsscientifiques et médicales Elsevier SAS

indazole / alkylation / stereostructure by NMR / antibacterial effect / in vitro toxicity

1. Introduction

Some thiosemicarbazides are known antibacterial com-pounds, like thiosemicarbazones of 5-nitrofurfurylidene-acetone [1] and dodecanone [2]. The 1-methylindole-2,3-dione 3-thiosemicarbazone, called Metisazone, was usedas an antiviral agent [3]. Because of their relatively hightoxicity these agents are not widely used drugs. Similarly,some members of the family of the fused pyrazolines andpyrazoles show antimicrobial effects [4–6]. Thisprompted us to find an effective antimicrobial agent ofrelatively low toxicity having a thioamide moiety at-tached to a pyrazoline ring. Recently we have publishedthe synthesis of several bi- and tricyclic pyrazolinesstarting from unsaturated ketones and hydrazine deriva-tives [7–9]. Some of these compounds can be consideredas cyclic thiosemicarbazide derivatives too. Our aim wasto prepare water-soluble thiosemicarbazide derivativesfor microbiological investigations. Therefore, 3,5-diarylidene-1-methyl-4-piperidones were also used asstarting ketones.

2. Chemistry

In order to obtain potentially antibacterial compoundsstarting from 2,6-dibenzylidenecyclohexanone, 3,5-diarylidene-1-methyl-4-piperidones or 2-arylidene-1-tetralones and semicarbazide or thiosemicarbazides, tennew bi- and tricyclic pyrazolines (4b, 8a–dand 11b–f)have been prepared (figures 1–3). The cyclizations per-formed with thiosemicarbazides under acidic conditionsyielded only one diastereoisomer of 3-H, 3a-Hcis. Whilethe reactions with semicarbazide afforded the mixture ofthe 3-H, 3a-Hcis andtransdiastereoisomers, which havebeen separated. The structure and the relative configura-tion of the compounds have been determined by1H-NMRand13C-NMR spectroscopic methods.

To increase the water solubility of the 2-isothio-carbamoyl pyrazolines they have been alkylated withalkyl halides. These reactions gave the correspondingS-alkyl derivatives (5, 6a–b, 9a–dand12a–g) depicted infigures 1–3. From 6a hydroiodide, the free base6c wasliberated to provide a compound of better solubility forNMR study. At the alkylation of8c–d pyrazolo[4,3-c]pyridines the corresponding hydrochloride was used toavoid the quaternarization of the pyridine ring. On the*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1009−1018 1009© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

contrary, allylic halides, like allyl bromide and benzylchloride, yielded the13 and 14 N-alkyl derivatives(figure 4). N-alkylation was observed at thioamides withalkylating agents favouring the carbonium ion forma-tion [10]. The structures of theN-alkylated andS-alkylated derivatives were proven by their1H- and13C-NMR spectra. Two of theS-alkyl derivatives (6a and9a) were prepared via an alternative route usingS-methylthiosemicarbazide hydroiodide and the corre-spondingα,â-unsaturated ketone. The products of thisone-step synthesis were identical in every respect withthose prepared in a two-step method. The physical data ofthe novel compounds are displayed intable I.

The1H- and13C-NMR data on the new compounds arelisted in tables II and III . As the spectra inequivocallyprove the expected constitutions, the following discus-sion is focused on the steric aspects of the structures, onthe configurations and conformations.

Two problems have to be considered: thecis or transconfiguration of the H-3, H-3a hydrogens and the hin-dered rotation of the carbamide or thiocarbamide moiety.

Concerning the first problem, the spectral data of thecis-transpair 8a andb can serve as a starting point. Thevicinal H-3, H-3a coupling is 9.3 and 11.5 Hz for the

isomers. While the values are in accordance with theexpected ratio3Jcis > 3Jtrans [11], the difference is far toosmall for a firm differentiation of thecis or transconfigurations in the case of single compounds withouttheir counterparts. Moreover, because of broadened sig-nals, it was not possible to determine the value of thiscoupling constant for all compounds and in most casesthe splitting was between the two values (about 10–11Hz) measured for the pair8a andb. On the contrary, the13C–NMR field effect [12, 13] arising in thecis isomers isa firm base to identify the C-3,3a configurations.

The C-3a chemical shifts for thecis andtrans isomers8a andb are 48.4 and 55.4 ppm. The significant upfieldshift (8.2 ppm) for C-3a in8a and the similarδC-3avalues (48.4–51.0 ppm) measured for the other com-pounds proved theircis-configuration unambiguously.

Signal splitting or broadening in the1H- and13C-NMRof most of the salts investigated refers to the equilibriumof rotamers or other differences in the structures. Becauseof the absence of these phenomena in the spectra of bases,the origin can also be found in different protonation sites:besides exocyclic, the basic cyclic sp2-N can also beprotonated. The double splitting of H-8 and C-7a signalsfor 12-type salts supports this latter assumption, whilesplitting or broadening of H-3 and C-3 signals can be

Figure 1. Synthesis of substituted hexahydroindazoles.

Figure 2. Synthesis of hexahydropyrazolo[4,3-c]pyridines.

1010

interpreted by hindered rotation only. Probably, the abovefacts have to arise from hindered rotation and/or siteisomers of protonated molecules, depending on the struc-ture as well as medium effects (solvent, concentration,temperature, pH, etc.).

Two further aspects of stereostructures should beconsidered: the geometrical isomerism of the benzylidenemoiety and the conformation of the partly saturatedcondensed ring. As regards the geometrical isomers, theZ-isomer (S-cis Ar and N-1) is to be precluded due tounfavourable steric structure. TheS-cis (to N-1) positionof the hydrogen in the exocyclic =CH- group reveals ahigh downfield shift of its1H-NMR signal (this appearsto overlap with the multiplets of the phenyl hydrogens)due to anisotropic deshielding of the non-bonded electronpair on N-1 [14].

The condensed, partly saturated six-membered ringcan exist in two preferred conformations. These chair-and boat-like forms with out-of-plane C-5 and C-7aatoms are both distorted approaching the sofa-

arrangement with coplanar C-7a. It is to be noted that theinversion of the six-membered ring hardly influences theenvelope form of the pyrazoline ring (with an out-of-plane C-3 atom) bearing the Ar group in quasi-equatorialposition, cf. Ref [7]. Nevertheless, the quartet-like signalof H-4ax (for 4b, 5, 6a–c, 11c, 13 and14) refers to threelarge H,H-splittings, one of which is due to geminalcoupling [15] 4ax,4eq and the two others to 3a,4ax and4ax,5ax diaxial vicinal interactions [16, 17] which can bepresent only in the chair-form. This conformation is alsopreferred in the case of8a and c, but the H-4ax signalshows a triplet splitting in the absence of H-5axhere. Dueto signal-overlap, it was not possible to identify themultiplicity of the H-4ax signal for9a andd, but in theTFA-d solution of 9a it is separated with triplet-likemultiplicity and the two large splits are proof of the sameconformation. In the1H-NMR spectra of12a,c andg thesignals are broadened and consequently, it was notpossible to identify the multiplicities. It is conceivablethat the signal-broadening is the consequence of a slowring inversion combined with the hindered rotation of theprotonated thiocarbamide group. That is, the conforma-tionally homogeneous quasi-rigid system of the othercompounds is substituted by a flexible equilibrium of ringisomers for9-type derivatives. Again, this equilibriumcan be influenced in12g by the incorporated sulphur inposition 5.

3. Biology

Antibacterial activity tests of the synthesized com-pounds were carried out on twenty cultures ofStaphylo-coccus aureusandEscherichia colistrains isolated fromdifferent clinical samples (see Experimental protocols).Eight of the most effective antibacterial compounds (5,

Figure 3. Synthesis of tricyclic pyrazolines.

Figure 4. N-alkyl derivatives13, 14 and monocyclic pyrazolederivative15.

1011

6a, 9a, 9d and12a–d) were examined for determinationof their minimum inhibitory concentration (MIC) valuesby the test tube dilution method (see Experimentalprotocols). In these experiments we used standard refer-ence strains. The MIC values of these five standardstrains were tested on eight well-known antibiotics tocompare our compounds with antibiotics used in therapyas well. In addition, the MIC values of compound9awere investigated on fiveE. coli strains cultivated fromurine. In vitro cytotoxicity tests of the compounds onHeLa cell-line OHIO, was carried out on microplates (seeExperimental protocols).

4. Results and discussion

Our aim was to study the structure-biological activityrelationship for these fused pyrazolines by varying the

type and size of the ring system, the aromatic substituentat position 3 and the substituents on the isothiocarbamoylgroup.

4.1. Antibacterial results

Our work provided us with information about thestructure-antimicrobial activity relationship connected tothe class of fused pyrazolines (table IV). Opposed to theMetisazone they are ineffective without anS-alkyl sub-stituent, like 8c. An arylidene substituent or a third(aromatic) ring is needed for effectiveness (see6a and12a versus5), in addition, without fused ring the pyra-zolines are ineffective, like15. Replacement of the 5-CH2group with sulfur in the benz[g]indazole series (12g)removed the antimicrobial effect. This can be explainedin part by the different conformation of compounds12aand 12g and the different electron structure. As for the

Table I. Physical data of compounds4b, 5, 6a–c, 8a–d, 9a–d, 11b–f, 12a–gand13–15.

Compound General formulaa M.p. (°C) Yield Method(%)

4b C22H23N3S 155 (dec., methanol) 62 –5 C15H19N3S×HI 172 (dec., acetone) 58 –6a C22H23N3S×HI 180–183 (acetone) 79 A, C6b C23H25N3S×HI 242 (dec., acetone) 47b A6c C22H23N3S 160 (dec., methanol) 92 –8ac C21H22N4O 163–165 (methanol) 33 –8bc C21H22N4O 130 (dec., methanol) 6 –8c C21H22N4S 174–177 (methanol) 70 –8c × HCl C21H22N4S×HCl 213–216 (methanol) 60 –8d C23H26N4O2S 181 (dec., methanol) 39 –8d × HCl C23H26N4O2S×HCl 189 (dec., methanol) 42 –9a C22H24N4S 2×HCl 200–204 (ethanol) 47 B, C9b C24H28N4O2S 2×HCl 195 (dec., ethanol) 39 B9c C23H27N5S 3×HCl 114 (dec., ethanol) 38 B9d C23H27IN4S×HI 198–202 (ethanol) 42 B11b C18H16ClN3S 225–227 (methanol) 90 –11c C18H16BrN3S 232–233 (methanol) 93 –11d C18H15Cl2N3S 118 (dec., methanol) 74 –11e C20H21N3O2S 130–132 (methanol) 40 –11f C16H15N3S2 190 (dec., methanol) 80 –12a C19H19N3S×HI 158–162 (acetone) 65 A12b C19H18ClN3S×HI 168–171 (acetone) 41 A12c C19H18BrN3S×HI 160 (dec., acetone) 43 A12d C19H17Cl2N3S×HI 170–173 (acetone) 65 A12e C21H23N3O2S×HI 158–160 (acetone) 69 A12f C17H17N3S2×HI 140–143 (acetone) 71 A12g C18H17N3S2×HI 183 (dec., acetone) 36 A13 C24H25N3S×HBr 176–180 (acetone) 75 A14 C25H23N3S×HCl 192–195 (acetone) 49 A15 C17H17N3S×HI 116 (dec., acetone) 25 A

aThe analytical values were within± 0.4% of the theoretical values for C, H and N.b36% of 4b recovered.cIR (KBr) 1 685 cm–1.

1012

3-aryl group, the best effect has been shown with the4≠-halogen substituted derivatives (12b–d). With respectto the S-alkyl group, the replacement of the nonpolarmethyl group removed the antibacterial effect as in thecase of9c. This effect requires anN-unsubstituted thioa-mide moiety (see6b). 9a proved to be the only effectivecompound against both Gram-negative and Gram-positive strains. Its quaternarization decreased its effectboth against the Gram-negative and Gram-positivestrains.

4.2. Conclusion

We have synthesized a series of 2-isothiocarbamoylsubstituted bi- and tricyclic pyrazolines and theirS-alkylderivatives as potent antibacterial compounds. For thisclass of compounds a certain substitution pattern isnecessary for the optimal antibacterial effect. Thusthe 7-benzylidene-3,3a,4,5,6,7-hexahydro-5-methyl-2-S-methylthiuronyl-3-phenyl-2H-pyrazolo[4,3-c]pyridinedihydrochloride (9a) is an acceptably potent compound

Table II. 1H-NMR data (chemical shifts, in ppm,δTMS = 0 ppm, and coupling constants in Hz) of compounds4b, 5, 6a–c, 8a–d, 9a, d, 11b–f,12a–gand13–15 in CDCl3 or DMSO-d6 solutiona at 250 MHzb.

Compoundc H-3 H-3a NMe SMe NHd (1H)d m (1H)e s (3H)f s (3H) 1 or 2s (2H)g,h

4b 6.08 3.45 3.18 – –i

5 5.75 3.75 – 2.57 8.9 9.56a 5.87 ≈ 3.9g – 2.65 9.2 9.76b 6.15 3.85 3.25 2.45 9.26c 5.70 3.45 – 2.28 6.58a 5.60 3.70 2.25 – 5.658b 4.82 3.05k 2.25 – 6.58c 6.09 3.77 2.25 – 6.2 7.18d 6.02 3.74 2.27 – –i

9a 6.07 4.70 2.70 2.76 10.05 12.49d 6.12 ≈ 4.7k 3.15 2.66l 9.5 10.011b 6.10 3.67 – – 6.3 7.111c 6.09 3.70 – – 6.2 7.111d 6.08 ≈ 3.7g – – 6.3 7.111e 5.93 3.73 – – 7.75 7.9511f 6.26 3.77 – – 7.74 8.0212a 5.95 4.18 – 2.67 9.2 9.712b 5.96g 4.13 – 2.63 9.15 9.7512c 6.02 4.18 – 2.67 9.2 9.712d ≈ 6.0g 4.14 – 2.64 9.2 9.812e 5.86 4.07 – 2.63 9.10 9.6512f 6.27 4.10 – 2.64 9.25 9.7012g ≈ 6.0g 4.50 – 2.64 9.5 10.013 6.05 3.95 4.00 – 8.35 8.914 5.75 3.98 4.42 – –i

15 5.95 – – 2.67 9.15 9.75

aSolvent was CDCl3 for 4b, 6c, 8a–d and 11b–d, TFA-d for 14. Compounds6b and 13 were measured in CDCl3 and DMSO-d6 (13 inDMSO-d6 at 100 °C),9a in DMSO-d6 and TFA-d. bMeasuring frequency was 500 MHz for8d, 11b andd–f, 12b andd–g and15. Furthersignals: CH2 (position 4–7), 2–8 signal (6H for4b, 6a–c, 13, 8H for 5, 2H for 12gand15 and 4H for all other compounds): 0.4–4.7, OCH3,s (3H): 3.78 and 3.82 (8d), 3.67 (6H,11e), 3.69 and 3.71 (12e), ArH + C(sp2)H, 1–6 m’s [11H or 5H (5), 8H (11cand12c), 9H (12aandg),10H (15) and 14H (14)]: 6.8–8.1, allylic group (13), C(sp3)H2: 4.10,≈ d and 4.25,≈ dqa, C(sp2)H2: 5.20,d and 5.30,d, C(sp2)H: 5.65 brand 5.90 br (doublet signals due to rotamers).cAssignments were proved by DR- (8a), 2D-COSY (8d, 11f) and DNOE-measurements (11c).A sample containing ca. 5% ofcis-isomer was measured for compound14,δH-3: 5.90 for the minor isomer.dJ: 11.0 (4b, 6a and8c), 9.4 (6bin DMSO-d6, 8a), ≈ 10 (6c, 9a andd, 13 and14), 11.5 (8b), 10.7 (8d, 11b andd), 8.3 (11c), 10.5 (11e), 10.2 (11f), broadened signal (5, 12a,candg and15). Doubled signals due to rotamers with the secondd at 6.05 (6a and12b), 6.62 (9a), 6.25 (9d), 6.1 (12a), 6.00 (12e) and 6.35(12f).edt (5, 6c, 8a andc, 12a, c andg), ddd (11c, 13 and14), broadened (6a). fd for 4b (J: 4.8),9a (J: 8.5) and13 (J: ≈ 7), NCH2 (intensity2H) for 13 and14 (N-allyl andN-benzyl group), two singlets (2×3H) for 9d with the second signal at 3.19.gBroad signals.hIntensity: 1H (5and6a–cand for the signal at 12.4 in9a). The split to two signals due to rotamers for5, 6a, 8c, 9a andd, 11b–f, 12a–g, 13 and15. Theintensity ratio is ca. 1:1, for5, 6a, 9d, 12a andd and15 ca. 2:1 for12eand f ca. 5:2.iNot identifiable.kIn overlap with the CH2-signals.lDoubled signal with the second singlet at 2.67.

1013

inhibiting both Gram-positive and Gram-negative bacte-ria. 12c was the best agent against the Gram-positivestrains (cf.table V; MIC values 12.5–25µg/mL.) Either afused aromatic ring or a planar arylidene substituent isalso crucial. It is very important that the thioamide moietycan not be substituted. In comparison with antibioticscommonly used in therapy our most effective compoundsshow similar or slightly less antibacterial activity (ta-bles VandVI). This class of compounds having relativelylow toxicity (cf. table VII; LDt50 > 250µg/mL) could benew potent antibacterial agents. Resistance to antimicro-bial drugs is increasing all over the world. Both Gram-negative and Gram-positive strains are involved in this

process. New types of compounds with antimicrobialactivity could diminish this negative tendency.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Boetius apparatusand are uncorrected. Microanalyses were carried out atthe Central Research Laboratory, University MedicalSchool, Pécs. IR spectra were run in KBr discs on aBruker IFS-55 FT-IR spectrometer controlled by Opus2.0 software.1H- and13C-NMR spectra were recorded in

Table III. 13C-NMR chemical shifts (δTMS= 0 ppm) of compounds4b, 5, 6a–c, 8a–d, 11b–f, 12a–g and15 in CDCl3 or DMSO-d6 solutionaat 63 or 126 MHzb.

Compound C–3 C–3a C–4 C–5 C–6 C–7 C–7a C–8 C(sp2)Xc XMed

4b 67.7 49.7 28.2 23.6 25.5 137.3 159.0 127.3e 176.4 31.15 65.7f 50.7 27.1e 22.9 24.7 27.6e ≈ 164f – ≈ 171f 14.5f

6a 66.7g 51.0 27.8 22.7 25.0 135.0 163.7 126.2 165.6g 14.2g

6b 67.5 50.6 27.8 22.9 24.7e 135.1 159.0 127.2e 176.1 25.4e

6c 66.2 49.9 28.0 23.7 25.5 136.5 155.4e 127.2e 159.9e 12.88a 62.6 48.4 55.7e – 54.2e 135.2 153.2 127.7 154.3 45.38b 66.5 55.4 56.6e – 56.0e 142.9 157.2 129.6i 152.8 45.18c 66.5 48.8 55.8e – 54.3e 135.0 158.0 127.7 176.0 45.58d 66.3 48.9 54.3 – 56.0 ?f 159.0 129.7 175.9 45.511b 66.6 49.2 29.3 24.3 140.3 126.5 157.6 126.9 176.1 –11c 66.4 48.9 29.1 24.1 140.1 126.3 157.5 126.7 175.8 –11d 66.1 49.1 29.2 24.3 140.2 126.3 157.6 126.9 176.1 –11e 67.2 49.1 29.4 24.7 141.1 127.7 157.1 125.7 176.6 –11f 63.7 49.3 29.3 24.0 141.2 ?f 157.0 125.8 176.6 –12a 66.5g 49.8 28.3 23.8 141.4 126.0 163.0g 127.0i 165.0 14.5g

12b 66.5g 50.6 29.2 24.7 142.4 126.8 163.3 127.8 164.8 15.5g

12c 66.1f 49.7 28.3 23.9 141.4 125.9 163.2g 127.0 ≈ 165f 14.7f

12d 66.3 50.6 29.2 24.7 142.3 126.5 ≈ 164f 127.8 ≈ 165f 15.412e 67.6g 50.6 29.2 24.6 142.4 126.8 ?f 126.1 164.4g 14.812f 64.2 50.7 29.1 23.9 142.4 126.0 ≈ 164f 127.7 ≈ 165f,g 15.512gk 67.3 49.5 26.5 – 137.8 123.7 159.4 126.4e 164.8 14.615 63.6 44.5 – – – – 162.5 – 165.1 14.6

aSolvent was CDCl3 for 4b, 6b andc, 8a,c andd and11b–dand DMSO-d6 for 5, 6a, 8b, 11eandf, 12a–g and15. Compounds8d, 11b andd–f, 12b andd–f and15 were measured at 126 MHz. Assignments were supported by DEPT (for8d, 11b andd–f, and12b andd–g) and2D-HSC (for8d and11eand f) measurements.bFuther lines, OMe: 55.2 and 55.3 (8d), 56.3 and 56.4 (11e), 56.4 and 56.6 (12e). Aromaticcarbons (because of poor quality of the13C-NMR spectra due to hindered rotation and bad solubility it was not possible to measure exactchemical shifts for some broad and weak lines of these carbons in the cases of6a andb, 12b andd–f and15), 3-phenyl/aryl or in11f and12f 3-(2-thienyl) and the conjugated phenyl/aryl groups in the side chain and in position 3 (for15): C-1≠: 135.0–137.3 (non-conjugated rings),126–130 (cr: conjugated rings), 141.8 (11f), C-2≠,6≠: 125.0–130.8, C-3≠,4≠,5≠: 126.0–130.0, except for the following cases:8d: C-3≠,5≠: 114.1,C-2≠,6≠: 131.7 (cr), C-4≠: 159.0, 159.7 (cr),11b and12b: C-4≠: 133.5,11cand12c: C-3≠,5≠: 131.7, 132.6, C-4≠: 121.4, 121.9,11d and12d:C-3≠: 132.9, 132.2, C-4≠: 131.8, C-5≠: 130.7,11eand12e: C-1≠: 131.2, C-2≠: 111.0, C-3≠: 149.6, C-4≠: 148.7, 149.7, C-5≠: 112.6, 113.0,C-6≠: 118.4,11f: C-3≠,5≠: 125.2, 125.5. Condensed benzene ring in11b–fand12a–g(the numbering is given in the figure), C-9: 125.1–126.6;C-10: 131.2–134.1; C-11: 129.0–130.4.cThiocarbamide (X = S) for4b, 5, 6a–c, 8c andd, 11b–f, 12a–g and15, carbamide (X = O) for8aandb. dX = N for 4b, 6band8a–d,X = S for 5, 6aandc, 12a–gand15. eInterchangeable assignments (with an aromatic line for C-8 in4b,6b andc and12g). fBroadened signal due to hindered rotation of theN-acyl group. Broadening was also observed for some aromatic carbonlines in12b, d andf and15. gSplit signals due to hindered rotation with the second line at 67.4 (C-3), 166.7 (Csp2S) and 14.8 (SMe) for6a,66.9 (C-3), 163.4 (C-7a) and 14.8 (SMe) for12a,67.0 (C-3) and 14.9 (SMe) for12b, 163.9 (C-7a) for12c,67.0 (C-3) and 164.9 (= CS) for12eandz 167f for 12f. iTwo overlapping lines, one originates from an aromatic carbon.kMeasurement temperature was 70 °C.

1014

different solutions (tables II and III ) in 5 mm tubes atroom temperature, on a Bruker WM-250 FT-spectrometerequipped with an Aspect 2000 computer at 250.13 (1H)and 62.89 (13C) MHz, respectively, using the deuteriumsignal of the solvent as the lock and TMS as internalstandard. Conventional CW irradiation of≈ 0.15 W was

used in the DR experiments. DEPT spectra [18] were runin a standard way [19], using only theθ = 135° pulse toseparate the CH/CH3 and CH2 lines phased up and down,respectively. For DNOE measurements [20, 21] the stan-dard Bruker microprogram ‘DNOEMULT.AU’ to gener-ate NOE was used. The 2D-HSC spectra [22] wereobtained by using the standard Bruker pulse program‘XHCORRD.AU’.

Analytical thin layer chromatography (TLC) was ap-plied to monitor the reactions using precoated plates(Silica gel 60 F-254, Merck), and spots were visualizedwith UV light. The synthesis of some starting tricyclicpyrazolines (3, 4a, 11a and 11g,) [7, 8], 2-isothio-carbamoyl-3,5-diphenyl-2-pyrazoline [23], 2-arylidene-1-tetralones [24, 25, 26 and unpublished data],2-benzylidene-1-thiochroman-4-one [27] and 3,5-diaryli-dene-1-methyl-4-piperidones [28] has been reported ear-lier. The analytical values were within± 0.4% of thetheoretical values for C, H and N.

5.1.1. General procedure for the preparation of 2-iso-thiocarbamoyl- or 2-carbamoyl- substituted compounds(4b, 8a–dand11b–f)

The mixture of thiosemiocarbazide or semicarbazide(30 mmol) and the corresponding unsaturated ketone(10 mmol) was refluxed in ethanol (110 mL) containing9% concentrated hydrochloric acid until the disappear-ance of the starting unsaturated ketone. The reactionmixture was cooled down, the precipitate was filtered andwashed with cold ethanol and water until neutral. As

Table IV. Microbial screening of the antimicrobial effects of thenovel compounds.

Compounda Number ofS. aureus Number ofE. coliinhibited / tested inhibited / tested

5 1/20 0/206a 19/20 0/206b 0/20 0/208a 0/20 0/208cb 0/20 0/209a 15/20 13/209b 0/20 0/209c 0/20 0/209d 20/20 0/2012a 20/20 0/2012b 18/20 0/2012c 19c/20 0/2012d 19/20 0/2012e 0/20 0/2012f 0/20 0/2012g 0/20 0/2015 0/20 0/20

aConcentration, 50µg/mL. bUsed as a hydrochloride.cThe effectwas bactericidal against some strains.

Table V. Minimum inhibitory concentration (MIC) values.

Concentrationµg/mL

STRAINS > 200 200 100 50 25 12.5

S. aureus 5 6a 9a 12cNIH HUNGARY 9d 12a 12b 12d118 003

S. saprophyticus 5 9d 9a 12c 6aNIH HUNGARY 12a 12d 12c120 008

M. luteus 9a 5 12bATCC 9341 9d 6a 12c

12a 12b12d

B. subtilis 9d 9a 5 6a 12d 12bATCC 6633 12a 12c

E. coli 5, 6a, 9a,ATCC 25922 9d, 12a, 12b,

12c, 12dE. coli

from 1–2 9aurine 3–5 9a

1015

for the preparation ofcis-7-Benzylidene-2-carbamoyl-3,3a,4,5,6,7-hexahydro-5-methyl-3-phenyl-2H-pyrazolo-[4,3-c]pyridine (8a), at the end of the reaction thesolution was made alkaline. The precipitate separated wasfiltered and washed with cold ethanol. Diastereoisomers8a and8b were separated by fractional recrystallizationfrom methanol.

5.1.2. General procedure for the preparation of 2-iso-thiocarbamoylpyrazolo[4,3-c]pyridine hydrochlorides(8c × HCl and8d × HCl)

5.51 mmol of the free base (8c or d) was dissolved inethanol and 5 mL of 6 N HCl in ethanol was added. Thesalt separated was filtered off and recrystallized frommethanol.

5.1.3. General procedure for the alkylation of 2-iso-thiocarbamoyl-substituted compounds(3, 4a–band11a-g)

Method A: an alkyl halide (9.3 mmol) was added to thesolution of the 2-thiocarbamoyl compound (8.4 mmol) inanhydrous ethanol (100 mL). The reaction mixture wasrefluxed for 2 h with the exclusion of moisture till thedisappearance of the starting 2-isothio-carbamoyl com-pound. The solvent was removed in vacuo and the residuewas recrystallized from acetone.

5.1.4.cis-7-Benzylidene-3,3a,4,5,6,7-hexahydro-2-S-methyl-thiuronyl-3-phenyl-2H-indazole(6c)

6a (0.50 g, 1 mmol) was dissolved in ethanol (100 mL)and it was treated with concentrated ammonia solution(5 mL). The solution was poured into water and theseparated precipitate was filtered off. It was recrystallizedfrom methanol.

Table VI. Minimum inhibitory concentration (MIC) values of standard commercial antibiotics measured on standard bacterial strains.

Concentration (µg/mL)

STRAINS > 100 50 25 12.5 6.25 3.12 1.56 0.78 0.39 0.20 < 0.20

S. aureus T C CRO CXM G OXANIH HUNGARY A P118 003

S. saprophyticus CRO C T A GNIH HUNGARY CXM OXA120 008 P

E. coli OXA G CXM C CROATCC 25922 P A T

M. luteus G C T CXMATCC 9341 CRO

AOXAP

B. subtilis C A CRO GATCC 6633 CXM T

OXAP

P = penicillin-G; OXA = oxacillin; T = oxytetracycline; A = ampicillin; CRO = ceftriaxone; CXM = cefuroxime; C = chloramphenicol;G = gentamicin.

Table VII. In vitro cytotoxicity test of compounds on a HeLacell-line.

Cell death %Compound at 250µg/mL LDt50

concentration µM/L

5 87 155.76a 35 > 510.89a 90 22.29d 73 96.712a 72 278.212b 29 > 516.712c 8 > 473.312d 17 > 482.4

1016

5.1.5. General procedure for the alkylation of 2-iso-thiocarbamoyl-substituted compounds(8a–d)

Method B: alkyl halide (1.93 mmol) was added to thesolution of the hydrochloride of the 2-isothiocarbamoylcompound (1.75 mmol) in anhydrous ethanol (65 mL).The reaction mixture was refluxed for 2.5 h with theexclusion of moisture until the disappearance of thestarting 2-isothiocarbamoyl compound. The reaction mix-ture was made alkaline by using concentrated ammoniasolution (0.7 mL) and poured into water. The precipitatewas filtered off and washed with water until neutral. Afterdrying the product was dissolved in ethanol and trituratedwith HCl gas.

5.1.6. General method for the preparation of S-methylderivatives from dibenzylidene ketones and S-methyl-thiosemicarbazide hydroiodide(6a and9a)

Method C: the unsaturated ketone (20 mmol) andS-methylthiosemicarbazide hydroiodide (30 mmol) weredissolved in the mixture of ethanol (300 mL) and con-centrated hydrochloric acid (10 mL). After 12 h boilingthe reaction mixture was cooled down. The precipitatewas filtered off, washed with cold ethanol and water untilneutral. It was recrystallized from the mixture of acetoneand methanol. The salt formed in the case of9a wasconverted to the corresponding base that was treated withHCl to yield a dihydrochloride. These samples were inevery respect identical with the product of the alkylation.

5.2. Biology

5.2.1. Microbial screensTwenty each ofS. aureusandE. coli isolates (the most

common representatives of Gram-positive and Gram-negative bacterial pathogens) of various clinical sourceswere selected for screening the antimicrobial effect of thecompounds. The strains were maintained on nutrient agarmedium. The test compounds were dissolved in nutrientbroth (Difco) medium at a concentration of 50µg/mL,and 2µL of a Nutrient Broth starter culture of thebacterial strain to be tested was added to achieve a finalinoculum of ca. 5 × 105 colony forming units permL [25]. The cultures were incubated overnight at 37 °C.Inhibition was shown by no change in optical density.Nutrient broth medium without the compound served ascontrol. Loopfuls of nutrient broth cultures were platedon nutrient agar to show if the effect of the compoundswas bacteriostatic or bactericidal [29].

5.2.2. Determination of minimum inhibitory concen-tration (MIC) value by test tube dilution method

Eight of the most effective antibacterial compounds (5,6a, 9a, 9d and12a–d) were tested on reference strains:S.

aureus NIH Hungary 118003, StaphylococcussaprophyticusNIH Hungary 120008,Micrococcus luteusATCC 9341,Bacillus subtilisATCC 6633,E. coli ATCC25922. The MIC values of these five standard strainswere tested on eight antibiotics commonly used intherapy: penicillin-G (P), Biogal, Debrecen, Hungary;oxacillin (OXA) and ampicillin (A), Bristol MyersSquibb Co., USA; oxytetracycline (T) and gentamicin(G), Chinoin, Budapest, Hungary; ceftriaxone (CRO) F,Hoffmann-La Roche AG, Basel, Switzerland; cefuroxime(CXM), GlaxoWellcome, Greenford, UK; chlorampheni-col (C), EGIS, Budapest, Hungary. In addition the9acompound was investigated on fiveE. coli strains isolatedfrom urine. The test conditions were similar as mentionedearlier. The exceptions: the compounds in 200µg/mLconcentrations were diluted in medium containing 2.5%DMSO. Using backwards dilution of DMSO all tubescontained the same concentration of DMSO. Controltubes without compounds were used to check the effect ofDMSO. Double dilution series of compounds were madein test tubes. After inoculation and 24 h incubation at37 °C, the MIC values were obtained from the lowestconcentration of compound where the tubes remainedclear, where the bacterial growth was inhibited. Allexperiments were performed in triplicate [29].

5.2.3. In vitro cytotoxicity tests of compounds on HeLacell-line OHIO

A microplate technique was used. The DMEM1(Sigma, Missouri, USA) growth medium contained 10%foetal bovine serum (Sigma, Missouri, USA). The testedcompounds were solved in growth medium containing2.5% DMSO. Double dilution series were set up from250µg/mL concentrations. 1.5× 104 cells/well wereincubated for 24 h at 37 °C. Cells killed by compoundswere washed out, living cells were fixed and stained bycrystal violet in methanol. The remnants of stain werewashed out. One hundredµL of 1% SDS/well solved thecells under slow shaking (20 min). The cell lysates weremeasured in a Dynatech MR7000 photometer at 595 nm.Controls without compounds (100% cell) were used indetermination of cell death percent. A graph of cell deathpercent vs. concentration was drawn. From this graph theLDt50 values were determined [30].

Acknowledgements

We are indebted to Mrs Gabriella Németh, Mr ZoltánNagy and Mr Attila Fürjes for their valuable technicalassistance and to Mrs Ildikó Meleg and Mr Antal Kovácsfor computer formulation of the manuscript.

1017

References

[1] Tornetta B., Arcoria A., Boll. Sci. Fac. Chim. Ind. Bologna 14(1956) 50–51.

[2] Manowitz M., Walter G., J. Pharm. Sci. 54 (1965) 650.

[3] Bauer D.J., Sadler P.W., Br. J. Pharmacol. 15 (1960) 101–110.

[4] Searle G.D., and Co. (1975) Brit. 1, 382, 773; Chem. Abstr. (1975)83 p58829g.

[5] Hamilton R.W., (1976) U.S. 3, 940, 418 Chem. Abstr. (1976) 85p5623v.

[6] Ramalingam K., Thyvelikakath G.X., Berlin K.D., Chesmut R.W.,Brown R.A., Durham N.N., Ealick S.E., Van der Helm D., J. Med.Chem. 20 (1977) 847–850.

[7] Lóránd T., Szabó D., Földesi A., Párkányi L., Kálmán A., NeszmélyiA., J. Chem. Soc. Perkin Trans. 1 (1985) 481–486.

[8] Tóth G., Szöllôsy Á., Lóránd T., Kónya T., Szabó D., Lévai A., J.Chem. Soc. Perkin Trans. 2 (1989) 319–323.

[9] Szöllôsy Á., Tóth G., Lóránd T., Kónya T., Aradi F., Lévai A., J.Chem. Soc. Perkin Trans. 2 (1991) 489–493.

[10] Walter W., Krohn J., Chem. Ber. 102 (1969) 3786–3794.

[11] Sohár P., Cyclopentane derivatives and their heteroanalogues, in:Nuclear Magnetic Resonance Spectroscopy Vol. 2, CRC Press, BocaRaton, Florida, 1983, pp. 22–23.

[12] Sohár P., The effect of structural parameters on the shielding ofcarbon nuclei, in: Nuclear Magnetic Resonance Spectroscopy Vol. 2,CRC Press, Boca Raton, Florida, 1983, pp. 154–155.

[13] Grant D.M., Cheney B.V., J. Am. Chem. Soc. 89 (1967) 5315.

[14] Sohár P., Fused and quasiaromatic heterocyclic systems; analoguescontaining more nitrogens, in: Nuclear Magnetic Resonance Spec-troscopy Vol. 2, CRC Press, Boca Raton, Florida, 1983, pp. 89–90.

[15] Sohár P., Geminal couplings, in: Nuclear Magnetic ResonanceSpectroscopy Vol. 1, CRC Press, Boca Raton, Florida, 1983,pp. 55–60.

[16] Karpus M.J., J. Chem. Phys. 30 (1959) 11.

[17] Karpus M.J., J. Chem. Phys. 33 (1960) 1842.

[18] Pegg D.T., Doddrell D.M., Bendall M.R., J. Chem. Phys. 77 (1982)2745.

[19] Bendall M.R., Doddrell D.M., Pegg D.T., Hull W.E., High Resolu-tion Multipulse NMR Spectra Editing and DEPT. Bruker, Karlsruhe(1982).

[20] Sohár P., Gated decoupling. Determination of NOE, and the absoluteintensities of13C-NMR signals, in: Nuclear Magnetic ResonanceSpectroscopy Vol. 1, CRC Press, Boca Raton, Florida, 1983, pp.196–197.

[21] Sanders J.K.M., Mersch J.D., Prog. Nucl. Magn. Reson. 15 (1982)353.

[22] Ernst R.R., Bodenhausen G., Wokaun A., Principles of NuclearMagnetic Resonance in One and Two Dimensions, Clarendon Press,Oxford, 1987 pp. 471–479.

[23] Balaban A.T., Zugravescu I., Avramovici S., Silhan W., Monatsh.Chem. 101 (1970) 704–708.

[24] El-Rayyes N.R., Al-Jawhary A., J. Heterocycl. Chem. 23 (1968)135–140.

[25] Al-Nakib T.M., Bezjak V., Meegan M.J., Chandy R., Eur. J. Med.Chem. 25 (1990) 455–462.

[26] Al-Nakib T.M., Perjési P., Varghese R., Meegan M.J., Med. Princ.Pract. 6 (1997) 14–21.

[27] Lévai A., Schág J.B., Pharmazie 34 (1979) 749.

[28] Krapcho J., Turk C.F., J. Med. Chem. 22 (1979) 207–210.

[29] Sahm D.F., Washington J.A., Antibacterial susceptibility tests:dilution methods, in: Ballows A., Hausler Jr. W.J., Herrmann K.L.,Isenberg H.D., Shadomy H.J. (Eds.), Manual of Clinical Microbi-ology 5th ed., ASM, Washington DC, 1991, pp. 1105–1116.

[30] Schlager S.I., Adams A.C., Use of dyes and radioisotopic markers incytotoxicity tests, in: Langone J.J., Vunakis H.V., (Eds.), Methods inEnzymology, Vol. 93, Immunochemical Techniques Part F, Conven-tional Antibodies, Fc Receptors and Cytotoxicity, Academic PressInc., San Diego, USA, (1983), pp. 233–245.

1018

Original article

Synthesis and structure-activity relationship study of the new set of trypsin-likeproteinase inhibitors

Pavol Zlatoidsky, Tibor Maliar*

Drug Research Institute, SK-90001 Modra, Slovak Republic

(Received 8 October 1998; revised 2 April 1999; accepted 6 April 1999)

Abstract – A new set of 25 trypsin-like proteinase inhibitors was prepared and the inhibiting activity on trypsin, thrombin, plasmin andurokinase was measured. The structure-activity relationship is discussed. High inhibiting activities were observed in 4-guanidinobenzoic acidesters only. The replacement of this moiety for N-formamidinyl-isonipecotic acid or an arginine moiety caused almost total loss of the activity.In the series of 4-guanidinobenzoic acid esters, any important influence of the ester-groups reactivity was observed. The trypsin-thrombinselectivity in the compounds with the guanidine-remote carboxylic function was also observed. © 1999 Éditions scientifiques et médicalesElsevier SAS

trypsine-like proteinases / inhibitors of proteinases / guanidine derivatives

1. Introduction

Trypsin-like proteinases are serine proteinases forwhich hyperactivity can cause a variety of damage tohealth [1]. Recently, the activity of the urokinase typeplasminogen activator (uPA) was discovered as an im-portant starting proteinase in the proteolytic cascade oftumour invasion and metastases [2]. Inhibitors of theseenzymes are of interest as potential therapeutics invarious diseases.

Recently, we have reported a set of 4-guanidinobenzoicacid esters with considerable inhibiting activity ontrypsin [3]. Continuing in this research we have synthe-sised a further set of compounds to demonstrate theinfluence of either ester group reactivity or guanidine-bearing function moiety.

2. Chemistry

Synthesis of the compounds2a–h was performed viaselective esterification of the hydroxyaromatic acid withthe corresponding halogen derivative, and followed byesterification of the produced compounds1a–h with

4-guanidinobenzoic acid mesylate via the dicyclohexylcarbodiimide (DCC) method as described [3].

Reaction of the 4-aminobenzoic acid with chloroacetylmorpholine in the presence of triethylamine in acetoni-trile gave a moderate yield of the corresponding ester1l.After diazotation and nucleophile substitution with potas-sium ethyl dithiocarbonate and amonolysis of the inter-mediate, the thiole1m was produced. This was esterifiedwith 4-GBA/DCC to get2l. Compound2i was preparedby hydrogenolysis of benzylester2c, compounds2j and2k were prepared by acidic hydrolysis of the tert. butylesters2g and2h (figure 1).

Compounds4a–ewere prepared via activation of thefree hydroxyaromatic acid with N-ethyl-N≠-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDI)followed by aminolysis of the activated complex with thecorresponding amine. The hydroxyaromatic amides3a–ewere esterified by 4-guanidinobenzoic acid mesylate andDCC (figure 2).

Synthesis of compounds7a and b is mentioned infigure 3. N-chloroacetyl morpholine was reacted withpotassium ethyl dithiocarbonate in ethanol. Followingamonolysis with ammonia yielded N-mercaptoacetylmorpholine 5a. BOC-glycine was activated with ethylchloroformate and reacted with morpholine. After acidic*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1023−1034 1023© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

hydrolysis of the BOC group, the free N-aminoacetylmorpholine was reacted with 4-acetoxybenzoyl chloride.Then, one-pot cleavage of the protective acetyl group viaaqueous methanolic sodium carbonate gave6a, as well as6b which was prepared via the same reactions. Esterifi-cation by 4-guanidinobenzoic acid mesylate/DCC gave7a andb (figure 3).

The sodium salt of 4-mercaptoanisole was the startingmaterial for synthesis of compounds10 and11. It reactedwith ethyl chloroacetate to give the corresponding esterwhich was deprotected with hydrobromic acid, both onphenolic and carboxylic hydroxyl, to give the acid8. Theacid 8 was esterified via the usual method with chloro-acetyl morpholine to give compound9. This was esteri-

Figure 1. i: Halogen-CH2COY/MeCN, TEA. ii: 4-Guanidinobenzoic acid mesylate/DCC/pyridide/RT. iii: H2/Pd-C. iv: CF3COOH/DCM/anisole. v: NaNO2, HCl, 2.KSCSOEt, 3.NH3.

1024

fied to get compound10. After oxidation of 10 withhydrogen peroxide-acetic acid, compound11 was pre-pared (figure 4).

N1-BOC-N6,8 di-Z-Arginine was coupled with com-pounds 1a and 1d to get esters12a and b. Afterhydrogenolysis of both Z-protective groups, compounds13a andb were obtained (figure 5).

Isonipecitic acid was protected by the BOC group toget 14 and coupled with compounds1a and 1b via theDCC method to give esters15a and b. After acidiccleavage of the BOC protective group, the free amineswere reacted with cyanamide to give compounds16aandb (figure 6).

3. Results and discussion

All inhibiting activities are presented intable I.There are no remarkable differences in the inhibiting

activity on trypsin, thrombin or urokinase in the firstgroup of compounds2a–h. After the benzene ring wasreplaced by the naphthalene ring, the increase of theselectivity for trypsin/thrombin is observable (compare

Figure 2. i: EDI/amine/TEA/MeCN. ii: 4-Guanidinobenzoicacid mesylate/DCC/pyridine.

Figure 3. i: 1. TEA, DCM, 2. Na2CO3, MeOH, H2O. ii: 4-Guanidinobenzoic acid mesylate/DCC/pyridine.

1025

2a versus2e). The extremely high activity (picomolarrange) of2c on plasmin is remarkable. Tert. butyl esters2g and2h didn’t show any remarkable activity. But afterremoval of the ester group, compounds2i, 2j and 2kshow remarkable selectivity for trypsin over thrombin.Introduction of the pyridine ring instead of the benzenering to increase the ester function reactivity didn’t lead toa considerable change in the activity, similarly to thecompound2l. Introduction of the thioester moiety wasintended to increase the ester reactivity towards nucleo-philes. No effect of the change of sulfide (10, no M+) andsulfone11 (M– effect and proposed enhancement of thenucleophile reactivity) has been observed.

Mild effects of conjugation on compounds of group4were observed (compare compounds4a and4b with 4dand4e). Reduced activity was observed in the compound7a (compare e.g. with2a) but the thioester7b shows thesimilar activity to the oxygen analogue2a. The com-

pounds with an arginine moiety13a, 13b and with anN-formamidinoisonipecotic acid moiety16aand16b areinactive. It is possible to conclude that the activity is notdependent on the ester function reactivity in nucleophilicsubstitution but only on the rate of hydrolysis of the4-guanidinobenzoyl-trypsine complex which is the com-monly accepted mechanism of the inhibiting mechanismof serine proteinases [4].

4. Experimental protocols

Melting points were measured on a Boetius micro-scope and are uncorrected. NMR spectra were run on aVarian 200 (200 MHz) using TMS as internal standard.All chemicals were supplied from Merck and Aldrich,solvents from Microchem (Slovakia) and enzymes andsubstrates from Sigma. HPLC was performed on aPye-Unicam system using Tessek C-18 columns (25×

Figure 4. i: 1. ClCH2COOEt/NaOET/EtOH, 2. HBr/AcOH. ii: N-Chloroacetyl morpholine/TEA/MeCN. iii: 4-Guanidinobenzoic acidmesylate/DCC/pyridine. iv: H2O2/AcOH.

1026

0.25 cm) for analytical purposes and Labio C-18 (25×5 cm) for semipreparations. The gradient MeCN-H2O(each contained 0.05% of trifluoroacetic acid) 0–90%was used at a flow rate of 1 mL/min (analytical) over50 min, or 10 mL/min (semipreparations). GC-MS analy-ses were run on Carlo Erba 1106.

N-Chloroacetyl-N≠-methylpiperazine hydrochlorideand 4-guanidinobenzoic acid mesylate were preparedaccording to the described procedure [3], as well asN2BOC-N6, N8-di-Z-Arginine [5].

4.1. N-chloroacetyl glycine benzyl ester

Glycine benzylester tosylate (16.85 g, 0.05 mol) weredissolved in 250 mL of dry dichloromethane and 26 mL(0.1 mol) of dry triethylamine were added. The mixturewas cooled to 0 °C, and 3.8 mL of chloroacetyl chloridein 15 mL of dry dichloromethane were added dropwise.The mixture was stirred for 1 h at room temperature,extracted with water, sat. NaHCO3 and brine, dried overNa2SO4, evaporated and used without further purifica-tion. The analytical sample was purified by high vacuum(0.0005 torr) molecular distillation (bath temperature,80 °C), to give the solidifying oil. Yield: 11.3 g, 74%(M.p. after solidification, 33–35 °C). MS: M+-H+ = 241,

Figure 5. i: 1a or 1d/DCC, DCM/MeCN. ii: H2/Pd-C/EtOH/HCl.

Figure 6. i: BOC2O/Na2CO3/wt. dioxane. ii: 1a or 1b/DCC/DMAP/MeCN-DCM. iii: 1. EtOAc/HCl, 2. cyanamide/concd./HCl/EtOH.

1027

other peaks: 192 (M+-CH2Cl), 164 (M+-ClCH2CO), 150(M+-tropylium), 91 (tropylium). 1H-NMR (CDCl3):2.38 s (2H, ClCH2), 3.68 d (2H, CH2NH), 3.88 s (2H,CH2O), 5.85 bs (1.3 H, NH), 6.49–6.56 m (5H, H arom).

4.2. N-chloroacetyl morpholine

Chloroacetyl chloride (7.5 mL, 0.1 mol), dissolved in50 mL of dry ether, was added dropwise to the stirredsolution of morpholine (22 mL, 0.2 mol) in 300 mL ofdry ether at –20 °C. The mixture was stirred for 30 min atroom temperature, the morpholinium chloride was fil-tered and washed with ether, the filtrate was evaporatedand the oily residue distilled in vacuo. B.p. 67–69 °C/0.6torr. Yield: 15.2 g, 93%. MS: 162 (M+-H+), other peaks:114 (M+-ClCH2), 86 (morpholinyl). 1H-NMR: 2.22 s(2H, ClCH2), 3.65 m (4H), 4.44 m (2H) and 4.56 m (2H).

4.3. Typical procedure for preparation of hydroxy-derivatives1a–hwas described in [3]

Physical-analytical data for each:

1a: m.p. 185–186 °C (iPr2O/EtOAc), yield: 1.48 g,56%. Elemental anal. for C13H15NO5 (C, H, N).1H-NMR(CDCl3): 3.43–3.46 (2H), 3.55–3.64 m (6H), 4.95 s (2H),6.74 d (2H),J = 7.56 Hz), 7.81 d (2H,J = 7.55 Hz).

1b: m.p. 201–203 °C (iPr2O/EtOAc), yield: 1.33 g,48%. Elemental anal. for C14H18N2O4 (C, H, N). 1H-NMR (CDCl3): 2.54 s (3H), 2.55 m (2H), 3.50 m (4H),3.65 m (2H), 6.78 d (2H), 7.68 d (2H).

1c: m.p. 65 °C (MeOH/iPr2O), yield: 1.12 g, 46%.Elemental anal. for C18H17NO6 (C, H, N). 1H-NMR(CDCl3): 3.5 bs (NH), 4.04 d (2H), 4.75 s (2H), 5.22 s(2H), 6.88 d (2H,J = 7.45 Hz), 7.43 m (5H), 7.97 d (2H,J = 7.44 Hz).

1d: m.p. 79–80 °C (iPr2O/EtOAc), yield: 1.36 g, 48%.Elemental anal. for C16H22N2O4 (C, H, N). 1H-NMR:2.54 s (3H, NCH3), 2.63 t (2H,J = 5.79 Hz), 2.89 m (2H),3.00 t (2H,J = 5.81 Hz), 3.45 m (6H), 6.75 d (2H,J =7.36 Hz), 7.43 d (2H,J = 7.38 Hz).

1e: m.p. 165–167 °C (hexane/EtOAc), yield: 1.61 g,51%. Elemental anal. for C17H17NO5 (C, H, N).1H-NMR(CDCl3): 3.65 m (8H), 4.89 s (2H), 7.24–7.31 m (broad,4H), 7.56 dd (1H), 7.80–7.86 m (4H).

1f: m.p. 212–214 °C (iPr2O/MeOH), yield: 1.75 g,63%. Elemental anal. for C15H18O5 (C, H). 1H-NMR(DMSO): 3.75 m (4H), 4.28 m (4H), 4.97 m (2H), 6.53 d(1H), 8.91 dd (1H), 8.24 d (1H,J = 6.56 Hz).

1g: m.p. 134 °C (hexane/EtOAc), yield: 1.80 g, 65%.Elemental anal. for C15H18O5 (C, H). 1H-NMR (CDCl3):1.52 s (9H), 4.66 s (2H), 6.19 d (1H,J = 9.78 Hz), 6.27 d,

Figure 7. i. N–chloroacetyl morpholin, MeCN, TEA, ii. 1.NaNO2, HCL, 2. KSCS (OEt), 3. wt. NH3, EtOH, iii. 4–gua-nidinobenzoic acid/DCC/pyridine.

Table I. Inhibiting activities. IC50 (nM) of compounds2a-l, 7a,7b, 10, 11, 13a, 13b, 16a, 16b on trypsin, thrombin, plasmin andurokinase.

Compound Trypsin Thrombin Plasmin Urokinase

2a 10 35 13 752b 10 94 28 552c 8 12 0.05 322d 65 475 1 163 1302e 26 1 030 446 7152f 5 40 8 1062g 32 78 336 7102h 375 435 5 025 1 0052i 10 726 38 432j 19 2 557 381 2132k 700 8 310 8 015 2 2352l 8 34 36 114a 32 447 343 1174b 140 1 018 1 450 1304c 30 192 79 NT4d 13 249 76 454e 13 218 49 217a 75 321 129 3847b 17 44 8 2210 15 22 507 3111 15 64 291 3313a 4.4 × 105 NT NT NT13b 4.72 × 105 NT NT NT16a 1.3 × 104 NT NT NT16b 1.7 × 104 NT NT NT

NT = not tested.

1028

6.84 d (2H,J = 7.55 Hz), 6.99 bs (1H), 7.33 d (2H,J =7.56 Hz) 7.61 d (1H,J = 7.78 Hz).

1h: m.p. 129 °C (iPr2O/EtOAc), yield: 1.19 g, 39%.Elemental anal. for C16H19NO5 (C, H, N). 1H-NMR(CDCl3): 1.59 s (9H), 2.28 s (2H), 4.28 s (2H, OCH2),6.18 m (1H), 6.29 dd (1H), 7.36 dd (1H), 7.77 m (1H),8.65 bs (1H, NH).

4.4. Typical procedure for preparation of hydroxyamides3a–e

4-hydroxyaromatic acid (0.01 mol) was suspended inacetonitrile (25 mL) and N-(3-dimethylaminopropyl)-N≠-ethylcarbodiimide (1.92 g, 0.01 mol) were added atonce. The acid dissolved immediately. Then, triethyl-amine (1.34 mL 0.01 mol) and the amine (0.01 mol) wereadded and the mixture was stirred for 18 h. After evapo-ration of volatile compounds in vacuo, the residue wasdistributed between water (50 mL) and ethyl acetate(20 mL) and the water phase was twice more extractedwith ethyl acetate. Joint organic extracts were washedwith brine, dried (Na2SO4) and evaporated in vacuo. Theresidue was recrystallised.

3a: m.p. 130 °C (iPr2O/EtOAc), yield: 1.79 g, 76%.Elemental anal. for C13H17NO3 (C, H, N). 1H-NMR(CDCl3): 2.56 t (2H,J = 5.55 Hz), 2.87 t (2H,J = 5.56Hz), 3.33 m (2H), 3.50 m (2H), 6.75 d (2H,J = 7.56 Hz,)7.03 d (2H,J = 7.55 Hz), 9.9 bs (1H).

3b: Syrup. Data for fumarate: m.p. 141–142 °C (THF),yield: 2.88 g 79%. Elemental anal. for C18H24N2O6 (C,H, N). 1H-NMR (DMSO): 2.18 s (3H), 2.25 t (2H), 2.52t (2H), 2.66 m (4H), 3.40 m (4H), 6.59 s (2H, fumarate),6.45 d (2H,J = 7.59 Hz), 7.01 d (2H,J = 7.60Hz).

3c: m.p. 65–67 °C (iPr2O/Hexane), yield: 2.72 g, 84%.Elemental anal. for C15H20N2O6 (C, H, N). 1H-NMR(CDCl3): 2.36 t (2H,J = 5.55 Hz), 3.83 t (2H,J = 5.57Hz), 3.36 m (2H), 3.76 m (6H), 5.01 s (2H), 6.75 d (2H,J = 7.28 Hz), 7.18 d (2H,J = 7.26 Hz).

3d: m.p. 213 °C (hexane/EtOAc), yield: 1.51 g, 65%.Elemental anal. for C13H15NO3 (C, H, N). 1H-NMR(CDCl3): 3.70 m (8H), 6.63 (1H,J = 10.27 Hz), 6.86 d(2H J = 7.45 Hz), 7.41 d (2H,J = 7.44 Hz), 7.67 d (1H,J = 10.26 Hz).

3e: m.p. 230–235 °C decomp. (hexane/EtOAc), yield:1.79 g, 71%. Elemental anal. for C16H15NO2 (C, H, N).1H-NMR (CDCl3): 4.56 s (2H), 6.31 d (1H,J = 11.0 Hz),7.1 d (2H,J = 7.67 Hz), 7.26–7.29 m (5H), 7.52 d (2H,J = 7.67 Hz), 7.69 d (1H,J = 10.9 Hz).

4.5. N-(Mercaptoacetyl)-morpholine(5b)

Chloroacetyl morpholin (2.95 g, 0.02 mol) was dis-solved in 35 mL of absolute ethanol and potassium

ethyldithiocarbonate (4 g, 0.025 mol) was added at once.The mixture was stirred under N2 and refluxed for 6 h,cooled and 35 mL of water ammonia were added andstirred overnight. The mixture was extracted three timeswith dichloromethane, dried (Na2SO4) and evaporated invacuo. After distillation under reduced pressure, the paleyellow oil was obtained. B.p. 72–74 °C/0.45 torr, yield:2.7 g, 86%. GC-MS: 161 (M+), others: 114 (N-morpholinylcarbonyl), 86 (morpholinyl), 46 (CH2=S+).1H-NMR (CDCl3): 1.11 s (1H), 2.36 s (2H), 3.75–3.79 m(8H).

4.6. Typical procedure for the preparation of compounds6a andb

In 35 mL of dry dichloromethane, N-(aminoacetyl)-morpholin or N-(mercaptoacetyl)-morpholin 5b(0.02 mol) were dissolved, triethylamine (2.7 mL,0.02 mol) were added and cooled to –10 °C. Then, thesolution of 4-acetoxybenzoyl chloride2 (3.95 g,0.02 mol) in 35 mL of dry DCM were added dropwise at–10 °C. The mixture was stirred for 2 h at room tempera-ture, evaporated in vacuo and the residue was dissolved in30 mL of methanol. Then, 30 mL of sat. Na2CO3 wereadded and the mixture was stirred for an additional 1 h. Itwas partly evaporated in vacuo, extracted three timeswith ethyl acetate, washed with brine, dried (Na2SO4),evaporated in vacuo and recrystallised.

6a: m.p. 196–197 °C (EtOAc/hexane), yield: 4.1 g,77%. Elemental anal. for C13H16N2O4 (C, H, N). 1H-NMR: 3.45–3.59 m (8H), 4.26 d (2H,J = 6.96 Hz,CH2NH), 6.79 d (2H,J = 7.56 Hz), 7.10 d (2H,J = 7.56Hz).

6b: m.p. 153–154 °C (EtOAc/hexane), yield: 3.8 g,72%. Elemental anal. for C13H15NO3S (C, H, N, S, % Scalcd. 12.08, found 12.49).1H-NMR: 3.38–3.63 m (8H),3.79 s (2H, CH2S), 6.83 d (2H,J = 7.49 Hz), 7.21 d (2H,J = 7.50 Hz).

4.7. Ethyl methoxyphenyl thioacetate

Sodium (2.3 g, 0.1 mol) was dissolved in 250 mL ofabsolute ethanol and 4-mercaptoanisole (14 g, 0.1 mol)were added. After being stirred for 1 h at room tempera-ture under N2, 11.4 g (0.1 mol) of ethyl chloroacetate wasadded. The mixture was stirred and refluxed under N2 for6 h, evaporated in vacuo and the residue was distributedbetween water (100 mL) and ether (50 mL). The waterphase was twice more extracted with ether, washed withbrine, dried (Na2SO4), evaporated in vacuo and distilledin vacuo. B.p. 72–74 °C/0.5 torr, yield: 13.9 g, 88%.GC-MS: 226 (M+), other peaks: 153 (MeO-C6H4-S=CH2), 107 (MeO-Ph+), 46 (CH2=SH+). 1H-NMR

1029

(CDCl3): 1.11 t (3H, ethyl), 3.43 d (2H, SCH2), 4.11 s(3H, CH3O), 4.76 q (2H, ethyl), 6.66 d (2H,J = 7.34 Hz),7.24 d (2H,J = 7.34 Hz).

4.8. 4-hydroxyphenyl-1-thioacetic acid(8)

Ethyl-4-methoxyphenyl thioacetate (12.9 g, 0.057 mol)was dissolved in 50 mL of glacial acetic acid and 25 mLof conc. hydrobromic acid was added. The mixture wasgently refluxed under N2 for 25 min, cooled and evapo-rated to dryness. The residue was crystallised fromwater-ethanol. M.p. 266–267 °C. Yield: 5.8 g, 56%. El-emental anal. for C8H8O3S (C, H, N, S, % S calc.17.40%, found 17.86%).1H-NMR (DMSO): 3.76 s (2H,CH2), 6.87 d (2H,J = 7.88 Hz), 7.87 d (2H,J = 7.87 Hz),10.11 bs, (1.1H, COOH).

4.9. 4-aminobenzoyloxy-acetylmorpholin(1l)

The compound was prepared according to the proce-dure described for compounds1a–h, starting from 1.25 g(0.01 mol) of 4-aminobenzoic acid. M.p. 177 °C (MeOH/diisopropyl ether), yield: 0.9 g, 36%. Elemental anal.C13H16NO4 (C, H, N). 1H-NMR (DMSO): 3.73 m (2H),3.83 m (6H), 4.76 s (2H), 6.55 d (2H,J = 7.23 Hz), 7.43d (2H, J = 7.34 Hz).

4.10. N-(4-Mercaptobenzoyloxy)-acetyl morpholine(1m)

N-(4-aminobenzoyl)-oxyacety morpholine1l (1.77 g,0.0072 mol) was dissolved in 20 mL of water and 1.5 mLof conc. HCl was added. The mixture was cooled to 0 °C,and 0.55 g (0.008 mol) of sodium nitrite in 5 mL waterwas added dropwise at 0 °C. Then, the mixture wasstirred for 15 min at the same temperature and 1.28 g(0.008 mol) of potassium ethyl dithiocarbonate in 10 mLof water was added portionwise. The mixture was stirredfor 10 min and heated gradually to 60 °C, and then stirredfor 2 h at room temperature. The precipitated compoundwas dissolved by addition of 20 mL of dioxane and20 mL of water ammonia was added and stirred for 2 hmore. The mixture was evaporated partly in vacuo,extracted with four portions of ethyl acetate, washed withsat. NaHCO3, brine, dried (Na2SO4), evaporated in vacuoand recrystallised from diisopropyl ether and methanol.M.p. 165–166 °C. Yield: 0.97 g, 48%. Elemental anal.C13H15NO4S (C, H, N, S, % S calc. 11.39, found 11.10).1H-NMR (CDCl3): 1.17 s (1.3H, SH), 3.56 m (2H), 3.93(6H), 4.55 s (2H, CH2O), 6.14 d (2H,J = 6.93 Hz), 7.25d (2H, J = 7.01 Hz).

4.11. Preparation of compounds2a–h, 2l, 4a–e, 7a, 7band10 (typical procedure)

4-Guanidinobenzoic acid mesylate (1.38 g, 0.005 mol)was dissolved in 40 mL of dry pyridine and 0.005 mol ofthe hydroxycompound were added. The mixture wascooled to 0 °C, and DCC (1.03 g 0.005 mol) were addedat once. The mixture was stirred overnight under chlor-calcium tube at room temperature, precipitated dicyclo-hexyl urea was filtered, washed with two small portionsof pyridine and the filtrate was poured into 350 mL ofether. The precipitated product was dissolved in methanol(30–40 mL) and passed through the column of 50 g ofneutral Al2O3, which was washed with methanol (ca.50 mL). The solution was evaporated in vacuo to ca.10 mL volume, and it was poured into 50 mL of dry ethylacetate. The product was filtered, washed with ethylacetate, dried in vacuo and the purity was monitored byHPLC.

2a: amorphous, yield: 1.06 g, 43%. Elemental anal. forC22H26N4O9S (C, H, N, S) % S calc. 6.14, found 5.79.1H-NMR (DMSO): 2.98 s (3H), 3.73 m (8H), 5.06 s(2H), 7.21 d (2H), 7.24 d (2H), 7.86 d (2H), 7.99 d (2H).13C-NMR (DMSO): 41.67, 44.36, 62.09, 65.96 (OCH2),122.43, 122.57, 124.89, 127.20, 129.47, 131.52, 131.14,155.45, 162.33, 163.57, 169.65.

2b: amorphous, yield: 1.04 g, 39%. Elemental anal. forC23H29N5O8S (C, H, N, S).1H-NMR (DMSO): 2.96 s(3H), 3.34 m (2H), 3.36 s (3H), 3.55 m (6H), 4.99 s (2H),7.11 d (2H), 7.23 d (2H), 7.76 d (2H), 8.06 d (2H).13C-NMR (DMSO): 38.46, 45.36, 45.76, 47.31, 47.26,51.31, 58.42, 122.49, 124.8, 127.1, 151.1, 131.3, 151.6,141.5, 155.6, 163.5, 164.6, 171.3.

2c: after semi-preparative HPLC purification, 250 mgof the compound was purified.

Amorphous, Elemental anal. for C27H29N4O10S (C, H,N, S). 1H-NMR (DMSO): 2.43 s (3H), 3.46 s (2H,OCH2Ar), 3.84 s (2H, NHCH2), 5.53 s (2H, OCH2CO),6.90 d (2H,J = 7.55 Hz), 7.29 m (7H), 7.96 d (2H,J =7.56 Hz), 8.25 d (2H,J = 8.91 Hz), 9.72 bs (3H).13C-NMR (DMSO): 39.50 (mesylate), 40.97 (OCH2Ar),61.31 (OCH2CO), 115.4, 117.0, 120.1, 122.6, 125.3,126.6, 127.6, 128.5, 128.2, 129.7, 130.3, 134.9, 155.4,163.2, 164.9, 165.3, 170.6.

2d: amorphous, yield: 1.40 g, 45%. Elemental anal. forC27H30N5O8S.1H-NMR (DMSO): 2.37 s (3H, mesylate),2.39 s (3H, CH3N), 2.75 t (2H,J = 4.78 Hz, CH2), 2.90t (2H, J = 4.78 Hz, CH2), 3.66 m (8H), 4.78 s (2H,OCH2), 6.90 d (2H,J = 8.76 Hz), 7.10 d (2H,J = 7.56Hz), 7.32 d (2H,J = 8.77 Hz), 7.90 d (2H,J = 8.76 Hz).13C-NMR (DMSO): 29.4 (CH2), 34.7 (CH2), 39.4 (me-sylate), 45.6, 54.11, 54.2, 61.3 (OCH2), 119.0, 121.7,

1030

121.8, 122.46, 129.18, 130.87, 137.8, 149.1, 153.7,156.5, 164.6, 171.7.

2e: amorphous, yield: 1.17 g, 45%. Elemental anal. forC24H28N4O9S (C, H, N, S) % S calc. 5.84, found 5.41.1H-NMR (DMSO): 2.44 (3H, mesylate), 3.45 m (8H),5.05 s (1H), 7.2–7.32 m (5H), 7.51–7.53 m (2H),7.77–7.79 m (1H), 8.64 m (1H).13C-NMR (DMSO):38.9 (mesylate), 40.7, 43.7, 60.4 (OCH2), 66.3, 118.3,119.2, 124.6, 126.2, 126.9, 129.5, 129.9, 130.5, 131.6,133.7, 136.48, 143.8, 146.3, 155.8, 165.8, 1265.9, 167.4.

2f: amorphous, yield: 1.19 g, 45%. Elemental anal. forC21H25N5O9S (C, H, N, S).1H-NMR (DMSO): 2.36 s(3H, mesylate), 3.55–3.74 2×d (8H), 5.05 s (2H), 6.57 d(1H), 7.36 d (2H,J = 8.76 Hz), 7.65 d (2H,J = 8.77 Hz),8.04 dd (1H), 8.29 d (1H).13C-NMR (DMSO): 39.5(mesylate), 43.43, 46.15, 62.68 (OCH2), 67.5, 67.6,111.3, 120.4, 121.5, 124.3, 130.34, 131.3, 135.4, 141.8,141.9, 155.7, 160.3, 165.8, 174.6.

2g: amorphous, yield: 1.03 g, 39%. Elemental anal. forC25H29N3O9S (C, H, N, S).1H-NMR, (DMSO): 1.49 s(9H, tBuO), 2.46 (3H, mesylate), 4.66 s (2H, OCH2),6.69 d (1H,J = 11.35 Hz, CH=), 7.19 d (2H,J = 7.45 Hz),7.45 d (2H,J = 8.1 Hz), 7.71 d (1H,J = 11.36 Hz), 7.89d (2H,J = 7.98 Hz), 8.17 d (2H,J = 7.44 Hz).13C-NMR(DMSO): 27.65, 39.65 (MsOH), 61.14 (OCH2), 81.58,117.3, 122.43, 122.6, 124.7, 130.5, 131.5, 131.9, 142.1,143.3, 152.2, 155.6, 163.8, 165.6, 166.8.

2h: amorphous, yield: 1.12 g, 40%. Elemental anal. forC25H30N4O9S (C, H, N, S) % S calc. 5.70, found 5.39.1H-NMR (DMSO): 1.42 s (9H, tBuO), 2.38 s (3H,MsOH), 4.58 s (2H, OCH2), 6.90 d (2H,J = 7.34 Hz),7.12 dd (1H), 7.36–7.40 m (2H), 7.86 d (2H,J = 7.83Hz), 8.88 s (1H, NH ind.).13C-NMR (DMSO): 27.5(tBu), 36.6 (MsOH), 37.2 (CH2Ar), 56.6 (OCH2), 111.1,118.69, 119.3, 121.6, 121.8, 121.9, 123.5, 123.9, 125.8,131.6, 136.3, 141.4, 142.3, 155.1, 161.3, 163.8, 170.9.

2l: amorphous, yield: 1.29 g, 48%. Elemental anal. forC22H26N4O8S2 (C, H, N, S).1H-NMR (DMSO): 2.34 s(3H, MsOH), 3.56 m (2H), 3.87 m (6H), 7.23 d (2H,J =7.56 Hz), 7.36 d (2H,J = 7.80 Hz), 7.85 d (2H,J = 7.55Hz), 7.98 d (2H,J = 7.58 Hz).13C-NMR (DMSO): 36.7(MsOH), 45.5, 46.1, 59.1 (OCH2), 65.5, 116.3, 117.5,117.9, 118.0, 118.2, 121.3, 138.3, 139.8, 155.8, 160.2,162.1, 165.3.

4a: amorphous, yield: 1.19 g, 46%. Elemental anal. forC24H28N4O7S (C, H, N, S).1H-NMR (DMSO): 2.36 s(3H, MsOH), 2.64 t (2H,J = 8.96 Hz), 2.86 t (2H,J =7.77 Hz), 3.43 m (8H), 7.11 d (2H,J = 6.98 Hz), 7.17 d(2H, J = 7.85 Hz), 7.37 d (2H,J = 7.00 Hz), 8.04 d (2H,J = 7.85 Hz).13C-NMR (DMSO): 33.3 (CH2), 33.7 (CH2),39.7 (MsOH), 40.4, 66.1 (OCH2), 121.6, 122.0, 122.7,129.4, 131.2, 138.9, 148.9, 149.5, 154.4, 164.2, 170.0.

4b: amorphous, yield: 1.14 g, 44%. Elemental anal. forC24H31N5O6S (C, H, N, S).1H-NMR (DMSO): 2.37 s(3H, Me-N), 2.46 s (3H, MsOH), 2.73 t (2H,J = 5.58Hz), 2.85 t (2H,J = 5.53 Hz), 3.32 m (6H), 3.64 m (2H),7.19 d (2H,J = 7.45 Hz), 7.36 d (2H,J = 8.24 Hz), 7.45d (2H,J = 7.46 Hz), 8.17 d (2H,J = 8.22 Hz).13C-NMR(DMSO): 29.9 (CH2), 33.6 (CH2), 39.3 (MsOH), 43.1(MeN), 44.1, 52.9, 53.1, 121.5, 122.5, 122.7, 125.3,129.5, 138.1, 148.7, 155.5, 164.1, 170.1.

4c: amorphous, yield: 1.17 g, 44%. Elemental anal. forC24H31N4O8S (C, H, N, S).1H-NMR (DMSO): 2.38 s(3H, MsOH), 2.56 t (2H,J = 5.87 Hz), 2.89 t (2H,J =5.88 Hz), 3.28 bs (2H), 3.87 m (6H), 4.22 d (2H,CH2NH), 6.66 bs (1H, NH), 6.87 d (2H,J = 7.81 Hz, 7.12d (2H,J = 7.56 Hz), 7.87 d (2H,J = 7.87 Hz), 7.99 d (2H,J = 7.53 Hz), 10.11 bs (2H, NH), 11.1 bs (1.5H, NH).13C-NMR (DMSO): 30.1 (CH2), 31.7 (CH2), 38.9(MsOH), 40.0, 45.7, 50.2, 53.3, 61.1 (OCH2), 120.1,122.21, 123.1, 123.8, 125.7, 126.6, 129.3, 138.3, 141.5,151.3, 158.7, 161.6, 170.0.

4d: amorphous, yield: 0.88 g, 38%. Elemental anal. forC20H26N4O7S (C, H, N, S), % S calc. 6.87, found 6.45.1H-NMR (DMSO): 2.38 s (3H, MsOH), 3.60 m (8H),6.67 d (1H,J = 10.26 Hz), 7.23 d (2H,J = 7.83 Hz), 7.36d (2H,J = 7.22 Hz), 7.66 d (1H,J = 10.3 Hz), 8.06 d (2H,J = 7.28 Hz), 8.11 d (2H,J = 7.81 Hz). 13C-NMR(DMSO): 39.91 (MsOH), 42.5, 45.7, 66.2, 66.3, 118.13,122.3, 122.5, 122.7, 122.8, 129.3, 131.4, 132.9, 140.7,147.1, 151.6, 154.7, 164.1, 164.5.

4e: amorphous, yield: 1.22 g, 48%. Elemental anal. forC25H26N4O6S (C, H, N, S).1H-NMR (DMSO): 2.39 s(3H, MsOH), 2.75 s (2H, CH2NH), 6.72 d (1H,J = 10.3Hz), 7.27 m (5H), 7.62 d (2H,J = 7.80 Hz), 7.71 d (2H,J = 7.21 Hz), 8.08 d (1H,J = 10.4 Hz), 8.17 d (1H,J =7.85 Hz), 8.24 d, (2H,J = 7.23 Hz).13C-NMR (DMSO):39.5 (MsOH), 44.37, 116.7, 122.1, 122.3, 123.3, 123.4,125.1, 128.3, 128.44, 129.3, 129.8, 130.2, 140.9, 145.3,155.8, 168.8, 178.2.

7a: amorphous, yield: 1.28 g, 49%. Elemental anal. forC22H27N5O8S (C, H, N, S).1H-NMR (DMSO): 2.41 s(3H, MsOH), 3.48 m (2H), 3.46 m (6H), 5.62 d (2H,CH2NH), 7.11 d (2H,J = 7.77 Hz), 7.49 d (2H,J = 7.42Hz), 7.58 d (2H,J = 7.77 Hz), 8.11 d (2H,J = 7.39 Hz).13C-NMR (DMSO): 33.3 (MsOH), 40.1, 41.8, 47.5,(NHCH2), 66.1, 114.8, 122.7, 124.8, 129.1, 131.5, 141.3,148.5, 155.4, 160.2, 165.1, 167.9.

7b: amorphous, yield: 1.00 g, 38%. Elemental anal. forC21H26N4O8S2 (C, H, N, S).1H-NMR (DMSO): 2.41 s(3H, MsOH), 3.40 m (2H), 3.57 m (6H), 4.14 s (2H,SCH2), 7.39 d (2H,J = 7.33 Hz), 7.59 d (2H,J = 7.87Hz), 8.05 d (2H,J = 7.34 Hz), 8.19 d (2H,J = 7.87 Hz).13C-NMR (DMSO): 39.5 (MsOH), 42.1, 45.9, 52.7

1031

(SCH2), 66.4, 115.3, 122.4, 128.5, 129.2, 130.7, 131.6,133.4, 155.9, 162.0, 163.6, 165.2.

10: amorphous, yield: 1.01 g, 37%. Elemental anal. forC23H28N4O9S2 (C, H, N, S).1H-NMR (DMSO): 2.42 s(3H, MsOH), 3.37 m (2H), 3.63 m (6H), 4.02 s (2H,SCH2), 4.87 s (2H, OCH2), 7.21 d (2H,J = 7.45 Hz), 7.32d (2H,J = 8.00 Hz), 7.51 d (2H,J = 7.56 Hz), 8.13 d (2H,J = 8.02 Hz).13C-NMR (DMSO): 39.7 (MsOH), 41.5,44.18, 44.27 (SCH2), 62.2 (OCH2), 65.8, 116.1, 122.6,122.9, 129.7, 130.9, 133.3, 134.4, 149.1, 155.3, 164.0,164.6, 168.9.

4.12. Preparation of compound2i

In the common apparatus for hydrogenation undernormal pressure, was dissolved 850 mg (1.12 mmol) ofthe compound2c in 10 mL of methanol, and 21 mg of10% Pd on charcoal were added. The hydrogenationproceeded for 4 h, the catalyst was filtered, washed withmethanol, the methanolic solution was partly evaporated,and the product was precipitated with an excess of ethylacetate. After purification by semi-preparative HPLC andlyophilisation, 420 mg of the product was gained as anamorphous powder (yield: 74%). Elemental anal. forC20H22N4O10S. (C, H, N, S).1H-NMR (DMSO): 2.80 s(3H, MsOH), 3.61 d (2H, NHCH2), 4.75 s (2H, OCH2),7.33 d (2H,J = 7.11 Hz), 7.45 d (2H,J = 7.34 Hz), 8.06 d(2H, J = 7.09 Hz), 8.21 d (2H,J = 7.14 Hz), 11.01 bs(3.5H, NH and COOH). 13C-NMR (DMSO): 20.4(MsOH), 41.4 (NHCH2), 62.4 (OCH2), 114.4, 122.1,122.3, 124.2, 126.8, 130.9, 131.1, 131.4, 132.9, 142.2,155.9, 163.5, 164.2, 165.9, 172.7.

4.13. Preparation of compounds2j and 2k (generalprocedure)

The tert. butyl ester2gor 2h (0.005 mol) was dissolvedin 40% trifluoroacetic acid in dry DCM (10 mL), con-taining 2 mL of anisole, and stirred for 15 min. Thevolatile compounds were removed in vacuo and the oilyresidue was treated with ethyl acetate. The precipitatewas filtered and purified by semi-preparative HPLC.After lyophilisation of the appropriate fraction, amor-phous product was obtained.

2j: yield: 354 mg, 56%. Elemental anal. forC22H20N3O9S (C, H, N, S) % S calc. 6.38, found 6.80.1H-NMR (DMSO): 2.42 s (3H, MsOH), 4.70 s (2H,OCH2), 6.78 d (1H,J = 10.3 Hz), 7.36 d (2H,J = 6.91Hz), 7.71 d (2H,J = 7.87 Hz), 7.84 d (1H,J = 10.1Hz),7.94 d (2H,J = 6.92 Hz) 8.28 d (2H,J = 7.87 Hz), 9.9 bs(4H, COOH and NH).13C-NMR: 39.9 (MsOH), 60.6

(OCH2), 117.5, 122.5, 122.7, 122.8, 129.8, 130.8, 130.9,131.0, 131.5, 131.8, 140.9, 144.5, 155.3, 161.7, 165.6,169.1.

2k: amorphous, yield: 372 mg, 59%. Elemental anal.for C21H22N4O9S (C, H, N, S).1H-NMR (DMSO): 2.38 s(3H, MsOH), 2.56 t (2H,J = 5.87 Hz), 2.89 t (2H,J =5.88 Hz), 3.28 bs (2H), 3.87 m (6H), 4.22 d (2H, CH2N),6.66 bs (1H, NH), 6.87 d (2H,J = 7.84 Hz), 7.12 d (2H,J = 7.56 Hz), 7.87 d (2H,J = 7.87 Hz), 7.99 d (2H,J =7.54 Hz), 10.11 bs (2H), 11.10 bs (1H).13C-NMR(DMSO): 30.1 (CH2), 31.7 (CH2), 38.9 (MsOH), 40.0,45.7, 50.2, 53.3, 61.11 (OCH2), 120.1, 122.2, 123.1,123.8, 125.7, 126.6, 129.3, 138.3, 141.5, 151.3, 158.7,161.6, 170.00.

4.14. Preparation of compound11

Compound10 (1.45 g, 2.55 mmol) were dissolved in15 mL of glacial acetic acid and cooled to 0 °C. Then,4.5 mL of 30% water hydrogen peroxide were added andthe mixture was stirred overnight at room temperature.The mixture was evaporated in vacuo (bath temp. max.30 °C), the gummy residue dissolved in methanol andprecipitated with ethyl acetate. 1.22 g (87%) of the crudecompound was gained. 350 mg of this compound werepurified by semi-preparative HPLC and proceeded to theanalyses and inhibiting activities. Elemental anal.C23H28N3O11S (C, H, N, S).1H-NMR (DMSO): 2.42 s(3H, MsOH), 3.36 m (4H), 3.56 m (4H), 4.87 s (2H),4.87 s (2H), 7.23 d (2H,J = 7.86 Hz), 7.36 d (2H,J = 8.11Hz), 7.87 d (2H,J = 7.88 Hz), 8.01 d (2H,J = 8.09 Hz).13C-NMR (DMSO): 25.4 (MsOH), 41.6, 60.1 (OCH2),62.4 (SO2CH2), 65.8, 122.4, 122.6, 122.8, 122.9, 124.9,126.1, 133.9, 141.9, 155.5, 162.4, 163.7, 164.8.

4.15. 1-BOC-isonipecotic acid14

Isonipecotic acid (12.9 g, 0.1 mol) was dissolved inwater (200 mL) containing 10.6 g (0.1 mol) of sodiumcarbonate and the mixture was cooled to 0 °C withstirring. Then, the solution of di-tert. butyl dicarbonate in200 mL of dioxane was slowly added at the sametemperature. The mixture was stirred overnight at roomtemperature, evaporated in vacuo to ca. 200 mL of thevolume, diluted with 200 mL of water, acidified with anexcess of the solid citric acid and extracted with threeportions of ethyl acetate. The ethyl acetate solution waswashed with sat. sodium carbonate, brine, dried(NA2SO4) and evaporated in vacuo.The product wasrecrystallised from hexane-EtOAc. M.p. 146–147 °C,yield: 19.9 g, 86%. Elemental anal. for C11H19NO4 (C, H,N). 1H-NMR (CDCl3): 1.03–1.45 m (4H), 2.79 s (9H,BOC), 3.11 m (4H), 4.33 m (1H, CH-N).

1032

4.16. Preparation of compounds12aandb and15aandb (typical procedure)

The corresponding acid (0.001 mol) was dissolved inthe mixture of dichloromethane-acetonitrile (both dry)(1:1 v/v, 20 mL) and the corresponding hydroxy deriva-tives (0.001 mol) were added. After both compoundswere dissolved, the solution was cooled to –20 °C underchlorcalcium cover and then DCC (2.01 g, 0.001 mol)was added at once. The mixture was stirred overnight andit was allowed to reach room temperature. The dicyclo-hexyl urea was removed by filtration, washed with twosmall portions of dichloromethane, the filtrate was evapor-ated and the product was crystallised.

12a: m.p. 145–146 °C (iPr2O-EtOAc). Yield: 490 mg,63%. Elemental anal. for C40H47N5O12 (C, H, N). 1H-NMR (CDCl3): 1.48 s (9H, BOC), 1.77 m (2H, CH2 Arg),3.87 t (2H, CH2 Arg.), 3.91–4.01 m (12H), 4.13 m (1H,CH-N), 4.9 s (2H, CH2O), 5.13 s (2H, CH2O), 5.25 s(2H, CH2O), 7.23 d (2H,J = 7.46 Hz), 7.25–7.36 m (10H,C6H5), 8.11 d (2H,J = 7.50 Hz). 13C-NMR (DMSO):24.8 (CH2 Arg), 28.2 (BOC), 33.9 (CH2N Arg), 42.1(CH2 Arg), 44.1 and 45.1 (OCH2Ph), 49.1 (CH-N), 53.5(OCH2), 61.7, 66.32, 69.0, (morpholine), 80.1 (BOC),115.43, 121.48, 127.82, 128.1, 128.3, 128.6, 129.1,129.2, 129.3, 129.4, 131.5, 132.1, 134.3, 136.7, 154.4,155.6, 155.7, 160.4, 161.6, 165.9, 170.7.

12b: m.p. 69–70 °C (iPr2O/EtOAc). Yield: 555 mg,68%. Elemental anal. for C42H51N5O12 (C, H, N). 1H-NMR (CDCl3): 1.47 s (9H, BOC), 2.70 t (2H,J = 5.7 Hz,Ph-CH2CH2-), 2.91 t (2H, J = 5.6 Hz, PhCH2CH2-),5.53 m (2H, CH2 Arg), 3.65–3.75 m (8H), 4.03 m (1H),4.70 s (2H, OCH2), 5.11 s (2H, OCH2), 5.25 s (2H,OCH2), 6.90 d (2H,J = 7.9 Hz), 7.18 d (2H,J = 7.8 Hz),7.22–7.26 m (10 H).13C-NMR (DMSO): 24.9 (CH2Arg),28.2 (BOC), 30.1, 33.9, 35.4, 42.1, 44.1, 44.9, 49.1, 53.4,61.1, 66.6, 66.9, 70.1, 79.9, 115.4, 121.3, 127.7, 127.9,128.1, 128.2, 128.3, 128.4, 128.6, 128.7, 129.1, 129.2,134.6, 136.8, 138.1, 148.8, 155.4, 155.7, 160.4, 163.7,163.8, 165.1.

15a: m.p. 111–112 °C (Hexane/EtOAC). Yield:280 mg, 58%. Elemental anal. for C24H32N2O8 (C, H, N).1H-NMR (CDCl3): 1.08 and 1.09 dd (1H), 1.4 s (9H,BOC), 2.00 and 2.03 dd (1H), 2.92 dd (1H), 3.45 bs (2H),3.58 m (3H), 3.70 bs (6H), 4.04 d (1H,J = 5.45 Hz),4.90 s (2H, CH2O), 7.15 d (2H,J = 7.62 Hz), 8.16 d (2H,J = 7.66 Hz).13C-NMR (DMSO): 27.8, 28.4, 41.5, 48.9,81.1, 127.8, 128.1, 128.5, 132.3, 133.1, 138.7, 154.2,165.3, 166.2, 172.4.

15b: m.p. 126–127 °C (iPr2O-EtOAc). Yield: 273 mg,56%. Elemental anal. for C25H35N3O7 (C, H, N). 1H-NMR (CDCl3): 1.43 s (9H), 1.72 dd (1H), 1.89 dd (1H),

2.0 dd (1H), 2.50 bs (2H), 3.33 s (3H, CH3N), 3.38 m(3H), 3.65 m (6H), 4.11 dd (1H), 4.96 s (2H, OCH2), 7.14d (2H,J = 7.41 Hz), 8.10 d (2H,J = 7.41 Hz).13C-NMR(DMSO): 28.3 (BOC), 34.7, 41.5, 41.8, 42.9, 43.1, 54.5,61.8, 79.7 (BOC), 121.4, 121.6, 127.0, 131.5, 154.5,164.7, 165.3, 172.5.

4.17. Preparation of compounds13a and b (typicalprocedure)

In the standard apparatus for hydrogenation undernormal pressure, were placed, 0.0005 mol of the com-pound16aor 16b dissolved in 10 mL of methanol, and 5µl of the 99% MeSO3H were added, followed by 50 mgof the 10% Pd on charcoal, and the hydrogenationproceeded over 2 h. The catalyst was removed by filtra-tion, washed with methanol, the methanolic solution waspartly evaporated in vacuo (to ca. 4–5 mL) and theproduct was precipitated with an excess of diisopropylether, filtered and purified via semi-preparative HPLC.

13a: m.p. amorphous. Yield: 121 mg (after HPLCpurification). Elemental anal. for C25H39N5O11S(CHNS). 1H-NMR (DMSO): 1.81 m (2H, CH2 Arg),2.87 s (9H, BOC), 3.25 t (2H, CH2 Arg), 3.37 m (2H,CH2N Arg), 3.59 bs (3H), 3.76 m (5H), 4.78 m (1H,CH-N), 4.87 s (2H, OCH2), 6.71 d (2H,J = 7.46 Hz),7.82 d (2H,J = 7.49 Hz).13C-NMR (DMSO): 25.7, 26.12(BOC), 35.7, 39.4, 42.1, 43.4, 46.4, 47.1, 48.1 (MsOH),50.4, 62.6, 81.1, 115.3, 118.4, 133.7, 142.5, 160.3, 168.4,170.1, 176.9.

13b: amorphous. Yield: 134 mg (after HPLC purifica-tion). Elemental anal. for C27H43N5O11S (C, H, N, S).1H-NMR (DMSO): 1.18 m (3H), 1.51 s (9H), 2.36 t (2H,J = 6.1 Hz), 2.71 t (2H,J = 6.0 Hz), 3.33 m (2H),3.781–3.75 m (6H), 3.91 m (2H), 4.2 m (1H), 4.81 s(2H), 7.11 d (2H,J = 7.11 Hz), 7.26 d (2H,J = 7.13 Hz),10.11 bs (3H).13C-NMR (DMSO): 24.9, 29.7 (BOC),31.3, 33.9, 35.6, 42.1, 44.1, 44.9, 50.1, 56.4, 60.3(OCH2), 81.3 (BOC), 118.3, 121.6, 127.3, 136.6, 151.1,155.6, 161.4, 168.6, 171.2.

4.18. Preparation of compounds16a and b (typicalprocedure)

The BOC protected compound16aor 16b (0.005 mol)was dissolved in 10 mL of ethyl acetate and cooled to0 °C. Then, 5 mL of 5 M HCl in EtOAc was addeddropwise over 5 min. The precipitate appeared immedi-ately. The mixture was standing for another 30 min, theprecipitate was filtered, dried in vacuo and dissolved in10 mL of ethanol. Cyanamide (0.27 g, 0.0065 mol) andconc. HCl (2 mL) were added and the mixture was stirredovernight. The product was precipitated after the reaction

1033

mixture was kept at –20 °C for 24 h. The precipitate wasfiltered, washed with isopropanol, dried and purified bysemi-preparative HPLC.

16a: amorphous. Yield: 127 mg (after HPLC purifica-tion). Elemental anal. for C20H26N4O6Cl (C, H, N, Cl, %Cl calc. 7.83, found 8.11).1H-NMR (DMSO): 1.92 m(2H), 2.15 m (2H), 2.93 m (2H), 3.28–3.61 m (9H),5.44 m (1H), 5.05 s (2H, CH2O), 7.39 d (2H, J =7.75 Hz), 8.09 d (2H,J = 7.73 Hz), 9.10–9.19 bs (4H).13C-NMR (DMSO): 22.7, 40.7, 41.6, 44.3, 44.4, 56.6,61.1, 61.4, 65.9, 110.5, 115.3, 120.1, 126.2, 131.8, 157.3,160.1, 160.4, 162.12, 165.2.

16b: amorphous. Yield: 111 mg (after HPLC purifica-tion). Elemental anal. for C20H32N5O5Cl2 (C, H, N, Cl, %Cl calc. 14.38, found 14.79).1H-NMR (DMSO): 2.61 m(2H), 2.81 m (2H), 2.81–3.51 m (10H), 3.51s (3H),4.51 m (1H), 5.11 s (2H, OCH2), 7.53 d (2H,J = 7.63Hz), 8.04 d (2H,J = 7.38 Hz).13C-NMR (DMSO): 16.19,24.3, 37.7 (MsOH), 38.6, 40.7, 41.9, 46.5, 46.6, 46.9,

57.0, 61.9, 122.2, 127.1, 130.9, 131.6, 154.2, 164.1,164.7, 171.6.

4.19. Inhibiting activity assays

Inhibiting activities were measured according to thedescribed procedure [3] using the following substrates:Bz-Arg-pNa for trypsin, Bz-Phe-Val-Arg-pNa for throm-bin, D-Val-Leu-Lys-pNa for plasmin and Z-Val-Gly-Arg-pNa for urokinase.

References

[1] De Clerck Y.A., Imren S., Eur. J. Cancer 30a (14) (1994)2180–2206.

[2] Das S., Mukopadhyay P., Acta Oncologica 33 (8) (1994) 859–876.

[3] Zlatoidsky P., Maliar T., Eur. J. Med. Chem. 31 (1996) 895–899.

[4] Ganu W.S., Shaw E., J. Med. Chem. 42 (1981) 698–700.

[5] Jetten M., Peters C.A.M., Van Nispen J.W.F.M., Ottenheijm H.J.C.,Tetrahedron Lett. 32 (1991) 6025–6028.

1034

Original article

Depsides as non-redox inhibitors of leukotriene B4 biosynthesis and HaCaT cellgrowth. 1. Novel analogues of barbatic and diffractaic acid

Sunil Kumar KCa, Klaus Müllerb*aInstitut für Pharmazie, Pharmazeutische Chemie I, Universität Regensburg, D-93040 Regensburg, Germany

bWestfälische Wilhelms-Universität Münster, Institut für Pharmazeutische Chemie, Hittorfstraße 58 – 62, D-48149 Münster, Germany

(Received 25 March 1999; accepted 7 May 1999)

Abstract – A series of barbatic and diffractaic acid analogues has been synthesized and evaluated as inhibitors of leukotriene B4 (LTB4)biosynthesis and as antiproliferative agents. The 4-O-demethyl barbatic and diffractaic acid derivatives were among the most activecompounds in both assays. In particular, ethyl 4-O-demethylbarbatate was the most potent LTB4 biosynthesis inhibitor of this series, with anIC50 value in the submicromolar range. Because the compounds did not show appreciable reactivity against a stable free radical,2,2-diphenyl-1-picrylhydrazyl, and did not produce appreciable amounts of deoxyribose degradation as a measure of their potency to generatehydroxyl radicals, a simple redox effect could not explain their biological activity. Also, there was no nonspecific cytotoxicity as documentedby the activity of lactate dehydrogenase released from the cytoplasm of keratinocytes, which was in the control range. © 1999 Éditionsscientifiques et médicales Elsevier SAS

barbatic acid / diffractaic acid / antiproliferative activity / keratinocytes / lactate dehydrogenase release / leukotriene biosynthesis

1. Introduction

Depsides are a distinct class of lichen-derived com-pounds which are formed by condensation of two or morehydroxybenzoic acids whereby the carboxyl group of onemolecule is esterified with a phenolic hydroxyl group ofa second molecule. One of the most common secondarymetabolites of many lichen species [1] is the didepsidediffractaic acid (1, figure 1). This compound has beenshown by several groups to exhibit antiviral [2], anti-tumour [3], analgesic and antipyretic [4] properties.Among several structurally dissimilar lichen-derived me-tabolites isolated fromParmelia species, we have iden-tified 1 as a non-redox inhibitor of the biosynthesis ofleukotriene B4 (LTB4) in bovine polymorphonuclearleukocytes (PMNL) [5]. Leukotrienes are derived fromthe biotransformation of arachidonic acid through theaction of 5-lipoxygenase (5-LO) and play an importantrole in a variety of pathophysiological states in man,particularly those involving inflammation [6]. Further-more, we found that1 is also a potent antiproliferative

agent against the growth of human keratinocytes [7].These combined inhibitory actions against 5-LO andkeratinocyte cell growth suggested a beneficial effectagainst inflammatory and hyperproliferative skin diseasessuch as psoriasis, since both pathological features weretargeted.

As part of our continuing search for agents suitable forthe treatment of inflammatory and hyperproliferative skin*Correspondence and reprints

Figure 1. Structures of diffractaic acid (1) and barbatic acid(2).

Eur. J. Med. Chem. 34 (1999) 1035−1042 1035© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

diseases such as psoriasis, we have synthesized severalnovel analogues of1 and its congener barbatic acid (2),modified at the carboxylic acid function and the4-methoxy group in the benzoyloxy moieties, to explorethe effect of increased lipophilicity and some redoxproperties on the biological activity of the compounds.The redox properties were evaluated in terms of reactivityagainst the stable free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH), in order to evaluate the antioxi-dant potential, and deoxyribose degradation was deter-mined as a measure of hydroxyl radical formation [8].The ability of the novel compounds to inhibit the growthof human keratinocytes was evaluated in HaCaT cells [9],and inhibition of LTB4 biosynthesis was assayed inbovine polymorphonuclear leukocytes [8].

2. Chemistry

Unambiguous syntheses of the lichen depsides1 and2have been reported [10]. The mononuclear precursors forthe novel depsides were obtained from ethyl 2,4-dihydroxy-3,6-dimethylbenzoate(3)followingthemethod-ology of Elix [10] which proved to be particularly suit-able for the large scale preparation of this startingmaterial (figure 2). Esters4a–4e were directly obtainedfrom 3 by transesterification in the presence of thecorresponding sodium alkoxides and alcohols [10]. Alkyl-ation of3 with a one molar proportion of dimethyl sulfateor the pertinent benzyl chlorides in the presence ofpotassium carbonate, selectively yielded the correspond-ing 4-methoxy derivative5 or 4-benzyloxy derivatives9aandb. Methylation of the second hydroxy group of thesederivatives with dimethyl sulfate gave the 2-methoxyderivatives6 and10aandb. Subsequent hydrolysis of theesters5, 6, 9a andb, and10aandb yielded the requisiteacids7, 8, 11aandb, and12aandb, respectively, for theA-ring of the desired depsides. Depside formation be-tween these acids and the phenolic esters3 and4a–e wasachieved by treatment with trifluoroacetic anhydride inanhydrous toluene and yielded the barbatic and diffractaicacid analogues13a–f, 14a–f, 15a–f and 17a–f, 18a–f,19a–f, respectively (figure 3). Hydrogenolytic cleavageof the benzyl ethers14a–e and 18a–e over palladium/carbon produced the phenolic analogues16a–e and20a–e, respectively, whereas benzyl esters14f and 18fwere cleaved to the acids16f and 20f, respectively.Analogously, benzyl esters13f and17f yielded the parentcompounds2 and1, respectively.

3. Biological results and discussion

Inhibition of LTB4 biosynthesis by the barbatic anddiffractaic acid analogues was determined in bovinepolymorphonuclear leukocytes. As shown intable I, ac-tivity of barbatic acid (2), with an IC50 value of 7.8µM,was similar to that of1. The biological activity wasreduced when the free carboxylic acid groups wereesterified. The esters13a–f of 2 were either moderatelyactive or inactive even at concentrations up to 20µM.Also, esterification of1 resulted in less active or inactivecompounds (17aandc–f), except for17b, where activitywas retained. In general, introduction of a 4-benzyl or4-methoxybenzyl group into1 and2, as in14a–f, 15a–f,18a–f, and19a–f dramatically reduced inhibitory actionagainst LTB4 biosynthesis. Most of these compoundswere inactive at 20 µM. Exceptions were the

Figure 2. Reagents: (a) Na, ROH,∆, 24 h; (b) Me2SO4,K2CO3, acetone,∆, 24 h; (c) 4-R≠C6H4CH2Cl, K2CO3, ace-tone,∆, 24 h; (d) KOH, H2O, DMSO, 90 °C.

1036

4-benzyloxydiffractates14c, e and f of 2 which showedcomparable activity.

Since a major class of leukotriene biosynthesis inhibi-tors often contains a hydroxylated aromatic ring [11], wehave speculated that demethylation of the 4-methoxygroup of 1 and 2 might improve their activity. Asexpected, 4-O-demethyl barbatic and diffractaic acidderivatives16a–f and20a–f, respectively, inhibited LTB4biosynthesis with IC50 values in the low micromolarrange. With the exception of the free acids16f and20f,these analogues approached the potency of their respec-tive parent compounds or were even more potent thanthese. In particular, the ethyl esters16b and20b were themost potent inhibitors of LTB4 biosynthesis of this series.Potency of compound16b, with an IC50 of 0.8 µM, wascomparable to that of the standard inhibitor nordihy-droguaiaretic acid.

One chemical feature of many inhibitors of LTB4

biosynthesis is their ability to remove free radicals, sincethe conversion of arachidonic acid into LTB4 is anoxidative process. Therefore, we have evaluated thedepsides for their ability to react with the stable freeradical DPPH to give the reduced 2,2-diphenyl-1-picrylhydrazine. Table I shows that no appreciableamount of reduced hydrazine was formed by thesecompounds, documenting their lack of reactivity againststable free radicals. This suggests that a simple redoxeffect does not explain their activity in the LTB4 assay.Rather, activity appears to be due to specific enzymeinteraction.

Moreover, the results obtained from the deoxyriboseassay (table I) also suggest that hydroxyl radicals are notinvolved in the mechanism of enzyme inhibition by thenovel depsides. The deoxyribose assay is a sensitive testfor the production of hydroxyl radicals [12]. The releaseof 2-thiobarbituric acid reactive material is expressed asmalondialdehyde (MDA) and reflects a measure forhydroxyl-radical generation. However, we did not ob-serve any deoxyribose degradation from depsides, evenfor compounds16a–f with three phenolic hydroxylgroups.

In vitro antiproliferative activities were determined in24-well culture dishes against the growth of HaCaT cells.This nontransformed human cell line can be used as amodel for highly proliferative epidermis [13]. The parentcompounds1 and2 were potent inhibitors of cell growthwith IC50 values of 2.6 and 4.1µM, respectively (table I).While the esters13a–c of 2 showed comparable orslightly improved activity, the corresponding esters17a–cof 1 were inactive. However, butyl ester17e displayedpotent antiproliferative activity. Furthermore, among the4-benzylated derivatives of1 were also some antiprolif-erative active agents. Similar to the results obtained in theLTB4 assay, 4-O-demethylation of barbatic and diffrac-taic acid (16a–f and 20a–f, respectively) generally pro-duced active compounds, although there were no im-provements as compared to the parent depsides.

There is little or no correlation between the in vitroantiproliferative activity of the compounds and theirability to inhibit LTB4 biosynthesis. With respect to bothfeatures, well-balanced representatives are found amongthe esters of diffractaic acids and the 4-O-demethylatedanalogues of1 and 2. Unfortunately, the potent LTB4biosynthesis inhibitor20b is inactive at 20µM. The mostpotent inhibitor of keratinocyte growth, ethyl diffractate(13b), is also an inhibitor of LTB4 biosynthesis. Like-wise, the most potent inhibitor of LTB4 biosynthesis,ethyl 4-O-demethylbarbatate (16b), also displays antipro-liferative activity.

Keratinocytes were also tested for their susceptibilityto the action of the most potent depsides on plasmamembrane integrity. As a measure of cytotoxicity, releaseof lactate dehydrogenase into the culture medium wasdetermined [14]. In these experiments, all potent inhibi-tors of keratinocyte growth showed values in the controlrange, documenting that their activity was due to cyto-static rather than cytotoxic effects. This may be advanta-geous as compared with the topical antipsoriatic agentanthralin, which is known to induce inflammation of thehealthy skin surrounding a psoriatic lesion. As a result ofthe strong hydroxyl radical generating activity of this

Figure 3. Reagents: (a) (CF3CO)2O, toluene, room tempera-ture; (b) Pd/C, EtOAc, room temperature. R1, R2, and R3 aredefined intable I.

1037

Table I. Redox properties, inhibition of LTB4 biosynthesis, antiproliferative activity and cytotoxicity against HaCaT cells of barbatic anddiffractaic acid derivatives.

kDPPHa DD (⋅OH)b LTB4

c AAd LDHe

Compound R1 R2 R3 (M–1 s–1) IC50 (µM) IC50 (µM) (mU)

1 Me Me H 0.63± 0.01 0.12± 0.07 7.6 2.6 1362 Me H H 0.14± 0.07 0.09± 0.01 7.8 4.1 14813a Me H Me 0.93± 0.01 0.16± 0.04 18.6 4.8 14813b Me H Et 0.31± 0.05 0.18± 0.06 11.3 1.9 14313c Me H Prop 0.29± 0.01 0.21± 0.03 10.5 3.2 14313d Me H CHMe2 ND ND > 20 > 20 ND13e Me H Bu ND ND > 20 > 20 ND13f Me H CH2Ph 0.45± 0.13 0.08± 0.01 19.2 > 20 ND14a PhCH2 H Me ND ND > 20 19.0 15014b PhCH2 H Et ND ND > 20 16.7 14214c PhCH2 H Prop 0.55± 0.04 0.19± 0.03 9.0 > 20 ND14d PhCH2 H CHMe2 0.78± 0.01 0.23± 0.01 16.2 > 20 ND14e PhCH2 H Bu 0.51± 0.01 0.11± 0.03 9.0 9.0 15014f PhCH2 H CH2Ph 0.96± 0.01 0.26± 0.01 7.9 8.4 14015a 4-MeOPhCH2 H Me ND ND > 20 > 20 ND15b 4-MeOPhCH2 H Et ND ND > 20 > 20 ND15c 4-MeOPhCH2 H Prop ND ND > 20 > 20 ND15d 4-MeOPhCH2 H CHMe2 0.72± 0.01 0.19± 0.01 15.0 > 20 ND15e 4-MeOPhCH2 H Bu ND ND > 20 > 20 ND15f 4-MeOPhCH2 H CH2Ph ND ND > 20 > 20 ND16a H H Me 0.34± 0.01 0.21± 0.02 2.1 8.2 14316b H H Et 0.52± 0.07 0.02± 0.01 0.8 8.4 14916c H H Prop 0.91± 0.60 0.12± 0.01 5.7 3.6 14916d H H CHMe2 0.43± 0.14 0.16± 0.01 2.6 8.0 14616e H H Bu 0.67± 0.09 0.72± 0.01f 5.0 3.8 14016f H H H 0.81± 0.03 0.01± 0.01 14.0 8.2 14017a Me Me Me 0.85± 0.03 0.19± 0.01 13.2 > 20 ND17b Me Me Et 0.57± 0.01 0.21± 0.01 5.3 > 20 ND17c Me Me Prop 0.93± 0.01 0.06± 0.02 19 > 20 ND17d Me Me CHMe2 ND ND > 20 14.0 170f

17e Me Me Bu ND ND > 20 4.1 167f

17f Me Me CH2Ph ND ND > 20 > 20 ND18a PhCH2 Me Me ND ND > 20 > 20 ND18b PhCH2 Me Et 0.97± 0.02 0.44± 0.02f 18.3 > 20 ND18c PhCH2 Me Prop ND ND > 20 ND ND18d PhCH2 Me CHMe2 ND ND > 20 ND ND18e PhCH2 Me Bu ND ND > 20 ND ND18f PhCH2 Me CH2Ph ND ND > 20 ND ND

a Reducing activity against 2,2-diphenyl-1-picrylhydrazyl with equimolar amounts of test compound.bDeoxyribose degradation as a measureof hydroxyl-radical formation. Indicated values areµmol of malondialdehyde/mmol of deoxyribose released by 75µM of test compound(controls< 0.1). cInhibition of LTB4 biosynthesis in bovine PMNL. Inhibition was significantly different with respect to that of the control;n = 3 or more,P < 0.01. Nordihydroguaiaretic acid was used as the standard inhibitor (IC50 = 0.4µM) [8]. dAntiproliferative activity againstHaCaT cells. Inhibition of cell growth was significantly different with respect to that of the control,n = 3, P < 0.01.eActivity of LDH (mU)release in HaCaT cells after treatment with 2µM test compound (n = 3, SD< 10%).fValues are significantly different with respect to vehiclecontrol (P < 0.05). ND = not determined.gPositive control [8].

1038

agent [15], LDH release by anthralin significantly ex-ceeded that of the vehicle control.

In conclusion, barbatic acid analogues were consis-tently more active against the biosynthesis of LTB4 andthe growth of HaCaT keratinocytes than the correspond-ing diffractaic acid analogues. Though this may be relatedto their additional phenolic hydroxyl group, determina-tion of the antioxidant and pro-oxidant potential of thecompounds did not reveal any appreciable redox activity.Barbatic acid analogue16bhas been identified as a potentnon-redox inhibitor of LTB4 biosynthesis which alsodisplays antiproliferative activity against keratinocytegrowth.

4. Experimental protocols

4.1. Chemistry

4.1.1. GeneralFor analytical instruments and methods see refer-

ence [16].Compounds 1–3 and 5–8 were prepared as de-

scribed [10].

4.1.2. Propyl 2,4-dihydroxy-3,6-dimethylbenzoate4bSodium (0.5 g, 21.73 mmol) was dissolved in absolute

propanol (50 mL) and stirred at room temperature. Then3 (3 g, 14.28 mmol) was added to the solution andrefluxed under nitrogen for 24 h. The solution wascooled, acidified with cold 10% HCl and extracted withether (3× 100 mL). The combined organic phase wasdried over MgSO4 and evaporated. The crude productwas purified by flash chromatography (SiO2, CH2Cl2) togive colourless crystals; FTIR 3 396, 2 980, 2 957,

1 630 cm–1; 1H-NMR (250 MHz, CDCl3) δ 12.14 (s, 1H),6.20 (s, 1H), 4.97 (s, 1H), 4.30 (t,J = 6.5 Hz, 2H), 2.48(s, 3H), 2.10 (s, 3H), 1.80 (m, 2H), 1.03 (t,J = 7.4 Hz,3H); Anal. (C12H16O4) C, H.

Analogously, compounds4a and c–e were preparedfrom 3 (table II).

4.1.3. Ethyl 2-hydroxy-4-(4-methoxybenzyloxy)-3,6-dimethylbenzoate9b

A suspension of3 (6 g, 28.57 mmol), anhydrous po-tassium carbonate (11.25 g, 81.39 mmol) and4-methoxybenzylchloride (4.47 g, 28.57 mmol) in dryacetone (75 mL) was refluxed for 24 h, then cooled,acidified with cold 10% HCl and extracted with ether (3× 200 mL). The combined organic phase was washedwith water, dried over MgSO4 and evaporated. The crudeproduct was purified by column chromatography (SiO2)using hexane/ethyl acetate (9:1) to afford colourlesscrystals; FTIR 3 438, 1 720, 1 636 cm–1; 1H-NMR(CDCl3) δ 11.91 (s, 1H), 7.35 (d,J = 9.5 Hz, 2H), 6.93 (d,J = 9.5 Hz, 2H), 6.34 (s, 1H), 5.03 (s, 2H), 4.38 (q,J =7.1 Hz, 2H), 3.82 (s, 3H), 2.52 (s, 3H), 2.11 (s, 3H), 1.41(t, J = 7.1 Hz, 3H). Anal. (C19H22O5) C, H.

4.1.4. Ethyl 2-methoxy-4-(4-methoxybenzyloxy)-3,6-dimethylbenzoate10b

A suspension of9b (2.20 g, 6.67 mmol), anhydrouspotassium carbonate (2.76 g, 20 mmol) and dimethylsulfate (0.6 mL, 6.67 mmol) in dry acetone (50 mL) wasrefluxed for 24 h, then cooled, acidified with cold 10%HCl and extracted with ether (3× 200 mL). The com-bined organic phase was washed with water, dried overMgSO4 and evaporated. The crude product was purifiedby column chromatography (SiO2) to afford a colourless

Table I. (continued).

kDPPHa DD (⋅OH)b LTB4

c AAd LDHe

Compound R1 R2 R3 (M–1 s–1) IC50 (µM) IC50 (µM) (mU)

19a 4-MeOPhCH2 Me Me ND ND > 20 > 20 ND19b 4-MeOPhCH2 Me Et ND ND > 20 > 20 ND19c 4-MeOPhCH2 Me Prop ND ND > 20 ND ND19d 4-MeOPhCH2 Me CHMe2 ND ND > 20 ND ND19e 4-MeOPhCH2 Me Bu ND ND > 20 ND ND19f 4-MeOPhCH2 Me CH2Ph ND ND > 20 ND ND20a H Me Me 0.93± 0.01 0.18± 0.01 7.8 9.0 13820b H Me Et 0.64± 0.01 0.10± 0.02 1.4 > 20 ND20c H Me Prop 0.89± 0.05 0.01± 0.01 8.5 > 20 ND20d H Me CHMe2 0.97± 0.01 0.09± 0.01 5.8 7.2 16820e H Me Bu 0.51± 0.01 0.16± 0.01 7.8 > 20 ND20f H Me H 0.24± 0.03 0.15± 0.01 11.0 9.8 114anthraling 24.2± 4.2f 2.89± 0.14f 37.0 0.7 294f

1039

Table II. Chemical data of starting materials, barbatic and diffractaic acid derivatives.

Compounda Formulab M.p. (°C) Yield (%) Solventc,d (vol%) Anal.e

4a C10H12O4 144; ref. [18] 145 84 MCc C, H4b C12H16O4 135; ref. [19] 139 93 MCc C, H4c C12H16O4 88; ref. [19, 20] 92 62 MCc C, H4d C13H18O4 120; ref. [19, 20] 123 88 MCc C, H4e C16H16O4 118; ref. [10] 113 55 MCc C, H9a C18H20O4 67; ref. [10] 67–68 65 H/EAc (9 + 1); Md C, H9b C19H22O5 80 85 H/EAc (9 + 1); Md C, H10a C19H22O4 oil; ref. [17] oil 91 H/EAc (9 + 1) C, H10b C20H24O5 oil 86 H/EAc (9 + 1) C, H11a C16H16O5 167; ref. [10] 165–167 84 H/EAc (1 + 1) C, H11b C17H18O5 198 92 H/EAc (1 + 1) C, H12a C17H18O4 120; ref. [17] 122 71 H/EAc (1 + 1) C, H12b C18H20O5 136 76 H/EAc (1 + 1) C, H13a C20H22O7 166; ref. [17] 170 77 H/EAc (9 + 1); M/Cd C, H13b C21H24O7 186; ref. [21] 189 71 H/EAc (9 + 1); M/Cd C, H13c C22H26O7 137; ref. [20] 138–139 89 H/EAc (9 + 1); M/Cd C, H13d C22H26O7 125; ref. [20] 128–129 64 H/EAc (9 + 1); M/Cd C, H13e C23H28O7 134; ref. [20] 133 90 H/EAc (9 + 1); M/Cd C, H13f C26H26O7 132; ref. [10] 136–138 69 H/EAc (9 + 1); M/Cd C, H14a C26H26O7 137; ref. [22] 133–134 63 H/EAc (9 + 1); M/Cd C, H14b C27H28O7 148 73 H/EAc (9 + 1); M/Cd C, H14c C28H30O7 108 73 H/EAc (9 + 1); M/Cd C, H14d C28H30O7 124 65 H/EAc (9 + 1); M/Cd C, H14e C29H32O7 114 61 H/EAc (9 + 1); M/Cd C, H14f C32H30O7 128; ref. [10] 131–132 78 H/EAc (9 + 1); M/Cd C, H15a C27H28O8 98 69 H/EAc (9 + 1); M/Cd C, H15b C28H30O8 107 73 H/EAc (9 + 1); M/Cd C, H15c C29H32O8 88 65 H/EAc (9 + 1); M/Cd C, H15d C29H32O8 97 68 H/EAc (9 + 1); M/Cd C, H15e C30H30O8 119 63 H/EAc (9 + 1); M/Cd C, H15f C33H32O8 103 71 H/EAc (9 + 1); M/Cd C, H16a C19H20O7 108; ref. [23] 108–112 87 H/EAd C, H16b C20H22O7 142 95 H/EAd C, H16c C21H24O7 146 90 H/EAd C, H16d C21H24O7 123 90 H/EAd C, H16e C22H26O7 153 89 H/EAd C, H16f C18H18O7 172; ref. [10] 136–138 91 H/EAd C, H17a C21H24O7 133; ref. [24] 127–128 69 H/EAc (9 + 1); M/Cd C, H17b C22H26O7 144; ref. [25] 141–144 73 H/EAc (9 + 1); M/Cd C, H17c C23H28O7 126; ref. [19] 127 82 H/EAc (9 + 1); M/Cd C, H17d C23H28O7 124; ref. [19] 127 66 H/EAc (9 + 1); M/Cd C, H17e C24H30O7 115; ref. [19] 115 88 H/EAc (9 + 1); M/Cd C, H17f C27H28O7 123; ref. [10] 119 83 H/EAc (9 + 1); M/Cd C, H18a C27H28O7 124 74 H/EAc (9 + 1); M/Cd C, H18b C28H30O7 124 67 H/EAc (9 + 1); M/Cd C, H18c C29H32O7 141 63 H/EAc (9 + 1); M/Cd C, H18d C29H32O7 113 74 H/EAc (9 + 1); M/Cd C, H18e C30H34O7 142 63 H/EAc (9 + 1); M/Cd C, H18f C33H32O7 118; ref. [17] 118–120 69 H/EAc (9 + 1); M/Cd C, H19a C28H30O8 130 82 H/EAc (9 + 1); M/Cd C, H19b C29H32O8 98 73 H/EAc (9 + 1); M/Cd C, H19c C30H34O8 122 62 H/EAc (9 + 1); M/Cd C, H19d C30H34O8 117 80 H/EAc (9 + 1); M/Cd C, H19e C31H32O8 111 77 H/EAc (9 + 1); M/Cd C, H19f C34H34O8 89 65 H/EAc (9 + 1); M/Cd C, H

1040

oil; FTIR 3 423, 1 700, 1 638 cm–1; 1H-NMR (CDCl3) δ7.32 (d,J = 9.5 Hz, 2H), 6.93 (d,J = 9.5 Hz, 2H), 6.53(s, 1H), 4.98 (s, 2H), 4.39 (q,J = 7.1 Hz, 2H), 3.82 (s,3H), 3.76 (s, 3H), 2.29 (s, 3H), 2.13 (s, 3H), 1.38 (t,J =7.1 Hz, 3H). Anal. (C20H24O5) C, H.

4.1.5. 2-Hydroxy-4-(4-methoxybenzyloxy)-3,6-dimethyl-benzoic acid11b

A solution of aqueous potassium hydroxide (2.86 g in7 mL H2O, 51.0 mmol) was added to a solution of9b(3.00 g, 8.82 mmol) in DMSO (40 mL) and heated on awater bath for 2.5 h (TLC control). The solution wascooled to room temperature, diluted with excess water(100 mL), acidified with cold 10% HCl, and extractedwith ether (3× 100 mL). The combined organic phasewas washed with water (3× 200 mL), dried over MgSO4and evaporated. The residue was purified by columnchromatography (SiO2) using hexane/ethyl acetate (1:1)to afford colourless crystals; FTIR 3 053, 1 700, cm–1;1H-NMR (CDCl3, DMSO-d6) δ 12.49 (s, 1H), 7.30(d, J = 9.5 Hz, 2H), 6.99 (d,J = 9.5 Hz, 2H), 6.32 (s, 1H),5.09 (s, 2H), 3.82 (s, 3H), 2.55 (s, 3H), 2.13 (s, 3H), Anal.(C17H18O5) C, H.

Analogously,12b was prepared from10b (table II).

4.1.6. General procedure for the condensation of benzoicacids with phenolic esters

4.1.6.1. Ethyl 4-(4-benzyloxy-2-hydroxy-3,6-dimethyl-benzoyloxy)-2-hydroxy-3,6-dimethylbenzoate14b

A solution of 11a (136 mg, 0.5 mmol) and3 [10](107 mg, 0.5 mmol) in anhydrous toluene (2 mL) andtrifluoroacetic anhydride (0.5 mL) was stirred at roomtemperature for 2.5 h (TLC control). The solvent wasremoved under reduced pressure and the residue waspurified by column chromatography (SiO2) using hexane/ethyl acetate (9:1). The product was recrystallized fromMeOH/CHCl3 to give colourless crystals; FTIR 3 430,

2 965, 2 900, 1 665, 1 618 cm–1; 1H-NMR (CDCl3) δ11.99 (s, 1H), 11.53 (s, 1H), 7.32–7.47 (m, 5H), 6.51 (s,1H), 6.44 (s, 1H), 5.17 (s, 2H), 4.44 (q,J = 7.1 Hz, 2H),2.66 (s, 3H), 2.55 (s, 3H), 2.17 (s, 3H), 2.08 (s, 3H), 1.43(t, J = 7.13 Hz, 3H); Anal. (C27H28O7) C, H.

Analogously,13a–f were prepared from7 [10] and 3and4a–c; 14a andc–f were prepared from11a[10] and4a–c; 15a–f were prepared from11b and 3 and 4a–c;17a–f were prepared from8 [10] and3 and4a–c; 18a–fwere prepared from12a[17] and3 and4a–c; 19a–f wereprepared from12b and3 and4a–c (table II).

4.1.7. General procedure for hydrogenolysis

4.1.7.1. Ethyl 4-(2,4-dihydroxy-3,6-dimethylbenzoyloxy)-2-hydroxy-3,6-dimethylbenzoate16b

A suspension of14b (112 mg, 0.25 mmol) and 10%palladium/carbon (25 mg) in dry ethyl acetate (2 mL) wasstirred in H2 for 2 h (TLC control). The suspension wasthen filtered through celite, and the filtrate was evapo-rated. The residue was purified by column chromatogra-phy (SiO2) using hexane/ethyl acetate (9:1) to givecolourless crystals; FTIR 3 456, 2 943, 1 665, cm–1;1H-NMR (CDCl3) δ 12.00 (s, 1H), 11.71 (s, 1H), 6.50 (s,1H), 6.30 (s, 1H), 5.29 (s, 1H), 4.43 (q,J = 7.1 Hz, 2H),2.61 (s, 3H), 2.54 (s, 3H), 2.12 (s, 3H), 2.08 (s, 3H), 1.43(t, J = 7.1 Hz, 3H); Anal. (C20H22O7) C, H.

Analogously,16aandc–f were prepared from14aandc–f; 20a–f were prepared from18a–f; 1 and 2 wereprepared from17f and13f, respectively (table II).

4.2. Biological assay methods

The procedures for the biological assays presented intable I were described previously in full detail: determin-ation of the reducing activity against 2,2-diphenyl-1-picrylhydrazyl [8], deoxyribose degradation [8], inhibi-

Table II. (continued).

Compounda Formulab M.p. (°C) Yield (%) Solventc,d (vol%) Anal.e

20a C20H22O7 161 98 H/EAd C, H20b C21H24O7 170 93 H/EAd C, H20c C22H26O7 143 97 H/EAd C, H20d C22H26O7 124 94 H/EAd C, H20e C23H28O7 112 91 H/EAd C, H20f C19H20O7 203; ref. [17] 207–209 95 H/EAd C, H

aAll compounds were obtained as colourless crystals except where stated otherwise.bAll new compounds displayed1H-NMR and FTIRconsistent with the assigned structure.cEluant used for column chromatography.dSolvent for recrystallization; C = chloroform; EA = ethylacetate; H = hexane; M = methanol; MC = methylene chloride.eElemental analyses were within± 0.4% of calculated values except wherestated otherwise.

1041

tion of LTB4 biosynthesis [8], inhibition of HaCaT cellproliferation [9], and release of LDH into culture me-dium [14].

Acknowledgements

We thank Mr K. Ziereis for his excellent technicalassistance. S.K. KC thanks the German Academic Ex-change Service for a scholarship.

References

[1] Culberson C.F., Chemical and Botanical Guide to Lichen Products,The University of North Carolina Press, Chapel Hill, 1969.

[2] Neamati N., Hong H., Mazumder A., Wang S., Sunder S., NicklausM.C., Milne G.W.A., Proksa B., Pommier Y., J. Med. Chem. 40(1997) 942–951.

[3] Yamamoto Y., Miura Y., Kinoshita Y., Higuchi M., Yamada Y.,Murakami A., Ohigashi H., Koshimizu K., Chem. Pharm. Bull. 43(1995) 1388–1390.

[4] Okuyama E., Umeyama K., Yamazaki M., Kinoshita Y., YamamotoY., Planta Med. 61 (1995) 113–115.

[5] Kumar K.C.S., Müller K., J. Nat. Prod. 62 (1999) in press.

[6] Brooks C.D.W., Summers J.B., J. Med. Chem. 39 (1996)2629–2654.

[7] Kumar K.C.S., Müller K., J. Nat. Prod. 62 (1999) in press.

[8] Müller K., Gürster D., Piwek S., Wiegrebe W., J. Med. Chem. 36(1993) 4099–4107.

[9] Müller K., Leukel P., Ziereis K., Gawlik I., J. Med. Chem. 37 (1994)1660–1669.

[10] Elix J.A., Norfolk S., Aust. J. Chem. 28 (1975) 1113–1124.

[11] Fitzsimmons B.J., Rokach J., in: Rokach J. (Ed.), Leukotrienes andLipoxygenases, Elsevier, Amsterdam, 1989, pp. 427–502.

[12] Halliwell B., Grootveld M., Gutteridge J.M.C., Methods Biochem.Anal. 33 (1988) 59–90.

[13] Boukamp P., Petrussevska R.T., Breitkreutz D., Hornung J.,Markham A., Fusenig N.E., J. Cell Biol. 761 (1988) 761–771.

[14] Müller K., Huang H.S., Wiegrebe W., J. Med. Chem. 39 (1996)3132–3138.

[15] Müller K., Biochem. Pharmacol. 53 (1997) 1215–1221.

[16] Müller K., Reindl H., Gawlik I., Eur. J. Med. Chem. 33 (1998)969–973.

[17] Elix J.A., Chester D.O., Gaul K.L., Parker J.L., Wardlaw J.H., Aust.J. Chem. 42 (1989) 1191–1199.

[18] Whalley W.B., J. Chem. Soc. (1949) 3278–3280.

[19] Fujikawa, Yakugaku Zasshi 60 (1940) 473–478, cited from BeilsteinCommander.

[20] Fujikawa, Shimamura, Tarui, Yakugaku Zasshi 61 (1941) 491–495,cited from Beilstein Commander.

[21] Robertson A., Stephenson R.J., J. Chem. Soc. (1932) 1675–1681.

[22] Elix J.A., Mahadevan I., Wardlaw J.H., Arvidsson L., JørgensenP.M., Aust. J. Chem. 40 (1987) 1581–1590.

[23] Yamamoto Y., Nishimura K.I., Kiriyama N., Chem. Pharm. Bull. 24(1976) 1853–1859.

[24] Asahina Y., Fuzikawa F., Ber. Dtsch. Chem. Ges. 65 (1932)583–584.

[25] Asahina Y., Hashimoto A., Ber. Dtsch. Chem. Ges. 66 (1933)641–649.

1042

Original article

Preparation and local anaesthetic activity of benzotriazinoneand benzoyltriazole derivatives

Giuseppe Caliendoa*, Ferdinando Fiorinoa, Paolo Griecoa, Elisa Perissuttia, Vincenzo Santagadaa,Rosaria Melib, Giuseppina Mattace Rasob, Angelina Zanescoc, Gilberto De Nuccic

aDipartimento di Chimica Farmaceutica e Tossicologica, Università degli Studi di Napoli “Federico II” Via D. Montesano 49,80131 Napoli, Italy

bDipartimento di Farmacologia Sperimentale, Università degli Studi di Napoli “Federico II” Via D. Montesano 49, 80131 Napoli, ItalycCartesius Analytical Unit, Department of Pharmacology, Biomedical Sciences Institute, University of Sao Paulo, Sao Paulo, Brazil

(Received 19 April 1999; revised 6 July 1999; accepted 15 July 1999)

Abstract – Two sets of benzotriazinone and benzoyltriazole derivatives were prepared and tested for local anaesthetic activity in comparisonwith lidocaine. Several of the prepared compounds exhibited a fairly good activity comparable or superior to that of lidocaine. The presenceof a benzotriazinone or a benzoyltriazole moiety as an aromatic system was quite profitable for both the intensity and duration of activity. Theacute toxicity in mice of the four most potent compounds of the series was also assessed. Compound1b, which has an anaesthetic activitycomparable to that of lidocaine, was also characterized by a more favourable therapeutic index. All compounds were tested in vitro to evaluatetheir negative chronotropic action in isolated rat right atria. © 1999 Éditions scientifiques et médicales Elsevier SAS

benzotriazinone / benzoyltriazole / local anaesthetic / negative chronotropic action

1. Introduction

In previous articles we described the synthesis of twosets of N-[2-(alkylamino)ethyl]benzotriazol-X-yl-acet-amides and of N-[2-(alkylamino)ethyl]benzotriazol-X-yl-isobutyramides designed as local anaesthetic agents [1,2]. The compounds of two sets fulfill the requirements ofthe pharmacofore scheme proposed by Löfgren [3] be-cause they are provided with an aromatic system, anintermediate chain and an aminic moiety ionizable underphysiological pH. They have been assayed in vivo fortheir local anaesthetic activity. Some of the investigatedcompounds showed anaesthetic activities comparablewith or higher than those exhibited by the reference druglidocaine.

Considering these results and the important role playedby the aromatic system in the interaction with a corre-sponding hydrophobic region, in order to increase the

activity and to evaluate the influence on activity by thedifferent aromatic systems, the benzotriazole nucleus wasreplaced by a 1,2,3-benzotriazin-4(3H)-one and by a4-benzoyl-1,2,3-triazole ring, thereby achieving generalstructures1, 2 and3 (figure 1).

The N-alkylamino groups of the intermediate acetyl-acetamido side chain are those previously reported [1, 2]displaying the highest anaesthetic activity: dimethyl-amino, diethylamino, pyrrolidine and piperidine substitu-ents.

Herein we report the synthesis of a series of 1,2,3-benzotriazinones (1a–d) and of 4-benzoyl-1,2,3-triazole(2a–d and 3a–d) derivatives (tables I and II ). The syn-thesized compounds were first tested as local anaestheticswith different in vivo assays and, successively consider-ing the relationships between local anaesthetic (Na+

channel block) and antiarrhythmic activities [4, 5], allcompounds were preliminary tested to evaluate theirnegative chronotropic action in rat isolated right atria.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1043−1051 1043© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

Compounds of general formula1 reported infigure 1were synthesized as illustrated infigure 2. They derivedfrom condensation of a 1,2,3-benzotriazin-4(3H)-one ring(4) with ethyl bromoacetate in butan-2-one in the pres-ence of potassium carbonate. The ethyl acetate deriva-

tives (5) were converted into the amide derivatives (1) byreaction in methyl alcohol solution with the appropriateamines.

The final compounds1a–d, reported intable I, werepurified by cromatography on a silica gel column andfurther by crystallization from an appropriate solvent(yields ranging between 52–66%).

Table I. Physicochemical properties of 1,2,3-benzotriazinone derivatives1a–d.

R Formula* MW Compound M.p. Cryst.** Yield

(°C) Solvent (%)

C13H17N5O2 275.31 1a 145–146 a + b 60

C15H21N5O2 303.37 1b 155–156 a 55

C15H19N5O2 301.35 1c 180–181 a + b 66

C16H21N5O2 315.38 1d 169–170 a 52

*Satisfactory microanalyses obtained: C, H, N values are within± 0.4% of theoretical values. **Crystallization solvents: a) diethyl ether; b)ethyl alcohol.

Figure 1. General structure of considered compounds.

1044

The 4-benzoyl-1,2,3-triazole derivatives (2a–d and3a–d) were synthesized following the steps outlined infigure 3.

The parent compound, 4-benzoyl-1,2,3-triazole (8),was obtained in 95% yield by a modified method alreadydescribed [6, 7]. Oxidation of the phenylethynylcarbinol

(6) with chromium trioxide and concentrated sulphuricacid produced phenyl ethynyl ketone (7) which wassuccessively treated with NaN3 in anhydrous dimethyl-acetamide providing the expected compound. Reaction of4-benzoyl-1,2,3-triazole with ethyl bromoacetate and po-tassium carbonate in butan-2-one gave a mixture of thetriazole isomers 1- and 2- with an overall yield of80% [8]. The two isomers were separated by columnchromatography on silica gel using diethyl ether/n-hexane (7:3 v/v) as eluent. The faster moving 2- isomer,ethyl 4-(benzoyl)-1,2,3-triazole-2-acetate (10), was col-lected with a higher yield (65%) with respect to theslower moving 1-substituted isomer (9) (35%). The ethylacetate derivatives (9 and 10) were converted into thecorresponding amide derivatives2 and 3 by reaction inmethyl alcohol solution with the appropriate amines.

All the final products (2a–d and 3a–d) reported intable II were further purified by crystallization from amixture of diethyl ether and methyl alcohol (yieldsranging between 40–70%).

The synthesized compounds listed intables I and IIwere characterized by1H-NMR spectroscopy whose datawere fully consistent with the described structures.

Table II. Physicochemical properties of 4-benzoyl-1,2,3-triazole derivatives2a–d and3a–d.

1-substituted 4-benzoyl-1,2,3-triazoles 2-substituted 4-benzoyl-1,2,3-triazoles

R Formula* MW Compound** M.p. (°C) Yield (%) Compound** M.p. (°C) Yield (%)

C15H19N5O2 301.35 2a 132–133 70 3a 96–97 60

C17H23N5O2 329.40 2b 136–137 40 3b 99–100 45

C17H21N5O2 327.39 2c 151–153 50 3c 98–99 55

C18H23N5O2 341.41 2d 159–161 40 3d 94–95 50

*Satisfactory microanalyses obtained: C, H, N values are within± 0.4% of theoretical values. **All compounds were crystallized by methylalchol/diethyl ether.

Figure 2. Synthesis of1a–d.

1045

1H-NMR of compounds2a–d and 3a–d (CDCl3)differentiated clearly between 1- and 2- substituted4-benzoyl-1,2,3-triazole derivatives. In fact there is adifference in chemical shift values among the protons inthe 5- position of the benzoyltriazole ring in the series of1- and 2- isomers. The triazole proton of the 1- isomerappears always as a singlet at lower field with respect tothe position of the same proton of analogues 2-substituted(chemical shift values ranging between 8.45–8.70δ and8.20–8.35δ, respectively) according with literature [8, 9].

Physicochemical data of all the final compounds aresummarized intables Iand II .

3. Pharmacology

The compounds reported intables IandII were testedin vivo for their anaesthetic activity by different tests:corneal anaesthesia in the rabbit, mouse tail anaesthesia(table III) and rat sciatic nerve block anaesthesia (ta-

ble IV). The ip acute toxicity and the therapeutic index ofthe more active compounds were also determined (ta-ble V). Successively the synthesized compounds weretested in vitro to evaluate their negative chronotropicaction in rat isolated right atria (table VI). The synthe-sized compounds were always compared for their activitywith lidocaine, taken as reference drug.

4. Results and discussion

Table III summarizes the results of the surface andinfiltration anaesthesia assays performed on the 1,2,3-benzotriazinone (1a–d) and on the 4-benzoyl-1,2,3-triazole (2a–d and 3a–d) derivatives. As for the 1,2,3-benzotriazinone derivatives (1a–d), in both tests, theresults indicate that the structures endowed with highestactivity are 1b and 1d, with values similar to thatmeasured for lidocaine.

Figure 3. Synthesis of 4-benzoyl-1,2,3-triazole de-rivatives.

1046

As for the 4-benzoyl-1,2,3-triazole derivatives (2a–dand 3a–d), the results indicate that the nature and theposition of the intermediate acetylacetamido side chainon the benzoyltriazole nucleus have significant influenceon activity. In both tests the 2-substituted isomers appearto be more active than the 1-substituted derivatives.Compounds3b and3d stand out in the data set for theirhigh activities (these values are comparable to thosedetermined for lidocaine). It appears that the position and

the nature of the side chain affect surface and infiltrationactivities in a similar way.

For all considered compounds, with respect to thenature of the intermediate acetylacetamido side chain,analogues with a terminal diethylamino group or apiperidine ring displayed the highest activity in accor-dance with previous results [1, 2]. Reducing the size froma six to a five-membered ring resulted in an improvementof anaesthetic activity.

In order to evaluate the duration of the local anaes-thetic activity, additional investigations were conductedby rat sciatic nerve block assay. In accordance with theprevious results, when a 2% solution was used, asreported intable IV,several compounds (1b, 1d, 3a, 3b,and3d) exhibited a better performance in blocking the ratsciatic nerve with respect to lidocaine (160, 160, 150, 210and 180 min for motor activity recovering, respectively,versus 117 min for lidocaine).

Finally, the acute toxicity and therapeutic index of themore active compounds in all anaesthetic tests (1b, 1d,

Table III. Rabbit corneal and mouse tail anaesthetic activities.

Compound Corneal anaesthetica Mouse tail anaestheticb

1a inactive 5.4 (± 0.38)10–2

1b 74.3± 9.3 0.69 (± 0.31)10–2

1c 12.1± 3.9 2.6 (± 0.43)10–2

1d 73.8± 8.9 1.2 (± 0.42)10–2

2a inactive 4.0 (± 0.36)10–2

2b 4.7± 1.5 3.0 (± 0.37)10–2

2c 3.2± 2.1 3.0 (± 0.33)10–2

2d 30.1± 5.6 3.6 (± 0.35)10–2

3a 22.5± 4.5 3.0 (± 0.29)10–2

3b 105.0± 3.5 1.4 (± 0.25)10–2

3c 48.5± 9.1 3.0 (± 0.31)10–2

3d 92.3± 8.0 1.4 (± 0.28)10–2

lidocaine HClc 100 0.68 (± 0.39)10–2

aAll compounds were in aqueous solution at 2% concentration. Thevalues expressed as % of the anaesthetic activity of lidocaine (=100), are means± SE of three determinations.bIC50 valuesexpressed as mol/L.cLidocaine hydrochloride was used for com-parison.

Table IV. Duration of local anaesthetic activity in rat sciatic nerveblock.

Compound Durationa (min)

1% 2%

1a 25 (± 10.5) 45 (± 17.0)1b 77 (± 6.8) 160 (± 18.2)1c inactive n.t.1d 90 (± 3.6) 160 (± 15.3)2a 40 (± 15.0) 72 (± 6.9)2b 35 (± 11.0) 63 (± 6.8)2c inactive n.t.2d 40 (± 12.0) 72 (± 7.0)3a 75 (± 7.0) 150 (± 17.2)3b 117 (± 8.0) 210 (± 16.9)3c 25 (± 12.3) 50 (± 6.5)3d 70 (± 6.6) 180 (± 20.3)

lidocaine HCl 65 (± 10.7) 117 (± 11.0)

aIn vivo duration of local anaesthetic activity in rat sciatic nerveblock (each rat received 0.2 mL of 1% and 2% anaesthetic solution,n.t. = not tested. The values are means± SE of three determina-tions.

Table V. Acute toxicity in mouse (LD50) and therapeutic index ofselected compounds1b, 1d, 3b and3d.

Com-pound

LD50a IC50 Therapeutic

indexb

LD50/IC50

1b 6.09 (± 0.49)10–2 M 0.69 (± 0.31)10–2 M 8.831d 5.24 (± 0.38)10–2 M 1.2 (± 0.42)10–2 M 4.373b 8.69 (± 0.53)10–2 M 1.4 (± 0.25)10–2 M 6.213d 4.46 (± 0.35)10–2 M 1.4 (± 0.28)10–2 M 3.19

lidocaineHCl

5.60 (± 0.39)10–2 M 0.68 (± 0.39)10–2 M 8.23

aMolar concentration of the injected solution (v = 0.2 mL/20 gbody weight).bEvaluated as ratio between LD50 and IC50 (mousetail test) expressed in mg/kg.

Table VI. Maximum response to different compounds in rat iso-lated right atria compared to lidocaine.

Compound Maximum response(beats/min)

n

1a –15± 6 61b –42± 8 61c –52± 8 61d –106± 27 52a –33± 13 62b –55± 12 62c –72± 5 62d –107± 27 63b –67± 9 63c –135± 31 63d –152± 12 6

lidocaine –180± 35 6

1047

3b, and 3d) were determined in mice. From the corre-sponding LD50 and LD50/IC50 values shown intable Vitis evident that1b is characterized by the most favourableratio between toxicity and mouse tail anaesthetic activity(its therapeutic index is slightly higher than that measuredfor lidocaine).

These studies showed that the presence of a piperidinering as a terminal N-alkylamino group gives an increasein toxicity, particularly evident on compound3d.

In conclusion, a comparison between the pharmaco-logical profile of benzotriazinone and benzoyltriazolederivatives versus that of the above mentioned benzotria-zole counterpartes [1, 2] clearly shows that the replace-ment of the benzotriazole moiety, found in the preceedingderivatives, by a benzotriazinone or by a benzoyltriazolering, seems to have a little influence on the anaestheticactivity. These results might be ascribed to an analoguelipophilic, steric, and electronic complementarity be-tween the benzotriazole residue compared to the benzo-triazinone or to the benzoyltriazole moiety.

As regards the results of chronotropic activity reportedin table VI, all compounds induced a negative chronotro-pic effect in rat isolated right atria at high doses, in aconcentration range greater thanµM. The intrinsic activ-ity of different compounds in decreasing spontaneousbeating rate in isolated right atria was distinct. Lidocainewas used as the pattern drug to compare the intrinsicactivity of all compounds; in fact it is well known thatlidocaine has local anaesthetic as well antiarrhythmicactions

In series1, compound1d determines a larger magni-tude of maximum response compared to drugs1b, 1c, and1a, whereas drug1a was less effective in producing anegative chronotropic response among these compounds(figure 4 and table VI). In series2, compound2d hadhigher intrinsic activity compared to compounds2a, 2b,and2c (figure 5). The intrinsic activity of compounds2aand2b were similar.

The intrinsic activity of compound3d was the best ofall studied compounds. Compound3c showed similar3dintrinsic activity even if less, whereas compound3binduced the lowest activity of the series (figure 6 andtable VI).

A comparison between ‘in vivo’ anaesthetic activityand ‘in vitro’ negative chronotropic activity data dis-played a similar pattern. In fact, clear negative chrono-tropic action was displayed by compound1d and 3d,which bear a piperidine ring, but also compound2d,which was poorly active on anaesthetic activity, showedan appreciable negative chronotropic action.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Kofler hot stageapparatus and are uncorrected. Structures described weresupported by1H-NMR spectra and microanalytical data.1H-NMR spectra were recorded on a Bruker WM 500spectrometer using DMSO and CDCl3 as solvent; chemi-cal shifts (δ) are reported as follows: s, singlet; d,

Figure 4. Negative chronotropic response for the compounds1a, 1b, 1c, 1d, and lidocaine in isolated right atria from rats.Data are means± SEM for 5–6 experiments.

Figure 5. Negative chronotropic response for the compounds2a, 2b, 2c, 2d, and lidocaine in isolated right atria from rats.Data are means± SEM for 6 experiments.

1048

doublet; t, triplet; m, multiplet. The spectra obtainedconfirmed the proposed structures. All pure compoundsgave a satisfactory analysis (C, H, N) within± 0.4% oftheoretical values.

Analytical TLC was performed on precoated silica-gel(0.2 mm GF 254, E Merck); the spots were located by UV(254 nm) light. Evaporation was performed in vacuo.Anhydrous sodium sulfate was used as a drying agent.Crude products were routinely passed through columns ofsilica gel (0.05–0.20 mm, Carlo Erba). Analytical data,melting points and crystallization solvents are reported intables Iand II.

5.1.1. 3-Carbethoxymethyl-1,2,3-benzotriazin-4(3H)-one(5)

To a magnetically stirred solution of 1,2,3-benzotriazin-4(3H)-one (4) (0.01 mol) and ethyl bromo-acetate (0.01 mol) in butan-2-one (50 mL) was addedK2CO3 (0.01 mol). The reaction mixture was heatedunder reflux for 10 h and monitored by TLC. Aftercooling, the butan-2-one was removed under reducedpressure and the residue was diluted with H2O andextracted with CHCl3. The organic layer was dried andevaporated to dryness. The crude product5 was purifiedby crystallization from methyl alcohol/diethyl ether.Yield 68%, m.p. 114–116 °C.1H-NMR (CDCl3) δ = 8.37(d, 1H, Ar-H,J = 7.5 Hz), 8.20 (d, 1H, Ar-H,J = 7.5 Hz),7.99 (t, 1H, Ar-H,J = 7.5 Hz), 7.82 (t, 1H, Ar-H), 5.19 (s,2H, -NCH2), 4.28 (m, 2H, OCH2), and 1.30 ppm (t, 3H,CH3).

5.1.2. General procedure for the preparation ofcompounds1a–d

To 0.01 mol of 3-carbethoxymethyl-1,2,3-benzo-triazin-4(3H)-one (5) dissolved in anhydrous methanol(50 mL) was added the appropriate amine (0.01 mol)dropwise. The reaction mixture was kept under refluxwith magnetic stirring for 8–12 h and monitored by TLCuntil the starting material had disappeared. After coolingthe solvent was removed under reduced pressure and theresidue was purified by silica-gel column chromatogra-phy using mixtures of diethyl ether/methanol 9:1 v/v.Further purification was obtained by crystallization fromthe appropriate solvent (table I).

Yields ranging between 52–66%. Spectral data of titlecompound1a: 1H-NMR (CDCl3), δ = 8.33 (d, 1H, Ar-H,J = 7.5 Hz), 8.15 (d, 1H, Ar-H,J = 7.5 Hz), 7.94 (t, 1H,Ar-H, J = 7.5 Hz), 7.78 (t, 1H, Ar-H,J = 7.5 Hz), 5.11 (s,2H, CH2-CO), 3.39 (t, 2H, NH-CH2, J = 7.5 Hz), 2.47 (t,2H, CH2-N, J = 7.5 Hz) and 2.24 ppm (s, 6H, N(CH3)2).Similar 1H-NMR data occur in all derivatives of generalformula 1.

5.1.3. Phenyl ethynyl ketone(7)A solution of chromium trioxide (0.10 mol) in water

(30 mL) and concentrated sulphuric acid (8.5 mL) wasslowly added to a stirred and cooled solution of phenyl-ethynylcarbinol (6) (0.15 mol) in acetone (50 mL). Thereaction mixture was carried out at 4 °C under a nitrogenatmosphere. After stirring for 7 h, water was added todissolve the precipitated chromium salts and the productwas extracted with chloroform. Evaporation of the or-ganic solution gave a yellow solid which was crystallizedfrom aqueous methanol to give 16.6 g (85%) of7 as paleyellow needles. The physical data are in agreement withthose given in ref. [6].

5.1.4. 4-benzoyl-1H-1,2,3-triazole(8)To a stirred and heated solution of NaN3 (0.10 mol) in

anhydrous dimethylacetamide (80 mL), phenyl ethynylketone7 (0.10 mol) dissolved in anhydrous dimethylacet-amide (80 mL) was slowly added. The reaction mixturewas kept at 100 °C for 2 h. After stirring for a further 12 hat room temperature, evaporation of the solvent underreduced pressure gave a liquid residue which was dilutedwith water. The aqueous layer was acidified (pH = 5) with10% HCl and extracted with ether (3× 200 mL). Thecombined organic layers were dried and evaporated togive a solid residue which was purified by crystallizationfrom ethyl alcohol: 16.4 g (95%) of8. The physical dataare in agreement with those given in ref. [7].

Figure 6. Negative chronotropic response for the compounds3b, 3c, 3d, and lidocaine in isolated right atria from rats. Dataare means± SEM for 6 experiments.

1049

5.1.5. 1-carbethoxymethyl-4-benzoyl-1H-1,2,3-triazole(9) and 2-carbethoxymethyl-4-benzoyl-2H-1,2,3-triazole(10)

Anydrous K2CO3 (0.04 mol) was added to a solutionof4-benzoyl-1,2,3-triazole (8) (0.04 mol)andethylbromo-acetate (0.04 mol) in 50 mL of butan-2-one. The mixturewas refluxed for 12 h and monitored by TLC. Aftercooling, the butan-2-one was removed under reducedpressure and the residue was diluted with H2O andextracted with CHCl3. The organic layer was washedwith 2 M NaOH, dried and evaporated to dryness.

The obtained residue, containing 1- and 2-substitutedisomers was finally fractionated by column chromatogra-phy using diethyl ether/n-hexane, 7:3 v/v, as eluent.Further purification of the isolated 1- and 2- isomers bycrystallization from diethyl ether gave the final products9 and 10. Characterization by1H-NMR spectra showedthat the first compound to be eluted was the 2-substituted1,2,3-triazole (compound10, yield 65%, m.p. 55–57 °C),whereas the 1-substituted isomer was eluted successively(compound9, yield 35%, m.p. 132–133 °C).9: 1H-NMR(DMSO), δ = 9. (s, 1H, H-triaz), 8.50 (d, 2H, Ar-H,J =7.5 Hz), 7.68 (t, 1H, Ar-H,J = 7.5 Hz), 7.55 (t, 2H, Ar-HJ = 7.5 Hz), 5.53 (s, 2H, CH2-CO), 4.27 (q, 2H, CH2O,J = 7.5Hz) and 1.28 (t, 2H, CH3, J = 7.5 Hz). 10: 1H-NMR (DMSO),δ = 8 (s, 1H, H-triaz), 8.58 (d, 2H, Ar-H,J = 7.5 Hz), 7.70 (t, 1H, Ar-H,J = 7.5 Hz), 7.58 (t, 2H,Ar-H, J = 7.5 Hz), 5.61 (s, 2H, CH2-CO), 4.26 (q, 2H,CH2O, J = 7.5 Hz) and 1.25 (t, 2H, CH3, J = 7.5 Hz).

5.1.6. General procedure for the preparation ofcompounds2a–dand3a–d

To 0.01 mol of the appropriate ethyl-4-benzoyl-1,2,3-triazolacetate derivative (9 or 10) dissolved in anhydrousmethanol was added dropwise the appropriate amine(0.01 mol). The reaction mixture was kept under refluxwith magnetic stirring for 8–12h and monitored by TLC,until the starting material had disappeared. After cooling,the solvent was removed by filtration under reducedpressure and the residue was purified by silica gel columnchromatography using methanol as eluent. Further puri-fication was obtained by crystallization from a mixture ofmethyl alcohol/diethyl ether. Yields ranging between 40and 70%. Spectral data of the title compound2a: 1H-NMR (CDCl3), δ = 8 (s, 1H, H-triaz), 8.27 (d, 2H, Ar-H,J = 7.5 Hz), 7.62 (t, 1H, Ar-H,J = 7.5 Hz), 7.51 (t, 2H,Ar-H, J = 7.5 Hz), 5.14 (s, 2H, CH2-CO), 3.35 (t, 2H,NH-CH2, J = 7.5 Hz), 2.40 (t, 2H, CH2-N(CH3)2, J = 7.5Hz) and 2.17 ppm (s, 6H, N(CH3)2). Similar 1H-NMRdata occur in all derivatives of the general formula2.Spectral data of the title compound3a: 1H-NMR(CDCl3), δ = 8.31 (s, 1H, H-triaz), 8.27 (d, 2H, Ar-H,

J = 7.5 Hz), 7.62 (t, 1H, Ar-H,J = 7.5 Hz), 7.52 (t, 2H,Ar-H, J = 7.5 Hz), 5.28 (s, 2H, CH2-CO), 3.33 (t, 2H,NH-CH2, J = 7.5 Hz), 2.35 (t, 2H, CH2-N(CH3)2, J = 7.5Hz) and 2.15 ppm (s, 6H, N(CH3)2). Similar 1H-NMRdata occur in all derivatives of the general formula3.

5.2. Pharmacology

5.2.1. Corneal anaesthesiaLocal anaesthetic activity was evaluated in male New

Zealand rabbits (Harlan-Nossan, Correzzana, Milan,weighing 2.4–2.8 kg) as local surface anaesthesia [10],by determining every 3 min the number of stimuli to thecornea, effected rhythmically with a Frey’s horse-hair,necessary to produce the blink reflex. If the reflex did notoccurr after 100 stimulations, anaesthesia was consideredtotal. At the beginning of the experiment care was takento ascertain that this reflex was normal in both eyes of therabbits. All compounds were dissolved in 0.1 N HCl andthe solution buffered to pH 6–7. The aqueous solutions(2%) of the compounds studied were dropped onto theconjunctival sac so that the space between the eyelidscontained a clearly visible film of solution for the set timeof 3 min. Lidocaine solution (2%) was used for compari-son.

5.2.2. Mouse tail anaesthesiaMale Swiss mice (Harlan-Nossan, Correzzana, Milan,

weighing 18–20 g) were used. The test was performedaccording to the method of Bianchi [11] in which theaqueous anaesthetic solution (0.1 mL) is injected sub-cutaneously about 1 cm from the base of the tail. Fifteenminutes after injection, the pain reflex of all injectedanimals was tested by applying a small artery clip to thezone where the compound was injected. The proportionof animals which did not show the usual pain reflexwithin 30 s was noted for each dose. Lidocaine solutionswere used for comparison. IC50 values were calculatedfor each compound by probit analysis using a computerprogram [12].

5.2.3. Rat sciatic nerve blockThis test was performed according to Al-Saadi and

Sneider [13] to determine conduction anaesthesia and itsduration. Triplicate sets of three groups of three maleWistar rats (Harlan-Nossan, Correzzana, Milan, weighing180–200 g) were used. Each rat received an injection(0.2 mL) of the aqueous anaesthetic solution (1% and2%) into the posterior side of the femur head. A positiveeffect of the drug resulted in a complete loss of motorcontrol of the injected limb. In order to assess theduration of the effect, the animals were observed from the

1050

time of onset of the motor paralysis, at 10 min intervalsover time, up to the first sign of motor activity.

5.2.4. Acute toxicityThe ip acute toxicity of the most active compounds1b,

1d, 3b, and 3d was determined in male Swiss mice(Harlan-Nossan, Correzzana, Milan, weighing 18–20 g)7 days after treatment. LD50 values were calculated foreach compound by probit analysis using a computerprogram [12].

5.2.5. Chronotropic activity: functional assay usingisolated rat right atria

This test was performed according to Hughes andSmith [14]. Wistar rats (250–300 g) were anesthetizedwith halothane and the hearts were rapidly removed andplaced in oxygenated Krebs Henseleit buffer (KHB). Theright atria were removed and mounted in a water jacketedtissue chamber (20 mL volume) containing KHB, pH7.3–7.5, at 37 °C and gassed with 95% O2/5% CO2. Thecomposition of the KHB was (mM): NaCl, 124; KCl,4.75; MgSO4, 1.30; CaCl2, 2.25; NaHCO3, 25.0;NaH2PO4, 0.6; dextrose, 10.0; sodium ascorbate, 0.3.

5.2.6. Construction and analyses of concentration-response curves

Concentration-response curves for all compounds wereconstructed by the cumulative variation of agonist con-centration at one-half log unit increments [15]. Allconcentration-response data were evaluated for a fit to alogistic function in the form:

E = Emax/((1 + (10c/10x)n) + U),

where E is the increase in rate above basal; Emax is themaximum response that the agonist can produce; c is thelogarithm of the EC50, the concentration of agonist thatproduces half-maximal response; x is the logarithm of theconcentration of agonist; the exponential term, n, is acurve fitting parameter that defines the slope of theconcentration-response line, andΦ is the response ob-

served in the absence of added agonist. Nonlinear regres-sion analyses to determine the parameters Emax, log EC50

and n were done using GraphPad Prism (GraphPadSoftware, San Diego, CA) with the constraint thatΦ = 0.

Acknowledgements

The NMR spectral data were provided by Centro diRicerca Interdipartimentale di Analisi Strumentale, Uni-versità degli Studi di Napoli “Federico II”. The assistanceof staff is gratefully appreciated.

References

[1] Caliendo G., Di Carlo R., Greco G., Grieco P., Meli R., NovellinoE., Perissutti E., Santagada V., Eur. J. Med. Chem. 30 (1995)603–608.

[2] Caliendo G., Greco G., Grieco P., Mattace Raso G., Meli R.,Novellino E., Perissutti E., Santagada V., Eur. J. Med. Chem. 31(1996) 99–110.

[3] Löfgren N., Studies on Local Anaesthetics, Hoeggstrom, Stockholm,1948.

[4] Morgan P.H., Mathison I.W., J. Pharm. Sci. 65 (1976) 635–648.

[5] Heymans L., Le Therizien L., Godfroid J., J. Med. Chem. 23 (1980)184–193.

[6] Bowden K., Heilbron I.M., Jones E.R.H., Weedon B.C.L., J. Chem.Soc. Part I (1946) 39–45.

[7] Nesmeyanov A.N., Rybinskaya M.I., Dolk. Akad. Nauk. SSSR 158(1964) 408–410.

[8] Biagi G., Ferretti M., Livi O., Scartoni V., Lucacchini A., MazzoniM., Il Farmaco Ed. Sc. 41 (1986) 388–400.

[9] Livi O., Biagi G., Ferretti M., Lucacchini A., Barili P.L., Eur. J. Med.Chem. -Chim. Ther. 18 (1983) 471–475.

[10] Regnier J., Bull. Sci. Pharm. 30 (1923) 580–586.

[11] Bianchi C., Br. J. Pharmacol. 11 (1956) 104–106.

[12] Tallarida R.J., Murray R.B., Manual of Pharmacological Calcula-tions with Computer Programs, 2nd Edition, Springer, New York,1981.

[13] Al Saadi D., Sneider W.E., Arzneim. -Forsch. 41 (1991) 195–198.

[14] Hughes I.E., Smith J.A., J. Pharm. Pharmacol. 30 (1978) 124–126.

[15] Van Rossum J.M., Arch. Int. Pharmacodyn. Ther. 143 (1963)299–33.

1051

Original article

Synthesis and in vitro antitumour evaluationof benzothiazole-2-carbonitrile derivatives

Valérie Bénéteaua, Thierry Bessona*, Jérôme Guillarda, Stéphane Léonceb, Bruno Pfeifferc

aLaboratoire de Génie Protéique et Cellulaire, UPRES 2001, Pôle Sciences et Technologie, Université de La Rochelle, avenue Marillac,F-17042 La Rochelle cedex 1, France

bInstitut de Recherche Servier, 11, rue des Moulineaux, 92150 Surennes, FrancecA.D.I.R., 1, rue Carle Hébert, 92415 Courbevoie cedex, France

(Received 20 May 1999; revised 7 July 1999; accepted 15 July 1999)

Abstract – Novel benzothiazole derivatives have been synthesised via the corresponding imino-1,2,3-dithiazoles. The cytotoxicity of someof these polyheterocyclic compounds was studied. Our results show that 2-cyano derivatives exhibit a medium in vitro antitumour activity.© 1999 Éditions scientifiques et médicales Elsevier SAS

imino-1,2,3-dithiazoles / antitumour activity / benzothiazoles / benzodioxines

1. Introduction

The benzothiazole ring is present in various marine orterrestrial natural compounds which have useful biologi-cal activities [1–4]. Because we are interested in hetero-cyclic systems with potential pharmacological value, wedecided to synthesise new benzothiazole derivativeswhich are related to synthetic thiazoles which haveshown antitumour activity [5]. Novel dioxinobenzothiaz-oles were also prepared with the aim to enhance theantiproliferative activity.

In this paper, we describe the biological evaluation of4,7-dimethoxybenzothiazoles and dioxinobenzothiazolesprepared viaN-arylimino-1,2,3-dithiazoles which haveproved to be highly versatile intermediates in hetero-cyclic chemistry [6–8].

2. Chemistry

2.1. 4,7-Dimethoxybenzothiazoles

4,5-Dichloro-1,2,3-dithiazolium chloride1 is a palegreenish yellow solid, insoluble in organic solvents. It is

completely stable in a dry inert atmosphere but reactsslowly with moisture to form 4-chloro-1,2,3-dithiazol-5-one. This compound is readily prepared from chloro-acetonitrile and disulfur dichloride [9], reacts rapidlywith 2,5-dimethoxy-aniline, in dichloromethane at roomtemperature, to give the stableN-arylimino-1,2,3-dithiazoles 2 in high yield (84%). Pyrolysis of theseimines gave 2-cyanobenzothiazoles3 by cyclisation ofthe ortho carbon onto sulfur with liberation of the othersulfur atom and hydrogen chloride [10] (figure 1).

Removal (hydrolysis and decarboxylation) of the cy-ano group in the thiazole ring was performed by vigorousheating of compound3a in concentrated hydrochloricacid. The decyanated benzothiazole4 was isolated ingood yield (70%) (figure 2). Using standard conditionsfor the transformation of cyano groups into carboxylicacid or amido groups [11], the starting thiazole3a wastreated with aqueous sodium hydroxide to give the acid5,or with sulfuric acid to provide the amide6, in quite goodyields. The substituted amide7 was prepared by treat-ment of the acid5 with N,N-dimethylethylenediamine inthe presence of 1-(3-dimethylaminopropyl)-3-ethyl car-bodiimide hydrochloride (EDCI) and 1-hydroxy-benzotriazole (HOBT), a method commonly used for thecoupling of amino acids (figure 2).*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1053−1060 1053© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Decyanation of compound3b, using the conditionsdescribed above, afforded the amino decyanated ben-zothiazole 8, in which the amino group was easilyreprotected, to give9, by treatment with acetic anhydridein the presence of pyridine. The quinone10 was obtainedby oxidation of3b with cerium ammonium nitrate (fig-ure 3).

Starting from the commercially available isothiocyan-ate 11, the 2-alkoxyderivative13 was synthesised by aJacobson process via the intermediate thiocarbamate12(figure 4). In this case, an electron releasing group ispresent in the 2-position of the benzothiazole ring.

2.2. Dioxinobenzothiazoles [12]

The chemistry of the salt1 described above also allowsa rapid access to the dioxinobenzothiazoles15–18, from

the starting 6-amino-2,3-dihydro-1,4-benzodioxin (fig-ure 5). The bromination (NBS, CCl4, AIBN)–debromina-tion (NaI, acetone) sequence previously described inseveral syntheses of benzodioxins [13], lead to the ex-pected products19and20 (figure 5), whilst heating of thebrominated intermediate21 in pyridine, in the presenceof one equivalent of copper iodide (CuI), allowed analternative access to the linear compound15 (the use ofthe standard method afforded the angular brominatedderivative22) (figure 6).

3. Pharmacology

Fifteen compounds were evaluated in vitro for theirantiproliferative activity using the murine L1210 leu-kaemia cell line [14]. The results expressed as IC50

Figure 1. Reactions and conditions: a) pyridine, dichloromethane, –15 °C, 3 h (2a, 40%) or room temperature, 3 h (2b, 84%);b) toluene, sealed tube, reflux, 4 h (3a, 70%); c) PyHBr3, pyridine, reflux, 1.5 h (3b, 72%).

Figure 2. Reactions and conditions: a) conc. HCl, reflux, 8 h, 70%; b) NaOH 10%, 80 °C, 3.5 h, 64%; c) conc. H2SO4, roomtemperature, 2 h, 72%; d)N,N-dimethylethylenediamine, EDCI, HOBT, DMF, 0 °C, 4 days, 43%.

1054

(concentration reducing the cell proliferation by 50%) arereported intable I. Cell cycle perturbations induced bythe most active compounds (IC50 < 20 µM) were alsoinvestigated.

4. Results and discussion

The 2-cyano-4,7-dimethoxybenzothiazole derivatives3a and 3b were first evaluated and found practically

Figure 3. Reactions and conditions: a) conc. HCl, reflux, 1.5 h, 83%; b) CAN, CH3CN, H2O, room temperature, 15 min, 56%;c) Ac2O, pyridine, room temperature, 12 h, 56%.

Figure 4. Reactions and conditions: a) PrOH, NaH, 83%; b) K3FeCN6, NaOH, 28%.

Figure 5. Reagents and conditions: a) PyHBr3 (1 eq.), DMF/pyridine, reflux, 90 min, 58% (29% for15 and 29% for16); b): HCl,reflux, 2.5 h, 80% (17) and 65% (18); c) i) NBS/AIBN, CCl4, hν, reflux, 10 h; ii) NaI, acetone, reflux, 1.5 h, 69% (19) and 27% (20);d) Br2 (1 eq.), CH3COOH, room temperature, 2 h, 98%.

1055

equipotent on cell proliferation with IC50’s of, respec-tively, 20.6µM and 25.2µM. Also, both were found to beable to almost totally block the cells in the G2 + M phaseof the cell cycle. Suppression of the 2-cyano substituent(compounds4 and9) or its replacement by a 2-carboxy(compound5), a 2-aminocarbonyl (compound6), a 2-[2-(N,N-dimethylamino)ethylaminocarbonyl] (compound7)or a 2-propoxy (compound13) substituent led to inactivederivatives. Compound10 which is a quinone derivativewas found significantly more active that its 4,7-dimethoxy counterpart3b with an IC50 of 5µM versus25.2 µM. This compound was unfortunately devoid ofany specific effect on the cell cycle.

All the 2-cyano-dioxinobenzothiazoles have alsoshown a relatively interesting antiproliferative activity.As for the 4,7-dimethoxybenzothiazoles, removal of thecyano substituent present in the 2-position of the dioxino-benzothiazole ring (compound17) involved the lost ofany activity (IC50 > 100 µM). The unsaturated com-pounds19 and 20 were found significantly more activethan their saturated conterparts15 and18, with IC50’s of,respectively, 18.2µM and 31.7 µM, confirming the

results already published on dioxinocoumarins [15, 16].The dioxinobenzothiazole19 was found able to blockL1210 cells in the G2 + M phase of the cell cycle.

5. Conclusion

In conclusion, we have described the synthesis ofnovel benzothiazoles and dioxinobenzothiazoles, amongwhich the 2-cyano derivatives exhibit interesting in vitroantitumour activity. Presence of unsaturation in the dioxinmoiety, in combination with the cyano group in the2-position of the thiazole ring, did not really involve anyspecific effect on the cell cycle. Our results suggest thatintroduction of the thiazolo-2-carbonitrile ring into moreextended and more complexe heterocyclic moieties couldopen the door to promising applications.

6. Experimental protocols

6.1. Chemistry

Melting points were determined using a Kofler bancand are uncorrected. IR spectra were recorded on aPerkin-Elmer Paragon 1000PC instrument.1H- and13C-NMR were recorded on a JEOL JNM LA400 (400 Mhz)spectrometer (Centre Commun d’Analyse, Université deLa Rochelle); chemical shifts (δ) are reported in parts permillion (ppm) downfield from tetramethylsilane (TMS),which was used as internal standard. Mass spectra wererecorded on a Varian MAT311 in the Centre de MesurePhysiques de L’Ouest (C.R.M.P.O.), Université deRennes. Chromatography was carried out on silica gel 60

Figure 6. Reagents and conditions: a) PyHBr3 (1 eq.), DMF/pyridine, reflux, 2 h, 25%; b) CuI, pyridine, reflux, 90 min,55%.

Table I. Characteristics and pharmacological activity of synthesised compounds.

Compound M.p. (°C) Formula IC50 (µΜ) % of L1210 cells in the G2 + M phasea (µM)

3a 174 C10H8N2O2S 20.6 64% (50)3b 173 C12H11N3O3S 25.2 77% (100)4 108 C9H9NO2S > 100 n.e.b

5 > 230 C10H9NO4S > 100 n.e.6 240 C10H10N2O3S > 100 n.e.7 144 C14H19N3O3S > 100 n.e.9 154 C11H12N2O3S > 100 n.e.10 242 C10H5N3O3S 5 non specific13 73 C12H15NO3S > 100 n.e.15 178 C10H6N2O2S 58.4 n.e.b

16 156 C10H6N2O2S 42.3 n.e.17 142 C9H7NO2S > 100 n.e.19 176 C10H4N2O2S 18.2 41% (50)20 150 C10H4N2O2S 31.7 n.e.22 250 C10H5BrN2O2S 80.8 n. e.

a24% of untreated control cells were in the G2 + M phase of the cell cycle.bn.e.: not evaluated (for IC50 > 30 µM).

1056

at medium pressure and the sample mixtures were appliedto the column preadsorbed onto silica. Light petroleumrefers to the fraction b.p. 40–60 °C. Further solvents wereused without purification. Thin-layer chromatographywas performed on Merck Kieselgel 60 F254 aluminiumbacked plates.

Spectral data for compounds14–22 are consistent withassigned structures as previously described in ref. [12].

6.1.1. 2,5-Dimethoxy-N-(4-chloro-5H-dithiazol-5-yli-dene)aniline2a

Under an inert atmosphere, 4,5-dichloro-1,2,3-dithiazolium chloride (15.42 g, 73.7 mmol) and pyridine(10.8 mL, 134 mmol) were added to a stirred solution of2,5-dimethoxyaniline (10.3 g, 67 mmol) in dichloro-methane (150 mL) cooled at –15 °C. After 2 h, themixture was filtered and the solvent removed in vacuo.The crude residue was purified by column chromatogra-phy (eluent: light petroleum/ethyl acetate: 9/1) to affordcompound2a as an orange oil (7.8 g, 40%); IR (film):ν2 938 and 2 833 (CH3), 1 586 (C=N), 1 498, 1 464,1 278, 1 223, 1 195, 1 165, 1 044 cm–1; 1H-NMR(CDCl3): δ 3.78 (s, 3H, OCH3), 3.82 (s, 3H, OCH3), 6.69(d, 1H, J = 2.9 Hz, 6-H), 6.75 (dd, 1H,J = 2.9 Hz andJ≠ = 8.8 Hz, 4-H), 6.93 (d, 1H,J≠ = 8.8 Hz, 3-H);13C-NMR (CDCl3): δ 56.65, 57.21, 106.11, 112.91,114.43, 141.71, 144.68, 148.48, 154.97, 161.02; MS (EI):m/z 288 (M+), 188, 148; HRMS: calc. forC10H9ClN2O2S2: 287.9794, found: 287.9801.

6.1.2. 4-Acetamido-2,5-dimethoxy-N-(4-chloro-5H-1,2,3-dithiazol-5-ylidene) aniline2b

Under an inert atmosphere, 4,5-dichloro-1,2,3-dithiazolium chloride (220 mg, 1.04 mmol) was added toa stirred solution of 4-acetamido-2,5-dimethoxyaniline(200 mg, 0.95 mmol) in dichloromethane (10 mL). After1 h, pyridine (110 mg, 1.39 mmol) was added and themixture stirred for 15 min. The solvent was removed invacuo and the crude residue purified by column chroma-tography (eluent: dichloromethane/ethyl acetate: 4/1) toafford compound2b (275 mg, 84%) as an orange powder;m.p. 155 °C; IR (KBr): ν 3 506 (NH), 1 671 (C=O),1 483, 1 402, 1 214, 1 041, 862 cm–1; 1H-NMR (CDCl3):δ 2.22 (s, 3H, NHCOCH3), 3.85 (s, 3H, OCH3), 3.87 (s,3H, OCH3), 6.72 (s, 1H, 6-H), 7.80 (s, 1H, NH), 8.28 (s,1H, 3-H); 13C-NMR (CDCl3): δ 25.01, 56.20, 56.24,102.98, 104.85, 126.62, 133.91, 141.64, 143.81, 147.80,158.88, 168.24; MS (EI):m/z345 (M+), 252 (M–CNSCl),195; HRMS: calc. for C12H12ClN3O3S2: 345.0009,found: 345.0016.

6.1.3. 4,7-Dimethoxybenzothiazole-2-carbonitrile3aIn a closed system, a solution of2a (871 mg,

3.02 mmol) in toluene (5 mL) was heated in an oil bath at200–210 °C for 8 h. After cooling, the toluene wasremoved in vacuo. The crude residue was purified bycolumn chromatography (eluent: dichloromethane/lightpetroleum: 6/4) and recrystallised from ethanol to afford3a (471 mg, 70%) as yellow needles; m.p. 174 °C; IR(KBr): ν 2 229 (nitrile), 1 595 (C=N), 1 501, 1 454,1 280, 1 140, 1 094, 1 046, 969 cm–1; 1H-NMR (CDCl3):δ 3.98 (s, 3H, OCH3), 4.05 (s, 3H, OCH3), 6.93 (s, 2H,5-H and 6-H); 13C-NMR (CDCl3): δ 56.36, 56.49,108.10, 108.44, 112.97, 126.32, 135.55, 143.75, 147.59,148.91; MS (EI):m/z 220 (M+), 205 (M+–CH3), 191;HRMS: calc. for C10H8N2O2S: 220.0306, found:220.0303.

6.1.4. 6-Acetamido-4,7-dimethoxybenzothiazole-2-carbo-nitrile 3b

Under an inert atmosphere, a mixture of2b (1.60 g,4.63 mmol) and pyridinium perbromide (1.63 g,5.10 mmol) in pyridine (20 mL) was heated at reflux for2.5 h. After cooling, pyridine was removed in vacuo andthe crude residue purified by column chromatography(eluent: dichloromethane/ethyl acetate: 95/5). Recrystal-lisation from ethanol afforded compound3b (927 mg,72%) as yellow needles; m.p. 173 °C; IR (KBr):ν 3 390(NH), 2 226 (nitrile), 1 683 (C=O), 1 600 (C=N), 1 523,1 448, 1 420, 1 240, 1 132, 1 050 cm–1; 1H-NMR(CDCl3): δ 2.30 (s, 3H, NHCOCH3), 3.96 (s, 3H, OCH3),4.08 (s, 3H, OCH3), 7.86 (bs, 1H, NH), 8.34 (s, 1H, 5-H);13C-NMR (CDCl3): δ 25.20, 56.68, 59.88, 101.65,112.78, 128.05, 132.37, 132.79, 134.56, 139.33, 151.25,168.62; MS (EI):m/z 277 (M+), 220 (M+–CNOCH3);HRMS: calc. for C12H11N3O3S: 277.0521, found:277.0524.

6.1.5. 4,7-Dimethoxybenzothiazole4A suspension of compound3a in concentrated hydro-

chloric acid (20 mL) was heated at reflux for 8 h. Themixture was cooled at 0 °C, neutralised to pH 8 usingsaturated aqueous sodium hydrogen carbonate and ex-tracted with dichloromethane. The combined extractswere dried over MgSO4 and the solvent removed invacuo. After recrystallisation from hexane, compound4(177 mg, 70%) was obtained as colourless needles; m.p.108 °C; IR (KBr):ν 3 074 (CHunsat.), 2 959 (CH3), 1 594(C=N), 1 500, 1 456, 1 439, 1 345, 1 263, 1 184,1 150 cm–1; 1H-NMR (CDCl3): δ 3.97 (s, 3H, OCH3),4.03 (s, 3H, OCH3), 6.78 (d, 1H,J = 8.6 Hz, Harom.), 6.86(d, 1H,J = 8.6 Hz, Harom.), 8.91 (s, 1H, 2-H);13C-NMR(CDCl3): δ 55.29, 55.63, 104.81, 106.39, 123.95, 144.11,

1057

147.59, 152.49, 152.53; MS (EI):m/z 195 (M+), 180(M+–CH3), 166; HRMS: calc. for C9H9NO2S: 195.0354,found: 195.0361.

6.1.6. 4,7-Dimethoxybenzothiaole-2-carboxylic acid5A suspension of compound3a (204 mg, 0.93 mmol) in

10% aqueous sodium hydroxide (10 mL) was heated at80 °C for 3.5 h. After cooling to room temperature, themixture was poured onto iced water and acidified to pH 1using 10% aqueous hydrochloric acid. The yellow pre-cipitate that separated was filtered under vacuum anddried. The product was purified by column chromatogra-phy (eluent: hexane/ethyl acetate/methanol: 1/1/0.5 thenmethanol). Recrystallisation from water afforded com-pound5 (141 mg, 64%) as yellow needles; m.p. 230 °C(dec); IR (KBr):ν 3 367 (broad: OH), 2 838 (CH3), 1 632(C=O), 1 499, 1 383, 1 263, 1 189, 1 097, 1 044,974 cm–1; 1H-NMR (DMSO-d6): δ 3.87 (s, 3H, OCH3),3.88 (s, 3H, OCH3), 6.90 (s, 2H, Harom.);

13C-NMR(DMSO-d6): δ 55.92, 56.11, 106.04, 107.53, 126.56,144.48, 147.54, 148.19, 162.07, 169.85; MS (EI):m/z209 (M–CH2O), 195 (M–CO2), 180 (M–CO2, CH3), 166;HRMS could not be measured due to the absence of themolecular pic.

6.1.7. 4,7-Dimethoxybenzothiazole-2-carboxamide6A solution of compound3a (703 mg, 3.19 mmol) in

concentrated sulfuric acid (5 mL) was stirred at roomtemperature for 2 h. After cooling at 0 °C, the mixturewas basified using 10% aqueous sodium hydroxide andthen extracted with ethyl acetate. The combined extractswere dried over MgSO4. Removal of the solvent in vacuofollowed by a recrystallisation from ethanol gives com-pound6 (548 mg, 72%) as amber needles; m.p. 240 °C;IR (KBr): ν 3 406 (NH), 1 682 (C=O), 1 600 (C=N),1 558, 1 505, 1 118, 799 cm–1; 1H-NMR (CDCl3): δ 3.97(s, 3H, OCH3), 4.04 (s, 3H, OCH3), 5.73 (bs, 1H, NH),6.84 (d, 1H,J = 8.6 Hz, Harom.), 6.89 (d, 1H,J = 8.6 Hz,Harom.), 7.41 (bs, 1H, NH);13C-NMR (DMSO-d6): δ55.98, 56.15, 107.50, 108.33, 126.58, 143.89, 147.44,148.33, 161.17, 163.60; MS (EI):m/z 238 (M+), 223(M+–CH3), 209, 192, 180; HRMS: calc. forC10H10N2O3S: 238.0412, found: 238.0415.

6.1.8. 4,7-Dimethoxybenzothiazole-2-[2-(N,N-dimethyl-amino)ethyl]carboxamide7

Under an inert atmosphere, 1-(3-dimethylamino-propyl)-3-ethyl-carbodiimide hydrochloride (1.22 g, 6.47mmol), 1-hydroxybenzotriazole (840 mg, 6.47 mmol)and N,N-dimethylethylenediamine (0.7 mL, 6.47 mmol)were added to a stirred solution of compound5 (703 mg,2.94 mmol) in DMF (20 mL) at 0 °C. The mixture wasstirred for 4 days, allowing the temperature to increase

slowly. Water was then added with cooling and the crudematerial extracted with ethyl acetate. The combinedextracts were washed with water (three times) and driedover MgSO4. After removal of the solvent in vacuo, theproduct was purified by column chromatography (eluent:dichloromethane/methanol: 95/5) and recrystallised fromhexane/ethanol to afford compound7 (393 mg, 43%) aspale yellow needles; m.p. 144 °C; IR (KBr):ν 3 168(NH), 2 950 (CH3), 2 824 (CH2), 1 662 (C=O), 1 540,1 276, 1 188, 1 143, 1 090, 1 045 cm–1; 1H-NMR(CDCl3): δ 2.24 (s, 6H, N(CH3)2), 2.50 (t, 2H,J = 6.25Hz, CH2–N), 3.54 (q, 2H,J = 6.25 Hz, NH–CH2), 3.91 (s,3H, OCH3), 3.99 (s, 3H OCH3), 6.77 (d, 1H,J = 8.7 Hz,Harom.), 6.82 (d, 1H,J = 8.7 Hz, Harom.), 7.73 (bs, 1H,NH); 13C-NMR (CDCl3): δ 37.47, 45.29 (×2), 56.05,56.34, 57.86, 106.40, 107.07, 128.11, 144.29, 148.29,148.39, 159.89, 163.77; MS (EI):m/z 309 (M+), 58((CH3)2N=CH2

+); HRMS: calc. for C14H19N3O3S:309.1147, found: 309.1155.

6.1.9. 6-Amino-4,7-dimethoxybenzothiazole8A solution of compound3b (93 mg, 0.34 mmol) in

concentrated hydrochloric acid (7 mL) was heated atreflux for 1.5 h. After cooling at room temperature, themixture was basified to pH 8 using saturated aqueoussodium hydrogen carbonate and extracted with dichloro-methane. The combined extracts were washed with waterand brine, and dried over MgSO4. The solvent wasremoved in vacuo and the crude residue was purified bycolumn chromatography (eluent: light petroleum/ethylacetate: 1/1) to afford compound8 (59 mg, 83%) as abrown powder; m.p. 153 °C; IR (KBr):ν 3 440/3 310(NH2), 1 614 (C=N), 1 500, 1 463, 1 398, 1 237, 1 043,987 cm–1; 1H-NMR (CDCl3): δ 3.86 (s, 3H, OCH3), 3.99(s, 3H, OCH3), 6.40 (s, 1H, 5-H), 8.60 (s, 1H, 2-H);13C-NMR (CDCl3): δ 56.09, 58.75, 98.02, 128.61,133.24, 137.35, 137.48, 148.11, 150.38; MS (EI):m/z210(M+), 195 (M+–CH3), 154 (M+–C3H4O).

6.1.10. 6-Acetamido-4,7-dimethoxybenzothiazole9Acetic anhydride (2.2 mL, 23.33 mmol) was added to a

stirred solution of8 (265 mg, 1.26 mmol) in pyridine(15 mL). The mixture was strirred overnight at roomtemperature. Then, water was added, the product wasextracted with ethyl acetate, purified by column chroma-tography (eluent: ethyl acetate/light petroleum: 7/3) andrecrystallised from dichloromethane/light petroleum toafford compound9 (179 mg, 56%) as pale yellowneedles; m.p. 154 °C; IR (KBr):ν 3 296 (NH), 2 963,1 668 (C=O), 1 606 (C=N), 1 532, 1 463, 1 386, 1 254,1 222, 1 042 cm–1; 1H-NMR (CDCl3): δ 2.25 (s, 3H,NHCOCH3), 3.93 (s, 3H, OCH3), 4.04 (s, 3H, OCH3),

1058

7.78 (bs, 1H, NH), 8.18 (s, 1H, 5-H), 8.79 (s, 1H, 2-H);13C-NMR (CDCl3): δ 25.09, 56.35, 59.74, 100.80,126.26, 128.94, 135.58, 140.60, 149.89, 150.81, 168.42;MS (EI): m/z 252 (M+), 237 (M+–CH3), 195; HRMS:calc. for C11H12N2O3S: 252.0569, found: 252.0567.

6.1.11. 6-Acetamido-2-cyano-4,7-dioxobenzothiazole10A solution of cerium ammonium nitrate (2.59 g,

4.72 mmol) in water (10 mL) was added to a stirredsolution of 3b (524 mg, 1.89 mmol) in acetonitrile(20 mL) over 10 min. After 15 min the mixture wasextracted with ethyl acetate, the combined extracts weredried over MgSO4 and the solvents removed in vacuo.The crude residue was purified by column chromatogra-phy (eluent: dichloromethane/ethyl acetate: 95/5) to af-ford compound10 (261 mg, 56%) as orange needles;m.p. 242 °C; IR (KBr):ν 3 266 (NH), 3 108 (CHunsat.),2 238 (nitrile), 1 708 (C=O), 1 522, 1 427, 1 331, 1 194,1 141, 1 014 cm–1; 1H-NMR (CDCl3): δ 2.32 (s, 3H,NHCOCH3), 7.94 (s, 1H, 5-H), 8.13 (bs, 1H, NH);13C-NMR (CDCl3): δ 25.04, 111.15, 115.47, 138.29,140.03, 142.88, 153.75, 169.02, 175.34, 178.42; MS (EI):m/z 247 (M+), 205; HRMS: calc. for C10H5N3O3S:247.0052, found: 247.0046.

6.1.12. 4,7-Dimethoxy-2-propyloxybenzothiazole13A mixture of 12 (698 mg, 2.73 mmol) and 30% aque-

ous sodium hydroxide (2.9 mL, 21.84 mmol) in ethanol(3 mL) was added dropwise to a solution of K3FeCN6

(3.6 g, 10.62 mmol) in water (5 mL) heated at 85 °C.After 1.5 h, the mixture was allowed to cool to roomtemperature and extracted with dichloromethane. Thecombined extracts were dried over MgSO4 and thesolvents removed in vacuo. The crude residue waspurified by column chromatography (eluent: lightpetroleum/ethyl acetate: 7/1). Recrystallisation from hex-ane afforded compound13 (191 mg, 28%) as colourlessneedles; m.p. 73 °C; IR (KBr):ν 2 964 and 2 833(CH2/CH3), 1 534, 1 500, 1 467, 1 382, 1 340, 1 262,1 097, 1 051 cm–1; 1H-NMR (CDCl3): δ 1.05 (t, 3H,J =7 Hz, CH3), 1.86 (sextuplet, 2H,J = 7 Hz, CH2), 3.89 (s,3H, OCH3), 3.96 (s, 3H, OCH3), 4.57 (t, 2H,J = 7 Hz,CH2), 6.63 (d, 1H,J = 8.5 Hz, Harom.), 6.77 (d, 1H,J =8.5 Hz, Harom.);

13C-NMR (CDCl3): δ 10.29, 22.20,55.93, 56.35, 73.68, 103.90, 107.25, 121.18, 139.71,146.32, 147.86, 173.04; MS (EI):m/z253 (M+), 211, 196;HRMS: calc. for C12H15NO3S: 253.0773, found:253.0774.

6.2. Antiproliferative activity

L1210 cells (murine leukaemia) provided by the NCI,Frederik, USA were cultivated in RPMI 1640 medium

(Gibco) supplemented with 10% foetal calf serum, 2 mML-glutamine, 100 units/mL penicillin, 100µg/mL strep-tomycin, and 10 mM HEPES buffer (pH = 7.4).

Cytotoxicity was measured by the microculture tetra-zolium assay as described in ref. [14]. Cells were exposedto graded concentrations of the compounds for 48 h andresults expressed as IC50 (concentration which reducedby 50% the optical density of treated cells with respect tountreated controls).

For the cell cycle analysis, L1210 cells (2.5× 105

cells/mL) were incubated for 21 h with various concen-trations of the compounds, then fixed by 70% ethanol(v/v), washed and incubated in PBS containing 100µg/mL RNAse and 25µg/mL propidium iodide for 30 min at20 °C. For each sample, 1× 104 cells were analysed on anATC3000 flow cytometer (Brucker, France) using anargon laser (Spectra-Physics) emitting 400 mW at488 nm. The fluorescence of propidium iodide was col-lected through a 615 nm long-pass filter.

Data are displayed as linear histograms and results areexpressed as the percentage of cells found is the G2 + Mphase of the cell cycle.

Acknowledgements

We thank the Communauté de Villes del’Agglomération de La Rochelle, the society ADIR(Groupe SERVIER) and the Comité de Charente-Maritime de la Ligue Nationale Contre le Cancer forfinancial support.

References

[1] Gunawardana G.P., Kohmoto S., Gunasekara S.P., McConnel O.J.,Koehn F.E., J. Am. Chem. Soc. 110 (1988) 4856–4858.

[2] Gunawardana G.P., Kohmoto S., Burres N.S., Tetrahedron Lett. 30(1989) 4359–4362.

[3] Gunawardana G.P., Koehn F.E., Lee A.Y., Clardy J., He H.Y.,Faulkner J.D., J. Org. Chem. 57 (1992) 1523–1526.

[4] Carroll A.R., Scheuer P.J., J. Org. Chem. 55 (1990) 4426–4431.

[5] Shi D.F., Bradshaw T.D., Wrigley S., McCall C.J., Lelieveld P.,Fichtner I., Stevens M.F.G., J. Med. Chem. 39 (1996) 3375–3384.

[6] Rees C.W., J. Heterocycl. Chem. 29 (1992) 639–651.

[7] Besson T., Emayan K., Rees C.W., J. Chem. Soc. Perkin Trans. 1(1995) 2097–2102.

[8] Besson T., Guillaumet G., Lamazzi C., Rees C.W., Synlett. (1997)704–706.

[9] Appel R., Janssen H., Siray M., Knoch F., Chem. Ber. 118 (1985)1632–1643.

[10] English R.F., Rakitin O.A., Rees C.W., Vlasova O.G., J. Chem. Soc.Perkin Trans. 1 (1997) 201–205.

[11] March J., in: Advanced Organic Chemistry, 4th edition, Wiley-Interscience Publication, 1992, p. 887.

1059

[12] Besson T., Guillard J., Tetrahedron 55 (1999) 5139–5144. In thispaper the multistep synthesis of such compounds was transposed toa focused microwave oven.

[13] Besson T., Ruiz N., Coudert G., Guillaumet G., Tetrahedron 51(1995) 3197–3204.

[14] Léonce S., Pérez V., Casabianca-Pignède M.R., Anstett M., Bisagni

E., Atassi, G., Invest. New Drugs 14 (1996) 169–180.

[15] Csik G., Besson T., Coudert G., Guillaumet G., Nocentini S., J.Photochem. Photobiol. B: Biol. 19 (1993) 119–124.

[16] Csik G., Ronto G., Nocentini S., Averbeck S., Averbeck D., BessonT., Coudert G., Guillaumet, G., J. Photochem. Photobiol. B: Biol. 24(1994) 129–139.

1060

Original article

Synthesis and aldose reductase inhibitory activityof a new series of 5-[[2-(ω-carboxyalkoxy)aryl]methylene]-4-oxo-2-

thioxothiazolidine derivatives

Makoto Murata*, Buichi Fujitani, Hiroyuki Mizuta

Department of Chemistry I. Discovery Research Laboratories, Dainippon Pharmaceutical CompanyLtd., Enoki 33-94, Suita/Osaka 564-0053, Japan

(Received 20 April 1999; accepted 15 July 1999)

Abstract – A new series of 5-[[2-(x-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothiazolidine derivatives was synthesized and evaluatedfor their potency as aldose reductase inhibitors (ARIs). Their activities were examined in terms of their inhibitory effect on rat lens aldosereductase in vitro and in terms of the preventive effect on sorbitol accumulation in the sciatic nerve of streptozotocin (STZ)-induced diabeticrats in vivo. Of these compounds, some of the naphthylmethylene thiazolidine derivatives were comparable to Zenarestat in the inhibitorypotency in vitro and in vivo. In particular, compound30was 1.5 times more potent than Zenarestat in the in vivo activity, and had an adequatepotency for clinical development. © 1999 Éditions scientifiques et médicales Elsevier SAS

5-[[2-(x-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothiazolidine derivatives / aldose reductase inhibitory activity / sorbitol accu-mulation inhibition

1. Introduction

Aldose reductase (AR), which is the rate limitingenzyme of the polyol pathway, is implicated in diabeticcomplications. Since AR has a low substrate affinity forglucose, the activity of the polyol pathway is very low atnormal physiological glucose concentrations [1]. How-ever, when there is elevated glucose concentration indiabetes, excessive sorbitol production from glucose byAR is thought to cause cellular damage as a result ofosmotic imbalance [2]. This effect leads to the develop-ment of diabetic complications such as neuropathy, ret-inopathy, nephropathy, and cataract formation [3–6].

Recently, numerous compounds have been selected aspotential aldose reductase inhibitors (ARIs) [7–11],whose representatives are shown infigure 1. These com-pounds possess an acidic proton which is attached to theimidic nitrogen or an acetic acid moiety in the mol-

ecule [12, 13]. Thus, we gave our attention to the com-pounds having a 2-(ω-carboxyalkoxy)aryl moiety andsynthesized those compounds. In this paper, we report thesynthesis and AR inhibitory activity of the 5-[[2-(ω-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxo-thiazoli-dine derivatives1 shown infigure 1.

2. Chemistry

The 4-substitued-2-formylphenoxyacetic acid deriva-tives were prepared according to the method of Emmottand Livingstone [14] with slight modifications, as shownin figure 2. The 5-substitued-salicylaldehydes2–4 werealkylated with ethyl bromoacetate in the presence offinely ground potassium carbonate, followed by hydroly-sis under the alkaline condition to give 4-substitued2-formylphenoxyacetic acids5–7.

The ω-(1-formyl-2-naphthyloxy)alkanoic acid deriva-tives9–10were prepared in a similar manner as describedabove.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1061−1070 1061© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

The 3-(1-formyl-2-naphthyloxy)propionic acid11 wasprepared by the alkylation of the aldehyde8 withâ-propiolactone in the presence of sodium hydrox-ide [15].

The 5-(2-carboxymethoxybenzylidene)thiazolidine de-rivatives13–24were prepared according to the method ofBaranov and Komaritsa [16], as shown infigure 3. Con-densation of the 4-substitued-2-formylphenoxyacetic ac-ids 12, 5, 6, or 7 with rhodanine under refluxing, in thepresence of 2 eq. anhydrous sodium acetate in acetic acid,gave compound13, 16, 19, or 22. Moreover, condensa-tion of those with N-methylrhodanine gave compound14, 17, 20, or 23, and with rhodanine-3-acetic acidlikewise gave compound15, 18, 19, or 24, respectively.

The 5-[(2-ω-carboxyalkoxy)naphthylmethylene]thi-azolidine derivatives25–30 were prepared in a similarmanner as described above.

The configuration of the exocyclic double bond ofthose thiazolidine derivatives were determined by1H-NMR and 13C-NMR spectroscopy according to themethod of Fresneau et al. and Isida et al. [17, 18].

The 2:3 mixture of compounds28 and 31 wereobtained by the photoirradation (fluorescent lamp) of28in methanol, as shown infigure 4. However, attemptedseparation of28 and31 failed since smooth reisomeriza-tion (31 to 28) occured during separation.1H-NMR and13C-NMR experiments on the mixture of compound28and31 disclosed the following results. (1) Although theC5 proton of compound28 showed a signal atδ 8.26 dueto the anisotropic effect by the carbonyl group (C4=O) ofthe thiazolidine ring, the same proton signal of31appeared atδ 7.91 in 1H-NMR spectroscopy. (2) Al-though the coupling constant value of the C4 carbon ofcompound28 hadJ = 6.5 Hz, the same value of31 had

Figure 1. Potential aldose reductase inhibitors.

1062

J = 12.4 Hz in 13C-NMR spectroscopy. From theseresults, we confirmed that the exocyclic double bond ofcompound28 had Z configuration.

Moreover, we postulated that the configuration of otherthiazolidine compounds had Z form on the basis of NMRdata.

3. Pharmacology

The AR inhibitory activity of the synthesized com-pounds was assessed by measuring the inhibition of theenzymatic activity in a partially purified rat lens prepa-ration [1]. The inhibitory activity was expressed as theconcentration (nM) of the test compound which inhibitedthe activity of AR by 50% (IC50). The in vivo ARinhibitory activity of the test compounds were alsoevaluated by measuring their ability to inhibit the sorbitolaccumulation in the sciatic nerve of STZ-induced diabeticrats [19].

4. Results and discussion

The in vitro AR inhibitory activity of the benzylidenethiazolidine derivatives13–24are shown intable I. WhenR1 is hydrogen or a carboxymethyl group, introduction ofthe substituent R resulted in an increase of AR inhibitoryactivity by 1 order of the potency as judged from IC50

values (i.e. compound13 vs. 16, 19 or 22; or compound15vs.18, 21or 24). When R1 is a methyl group however,introduction of the substituent R (i.e. compound14 vs.17, 20 or 23) had no influence on the AR inhibitoryactivity. Compounds14 and 19–23, which inhibited therat AR in vitro at IC50 values of 10 nM order, wereequipotent compared with the reference compound Ze-narestat. Furthermore, compounds16, 17 and18 were 2times more potent than Zenarestat. However, these com-pounds were inactive in inhibiting the sorbitol accumu-lation in the sciatic nerve of STZ-induced diabetic rats at100 mg/kg p.o. (table II).

Reagents: a) K2CO3/toluene, tris(dioxa-3,6-heptyl)amine, BrCH2CO2Et; b) 1 N NaOH/dioxane; c) â-propiolactone, NaOH/H2O.

Figure 2. Synthesis of compounds5–7, 9, 10 and11.

1063

The in vitro AR inhibitory activity of the naphthyl-methylene thiazolidine derivatives25–30 are shown intable III. When R1 is hydrogen, the increase in thenumber of methylene groups gave a remarkable decreaseof the in vitro activity (i.e. compound25 and26 vs. 27).When R1 is a methyl group, the increase in the number of

methylene groups gave a moderate decrease of the invitro activity (i.e. compound28 vs. 29 vs. 30). Of thesecompounds, compound25 was equipotent to Zenarestat,and compound26 was 3 times more potent than Zenar-estat.

Figure 3. Synthesis of thiazolidine derivatives13–24 and25–30.

Figure 4. Photoirradiation of derivative28.

1064

The in vivo inhibitory activity of the naphthylmethyl-ene thiazolidine derivatives on the sorbitol accumulationin the sciatic nerve of STZ-induced diabetic rats is shownin table II. When R1 is hydrogen, the increase in thenumber of methylene groups gave a moderate decrease ofthe in vivo activity (i.e. compound25 vs. 26). When R1

is a methyl group, the increase in the number of methyl-ene groups unexpectedly gave a moderate increase of thein vivo activity. Of these, compound25had equipotent invivo activity to Zenarestat. Of particular interest is thatthe in vivo activity of compound30was more potent thanthat of Zenarestat, although it was 20 times less potent

than Zenarestat in vitro. Therefore, we postulate thatcompound30 possessed good oral absorption and effi-cient tissue penetration properties because of the in-creased number of methylene groups.

In conclusion, we have reported in this article 5-[[2-(ω-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothia-zolidine derivatives having biological activities whichwere comparable to the in vitro and in vivo inhibitoryactivities of Zenarestat. Of these, compound30 was 1.5times more potent than Zenarestat in the in vivo activity,and had an adequate potency for clinical development.

Table I. Chemical and biological data, in vitro, of benzylidene thiazolidine derivatives.

AR inhibition in vitro

rat lens ARa

Compound M.p. (°C) Recryst. solvent Formula IC50 (nM)b

13 > 250 EtOH-H2O C12H9NO4S⋅0.2H2O 56014 220–221 AcOH-H2O C13H11NO4S2 2715 > 250 EtOH-H2O C14H10NNaO6S2⋅H2O 17016 > 250 (CH3)2CO-H2O C12H8BrNO4S2 1717 243–246 (CH3)2CO-H2O C13H10BrNO4S2 1618 204–208 AcOH-H2O C14H10BrNO6S2 1819 > 250 (CH3)2CO-H2O C12H8ClNO4S2 2920 237–239 (CH3)2CO-H2O C13H10ClNO4S2 1821 209–212 AcOH-H2O C14H10ClNO6S2⋅0.5H2O 2122 226–228 MeOH-H2O C13H11NO5S2⋅0.25H2O 3823 215–216 (CH3)2CO-H2O C14H13NO5S2 1924 213–215 AcOH-H2O C15H13NO7S2 86Zenarestat 36

aAR: aldose reductase.bThe concentration of test compounds required for 50% inhibition of AR.

Table II. Biological data, in vivo, of benzylidene thiazolidine and naphthylmethylene thiazolidine derivatives.

Aldose reductase inhibition

Compound Dose (mg/kg) in vivo % inhibitiona

13 100 5.8± 3.116 100 NS17 100 NS18 100 NS25 100 20.1± 4.826 100 8.0± 4.828 30 11.7± 5.829 100 8.6± 7.130 100 34.6± 6.4Zenarestat 100 21.8± 7.2

aPercent inhibition of sorbitol accumulation in the sciatic nerve of streptozotocin-induced diabetic rats. Test compounds were orally given atthe single dose indicated. Values are mean± SEM; mean of 4-6 rats. NS, no significant inhibition.

1065

5. Experimental protocols

5.1. Chemistry

Melting points (m.p.) were determined on aYanagimoto micro-melting apparatus and are uncor-rected. Proton nuclear magnetic resonance (1H-NMR)spectra were obtained on a Varian Gemini-200 spectrom-eter with tetramethylsilane as an internal standard.Chemical shifts are reported inδ values from internaltetramethylsilane. Splitting patterns are designated asfollows: s, singlet; d, doublet; dd, double doublet; t,triplet; q, quartet; br s, broad singlet; m, multiplet.Coupling constants are reported in hertz (Hz). Infrared(IR) spectra were recorded on a Shimazu FTIR-8200PCspectrophotometer. Elemental analyses were carried outon a Perkin-Elmer 2400 element analyzer and resultsobtained for specified elements were within± 0.4% of thetheoretical values. Visualization was accomplished withUV light. Unless otherwise noted, all commerciallyavailable materials were used without further purifica-tion.

5.1.1. Typical procedure for the preparation of 2-formylphenoxy acetic acids5–7

5.1.1.1. (4-Chloro-2-formylphenoxy)acetic acid65-Chlorosalicylaldehyde (3; 6.3 g, 40 mmol) was dis-

solved in dry toluene (90 mL) and finely ground potas-sium carbonate (5.6 g, 40 mmol) were added. After themixture was heated at 100 °C for 4 h, tris(dioxa-3,6-heptyl)amine (1.3 mL, 4.1 mmol) and ethyl bromoacetate(5.6 mL, 50 mmol) were added. The mixture was heatedat 100 °C for 6 h, filtered through Celite and washed withtoluene. The filtrate was washed with water, and brine,dried (anhydrous MgSO4), and evaporated. The residuewas dissolved in dioxane (50 mL), and 1 N NaOH(50 mL) was added. The mixture was heated at 100 °C for

1 h. After cooling to room temperature, the reactionmixture was diluted with water (150 mL) and thenacidified with concentrated HCl on an ice bath, and theprecipitated solid was collected by filtration. The solidwas recrystallized from acetone/H2O to give the titlecompound as a solid (5.3 g, 62%). M.p. 174–175 °C (lit.173–174 °C) [20].

The following compounds were prepared in the sameway as described above:

(4-Bromo-2-formylphenoxy)acetic acid5 (preparedfrom 5-bromosalicylaldehyde2; recrystallized fromacetone/H2O). M.p. 172–174 °C (lit. 174–176 °C) [20].(2-Formyl-4-methoxyphenoxy)acetic acid7 (preparedfrom 5-methoxysalicylaldehyde4; recrystallized fromacetone/H2O). M.p. 156–159 °C (lit. 157–159 °C) [20].

5.1.2. (1-Formyl-2-naphthyloxy)acetic acid92-Hydroxy-1-naphthylaldehyde (8; 32 g, 0.2 mol) was

dissolved in dry toluene (200 mL) and finely groundpotassium carbonate (15 g, 0.11 mol) was added. Afterthe mixture was heated at 100 °C for 4 h, tris(dioxa-3,6-heptyl)amine (4 mL, 12.5 mmol) and ethyl bromoacetate(28 mL, 0.25 mmol) were added. The mixture was heatedat 100 °C for 4 h, filtered through Celite and washed withtoluene. The filtrate was washed with water and brine,dried (anhydrous MgSO4), and evaporated. The residuewas dissolved in dioxane (250 mL) and 1 N NaOH(250 mL) was added. The mixture was heated at 100 °Cfor 1 h. After cooling to room temperature, the reactionmixture was diluted with water (700 mL) and thenacidified with concentrated HCl on an ice bath, and theprecipitated solid was collected by filtration. The solidwas recrystallized from AcOEt/n-hexane to give the titlecompound as a solid (15 g, 65%): M.p. 176–179 °C (lit.176–177 °C) [14].

Table III. Chemical and biological data, in vitro, of naphthylmethylene thiazolidine derivatives.

AR inhibition in vitro

rat lens ARa

Compound M.p. (°C) Recryst. solvent Formula IC50 (nM)b

25 231–234 AcOH-H2O C16H11NO4S2 3326 212–216 AcOH-H2O C17H13NO4S2 1127 195–198 AcOEt-hexane C18H15NO4S2 1 10028 206–209 (CH3)2CO-H2O C17H13NO4S2 18029 165–168 (CH3)2CO-H2O C18H15NO4S2 63030 93–95 AcOEt-hexane C19H17NO4S2⋅0.25H2O 700Zenarestat 36

aAR: aldose reductase.bThe concentration of test compounds required for 50% inhibition of AR.

1066

5.1.3. 4-(1-Formyl-2-naphthyloxy)butyric acid10Compound10 was synthesized in the same way as in

5.1.2. Yield 34%. M.p. 179–180 °C (recrystallized fromAcOEt/n-hexane); 1H-NMR (DMSO-d6) d 2.00–2.18(2H, m), 2.47 (2H, t,J = 7.5 Hz), 4.35 (2H, t,J = 7.5 Hz),7.40–7.70 (3H, m), 7.90–8.00 (1H, m), 8.27 (1H, d,J = 9Hz), 9.07–9.17 (1H, m), 10.7 (1H, s), 12.18 (1H, br s); IR(KBr, cm–1) 1 705 (CO), 1 672 (CO). Anal. calcd. forC15H14O4: C 69.76; H 5.46; found: C 69.65; H 5.40.

5.1.4. 3-(1-Formyl-2-naphthyloxy)propionic acid11A solution ofâ-propiolactone (7.6 g, 0.1 mol) in water

(10 mL) was added dropwise to a stirred solution of2-hydroxy-1-naphthylaldehyde (8; 17.2 g, 0.1 mol) andNaOH (4 g, 0.1 mol) in water (50 mL) at 100 °C, and themixture was stirred for a further 30 min at 100 °C. Aftercooling to room temperature, the reaction mixture wasacidified with concentrated HCl on an ice bath, and thenextracted three times with diethyl ether. The combinedether extracts were shaken three times with saturatedNaHCO3. The solid, which was precipitated on acidifi-cation of the alkaline solution, was filtered, washed withwater, dried, and recrystallized from AcOEt/n-hexane togive the title compound as a solid (4.2 g, 17%). M.p.157–159 °C (recrystallized from AcOEt/n-hexane);1H-NMR (DMSO-d6) d 2.84 (2H, t,J = 6 Hz), 4.52 (2H, t,J = 6 Hz), 7.42–7.52 (1H, m), 7.58–7.72 (2H, m),7.92–8.02 (1H, m), 8.30 (1H, d,J = 9 Hz), 9.07–9.17 (1H,m), 10.72 (1H, s), 12.48 (1H, br s); IR (KBr, cm–1) 1 705(CO), 1 672 (CO). Anal. calcd. for C14H12O4: C 68.85; H4.95; found: C 68.74; H 4.88.

5.1.5. 5-(5-Bromo-2-carboxymethoxybenzylidene)-4-oxo-2-thioxothiazolidine16

A mixture of 4-Bromo-2-formylphenoxyacetic acid (5;6.0 g, 23.2 mmol), rhodanine (3.7 g, 27.8 mmol) andanhydrous sodium acetate (3.8 g, 46.3 mmol) in aceticacid (80 mL) was refluxed for 17 h. Upon cooling, thereaction mixture was diluted with water and stirred atroom temperature for a further 1 h. The solid wasfiltrated, and added to dilute HCl, then the mixture wasstirred at room temperature for 1 h. After filtration, thesolid was recrystallized from acetone/H2O to give thetitle compound as a yellow solid (5.9 g, 68%). M.p.>250 °C; 1H-NMR (DMSO-d6) d 4.90 (2H, S), 7.06 (1H,d, J = 9 Hz), 7.49 (1H, d,J = 2.5 Hz), 7.63 (1H, dd,J =2.5, 9 Hz), 7.76 (1H, s), 13.20 (1H, br s), 13.88 (1H, brs); IR (KBr, cm–1) 1 714 (CO). Anal. calcd. forC12H8BrNO4S2: C 38.51; H 2.15; N 3.74; S 17.14; Br21.35; found: C 38.57; H 1.92; N 3.73; S 17.29; Br;21.50.

5.1.6. 5-(2-Carboxymethoxybenzylidene)-4-oxo-2-thioxo-thiazolidine13

Compound13 was synthesized in the same way as in5.1.5. Yield 32%. M.p.> 250 °C (recrystallized from aq.EtOH); 1H-NMR (DMSO-d6) d 4.88 (2H, S), 7.03–7.18(2H, m), 7.38–7.52 (2H, m), 7.90 (1H, s), 13.20 (1H, brs), 13.74 (1H, br s); IR (KBr, cm–1) 1 705 (CO), 1 672(CO). Anal. calcd. for C12H9NO4S2⋅0.2H2O: C 48.21; H3.17; N 3.74; S 21.45; found: C 48.49; H 2.93; N 4.74; S21.27.

5.1.7. 5-(2-Carboxymethoxybenzylidene)-3-methyl-4-oxo-2-thioxothiazolidine14

Compound14 was synthesized in the same way as in5.1.5. Yield 77%. M.p. 220–221 °C (recrystallized fromAcOH/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 4.90(2H, s), 7.04–7.20 (2H, m), 7.43–7.55 (2H, m), 8.07 (1H,s), 13.18 (1H, br s); IR (KBr, cm–1) 1 753 (CO), 1 689(CO). Anal. calcd. for C13H11NO4S2: C 50.47; H 3.58; N4.53; S 20.73; found: C 50.17; H 3.56; N 4.62; S 21.00.

5.1.8. [5-(2-Carboxymethoxybenzylidene)-4-oxo-2-thio-xothiazolidin-3-yl]-acetic acid mono sodium salt15

Compound15 was synthesized in the same way as in5.1.5. Yield 11%. M.p.> 250 °C (recrystallized from aq.EtOH);1H-NMR (DMSO-d6) d 4.50 (2H, s), 4.69 (2H, s),6.95–7.15 (2H, m), 7.41–7.51 (2H, m), 8.06 (1H, s); IR(KBr, cm–1) 1 716 (CO). Anal. calcd. forC14H10NNaO6S2⋅H2O: C 42.75; H 3.07; N 3.56; S 16.30;Na 5.84; found: C 42.82; H 3.07; N 3.59; S 16.01; Na5.54.

5.1.9. 5-(5-Bromo-2-carboxymethoxybenzylidene)-3-methyl-4-oxo-2-thioxothiazolidine17

Compound17 was synthesized in the same way as in5.1.5. Yield 67%. M.p. 243–246 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 3.41 (3H, s), 4.93(2H, s), 7.07 (1H, d,J = 9 Hz), 7.53 (1H, d,J = 2.5 Hz),7.65 (1H, dd,J = 2.5, 9 Hz) 7.91 (1H, s), 13.24 (1H, brs); IR (KBr, cm–1) 1 705 (CO). Anal. calcd. forC13H10BrNO4S2: C 40.22; H 2.60; N 3.61; S 16.52; Br20.58; found: C 40.15; H 2.61; N 3.49; S 16.58; Br 20.55.

5.1.10. [5-(5-Bromo-2-carboxymethoxybenzylidene)-4-oxo-2-thioxo-thiazolidin-3-yl]-acetic acid18

Compound18 was synthesized in the same way as in5.1.5. Yield 55%. M.p. 204–208 °C (recrystallized fromAcOH/H2O); 1H-NMR (DMSO-d6) d 4.75 (2H, s), 4.95(2H, s), 7.09 (1H, d,J = 9 Hz), 7.60 (1H, d,J = 2.5 Hz),7.96 (1H, dd,J = 2.5, 9 Hz) 7.96 (1H, s), 13.36 (2H, brs); IR (KBr, cm–1) 1 712 (CO). Anal. calcd. forC14H10BrNO6S2: C 38.90; H 2.33; N 3.24; S 14.84; Br18.49; found: C 38.73; H 2.39; N 3.15; S 14.59; Br 18.26.

1067

5.1.11. 5-(5-Chloro-2-carboxymethoxybenzylidene)-4-oxo-2-thioxothiazolidine19

Compound19 was synthesized in the same way as in5.1.5. Yield 62%. M.p.> 250 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 4.90 (2H, s), 7.14(1H, d,J = 9 Hz), 7.44 (1H, d,J = 2.5 Hz), 7.52 (1H, dd,J = 2.5, 9 Hz) 7.84 (1H, s), 13.24 (1H, br s), 13.88 (1H,br s); IR (KBr, cm–1) 1 733 (CO), 1 683 (CO). Anal.calcd. for C12H8ClNO4S2: C 43.71; H 2.45; N 4.25; S19.45; Cl 10.75; found: C 43.79; H 2.42; N 4.19; S 19.12;Cl 10.75.

5.1.12. 5-(5-Chloro-2-carboxymethoxybenzylidene)-3-methyl-4-oxo-2-thioxothiazolidine20

Compound20 was synthesized in the same way as in5.1.5. Yield 58%. M.p. 237–239 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 4.93(2H, s), 7.14 (1H, d,J = 9 Hz), 7.42 (1H, d,J = 2.5 Hz),7.54 (1H, dd,J = 2.5, 9 Hz) 7.93 (1H, s), 13.24 (1H, brs); IR (KBr, cm–1) 1 705 (CO). Anal. calcd. forC13H10ClNO4S2: C 45.42; H 2.93; N 4.07; S 18.65; Cl10.31; found: C 45.43; H 2.89; N 4.02; S 18.92; Cl 10.11.

5.1.13. [5-(5-Chloro-2-carboxymethoxybenzylidene)-4-oxo-2-thioxothiazolidin-3-yl]-acetic acid21

Compound21 was synthesized in the same way as in5.1.5. Yield 64%. M.p. 209–212 °C (recrystallized fromAcOH/H2O); 1H-NMR (DMSO-d6) d 4.75 (2H, s), 4.95(2H, s), 7.15 (1H, d,J = 9 Hz), 7.49 (1H, d,J = 2.5 Hz),7.55 (1H, dd,J = 2.5, 9 Hz) 7.97 (1H, s), 13.36 (1H, brs); IR (KBr, cm–1) 1 716 (CO). Anal. calcd. forC14H10ClNO6S2⋅0.5H2O: C 42.38; H 2.79; N 3.53; S16.16; Cl 8.93; found: C 42.66; H 2.65; N 3.45; S 15.83;Cl 8.76.

5.1.14. 5-(2-Carboxymethoxy-5-methoxybenzylidene)-4-oxo-2-thioxothiazolidine22

Compound22 was synthesized in the same way as in5.1.5. Yield 8%. M.p. 226–228 °C (recrystallized fromMeOH/H2O); 1H-NMR (DMSO-d6) d 3.78 (3H, s), 4.83(2H, s), 6.88 (1H, d,J = 2.5 Hz), 7.03 (1H, d,J = 9 Hz),7.08 (1H, dd,J = 2.5, 9 Hz), 7.87 (1H, s), 13.10 (1H, brs), 13.84 (1H, br s); IR (KBr, cm–1) 1 751 (CO), 1 695(CO). Anal. calcd. for C13H11NO5S2⋅0.25H2O: C 47.34;H 3.51; N 4.25; S 19.44; found: C 47.39; H 3.19; N 4.38;S 19.22.

5.1.15. 5-(2-Carboxymethoxy-5-methoxybenzylidene)-4-oxo-2-thioxothiazolidine23

Compound23 was synthesized in the same way as in5.1.5. Yield 35%. M.p. 215–216 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 3.80(3H, s), 4.83 (2H, s), 6.92 (1H, d,J = 2.5 Hz), 7.05 (1H,

d, J = 9 Hz), 7.10 (1H, dd,J = 2.5, 9 Hz) 8.03 (1H, s),13.12 (1H, br s); IR (KBr, cm–1) 1 712 (CO). Anal. calcd.for C14H13NO5S2: C 49.55; H 3.86; N 4.13; S 18.90;found: C 49.36; H 3.59; N 4.03; S 18.73.

5.1.16. [5-(2-Carboxymethoxy-5-methoxybenzylidene)-4-oxo-2-thioxo-thiazolidin-3-yl]-acetic acid24

Compound24 was synthesized in the same way as in5.1.5. Yield 40%. M.p. 213–215 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 3.80 (3H, s), 4.75(2H, s), 4.84 (2H, s), 6.97 (1H, d,J = 2.5 Hz), 7.06 (1H,d, J = 9 Hz), 7.12 (1H, dd,J = 2.5, 9 Hz), 8.07 (1H, s),13.30 (2H, br s); IR (KBr, cm–1) 1 716 (CO). Anal. calcd.for C15H13NO7S2: C 46.99; H 3.42; N 3.65; S 16.73;found: C 47.09; H 3.26; N 3.75; S 16.67.

5.1.17. 5-(2-Carboxymethoxy-1-naphthylmethylene)-4-oxo-2-thioxothiazolidine25

Compound25 was synthesized in the same way as in5.1.5. Yield 50%. M.p. 231–234 °C;1H-NMR (DMSO-d6) d 5.03 (2H, s), 7.40 (1H, d,J = 9 Hz), 7.43–7.53 (1H,m), 7.57–7.67 (1H, m), 7.82–7.90 (1H, m), 7.92–8.00(1H, m), 8.08 (1H, d,J = 9 Hz), 8.09 (1H, s), 13.20 (1H,br s), 13.70 (1H, br s); IR (KBr, cm–1) 1 712 (CO), 1 662(CO). Anal. calcd. for C16H11NO4S2: C 55.64; H 3.21; N4.06; S 18.57; found: C 55.74; H 3.49; N 3.97; S 18.49.

5.1.18. 5-[2-(2-carboxyethoxy)-1-naphthylmethylene]-4-oxo-2-thioxothiazolidine26

Compound26 was synthesized in the same way as in5.1.5. Yield 31%. M.p. 212–216 °C (recrystallized fromAcOH/H2O); 1H-NMR (DMSO-d6) d 2.78 (2H, t,J = 6Hz), 4.48 (2H, t,J = 6 Hz), 7.42–7.66 (2H, m), 7.57 (1H,d, J = 9 Hz), 7.78–7.86 (1H, m), 7.94–8.02 (1H, m), 8.00(1H, s), 8.11 (1H, d,J = 9 Hz), 12.38 (1H, br s), 13.66(1H, br s); IR (KBr, cm–1) 1 716 (CO), 1 683 (CO). Anal.calcd. for C17H13NO4S2: C 56.81; H 3.65; N 3.90; S17.84; found: C 56.55; H 3.82; N 3.71; S 17.49.

5.1.19. 5-[2-(3-carboxypropoxy)-1-naphthylmethylene]-4-oxo-2-thioxothiazolidine27

Compound27 was synthesized in the same way as in5.1.5. Yield 58%. M.p. 195–198 °C (recrystallized fromAcOEt/n-hexane); 1H-NMR (DMSO-d6) d 1.93–2.10(2H, m), 2.40 (2H, t,J = 7.5 Hz), 4.31 (2H, t,J = 7.5 Hz),7.41–7.66 (2H, m), 7.55 (1H, d,J = 9 Hz) 7.79–7.87 (1H,m), 7.93–8.00 (1H, m), 8.05 (1H, s), 8.10 (1H, d,J = 9Hz), 12.15 (1H, br s), 13.68 (1H, br s); IR (KBr, cm–1)1 705 (CO). Anal. calcd. for C18H15NO4S2: C 57.89; H4.05; N 3.75; S 17.17; found: C 57.94; H 4.11; N 3.71; S16.88.

1068

5.1.20. 5-(2-carboxymethoxy-1-naphthylmethylene)-3-methyl-4-oxo-2-thioxothiazolidine28

Compound28 was synthesized in the same way as in5.1.5. Yield 60%. M.p. 206–209 °C;1H-NMR (DMSO-d6) d 3.43 (3H, s), 5.02 (2H, s), 7.42 (1H, d,J = 9 Hz),7.43–7.53 (1H, m), 7.57–7.67 (1H, m), 7.82–7.91 (1H,m), 7.93–8.01 (1H, m), 8.10 (1H, d,J = 9 Hz), 8.26 (1H,s), 13.22 (1H, br s); IR (KBr, cm–1) 1 757 (CO), 1 668(CO). Anal. calcd. for C17H13NO4S2: C 56.81; H 3.65; N3.90; S 17.84; found: C 56.67; H 3.54; N 3.76; S 18.14.

5.1.21. 5-[2-(2-carboxyethoxy)-1-naphthylmethylene]-3-methyl-4-oxo-2-thioxothiazolidine29

Compound29 was synthesized in the same way as in5.1.5. Yield 28%. M.p. 165–168 °C (recrystallized fromacetone/H2O); 1H-NMR (DMSO-d6) d 2.78 (2H, t,J = 6Hz), 3.43 (3H, s), 4.48 (2H, t,J = 6 Hz), 7.43–7.66 (2H,m), 7.59 (1H, d,J = 9 Hz), 7.79–7.87 (1H, m), 7.94–8.02(1H, m), 8.13 (1H, d,J = 9 Hz), 8.18 (1H, s), 12.38 (1H,br s); IR (KBr, cm–1) 1 716 (CO). Anal. calcd. forC18H15NO4S2: C 57.89; H 4.05; N 3.75; S 17.17; found:C 57.77; H 4.21; N 3.56; S 16.70.

5.1.22. 5-[2-(3-carboxypropoxy)-1-naphthylmethylene]-3-methyl-4-oxo-2-thioxothiazolidine30

Compound30 was synthesized in the same way as in5.1.5. Yield 43%. M.p. 93–95 °C (recrystallized fromAcOEt/n-hexane); 1H-NMR (DMSO-d6) d 1.80–2.20(2H, m), 2.40 (2H, t,J = 7.5 Hz), 3.44 (3H, s), 4.30 (2H,t, J = 7.5 Hz), 7.42–7.66 (2H, m), 7.56 (1H, d,J = 9 Hz),7.80–7.88 (1H, m), 7.94–8.01 (1H, m), 8.12 (1H, d,J = 9Hz), 8.23 (1H, s), 12.14 (1H, br s); IR (KBr, cm–1) 1 716(CO), 1 695 (CO). Anal. calcd. for C19H7NO4S2⋅0.25H2O: C 58.22; H 4.50; N 3.57; S 16.36; found: C58.31; H 4.35; N 3.60; S 16.67.

5.2. Pharmacology

5.2.1. Preparation of aldose reductaseRat lens AR was prepared according to the method of

Hayman and Kinoshita [1] with slight modifications. Thelenses were homogenized in 250 mM phosphate buffer(pH 7.4) containing 2 mM mercaptoethanol at 0–4 °C,and the homogenate was centrifuged at 20 000g for30 min. The supernatant was subjected to a 40–60%ammonium sulfate fractionation. The resultant precipitatewas dissolved in 5 mM phosphate buffer (pH 7.4) con-taining 2 mM mercaptoethanol and used for enzymeassay. One unit (U) of AR was defined as the enzymeactivity which oxidizes 1µmol of NADPH in 1 minunder the assay conditions described below.

5.2.2. Inhibition of aldose reductase in vitroThe reaction mixture consisted of 100 mM phosphate

buffer (pH 6.5), 0.2 mM NADPH, 1.5 mM D,L-glyceraldehyde, 0.4 M lithium sulfate, 7.0 mU/mL ofenzyme and test compounds at various concentrations.The reaction mixture was incubated at 37 °C, and theabsorbance at 340 nm was measured with a spectropho-tometer (Model 150-20, Hitachi Ltd., Japan). The enzymeactivity was estimated on the basis of its decrease in theabsorbance over a period of 1 min. The concentration ofcompounds required for 50% inhibition of enzyme activ-ity (IC50) was estimated graphically from the logconcentration-inhibition curve.

5.2.3. Inhibition of sorbitol accumulation in vivoMale Wistar rats (200–250 g) were rendered diabetic

by an intravenous injection of streptozotocin (40 mg/kg),which had been freshly dissolved in physiological saline.After 7 days, the rats were divided into various groupswith 4–6 animals/group, and orally given test compoundssuspended in 0.5% tragacanth, or an equivalent volume of0.5% tragacanth. The rats were sacrificed 4 h after theadministration of the test compounds. Tissue sorbitollevels were determined according to the method ofClements et al. [17] with slight modifications.

The sciatic nerve sample (30–60 mg) was quicklydissected from the hind limb, placed into water (1.0 mL/40 mg of tissue), heated in a boiling bath for 2 min, andthen homogenized with a Polytron instrument in 6%perchloric acid (1 mL/10 mg of tissue). The homogenatewas centrifuged at 1 050g for 15 min at 4 °C. Thesupernatant was neutralized with 2 M K2CO3 and used astissue extract for the assaying of sorbitol. Sorbitol wasassayed by an enzymatic method in which sorbitoldehydrogenase catalyses the stoichiometric conversion ofNAD by sorbitol to a fluorogenic product, NADH. Thereaction mixture contained 30 mM glycine buffer (pH9.4), 1.3 mM NAD, 1.3 U/mL sorbitol dehydrogenaseand 1.0 mL of the tissue extract in a total volume of 3 mL.After the mixture was allowed to stand for 60 min atroom temperature, the fluorescence intensity was mea-sured at 365 nm excitation wavelength and 430 nm emis-sion wavelength using a fluorospectrophotometer (F3000,Hitachi Ltd., Japan). The sorbitol concentration wasquantitated by comparison with standards of sorbitol. Thesorbitol content in the sciatic nerve of each animal wasexpressed as nmole/wet weight.

The activity of test compounds was expressed as thepercent inhibition of sorbitol accumulation at a givendose, which was calculated according to the followingequation:

% inhibition = (S – T) / (S – N)× 100

1069

where S is the sorbitol content in the sciatic nerve ofuntreated diabetic control rats, T is the sorbitol content inthe sciatic nerve of diabetic rats given test compoundsand N is the sorbitol content in the sciatic nerve ofage-matched non-diabetic control rats.

Acknowledgements

We thank Dr S. Naruto and Dr T. Kadokawa forencouragement throughout this work, and members of theanalytical section for elemental analyses and spectralmeasurements. We also thank Ms. M. Furuichi for mea-surements of pharmacological studies.

References

[1] Hayman S., Kinoshita J.H., J. Biol. Chem. 240 (1965) 877–882.

[2] Kador P.F., Akagi Y., Kinoshita J.H., Metabolism 35 (1986) (suppl.1) 15–19.

[3] Gabbay K.H., New Engl. J. Med. 288 (1973) 831–836.

[4] Akagi Y., Kador P.F., Kuwabara T., Kinoshita J.H., Invest. Ophthal-mol. Vis. Sci. 24 (1983) 1516–1519.

[5] Cohen M.P., Metabolism 35 (1986) (suppl. 1) 55–59.

[6] Varma S.D., Schocket S.S., Richards R.D., Invest. Ophthalmol. Vis.Sci. 18 (1979) 237–241.

[7] Canal N., Comi G., Trends Pharmacol. Sci. 6 (1985) 328–330.

[8] Sestanj K., Bellini F., Fung S., Abraham N., Treasurywala A.,Humber L., Simard-Duquesne N., Dvornik D., J. Med. Chem. 27(1984) 255–256.

[9] Kikkawa R., Hatanaka I., Yasuda H., Kobayashi N., Shigeta Y.,Terashima H., Morimura T., Tsuboshima M., Diabetlogia 24 (1983)290–292.

[10] Mizuno K., Kato N., Matsubara A., Nakano K., Kurono M.,Metabolism 41 (1992) 1081–1086.

[11] Ao S., Shingu Y., Kikuchi C., Takano Y., Nomura K., Fujiwara T. etal., Metabolism 40 (1991) 77–87.

[12] Kador P.F., Kinoshita J.H., Sharpless N.E., J. Med. Chem. 28 (1985)841–849.

[13] Lee Y.S., Pearlstein R., Kador P.F., J. Med. Chem. 37 (1994)787–792.

[14] Emmott P., Livingstone R., J. Chem. Soc. (1957) 3144–3148.

[15] Gresham T.L., Jansen J.E., Shaver F.W., Bankert R.A., Beears W.L.,Prendergast M.G., J. Am. Chem. Soc. 71 (1949) 661–663.

[16] Baranov S.N., Komaritsa I.D., Khim. Geterotsikl. Soedin. Akad.Nauk Latv. SSR 1 (1965) 69–73.

[17] Fresneau P., Cussac M., Morand J.M., Szymonski B., Tranqui D.,Leclerc G., J. Med. Chem. 41 (1998) 4706–4715.

[18] Ishida T., In Y., Inoue M., Ueno Y., Tanaka C., Tetrahedron Lett. 30(1989) 959–962.

[19] Clements R.S., Morisson Jr. A.D., Winegrad A.I., Science 166(1969) 1007–1008.

[20] Hullar T.L., Failla D.L., J. Med. Chem. 12 (1969) 420–424.

1070

Short communication

Synthesis and antimycobacterial activity of some isonicotinoylhydrazones

Maria T. Coccoa*, Cenzo Congiua, Valentina Onnisa, Maria C. Pusceddub, Maria L. Schivob,Alessandro De Logub

aDipartimento di Tossicologia, Università di Cagliari, Viale A. Diaz 182 - 09126 Cagliari, ItalybDipartimento di Scienze Chirurgiche e Trapianti d’Organo, Sezione di Microbiologia e Virologia, Università di Cagliari, Via Palabanda

14 - 09123 Cagliari, Italy

(Received 15 February 1999; accepted 8 June 1999)

Abstract – A series of isonicotinoylhydrazones2 were prepared by addition of some aryloxyacetonitriles with isonicotinoylhydrazine in basicmedium. These compounds have been further reacted with pyridinecarboxaldehydes to give the corresponding pyridylmethyleneaminoderivatives3–5. The new synthesized hydrazones and their pyridylmethyleneamino derivatives were tested for their activity againstmycobacteria, Gram-positive and Gram-negative bacteria. The cytotoxicity was also tested. Several compounds showed a good activity againstMycobacterium tuberculosisH37Rv and some isonycotinoylhydrazones2 showed a moderate activity against a clinically isolatedM.tuberculosiswhich was isoniazid resistant. © 1999 Éditions scientifiques et médicales Elsevier SAS

aminohydrazone derivatives / synthesis / antimycobacterial activity

1. Introduction

At present, tuberculosis is considered, by the WorldHealth Organisation, to be the most important chroniccommunicable disease in the world [1, 2, 3]. Over thepast decade, tuberculosis has re-emerged both in indus-trial and developing countries. The emergence of AIDS,decline of socioeconomic standards and a reduced em-phasis on tuberculosis control programs contribute to thedisease’s resurgence in industrialised countries [4]. Inmost developing countries, although the disease hasalways been endemic, its severity has increased becauseof the global HIV endemic and extensive social restruc-turing due to rapid industrialisation and conflicts.

Further contributing to the increased morbidity is theemergence of new strains ofM. tuberculosisresistant tosome or all current antitubercular drugs [5, 6]. Amongthese, isoniazid (INH), an inexpensive and relatively safedrug, continues to be well established for the treatment oftuberculosis. The mechanism of action of INH, as well asthe mechanism conferring INH resistance, are complexand not completely understood. However, several studies

suggest that INH inhibits the biosynthesis of cell wallmycolic acids, thereby making the mycobacteria suscep-tible to reactive oxygen radicals and other environmentalfactors [7]. INH is active against theMycobacteriumcomplex (Mycobacterium tuberculosis, M. bovis, M.africanum and M. microti) at MICs ranging from0.025–0.05µg/mL [8], while at a higher concentration(MIC of 500 µg/mL) it inhibits the growth of othermicroorganisms such as the opportunistsKlebsiella, Ser-ratia andEnterobacter.

For these reasons the antimycobacterial pharmacoph-ore moiety of INH is introduced in several molecules toimprove their activity against Mycobacteria.

On the other hand aminohydrazone derivatives, struc-turally correlated to INH, have been described for their invitro antimycobacterial activity and some of these com-pounds exhibited inhibitory activity toward a humanstrain of M. avium resistant to the primary drugs INH,rifampicin and ofloxacin [9, 10]. As a part of our studieson aminohydrazone derivatives [11, 12], we becameinterested in a new series of 2-amino-2-(iso-nicotinoylhydrazono)ethyl aryl ethers, assuming that theisonicotine hydrazonic moiety is an important pharma-cophore to antimycobacterial activity. Here, we report the*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1071−1076 1071© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

synthesis of these aminohydrazone derivatives and theevaluation of their activity against mycobacteria. Activityagainst Gram-positive and Gram-negative bacteria aswell as their cytotoxicity were also evaluated.

2. Chemistry

Isonicotinoylhydrazones2 and their pyridylmethylene-amino derivatives3, 4 and5 described in this study areshown intables IandII , and a reaction sequence for theirpreparation is outlined infigure 1. The starting aryloxy-acetonitriles1 were prepared according to known proce-dures [13]. Aryloxyacetonitriles1 were converted to hy-drazones2 by addition of isonicotinoylhydrazine in thepresence of catalytic amounts of sodium ethoxide inanhydrous ethanolic solution.

Heating compounds2 with pyridinecarboxaldehyde inethanolic solution in the presence of piperidine affordedpyridylmethyleneamino derivatives3–5.

The structures for all new obtained compounds2–5were determined by examining their IR and1H-NMRspectra as well as by elemental analyses (tables IandII ).

3. Microbiology

Antimycobacterial activity was investigated againstM.tuberculosis H37Rv ATCC 25584,M. avium ATCC19421 andM. FortuitumATCC 9820, andM. tuberculosisresistant to isoniazid (INH-R) isolated from a patient withan active clinical infection treated at the Università diCagliari. The antimicrobial activity of compounds2–5was also evaluated against Gram-positive and Gram-negative bacteria isolated from clinical specimens andCandida albicansATCC E10931. At the same time, cellcytotoxicity of all compounds was tested in vitro on Verocells.

4. Results and discussion

With the exception of compound2f which showedtoxicity at concentrations higher than 62.5µg/mL, thetested compounds exhibited high values of maximumnon-toxic dose (MNTD) on Vero cells, ranging from250µg/mL for 4j, to 500–1 000µg/mL for the othermembers of the series.

Table I. Chemical and spectral data of compounds2.

Compound R Yield(%)

M.p. (°C)(recryst. solvent)

IR (Nujol)ν (cm–1)

1H-NMR (DMSO-d6/TMS) δ (ppm)

2a H 77 174–176 3 410, 3 280,3 020, 1 680,1 590, 1 570

4.50 (s, 2H, CH2), 6.63 (s, 2H, NH2), 6.97–7.24 (m, 5H, Ar ), 7.72, 8.63(m, 4H, Py), 10.11 (br s, 1H, NH)

(CH3CN)

2b 2-CH3 74 178–180 3 270, 3 310,1 660, 1 605

2.17 (s, 3H, CH3), 4.51 (s, 2H, CH2), 6.59 (s, 2H, NH2), 6.82–7.10 (m,4H, Ar), 7.73, 8.63 (m, 4H, Py), 10.13 (br s, 1H, NH)(EtOH)

2c 3-CH3 68 162–164 3 170, 3 060,1 680, 1 650

2.23 (s, 3H, CH3), 4.49 (s, 2H, CH2), 6.61 (s, 2H, NH2), 6.75–7.13 (m,4H, Ar), 7.73, 8.60 (m, 4H, Py), 10.11 (br s, 1H, NH)(2-PrOH)

2d 4-CH3 66 208–210 3 270, 3 080,1 680, 1 650

2.17 (s, 3H, CH3), 4.46 (s, 2H, CH2), 6.60 (s, 2H, NH2), 6.72–7.06 (m,4H, Ar), 7.72, 8.62 (m, 4H, Py), 10.09 (br s, 1H, NH)(CH3CN)

2e 2-OCH3 72 175–177 3 260, 3 100,3 020, 1 670

3.73 (s, 3H, CH3), 4.47 (s, 2H, CH2), 6.58–6.94 (m, 6H, Ar + NH2), 7.72,8.63 (m, 4H, Py) 10.11 (s, 1H, NH)(EtOH)

2f 3-OCH3 99 158–160 3 270, 3 120,1 675, 1 635

3.68 (s, 3H, CH3), 4.49 (s, 2H, CH2), 6.49–7.18 (m, 6H, Ar + NH2), 7.72,8.61 (m, 4H, Py), 10.10 (brs, 1H,NH)(CH3CN)

2g 4-OCH3 68 193–195 3 360, 1 675,1 590, 1 570

3.64 (s, 3H, CH3), 4.44 (s, 2H, CH2), 6.60 (s, 2H, NH2), 6.80–6.94 (m,4H, Ar), 7.72, 8.62 (m, 4H, Py), 10.08 (br s, 1H, NH)(CH3CN)

2h 2-Cl 68 188–190 3 410, 3 320,1 690, 1 620

4.58 (s, 2H, CH2), 6.62 (s, 2H, NH2), 6.94–7.38 (m, 4H, Ar), 7.72, 8.64(m, 4H, Py), 10.16 (br s, 1H, NH)(EtOH)

2i 3-Cl 64 180–182 3 410, 3 080,1 680, 1 585

4.54 (s, 2H, CH2), 6.64 (s, 2H, NH2), 6.95–7.31 (m, 4H, Ar), 7.72, 8.63(m, 4H, Py), 10.12 (br s, 1H, NH)(EtOH)

2j 4-Cl 99 236–238 3 240, 3 080,1 680, 1 640

4.51 (s, 2H, CH2), 6.64 (s, 2H, NH2), 7.01–7.31 (m, 4H, Ar), 7.72, 8.63(m, 4H, Py), 10.11 (br s, 1H, NH)(CH3CN)

2k 4-NO2 77 228–230 3 360, 3 280,3 160, 3 080,1 680, 1 640

4.70 (s, 2H, CH2), 6.74 (s, 2H, NH2), 7.18–7.73 (m, 4H, Ar), 8.20, 8.63(m, 4H, Py), 10.16 (br s, 1H, NH)

(EtOH)

2l 4-NHCOCH3 73 191–193 3 220, 3 060,1 660, 1 595

1.95 (s, 3H, CH3), 4.46 (s, 2H, CH2), 6.61 (s, 2H, NH2), 6.90–7.42 (m,4H, Ar), 7.72, 8.62 (m, 4H, Py), 9.77 (s, 1H, NH), 10.09 (br s, 1H, NH).(EtOH)

1072

The tested compounds showed activity against myco-bacteria with MIC values ranging from 6.25µg/mL toover 100µg/mL. MIC values as well as the results ofcytotoxicity assays are reported intable III. The mostactive compounds againstM. tuberculosisH37Rv were2d, 2j and2k (MIC 6.25µg/ml). However, a significant

activity was also exhibited by compounds2c, 2f, 2g and2i (MIC 12.5µg/mL) and 2b, 2h and 4a (MIC 25 µg/mL). The most effective compounds in the inhibition ofgrowth of M. tuberculosisINH-R were 2g, 2j and 2k(MIC 25 µg/mL). All the tested compounds showed ahigher activity againstM. tuberculosisH37Rv thanM.

Table II. Chemical and spectral data of compounds3, 4 and5.

Compound R Pyr Yield(%)

M.p. (°C)(recryst. solvent)

IR (Nujol)ν (cm–1)

1H-NMR (DMSO-d6 /TMS) δ (ppm)

3a H 2-pyridyl 42 168–169 3 050, 1 610,1 590, 1 540

4.74 (s, 2H, CH2), 6.64 (s, 1H, CH), 6.93–7.81 (m, 10H, Ar), 8.54(s, 1H, NH), 8.63 (m, 3H, Py)(Benzene)

4a H 3-pyridyl 55 164–165 3 180, 3 030,1 615, 1 600,1 550

4.81 (s, 2H, CH2), 6.75 (s, 1H, CH), 6.93–8.62 (m, 13H, Ar), 8.68(s, 1H, NH)

(2-PrOH)

5a H 4-pyridyl 63 190–191 3 070, 1 620,1 595

4.80 (s, 2H, CH2), 6.69 (s, 1H, CH), 6.93–8.64 (m, 13H, Ar), 8.71(s, 1H, NH)(Benzene)

5b 2-CH3 4-pyridyl 48 162–163 3 360, 3 020,1 620, 1 590

2.13 (s, 3H, CH3), 4.81 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.83–8.64(m, 13H, Ar + NH)(Toluene)

4c 3-CH3 3-pyridyl 57 124–125 3 060, 1 610,1 540

2.24 (s, 3H, CH3), 4.81 (s, 2H, CH2), 6.78 (s, 1H, CH), 6.80–8.68(m, 13H, Ar + NH)(Toluene)

5c 3-CH3 4-pyridyl 59 174–175 3 060, 1 620,1 540

2.23 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.78–8.65(m, 12H, Ar), 8.71 (s, 1H, NH)(Toluene)

4d 4-CH3 3-pyridyl 64 152–153 3 020, 1 620,1 600, 1 540

2.19 (s, 3H, CH3), 4.77 (s, 2H, CH2), 6.74 (s, 1H, CH), 6.86–8.74(m, 13H, Ar + NH)(Toluene)

5d 4-CH3 4-pyridyl 56 159–160 3 060, 1 620,1 560

2.22 (s, 3H, CH3), 4.78 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.88–8.66(m, 12H, Ar), 8.72 (s, 1H, NH)(Toluene)

4e 2-OCH3 3-pyridyl 68 134–135 3 080, 3 020,1 630, 1 590

3.74 (s, 3H, CH3), 4.80 (s, 2H, CH2), 6.79 (s, 1H, CH), 6.87–8.65(m, 13H, Ar + NH)(Toluene)

5e 2-OCH3 4-pyridyl 67 129–130 1 630, 1 600 3.72 (s, 3H, CH3), 4.76 (s, 2H, CH2), 6.71 (s, 1H, CH), 6.83–8.64(m, 13H, Ar + NH)(Toluene)

4f 3-OCH3 3-pyridyl 50 137–138 3 160, 3 060,1 640, 1 590

3.66 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.51–9.13 (m, 14H, Ar +NH + CH)(Toluene)

5f 3-OCH3 4-pyridyl 52 157–158 3 140, 3 080,1 620, 1 600

3.66 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.70–8.63 (m, 13H, Ar +CH), 8.71 (s, 1H, NH)(Toluene)

4g 4-OCH3 3-pyridyl 72 123–124 3 150, 3 020,1 620, 1 580

3.67 (s, 3H, CH3), 4.75 (s, 2H, CH2), 6.75 (s, 1H, CH), 6.83–8.64(m, 12H, Ar), 8.67 (s, 1H, NH)(Toluene)

5g 4-OCH3 4-pyridyl 60 129–130 3 100, 1 630,1 620, 1 595

3.65 (s, 3H, CH3), 4.73 (s, 2H, CH2), 6.68 (s, 1H, CH), 6.81–8.64(m, 12H, Ar), 8.69 (s, 1H, NH)(Toluene)

4h 2-Cl 3-pyridyl 60 173–174 3 350, 1 630,1 610, 1 550

4.92 (s, 2H, CH2), 6.78 (s, 1H, CH), 6.95–8.65 (m, 13H, Ar + NH)(Toluene)

5h 2-Cl 4-pyridyl 50 184–185 3 340, 3 020,1 620, 1 590,1 540

4.90 (s, 2H, CH2), 6.71 (s, 1H, CH), 6.95–8.63 (m, 12H, Ar), 8.68(s, 1H, NH)

(Toluene)

4i 3-Cl 3-pyridyl 51 139–140 3 060, 1 620,1 590, 1 550

4.88 (s, 2H, CH2), 6.76 (s, 1H, CH), 6.96–8.73 (m, 13H, Ar + NH)(Toluene)

5i 3-Cl 4-pyridyl 67 139–140 3 060, 1 620,1 590, 1 540

4.88 (s, 2H, CH2), 6.72 (s, 1H, CH), 6.97–8.66 (m, 12H, Ar), 8.74(s, 1H, NH)(Toluene)

4j 4-Cl 3-pyridyl 20 164–165 3 100, 3 020,1 630, 1 610,1 550

4.84 (s, 2H, CH2), 6.75 (s, 1H, CH), 7.00–8.62 (m, 12H, Ar), 8.67(s, 1H, NH)

(Ethyl acetate)

5j 4-Cl 4-pyridyl 51 159–160 1 600, 1 550 4.82 (s, 2H, CH2), 6.69 (s, 1H, CH), 6.99–8.64 (m, 12H, Ar), 8.71(s, 1H, NH)(Toluene)

4k 4-NO2 3-pyridyl 65 189–190 1 620, 1 580,1 500

5.04 (s, 2H, CH2), 6.79 (s, 1H, CH), 7.20–8.63 (m, 12H, Ar), 8.74(s, 1H, NH)(2-PrOH)

5k 4-NO2 4-pyridyl 62 184–185 1 600, 1 580,1 540

5.02 (s, 2H, CH2), 6.73 (s, 1H, CH), 7.19–8.64 (m, 12H, Ar), 8.78(s, 1H, NH)(2-PrOH)

1073

tuberculosisINH-R, with the exception of compounds4j,5g and 5j that were inactive againstM. tuberculosisH37Rv (MIC > 100µg/mL) but showed a weak activityagainstM. tuberculosisINH-R (MIC 50µg/mL). Isony-cotinoylhydrazones2g, 2j and 2k were the most activecompounds in the whole series, with potencies superior toINH againstM. tuberculosisINH-R. Although none ofthe tested compounds exhibited a significant activityagainstM. fortuitum, we observed a poor activity by2gand2i (MIC 100 µg/mL) againstM. avium.

Compounds2–5were inactive at inhibiting the growthof the Gram-negative species tested. A weak activityagainstStaphyloccocus aureuswas shown by compounds2h, 2i and 2j (MIC 100 µg/mL) and by compound3aagainstC. albicans(MIC 100 µg/mL).

From a first examination of these results, it appears thatcompounds2, containing the NH2 group, show betteractivity againstM. tuberculosisH37Rv andM. tubercu-losisINH-R with respect to derivatives3, 4 and5. Furtherstudies to acquire more information about structure-activity relationships are in progress in our laboratories.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Kofler hot stageand are uncorrected. IR spectra were recorded on Nujolmulls between salt plates in a Perkin-Elmer 398 spectro-photometer.1H-NMR spectra were recorded on a VarianUnity 300 spectrometer. Elemental analyses were carriedout with a Carlo Erba Model 1106 Elemental Analyzer.

5.1.1. 2-Amino-2-(isonicotinoylhydrazono)ethyl arylethers2

General Procedure: to a stirred solution of NaOEt(0.005 mol) in dry EtOH (5 mL) the appropriate aryloxy-acetonitrile 1 (0.05 mol) was added. The mixture wasstirred at room temperature for 1 h and after cooling to0 °C, an ethanolic solution of INH (6.8 g, 0.05 mol) wasadded dropwise. The resulting solution was stirred at the

Figure 1. Synthetic pathways to compounds2, 3, 4 and5.

Table III. Cytotoxicity and antimycobacterial activity of com-pounds2–5.

Com-pound

MNTD(µg/mL)

MIC (µg/mL)

Verocells

M. tuberculosisH37Rv

M. tuberculosisINH-R

M. aviumATCC 19421

ATCC 25584

2b 1 000 25 100 > 1002c 1 000 12.5 50 > 1002d 500 6.25 50 > 1002f 62.5 12.5 100 > 1002g 500 12.5 25 1002h 500 25 100 > 1002i 1 000 12.5 50 1002j 500 6.25 25 > 1002k 1 000 6.25 25 > 1002l 1 000 50 > 100 > 1004a 1 000 25 50 > 1004j 250 > 100 50 > 1004k 1 000 50 100 > 1005c 1 000 50 > 100 > 1005g 500 > 100 50 > 1005i 500 50 50 > 1005j 500 > 100 50 > 100INH 1 000 0.06 50 100

1074

same temperature for 2 h and then allowed to reach roomtemperature and to stand overnight. The formed precipi-tate was collected by filtration and recrystallised from thesolvent shown intable I.

5.1.2. 2-(Pyridylmethyleneamino)-2-(isonicotinoylhydra-zono)ethyl aryl ethers3, 4 and5

General procedure: to a solution of compounds2(1.85 mmol) and 2-, 3- or 4-pyridinecarboxaldehyde(1.85 mmol) in EtOH (15 mL), a few drops of piperidinewere added. After heating at reflux for 6 h the solvent wasremoved in vacuo and the residue treated with H2O(20 mL) and extracted with CHCl3 (3 × 10 mL). Theorganic layers were dried (Na2SO4) and the solventevaporated to give the crude3, 4 and5 that were purifiedby crystallisation from the solvent indicated intable II.

5.2. Microbiology

For the antimicrobial and cytotoxicity studies, com-pounds were dissolved in DMSO at 10 mg/mL and keptat –20 °C. The working solutions were prepared in thesame medium used for the tests. To avoid interference bythe solvent [14], the highest DMSO concentration was1%. The MICs of various compounds against Gram-positive, Gram-negative andC. albicanswere determinedby a standard broth serial dilution method [15, 16].

5.2.1. Antibacterial assayThe antimicrobial activity of compounds2–5 was

evaluated against six Gram-positive and five Gram-negative species isolated at the Sezione di microbiologiae virologia, Dipartimento di Scienze Chirurgiche e Tra-pianti d’Organo, from clinical specimens obtained frompatients treated at the Università di Cagliari. In particular,compounds2–5 were tested againstStaphylococcus au-reus (isolated from urine),S. epidermidis(isolated fromurine by suprapubic aspirate),Streptococcus agalactiae(isolated from vaginal swab),S. faecalis(isolated fromurine), Bacillus licheniformis(isolated from blood),B.subtilis (isolated from eye swab),Escherichia coli(iso-lated from urine),Pseudomonas aeruginosa(isolatedfrom urine), Salmonella typhi(isolated from faeces),Proteus mirabilis (isolated from urine) andKlebsiellapneumoniae(isolated from urine). Tests with Gram-positive and Gram-negative bacteria were carried out inMueller Hinton broth (Difco). The compounds werediluted in the test medium to obtain final concentrationsranging from 100–0.19µg/mL. Tubes were inoculatedwith 1 × 105 cells/mL and were incubated at 37 °C for 18or 24 h. The effects on the growth of mycobacteria wasinvestigated againstM. tuberculosis H37Rv ATCC25584,M. aviumATCC 19421,M. fortuitumATCC 9820

and M. tuberculosisresistant to isoniazid (INH-R) iso-lated from a patient with an active clinical infection. Thedetermination of MICs against mycobacteria were carriedout by the two-fold agar dilution method [17] using 7H11agar (Difco Laboratories) containing compounds2–5 atconcentrations that ranged between 100–0.19µg/mL, onwhich 100µL of the test bacterial suspension werespotted. Suspensions to be used for drug susceptibilitytesting were prepared from 7H9 broth cultures supple-mented with 10% OADC (oleic acid-albumin-dextrose-catalase) enrichment (Difco Laboratories) and 0.05%(v/v) Tween 80 to avoid clumping. Cells were thenwashed, suspended in saline, shaken and sonicated in abath type ultrasonicator (output power 80 W) untilvisibile clumps were disrupted (usually from 15–30 s).Suspensions were then diluted in saline to a turbidity ofno. 1 McFarland (M. tuberculosis) or no. 0.5 McFarland(M. aviumand M. fortuitum) and then diluted to obtaininocula of 3× 105 cells per well ofM. tuberculosisand1.5 × 104 cells per well ofM. aviumand M. fortuitum.The MICs of the compounds were determined after 7(M. fortuitum) or 21 (slow growers) days of cultivation at37 °C in a CO2 (5% CO2/95% humidified air) incubator.

5.2.2. Antifungal assayFor the evaluation of the antifungal activity,Candida

albicansATCC E10931 was employed. Antifungal activ-ity againstC. AlbicansATCC E10231 was evaluated inyeast extract peptone dextrose medium (Difco) [18].

5.2.3. In vitro cytotoxicity assayCell cytotoxicity of compounds2–5 was tested in vitro

by two methods. In the first method, RPMI 1640 medium(Gibco) with 2% foetal calf serum (FCS, Gibco) alone, orRPMI 1640 with 2% FCS containing compounds atconcentrations ranging from 1 000–62.5µg/mL, wereinoculated onto cultures of Vero cells in 6 well tissueculture plates. The cells were observed daily for 6 daysfor any sign of cell cytotoxicity compared with thecontrols. In the second method, a cell viability assaypreviously reported [19, 20] was used. Monolayers ofVero cells in 96 multiwell plates were incubated with thetesting compounds at concentrations of 1 000–62.5µg/mL in RPMI 1640 with 5% FCS for 48 h and the mediumreplaced with 50µL of 1 mg/mL solution of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT) in RPMI without phenol red. Cells were incubatedat 37 °C for 3 h, the untransformed MTT removed and0.04 N HCl isopropanolic solution (50µL) was added toeach well. After a few minutes at room temperature toensure that all crystals were dissolved, the plates were

1075

read using an automatic plate reader with a 650 nm testwavelength and a 690 nm reference wavelength.

Acknowledgements

Financial support from the Ministero della Università eRicerca Scientifica e Tecnologica is gratefully acknowl-edged.

References

[1] Bloom B.R., Murray C.J.L., Science 257 (1992) 1055–1064.

[2] World Health Organization, Bridging the gaps: the world healthreport, World Health Organization, Geneva, 1995.

[3] World Health Organization report on TB epidemic, Global TBprogramme, World Health Organization, Geneva, 1997.

[4] Barnes P., Blotch A.B., Davidson B.T., Snyder Jr. D.E., N. Engl. J.Med. 324 (1991) 1644–1650.

[5] Snyder Jr. D.E., Roper W.L., N. Engl. J. Med. 326 (1992) 703–705.

[6] Freiden T.E., Sterling T., Pablos-Mendez A., Kilburn J.O., CauthenJ.O., Dooley S.W., N. Engl. J. Med. 328 (1993) 521–526.

[7] Sherman D.R., Mdluli K., Hickey M.J., Arain T.M., Morris S.L.,Barry III C.E., Kendall Stover C., Science 272 (1996) 1641–1643.

[8] Mandell G.L., Sande M.A., in: Goodman and Gilman’s – ThePharmacological Basis of Therapeutics, 8th Edition, PergamonPress, New York, 1984, p. 1147.

[9] Banfi E., Mamolo M.G., Vio L., Predominato M., J. Chemother. 5(1993) 164.

[10] Mamolo M.G., Vio L., Banfi E., Farmaco 51 (1996) 65–70.

[11] Bernard A.M., Cocco M.T., Congiu C., Onnis V., Piras P.P., J.Heterocycl. Chem. 34 (1997) 1283–1290.

[12] Bernard A.M., Cocco M.T., Congiu C., Onnis V., Piras P.P.,Synthesis (1998) 317–320.

[13] McManus J.R., Herbst R.M., J. Org. Chem. 24 (1959) 1464–1467.

[14] Jagannath C., Reddy V.M., Gangadharam P.R., J. Antimicrob.Chemother. 35 (1995) 381–390.

[15] Trhupp L.D., in: Lorian V. (Ed.), Antibiotics in Laboratory Medi-cine, Williams & Wilkins, Baltimore, 1986, pp. 93–150.

[16] Washington II J.A., Sutter V.L., in: Lennette E.H., Spaulding E.H.,Truant J.I. (Eds.), Manual of Clinical Microbiology, AmericanSociety of Microbiology, Washington, 1980, pp. 533–544.

[17] Hopewell P., Cynamon M., Starke J., Iseman M., O’Brein R., Clin.Infec. Dis. 15 (suppl 1) (1992) S282–S295.

[18] Pfaller M.A., Rinaldi M.G., Galgiani J.N., Bartlett M.S., Body B.A.,Espinel-Ingroff A. et al., Antimicrob. Agents Chemother. 34 (1990)1648–1654.

[19] Denizot F., Lang R., J. Immunol. Methods 89 (1986) 271–277.

[20] Mosmann T., J. Immunol. Methods 65 (1983) 55–63.

1076

Short communication

Structure-activity relationships of quaternary protoberberine alkaloids having anantimalarial activity

Kinuko Iwasaa*, Yumi Nishiyamaa, Momoyo Ichimarua, Masataka Moriyasua, Hye-Sook Kimb,Yusuke Watayab, Takao Yamoric, Turuo Takashid, Dong-Ung Leee

aKobe Pharmaceutical University, 4-19-1 Motoyamakita, Higashinada-ku, Kobe 658-8558, JapanbFaculty of Pharmaceutical Sciences, Okayama University, Tsushima, Okayama 700-8530, Japan

cDivision of Experimental Chemotherapy, Cancer Chemotherapy Center, 1-37-1 kamiikebukuro, Toyoshima-ku, Tokyo 170-8455, JapandInstitute of Molecular and Cellular Biosciences, University of Tokyo, 1-1-1 yayoi, bunkyo-ku, Tokyo 113-0000, Japan

eDepartment of Biochemistry, College of Natural Science, Dongguk University, Kyongju Kyongbuk 780-714, Korea

(Received 20 April 1999; revised 28 June 1999; accepted 30 June 1999)

Abstract – Seventeen quaternary protoberberine alkaloids related to berberine1 were tested for antimalarial activity in vitro againstPlasmodium falciparumand structure-activity relationships are proposed. The activity of the protoberberine alkaloids was influenced by thetype of the oxygen substituents on rings A, C and D and the position of the oxygen functions on ring D. The position of the oxygen functionson ring D and the type of the oxygen substituents at the C-13 position (ring C) strongly influenced the activity. Shifting the oxygen functionsat C-9 and C-10 to C-10 and C-11 on ring D resulted in a significant increase in the activity. Compounds bearing a methylenedioxy functionat C-2 and C-3 (ring A) or C-9 and C-10 (ring D) showed higher activity than those which have methoxy groups at the same positions.Introduction of a methoxy group into the C-1 position (ring A) decreased the activity. Replacement of a hydroxy group at C-2 or C-3 (ringA) by a methoxy group led to a reduction in the activity. Displacement of a hydroxy function at C-13 (ring C) by the oxygen substituents suchas OMe, OEt, OCOOEt, and OCON(Me)2 reduced the activity. In the same replacement at C-9 (ring D), the activity depended upon the typeof the oxygen function. Six protoberberines displayed more potent activity than berberine1. The activity decreased in the order:10, 11, 17and18 > 7 and8 > 1. © 1999 Éditions scientifiques et médicales Elsevier SAS

in vitro antimalarial activity / structure-activity relationships / protoberberinium salts

1. Introduction

In Japan, and other Asian countries, berberine1 and theextracts ofPhellodendri cortex(Obaku in Japanese) andCoptidis rhizoma(Oren in Japanese) are used in thetreatment of diarrhoea and other gastrointestinal diseases.Berberine and its relatives exhibit several types of bio-logical activity [1]. We found that some of the 8- and13-alkylderivatives of berberine1 and palmatine2 pos-sessed antimicrobial and antimalarial activities [2–6]. Wehave recently shown that a 13-hydroxy derivative of1had similar antimalarial activity to that of1 [6].

In the present study, the antimalarial activity of 17protoberberinium salts, in which the type and/or positionof the oxygen substituents such as hydroxy, methylene-dioxy, methoxy, ethoxycarbonyloxy, andN, N-dimethyl-carbamoyloxy functions on the aromatic rings A, C and Dare different from1 or 2 (table I), were examined usingthe selectivity indices as an indication of the activity. Theresults on the study of the structure-antimalarial activityrelationships of the tested alkaloids are discussed.

2. Chemistry

The protoberberinium salts3–13 and17–20 have beenprepared or isolated from natural sources previously [2,5, 7, 8]. Palmatrubine14 was prepared according to themethod applied to the synthesis of5 [2]. The carboethoxy*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1077−1083 1077© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

and N,N-dimethylcarbamoyl derivatives of berberrubine5 and 13-hydroxyberberine17 (15and21, and16and22,respectively) were prepared by the reaction with ethylchloroformate andN,N-dimethylcarbamoyl chloride, re-spectively, in the presence of triethylamine or pyridine.1H-NMR (table II), LSIMS (table III), and HR-LSIMS(table III) data were consistent with the assigned struc-tures.

3. Results and discussion

Protoberberinium salts6–22 were tested in vitroagainst human malariaPlasmodium falciparumFCR-3.The antimalarial activity of each compound was deter-mined as a percentage reduction. The compound concen-tration required to inhibit cell growth by 50% wasexpressed as IC50 (table IV). From the evaluation of thetoxicity of the compounds for mammalian cells, theconcentration causing a 50% reduction of cell growth(GI50) of mouse mammary FM3A cells, a model of thehost, was determined (table IV). The GI50/IC50 ratios forthe compounds were calculated as selectivity indices(table IV). These ratios were used as an evaluation ofantimalarial activity. The results are presented intable IV.

Table I. Protoberberinium salts.

R1 R2 R3 R4 R5 R6 R7

1 H OCH2O OMe OMe H H2 H OMe OMe OMe OMe H H3 H OCH2O OMe OMe H Me4 H OMe OMe OMe OMe H Me5 H OCH2O OH OMe H H6 OMe OCH2O OMe OMe H H7 H OCH2O OCH2O H H8 H OCH2O OCH2O H Me9 H OMe OMe OCH2O H Me10 H OCH2O H OMe OMe H11 H OCH2O H OMe OMe Me12 H OMe OH OMe OMe H H13 H OH OMe OMe OMe H Me14 H OMe OMe OH OMe H H15 H OCH2O OCOOEt OMe H H16 H OCH2O OCON(Me)2 OMe H H17 H OCH2O OMe OMe H OH18 H OCH2O OMe OMe H O–

19 H OCH2O OMe OMe H OMe20 H OCH2O OMe OMe H OEt21 H OCH2O OMe OMe H OCOOEt22 H OCH2O OMe OMe H OCON(Me)2

Table II. 1H-NMRa data of the protoberberinium salts14–16, 21 and22.C-9 or C-13 C-9 or C-13 H-5 H-6 2-OMe 3-OMe 10-OMe 9-OMe OCH2O H-4 H-1 H-11 H-12 H-13 H-8

OCOOCH2CH3 OCON(CH3)2

14 3.25 4.87 4.00 3.92 4.05 7.00 7.58 7.96 7.71 8.65 9.732H, t (6.0) 2H, t (6.0) 3H, s 3H, s 3H, s 1H, s 1H, s 1H, d (9.0) 1H, d (9.0) 1H, s 1H, s

15 4.40 1.43 3.26 4.94 4.11 6.11 6.97 7.69 8.24* 8.21* 8.82 9.792H, q (7.0) 3H, t (7.0) 2H, t (6.5) 2H, t (6.5) 3H, s 2H, s 1H, s 1H, s 1H, d (9.0) 1H, d (9.0) 1H, s 1H, s

16 3.09 3.31 3.26 4.95 4.09 6.11 6.97 7.68 8.19* 8.18* 8.80 9.733H, s 3H, s 2H, t (6.5) 2H, t (6.5) 3H, s 2H, s 1H, s 1H, s 1H, d (9.0) 1H, d (9.0) 1H, s 1H, s

21 4.31 1.31 3.21 4.95 4.13* 4.24* 6.12 7.03 7.67 8.19* 8.01* 9.882H, q (7.0) 3H, t (7.0) 2H, t (6.0) overlap 3H, s 3H, s 2H, s 1H, s 1H, s 1H, d (9.0) 1H, d (9.0) 1H, s

22 2.98 3.42 3.20 4.92 4.12* 4.22* 6.11 7.00 7.63 8.17 7.96 9.843H, s 3H, s 2H, t (6.0) overlap 3H, s 3H, s 2H, s 1H, s 1H, s 1H, d (9.0) 1H, dd (9.0, 0.5) 1H, d (0.5)

aCoupling constants (Hz in parentheses).*Assignments may be interchanged.

Table III. Mass spectral data and melting points of the protober-berinium salts (15, 16, 21 and22).

Com-pound

M.p. °C(dec)

Formula LSIMS HR-LSIMSm/z [M – Cl]+ calcd. found

15 160–166 C22H20NO6 394 394.1289 394.131016 210–214 C22H21N2O5 393 393.1449 393.145821 155–158 C23H22NO7 424 424.1395 424.140422 194–200 C23H23N2O6 423 423.1556 423.1565

1078

Introduction of a methoxy group at the C-1 position(ring A) of 1 caused a decrease in antimalarial activity(compare6 with 1) (table IV). Replacement of methoxygroups at the C-9 and C-10 positions (ring D) of1 or 3 bya methylenedioxy group increased the activity (compare7 with 1 or 8 with 3) (table IV) and the same replacementat the C-2 and C-3 positions (ring A) of9 resulted in anincrease in the activity (compare9 with 8) (table IV).

Shifting methoxy groups at C-9 and C-10 of1 or 3 toC-10 and C-11 caused a significant increase in the activity(compare10 with 1 or 11 with 3) (table IV). Displace-ment of a methoxy group at C-3 (ring A) of2 by ahydroxy function increased the activity and the samereplacement at the C-2 position (ring A) of4 alsoincreased the activity (compare12 with 2 and13 with 4)(table IV). Substitution of a methoxy group at C-9 (ringD) of 2 by a hydroxy group increased the activity(compare14with 2) (table IV), though the same displace-ment at C-9 of1 resulted in a decrease in the activity(compare5 with 1) as shown in the previous data [6].Replacement of a hydroxy group at C-9 (ring D) of5 by

OCOOEt did not notably change its activity (compare15with 5) (table IV), whereas replacing a hydroxy functionat C-9 of 5 by OCON(Me)2 caused an increase in theactivity (compare16 with 5) (table IV).

We have found that the inhibitory effect of 13-hydroxyberberine17 was comparable to that of berberine1 [6]. However, the inhibitory effects of phenolbetaineform 18 have been reported to be much less than that of1 [9]. Thus, a comparison of the inhibitory effect between17 and 18 was carried out. As a result, both17 and 18displayed a similar inhibitory effect in vitro. Substitutionof a hydroxy group at the C-13 position of17 by OMe orOEt or OCOOEt or OCON(Me)2 showed a reduction inthe activity (compare19–22 with 17) (table IV).

Among the tested salts,7, 8, 10, 11, 17 and 18exhibited much more potent activity (selectivity indexover 50) than berberine1.

The selectivity indexes (table IV) of 1, 7, 10 and 17were also determined using the 50% growth-inhibitoryconcentration (GI50, in tableV) derived from the re-sults [10] of in vitro cytotoxity tests on 38 human tumour

Table IV. In vitro antimalarial activity of protoberberinium salts1–22.

50% growth inhibition (µM) Selectivity indexes Distribution of selectivity indexesPlasmodium falciparum mouse mammary

cellsmouse mammary cells/Plasmodium falciparum

human tumour cells/Plasmodium falciparum

FCR-3 FM3A the number (selectivity indexes: GI50a/IC50)

IC50 GI50 GI50/IC50 < 150 150–300> 300

1 0.27b > 12b > 44b 27 7 4 (> 370)2 6.4b > 27b > 4.0b 383 2.1b > 11b > 5.2b 384 7.0b > 25b > 3.5b 385 2.2b 5.0b 2.3b 386 0.60 12 207 0.25 14 56 23 7 2 (300–400) 6 (> 400)8 0.26 > 14 > 549 0.50 > 14 > 2810 0.18 > 19 >106 6 7 1 (300–400) 5 (400–550) 19

(> 634)11 0.29 > 15 > 5212 0.67 > 17 > 2513 1.0 > 16 > 1614 0.90 5.0 5.615 2.2 4.0 2.016 1.9 > 41 > 2217 0.56 > 38 > 68 38 (> 454)18 0.78 > 41 > 5319 2.4 20 1220 0.084 0.16 221 0.68 17 2522 5.3 37 7Quinine 0.11 100 910 20 (< 38) 17 (> 1100)

aGI50 values presented intable V. bThese data have been previously reported [6].

1079

cell lines (seven lung, six colon, six CNS, six stomach,five ovarian, five breast, two renal, and one melanoma),the hosts for FM3A cells. The selectivity indexes of

inactive2–5 and the antimalarial drug quinine were alsodetermined as the control experiments. 13-Hydroxyberberine17 and compound10 displayed a

Table V. Cytotoxicities of protoberberinium salts1–5, 7, 10 and17 against various human cell lines.

Human tumourcell line

Compounds / Cytotoxicity (GI50 in µM)a

1 2 3 4 5 7 10 17 Quinine

BreastHBC-4 4.4 76 5.1 33 35 19 17 b bBSY-1 5.2 27 2.8 12 23 10 7.1 b 35HBC-5 30 b 29 99 39 11 79 b bMCF-7 24 b 10 50 20 12 27 b 55MDA-MB-231 50 b 47 b 56 39 b b 12CNSU251 6.9 55 5.7 76 28 39 35 b bSF-268 7.7 48 14 91 30 27 27 b bSF-295 4.8 b 1.8 b 27 b b b bSF-539 23 b 17 81 25 37 b b 37SNB-75 30 b 30 b 22 5.9 42 b bSNB-78 35 b 37 55 26 52 b b bColonHCC2998 20 b 16 96 51 38 44 b 76KM-12 4.3 71 4.6 35 27 7.6 33 b bHT-29 41 b 31 b 43 79 b b 41WiDr 42 b 34 b 35 53 b bHCT-15 b b b b 50 b b b bHCT-116 30 b 27 b 32 37 b b 43LungNCI-H23 8.2 b 7.1 62 48 18 81 b 87NCI-H226 8.1 58 14 40 19 36 b b 29NCI-H522 5.8 28 4.3 28 37 12 16 b 100NCI-H460 12 b 16 b 11 19 b b 86A549 41 b 77 b 22 48 b b bDMS273 5.9 b 4.2 40 25 20 56 b 52DMS114 3.9 31 2.1 25 42 6.0 20 b 85MelanomaLOX-IMVI 33 b 25 b 32 40 b b 31OvarianOVCAR-3 6.2 46 5.2 34 44 21 29 b bOVCAR-4 31 b 22 b 47 36 77 b bOVCAR-5 39 b 29 b 31 31 b b 72OVCAR-8 17 b 14 b 40 32 91 b bSK-OV-3 b b b b 49 b b b bRenalRXF-631L 59 b 64 b 26 b b b bACHN b b b b 17 b b b 79StomachSt-4 b b 83 b 36 b b b 30MKN1 76 b 42 b 38 93 b b 74MKN7 26 b 32 84 31 4.2 39 b 38MKN28 14 95 13 76 27 11 b b 79MKN45 28 b 28 b 35 24 34 b bMKN74 16 53 16 b 31 9.8 96 b 71

aThe cytotoxicity GI50 values are the concentrations corresponding to 50% growth inhibition. GI50 values were not presented previously [10].bGI50 value is> 100 µM.

1080

significant activity, coptisine7 less potent activity, ber-berine1 much less activity, and compounds2–5 inactiv-ity.

4. Conclusion

From the structure-activity point of view, some fea-tures can be pointed out. On the basis of the resultsderived from examinations with protoberberinium salts, itappears that the position and the type of the oxygensubstituents on rings A, C and D had an influence on theantimalarial activity. The activity was strongly influencedby the position of the oxygen functions on ring D and thetype of the oxygen substituents at the C-13 position (ringC). Compounds with the oxygen functions at C-10 andC-11 on ring D displayed more activity than those whichbear the same substituents at C-9 and C-10. Compoundsbearing a methylenedioxy function at C-2 and C-3 or C-9and C-10 on ring A or D showed higher activity thanthose which have methoxy groups at the same position.Introduction of a methoxy group into the C-1 position onring A caused a decrease in the activity. Replacement ofa hydroxy group at C-2 or C-3 on ring A by a methoxygroup led to a reduction in the activity. Displacement ofa hydroxy function at C-13 on ring C by an OMe, OEt,OCOOEt or OCON(Me)2 function decreased the activity.In the same replacement at C-9 on ring D, the activitydepended on the type of the oxygen substituents. Amongthe potent protoberberinium salts under investigation, theactivity decreased in the order:10, 11, 17 and18 > 7 and8 > 1. The most active compound may be comparablewith the antimalarial drug quinine in activity. Studies onin vivo antimalarial activity of the most potent alkaloidsare in progress.

5. Experimental protocols

5.1. Chemistry

Melting points were determined on a Yanako Mi-cromelting Point apparatus and are uncorrected.1H-NMRspectra were recorded on a Varian VXR-500 (500 MHz)spectrometer using tetramethylsilane as an internal stan-dard and CD3OD as solvent. Mass spectra were deter-mined on a Hitachi M-4100 instrument. The secondaryion mass spectra (LSIMS) were measured using glycerolas matrix. Preparative HPLC and HPLC analyses wereperformed on a Hitachi M-6250 Intelligent Pump (6 mL/min) and a Hitachi M-6200 Intelligent Pump (1 mL/min),respectively, and a Hitachi L-4000 UV detector (280 nm).

Analyses were made on a Cosmosil 5C18-AR reversed-phase column (20 i.d.× 250 mm or 4.6 i.d.× 150 mm)eluting with H2O (0.05% TFA)/MeOH (0.05% TFA),A/B, initial (30% of B), 20 min (100% of B).

Berberine1 and palmatine2 were purchased. Proto-berberinium salts3 [7], 4 [7], 5 [2], 6 [5], 7 [5], 9–11 [5],and17–20 [5] had previously been prepared. Corysamine8 was obtained by oxidation of tetrahydrocorysamine [11]isolated fromCorydalis incisaPers.

Columbamine12 [8] and dehydrocorybulbine13 [8]were natural products isolated fromC. turtschaninoviiBesser formayanhusuoY.H. Chou and C.C. Hsu.

5.1.1. Preparation of palmatrubine14Palmatine (1 g) was heated at 240–250 °C in a dry

oven under vacuum (20–30 mm Hg) for 10 min. Thecrude product was recrystallized from EtOH to give14(870 mg, 90%), m.p. 288–295 °C (dec.); LSIMSm/z338.For 1H-NMR data seetable II.

5.1.2. Preparation of the ethoxycarbonyl derivative ofberberrubine15

A solution of ClCOOEt (1.2 g) in CHCl3 (3 mL) wasadded dropwise to a mixture of berberrubine5(300 mg) [2] and (Et)3N (900 mg) in CHCl3 (50 mL).The mixture was stirred at room temperature for 30 minand evaporated in vacuo. (Me)2CO was added to theresidue and the resulting crystals were collected, whichrecrystallized from MeOH-(Me)2CO to give15 (229 mg,64%). For1H-NMR, LSIMS, and HR-LSIMS data andmelting point seetables II and III .

5.1.3. Preparation of N,N-dimethylcarbamoyl derivativeof berberrubine16

A solution of (Me)2NCOCl (0.7 mL) in CHCl3 (2 mL)was added dropwise to a mixture of berberrubine5(200 mg) and (Et)3N (0.5 mL) in CHCl3 (50 mL). Themixture was stirred at room temperature for 5 h. To thereaction mixture (Me)2NCOCl (0.5 mL) was furtheradded and the mixture was allowed to stir at roomtemperature overnight. (Me)2NCOCl (0.5 mL) was addedto the solution which was further stirred overnight.Solvent was removed under reduced pressure and(Me)2CO was added to the residue. The resulting crystalswere recrystallized from MeOH-(Me)2CO to give 16(196 mg, 82%). For1H-NMR, LSIMS, and HR-LSIMSdata and melting point seetables II and III .

5.1.4. Preparation of ethoxycarbonyl derivative of 13-hydroxyberberine21

A solution of ClCOOEt (0.7 mL) in CHCl3 (2 mL) wasadded dropwise to a mixture of 13-hydroxyberberine17(200 mg) [5] and (Et)3N (0.5 mL) in CHCl3 (20 mL). The

1081

mixture was stirred at room temperature for 1 h andevaporated in vacuo. (Me)2CO was added to the residueand the resulting crystals were recrystallized from H2O-MeOH to give21 (100 mg, 46%). For1H-NMR, LSIMS,and HR-LSIMS data and melting point seetables II andIII .

5.1.5. Preparation of N,N-dimethylcarbamoyl derivativeof 13-hydroxyberberine22

A solution of (Me)2NCOCl (0.5 mL) in CHCl3 (2 mL)was added dropwise to a mixture of 13-hydroxyberberine17 (50 mg) and (Et)3N (0.5 mL) in CHCl3 (20 mL). Themixture was stirred at room temperature overnight. Thereaction was followed by HPLC. To the solution 3 dropsof pyridine was added and the mixture was allowed to stirat room temperature for 3 days. The solvent was evapo-rated under reduced pressure. The residue was purified bypreparative HPLC to give22 (20 mg, 36%). For1H-NMR, LSIMS, and HR-LSIMS data and melting pointseetables II and III .

5.2. In vitro antimalaria screening

5.2.1. Parasites and mammalian cellsPlasmodium falciparum(ATCC 30932, FCR-3 strain)

was used in our study.P. falciparumwas cultivated by amodification of the method of Trager and Jensen [12].Mouse mammary tumour FM3A cells (wild-type, sub-clone F28-7) in culture were used as a control formammalian cell cytotoxicity [13].

5.2.2. Drug testingThe following procedures were used for assaying

antimalarial activity [6, 14, 15].Five microlitres of each solution were added to indi-

vidual wells of a 24 well dish. Erythrocytes with 0.3%parasitaemia were added to each well containing 995µLof culture medium to give a final haematocrit level of 3%.The plates were incubated at 37 °C for 72 h in a 5%CO2/5% O2/90% N2 incubator. To evaluate the antima-larial activity of the test compound, a total of 1× 104

erythrocytes/L of thin blood film were examined undermicroscopy. The IC50 value refers to the concentration ofthe compound necessary to inhibit the increase in parasitedensity at 72 h by 50% of control.

5.2.3. Toxicity to mammalian cellsFM3A cells grew with a doubling time of about 12 h.

Prior to exposure to drugs, cell density was adjusted to 5× 104 cells/mL. A cell suspension of 995µL was dis-pensed into the test plate, and compounds were added toindividual wells of a 24 well dish. The plates wereincubated at 37 °C in a 5% CO2 atmosphere for 48 h. Cell

numbers were measured using a microcell counter CC-130 (Toa Medical Electric Co., Japan). The GI50 valuerefers to the concentration of the compound necessary toinhibit the increase in cell density at 48 h by 50% ofcontrol. Selectivity refers to the mean of the GI50 valuefor FM3A cells per the mean of the IC50 value for P.falciparum.

5.2.4. Selective toxicity using human tumour cellsThe in vitro cytotoxicity assay was carried out accord-

ing to the method [10, 16] of a modified National CancerInstitute protocol [17]. We screened eight compounds,1–5, 7, 10, and17against a panel of 38 human cancer celllines (seven lung, six CNS, six colon, six stomach, fivebreast, five ovarian, two renal, and one melanoma). The50% growth-inhibitory concentration (GI50) for anysingle cell line is an index of cytotoxicity or cytostasis.The selective toxicity was estimated from the GI50/IC50

ratio between the malaria parasites and human tumourcells which served as a model host.

Acknowledgements

This work was supported in part by a Grants-in-Aid forScientific Research on Priority Areas (08281105) fromthe Ministry of Education, Science, Culture and Sports,Japan.

References

[1] Bhukuni D.S., Jain S., in: Brossi A. (Ed.), The Alkaloids Vol. 28,Academic Press, New York, 1986, pp. 95–171.

[2] Iwasa K., Kamigauchi M., Ueki M., Taniguchi M., Eur. J. Med.Chem. 31 (1996) 469–478.

[3] Iwasa K., Kamigauchi M., Sugiura M., Nanba H., Planta Med. 63(1997) 196–198.

[4] Iwasa K., Lee D.U., Kang S.I., Wiegrebe W., J. Nat. Prod. 61 (1998)1150–1153.

[5] Iwasa K., Nanba H., Lee D.U., Kang S.I., Planta Med. 64 (1998)1–4.

[6] Iwasa K., Kim H.S., Wataya Y., Lee D.U., Eur. J. Med. Chem. 33(1998) 65–69.

[7] Iwasa K., Kondoh Y., Kamigauchi M., J. Nat. Prod. 58 (1995)379–391.

[8] 27th Symposium on Natural Drug Analysis, Higashiosaka, Osaka,(1998); p. 69.

[9] Jonathan L.V., Daniel L.K., J. Med. Chem. 31 (1988) 1084–1087.

[10] Yamori T., Jpn. J. Cancer Chemother. 25 Supplement II (1998)373–381.

[11] Nonaka G., Okabe H., Nishioka I., Takao N., J. Pharm. Soc. Jpn. 93(1973) 87–93.

[12] Trager W., Jensen J.B., Science 193 (1976) 673–675.

[13] Yoshioka A., Tanaka S., Hiraoka O., Koyama Y., Hirota Y., AyusawaD., Seno T., Garrett C., Wataya Y., J. Biol. Chem. 262 (1987)8235–8241.

1082

[14] Kim H.S., Miyake H., Arai M., Wataya Y., Parasitol. Int. 47 (1998)59–67.

[15] Takaya Y., Kurumada K., Takeuji Y., Kim H.S., Shibata Y., IkemotoN., Wataya Y., Oshima Y., Tetrahedron Lett. 39 (1998) 1361–1364.

[16] Yamori T., Jpn. J. Cancer Chemother. 24 (1997) 129–135.

[17] Weinstein J.N., Myers T.G., O’Connor P.M., Friend S.H., FornaceJ.R.A.J., Kohn K.W. et al., Science 275 (1997) 343–349.

1083

Short communication

Synthesis and CNS activity of tricyclic theophylline derivatives.8-substituted imidazo[2,1-f]theophyllines

Maciej Pawłowskia*, Anna Drabczyn´skab, Jacek Katlabia, Maria Gorczycaa, Danuta Malecc,Jerzy Modzelewskic

aDepartment of Pharmaceutical Chemistry, Collegium Medicum of Jagiellonian University, 9 Medyczna St., 30-688 Kraków, PolandbDepartment of Chemical Technology of Drugs, Collegium Medicum of Jagiellonian University, 9 Medyczna St., 30-688 Kraków, Poland

cDepartment of Pharmacodynamics School of Medicine, 4 Staszica St., 20-081 Lublin, Poland

(Received 26 April 1999; revised 13 June 1999; accepted 15 July 1999)

Abstract – Based on previously described results of pharmacological in vitro and in vivo tests of some series of pyrimido- anddiazepino-[2,1-f]theophylline derivatives, N8-unsubstituted, N8-benzyl and N8-arylpiperazino-alkyl derivatives of imidazo[2,1-f]theo-phyllines were synthesized and tested for their CNS activity. It has been shown that tricyclic [2,1-f]theophyllines possess sedative andanalgesic activity. N8-Phenylpiperazinopropyl-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline6 showed significant anti-serotonin andlong-lasting hypothermic effects. N8-Benzyl-1,3,6,8-tetrahydroimidazo-7-on[2,1-f]theophylline8 possess anticonvulsant properties. © 1999Éditions scientifiques et médicales Elsevier SAS

imidazo[2,1-f]theophyllines / CNS activity of fused [2,1-f]theophyllines / arylpiperazinoalkylimidazo[2,1-f]theophyllines acting on CNS

1. Introduction

In this paper we would like to summarize the results ofour investigation on chemical and pharmacological prop-erties of derivatives in which the third heterocyclic ring isbound at position 7,8- of theophylline. In our initialstudies we found that some benzyl or dialkylamino-alkylderivatives of 1,3,6,7,8,9-hexahydropyrimido[2,1-f]theo-phyllines and 1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1-f]theophyllines (figure 1) changed the pharmacologicalprofile of the mother compound (theophylline) into seda-tive, hypothermic, analgesic and neuroleptic-like activ-ity [1–3].

These unexpected results seem to be related to thepresence of a heterocyclic six membered ring, as well asto the basic alkylamino side chain substituted at the N9position. Moreover, in preliminary screening it appearedthat among others, the most active are the derivativeswhich contain arylpiperazino-propyl or its butyl homo-logue as an N9-substituent. This observation prompted usto synthesize the series of new compounds with seven, six

and five membered rings bound at position 7,8- oftheophylline and a piperazino-alkyl substituent at the N8,N9 or N10 position, to study the influence of ring size andkind of substituent in fused heterocyclic moieties on CNSactivity of investigated compounds.

The enlargement of the third heterocyclic ring lead to1,3-diazepino- and 1,3-diazepin-9-on-[2,1-f]-theophylli-nes [4] (figure 2).*Correspondence and reprints

Figure 1. Structure of 9-substituted 1,3,6,7,8,9-hexahydro-(9H)-pyrimido[2,1-f]theophyllines and 1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1-f]theophyllines.

Eur. J. Med. Chem. 34 (1999) 1085−1091 1085© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Both investigated groups exhibit high CNS activity.Particularly, N10-phenyl-piperazinopropyl-1,3,6,7,8,9-hexahydro-(10H)-1,3-diazepino[2,1-f]theophylline showedstrong analgesic and sedative activity, but only slighthypothermic and antiamphetamine effects [4].

A similar mode of action was observed in the group ofpyrimidin-8-on[2,1-f]theophyllines with a double bondbetween C6–C7 in the third ring [5] (figure 3).

Particularly, the N9-phenylpiperazinopropyl-derivateof 1,3-dihydro-(9H)-pyrimid-8-on[2,1-f]theophylline in-duced hypothermia and strongly decreased the spontane-ous locomotor activity and amphetamine induced hyper-activity [5].

Taking into account that various 1-aryl(heteroaryl)-substituted piperazines have been reported to possessagonistic activity on serotonin receptors [6], we alsoinvestigated, in behavioral tests, the compounds with a1-(2-pyrimido)-piperazinoalkyl substituent connectedwith a non-lactam tricyclic system [7] and some of theirlactam analogues (1 and2) (figure 4).

Following our research to verify the influence of thethird ring in tricyclic componds on the pharmacologicalprofile of CNS-activity, we undertook the synthesis ofN8-unsubstituted (3) and 8-arylpiperazinoalkyl deriva-tives (4–7) of 1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theo-phylline.

In order to compare CNS-activity, previously obtainedN9-benzyl-1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1-f]theophylline [1] with its analogue in which the fivemember lactam ring is fused in the 7,8- position oftheophylline, N8-benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f]theophylline8 was synthesized.

2. Chemistry

9-{3-[4-(2-Pyrimidynyl)-1-piperazinyl]-propyl}-1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1-f]theophylline (1)and 10-{3-[4-(2-pyrimidynyl)-1-piperazinyl]-propyl}-1,3,6,7,8,10-hexahydro-1,3-diazepin-9-on[2,1-f]theophylli-ne (2) were prepared according to the procedures de-scribed previously [6].

In the reaction of 7-â-bromoethyl-8-bromotheo-phylline [8] with ammonia 8-unsubstituted 1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline, (3) was ob-tained. Arylpiperazino-alkyl derivatives of 1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline (4–7) wereobtained in a several step synthesis (figure 5).

8-â-Hydroxyethyl-1,3,6,7-tetrahydro-(8H)-imidazo-[2,1-f]theophylline (I ) and its homologue (Ia) wereprepared by condensation of 7-â-bromoethyl-8-bromo-theophylline with 2-aminoethanol or 3-aminopropan-1-ol. CompoundsI and Ia, upon bromination with PBr3 inCHCl3, yielded II and IIa . Aminolysis of compoundIIand IIa with a double amount of phenylpiperazine or2≠-pyrimidyl-piperazine in anhydrous toluene in the pres-ence of K2CO3 gave the compounds4–7.

N8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f]theo-phylline 8 was synthesized via cyclization of 8-benzyl-aminotheophylline-7-acetic acid [9] in boiling acetic an-hydride (figure 6).

3. Pharmacological results and discussion

The compounds1–8were evaluated for their influenceon the CNS. Their acute toxicity assessed in mice as theLD50 (confidence limits) was as follows:

Compound1: 235 mg/kg (178–310)

Figure 2. Structure of 10-substituted 1,3,6,7,8,9-hexahydro-(10H)-1,3-diazepino[2,1-f]theophyllines and 1,3,6,7,8,10-hexahydro-1,3-diazepin-9-on[2,1-f]theophyllines.

Figure 3. Structure of 9-substituted 1,3-dihydro-(9H)-pyrimidin-8-on-[2,1-f]theophyllines.

Figure 4. Structure of 1-(2-pyrimido)-piperazinoalkyl deriva-tives of tricyclic non-lactam and lactam system.

1086

Compound2: 560 mg/kg (478–655)Compound3: 320 mg/kg (269–381)Compound4: 175 mg/kg (117–263)Compound5: 320 mg/kg (264–387)Compound6: 540 mg/kg (439–664)Compound7: 2 000 mg/kgCompound8: 2 000 mg/kgThe CNS effects of all these substances (1/10 LD50)

were studied in several tests, and it was observed thatsubstances2, 3, 4, 6, 7 and 8 induced marked sedativeeffects in the locomotor activity test (figure 7).

Compound6 had the strongest action, which reducedthe activity of mice by about 90% (1/10 and 1/20 LD50).These doses of substance6 also antagonizedamphetamine-induced hyperactivity in mice, but did notchange apomorphine-induced stereotypy or haloperidol-induced catalepsy in rats. Compounds4 and 6 also hadantiserotonin effects, as they diminished 5HTP-inducedhead-twitch reactions of mice (table I). Hypothermicaction was observed after the administration of com-pounds1, 6, 7 and 8 (figure 8). The most apparent andlong-lasting effects were observed for compounds6 and8. Analgesic action in the hot plate test was induced inmice by nearly all substances, except for compound8(table II).

These effects were of rather a short duration (up to90 min after injection). Compound8 had anticonvulsantaction in the pentetrazole test in mice, as it prevented thetonic phase of the seizures and lethality of the animals.Reserpine-induced hypothermia was slightly reversed bysubstances1, 3 and 4, and this action suggests someantidepressant properties of these three substances. How-

Table I. The influence of the compounds (0.1 LD50) on the headtwitch responses of mice receiving L-5HTP (180 mg/kg i.p.).

Compound Mean± SEnumber of head twitch responses

Tylose 11.4± 2.91 18.0± 4.32 14.5± 4.03 15.0± 4.24 3.6* ± 1.35 12.3± 5.56 1.8** ± 0.77 9.3± 3.28 12.7± 3.7

The compounds were injected 60 min before L-5HTP.* P < 0.05, ** P < 0.01 (Student’st-test).

Figure 6. N8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f]theophylline8.

Figure 5. Arylpiperazino-alkyl derivatives of 1,3,6,7-tetra-hydro-(8H)-imidazo[2,1-f]theophylline (4–7).

Figure 7. The influence of compounds (0.1 LD50) on sponta-neous locomotor activity of mice.

1087

ever, we observed neurotoxic effects in the rota-rod test inmice following injection of nearly all compounds tested.

4. Conclusion

Analysing the results of our preliminary investigationson chemical and pharmacological properties of tricyclictheophylline derivatives from the point of view ofstructure-activity relationships, it may be concluded that:

1. In contrast to mother theophylline, all investigatedtricyclic compounds generally exhibited a strong sedativeaction, particularly those which contained the basicaminoalkyl side chain, substituted on the third bound ringat N8 of imidazole, N9 of pyrimidine and N10 ofdiazepine. Their sedative properties were confirmed in anumber of behavioral tests (in mice).

2. The replacement of the methylene group with acarbonyl one in pyrimido[2,1-f]theophyllines or diaze-

pino[2,1-f]theophyllines (lactam structure) did not have asignificant influence on the direction of pharmacologicalactivity.

3. The potency of sedative action depends significantlyon the kind of N-substituted basic side chain. Variation ofthe alkylamino substituents confirmed unequivocally thatthe most apparent CNS activity was exerted by thecompounds with phenylpiperazinopropyl- or phenyl-piperazinobutyl- substituents.

4. Length reduction of the alkylene chain betweendialkylamino substituents and tricyclic substituents fromthree or four to two carbon atoms resulted mostly in anincrease of toxicity.

5. The replacement of the phenylpiperazinoalkyl sub-stituent with an (ω-4-pyrimidinyl)-piperazino-alkyl sidechain generally decreased the activity and revealed someneurotoxic effects.

Figure 8. The influence of compounds (0.1 LD50) on body temperature of normothermic mice.

Table II. The influence of the tested compounds on the reactivity of mice in the hot plate test.

Compound 0.1 LD50 Reaction time (s) of mice after injection (min)

30 60 90 120

Vehicle 8.9± 0.6 9.3± 0.6 9.5± 0.7 9.1± 0.71 14.5*** ± 0.9 13.4** ± 0.8 15.1** ± 1.2 11.5± 0.92 13.0*** ± 0.7 11.7± 0.9 10.8± 0.6 10.4± 0.53 10.8± 0.6 13.8** ± 1.1 10.4± 0.6 10.8± 0.74 14.7*** ± 1.0 12.3*± 0.8 13.8*± 1.1 13.4*± 1.25 13.4** ± 1.1 12.9*± 1.0 11.2± 0.7 11.2± 0.86 13.0** ± 1.0 14.3** ± 1.2 12.3± 1.2 11.5± 0.97 14.7*** ± 1.1 13.2*** ± 1.0 10.6± 0.6 11.2± 0.88 10.1± 0.6 10.9± 0.7 9.3± 0.5 10.2± 0.6

* P < 0.05; ** P < 0.02; *** P < 0.001 (Student’st-test).

1088

6. All phenylpiperazinoalkyl substituted tricyclic [2,1-f]theophyllines reduced the amphetamine induced hyper-activity.

Pyrimido[2,1-f]theophyllines with phenylpiperazino-alkyl side chains showed mainly analgesic and neuro-leptic-like properties. The pyrimidinyl-piperazino ana-logues exhibited agonistic activity on 5-HT1A receptors.Their diazepino homologues exerted apparent analgesicactivity.

7. Some of previously described tricyclic theophyllinederivatives, evaluated for the affinities for 5-HT recep-tors, were classified as postsynaptic antagonists or partialagonists of 5-HT1A receptors [6].

Pyrimido[2,1-f-]theophyllines were also tested for af-finity to adenosine receptors [10]. They were classified asadenosine receptor subtype A1 antagonists.

Some of the presented data on tricyclic [2,1-f-]theo-phyllines summarized in this article remains open tofurther investigation.

5. Experimental protocol

5.1. Chemistry

Melting points were determined on a Boetius apparatusand are uncorrected. UV spectra were obtained on theUV-VIS Lambda 12 spectrophotometer (Perkin Elmer)(conc. 5 × 10–5 M/dm3 in methanol). IR spectra weredetermined on the Specord M 80 spectrophotometer (CarlZeiss Jena) in KBr pellets.1H-NMR spectra were ob-tained on a Brucker AC200 F spectrometer in CDCl3

(except I measured in DMSO) solution using TMS asinternal standard. Chemical shifts are reported inδ units.Coupling constants are reported in Herz. Elementalanalyses indicated by the symbols were within± 0.4% ofthe theoretical value. TLC was performed on Merckplates (Kieselgel 60 F254 with S = benzene:acetone:methanol (1:1:1) + 2 drops NH4OH.

5.1.1. 8-â-Hydroxyethyl-1,3,6,7-tetrahydro-(8H)-imidazo-[2,1-f]theophyllineI

A mixture of 7-â-bromoethyl-8-bromotheophylline(5.0 g; 0.014 mol), 2-aminoethanol (5 cm3; 0.08 mol) and40 cm3 of ethanol was refluxed for 5 h. After refrigerationthe product was filtered off and recrystallized. Yield 69%;m.p. 238–240 °C (ethanol). Rf 0.74, UV λmax nm 299.4(log e 4.30).1H-NMR 3.17 (s, 3H, N3CH3), 3.34 (s, 3H,N1CH3), 3.30–3.36 (m, 2H, CH2OH), 3.57–3.64 (t, 2H,N8CH2), 3.91–3.98 (t, 2H, C7H2), 4.05–4.12 (t, 2H,C6H2), 4.82–4.86 (t, 1H, OH). IR (cm–1) 3 300 (OH),2 800–2 950 (CH2)n, 1 650 (CO), 1 450 (CH2–N). Anal.C11H15O3N5; C, H, N.

5.1.2. 8-γ-Hydroxypropyl-1,3,6,7-tetrahydro-(8H)-imida-zo[2,1-f]theophyllineIa

The title compound was obtained similarly to com-pound I using 3-amino-1-propanol and isobutanol as asolvent. The mixture was refluxed for 10 h. After refrig-eration, the separated product was recrystallized. Yield54%; m.p. 186–188 °C (isopropanol). Rf 0.74, UV λmax

nm 298.5 (log e 4.19). 1H-NMR 1.80–1.90 (m, 2H,CH2CH2CH2) 3.34 (s, 3H, N3CH3), 3.48 (s, 3H, N1CH3),3.49–3.55 (m, 2H, CH2OH), 3.66–3.71 (t, 2H, N8CH2),3.92–3.98 (t, 2H, C7H2), 4.22–4.28 (t, 2H, C6H2). IR(cm–1) 3 300–3 450 (OH), 2 800–2 950 (CH2)n, 1 650(CO), 1 450 (CH2–N). Anal. C12H17O3N5; C, H, N.

5.1.3. 8-(â-Bromoethyl)-1,3,6,7-tetrahydro-(8H)-imida-zo[2,1-f]theophylline II and 1,3-8-(γ-bromopropyl)-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophyllineIIa

These two compounds were obtained similarly. To0.01 mol of compoundI or Ia, fourfold amounts of PBr3in 25 cm3 of CHCl3 was added and refluxed for 5 h. Thenthe solvent and the excess of PBr3 were distilled off underreduced pressure and to the residue, water was added.The mixture was alkalized with 10% NaOH (to pH 8).The crude products were separated and recrystallized.

CompoundII . Yield 92%; m.p. 210–212 °C (ethanol).Rf 0.87, UV λmax nm 298.5 (loge 4.20).1H-NMR 3.36(s, 3H, N3CH3), 3.51 (s, 3H, N1CH3), 3.60–3.65(t, 2H, CH2Br), 3.76–3.81 (t, 2H, N8CH2), 4.02–4.08(t, 2H, C7H2), 4.24–4.31 (t, 2H, C6H2). IR (cm–1)2 800–2 950 (CH2)n, 1 650 (CO), 1 450 (CH2–N). Anal.C11H14O2N5Br; C, H, N.

Compound IIa . Yield 94%; m.p. 162–164 °C (50°ethanol). Rf 0.87, UV λmax nm 299.4 (log e 4.11).1H-NMR 2.21–2.32 (m, 2H, CH2CH2CH2) 3.37 (s, 3H,N3CH3), 3.52 (s, 3H, N1CH3), 3.47–3.54 (t, 2H, CH2Br),3.88–3.95 (t, 2H, C7H2), 4.05–4.12 (t, 2H, C6H2). IR(cm–1) 2 800–2 950 (CH2)n, 1 650 (CO), 1 450 (CH2–N).Anal. C12H16O2N5Br; C, H, N.

5.1.4. 1,3,6,7-Tetrahydro-(8H)-imidazo[2,1-f]theophyl-line 3

A mixture of 7-â-bromoethyl-8-bromotheophylline(4.0 g; 0.011 mol), 5 cm3 of 25% ammonia and 15 cm3 ofethanol was heated in an autoclave at 180–190 °C for 5 h.Then the solution was evaporated and the residue wasrefluxed with isopropanol to separate the inorganic salts.The hot mixture was filtered off and the filtrate wasdistilled off under reduced pressure. The crude productwas recrystallized from ethanol; m.p. 270–272 °C, yield60%. Compound3 was obtained by Eckstein et al. [11]using another method.

1089

5.1.5. General procedure for compounds4–7Compounds4–7were obtained by aminolysis of com-

poundsII or IIa with double the amount of phenylpip-erazine or 2≠-pyrimidynyl-piperazine. The reaction wascarried out in anhydrous toluene in the presence of astoichiometric amount of K2CO3 and refluxed for 5 h.The inorganic salt was filtered off from the hot mixtureand the filtrate was refrigerated. The crystalline productswere filtered off and recrystallized.

5.1.5.1. 8-{2-[4-(Phenyl)-1-piperazinyl]-ethyl}-1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline4

Yield 52%; m.p. 164–166 °C (ethanol + butyl acetate).Rf 0.80, UV λmax nm 244.0 (loge 4.09), 296.7 (loge4.23). 1H-NMR 2.70–2.75 (m, 10H, CH2N(CH2)4N),3.19 (t,J = 4.8, 2H, N8CH2), 3.37 (s, 3H, N3CH3), 3.51(S, 3H, N1CH3), 3.99 (t,J = 7.7, 2H, C7H2), 4.22 (t,J =7.8, 2H, C6CH2), 6.86–6.93 (m, 3H, 3≠,4≠,5≠-phenyl),7.26 (m, 2H, 2≠,6≠-phenyl). Anal. C21H27O2N7, C, H, N.

5.1.5.2. 8-{2-[4-(2-Pyrimidynyl)-1-piperazinyl]-ethyl}-1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline5

Yield 91%; m.p. 194–195 °C (toluene). Rf 0.50, UVλmax nm 242.0 (loge 4.14), 301.0 (loge 3.99).1H-NMR2.55–2.70 (m, 10H, CH2N(CH2)4N), 3.37 (s, 3H,N3CH3), 3.51 (s, 3H, N1CH3), 3.78 (t, J = 5.0, 2H,N8CH2), 3.35–4.03 (m, 2H, C7H2), 4.18–4.26 (m, 2H,C6CH2), 6.48 (m,J = 4.7, 1H, 5≠-pyrimidyl), 8.30 (d,J= 4.8, 2H, 4≠,6≠-pyrimidyl). Anal. C19H25O2N9, C, H, N.

5.1.5.3. 8-{2-[4-(Phenyl)-1-piperazinyl]-propyl}-1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline6

Yield 79%; m.p. 192–194 °C (toluene). Rf 0.75, UVλmax nm 246.0 (loge 4.09), 299.0 (loge 4.09).1H-NMR1.84–1.92 (m, 2H, CH2CH2CH2), 2.44–2.63 (m, 10H,CH2N(CH2)4N), 3.19 (t,J = 4.7, 2H, N8CH2), 3.35 (s,3H, N3CH3), 3.47 (s, 3H, N1CH3), 3.86 (t,J = 8.0, 2H,C7H2), 4.20 (t,J = 8.0, 2H, C6CH2), 6.84–6.93 (m, 3H,3≠,4≠,5≠-phenyl), 7.24 (t,J = 7.8, 2H, 2≠,6≠-phenyl).Anal. C22H29O2N7, C, H, N.

5.1.5.4. 8-{2-[4-(2-Pyrimidynyl)-1-piperazinyl]-propyl}-1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline7

Yield 79%; m.p. 192–194 °C (toluene). Rf 0.59, UVλmax nm 242.0 (loge 4.15), 301.0 (loge 3.94).1H-NMR1.84–1.92 (m, 2H, CH2CH2CH2), 2.43–2.53 (m, 10H,CH2N(CH2)4N), 3.35 (s, 3H, N3CH3), 3.41 (t,J = 4.7,2H, N8CH2), 3.49 (s, 3H, N1CH3), 3.75–3.92 (m, 2H,C7H2), 4.21 (t,J = 8.0, 2H, C6CH2), 6.47 (t,J = 4.8, 1H,5≠-pyrimidyl), 8.30 (d, J = 4.8, 2H, 4≠,6≠-pyrimidyl).Anal. C20H27O2N9, C, H, N.

5.1.5.5. 8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f]theophylline8

A mixture of 8-benzylaminotheophyllino-7-acetic acid(3.43 g; 0.01 mol) and 15 cm3 of acetic anhydridewas refluxed for 3 h. After cooling, the product wasseparated by filtration, washed with a small amount ofmethanol and recrystallized. Yield 50%, m.p. 216–218 °C (methanol). Rf 0.92 (chloroform:methanol: 25%NH4OH (8:2:0.25)), UV λmax nm 300.0 (loge 4.01)(conc. 5 × 10–5 M/dm3 in chloroform). 1H-NMR 3.33(s, 3H, N3CH3), 3.58 (s, 3H, N1CH3), 4.71 (s, 2H,N8CH2C6H5), 4.93 (s, 2H, C6CH2), 7.33–7.48 (m, 5H,C6H5). Anal. C16H15O3N5, C, H, N.

5.2. Pharmacology

The experiments were performed on male Swiss mice(19–26 g), and on Wistar rats (180–200 g), kept at ambi-ent temperature on a 12 h light/dark schedule, with freeaccess to food and water before the experiments. Theexperimental groups consisted of 5–8 animals each. Allinvestigated compounds were injected i.p., 1 h before thetest, as a suspension in 0.5% tylose solution.

The doses corresponding to 0.1 LD50 of the com-pounds were administered in all tests. The acute toxicityof the compounds was assessed in mice as the LD50

calculated on the basis of mortality within 48 h [12].Spontaneous locomotor activity and amphetamine-

induced hyperactivity in mice were observed in a photo-resistor actometer for 30 min. Analgesic effects wereassessed in the hot plate (56 °C) test in mice, and bodytemperature was measured with a thermistor thermometerin the rectum of mice. Antipsychotic action was evaluatedin the tests of amphetamine (5 mg/kg s.c.)-inducedhyperactivity of mice and apomorphine (3 mg/kg s.c.)-induced stereotypy in rats. Antiparkinsonian effects weremeasured by investigation of the influence of the com-pounds on haloperidol (1 mg/kg)-induced catalepsy inrats. Anticonvulsant activity was assessed in pentetrazole(90 mg/kg s.c.)-induced seizures in mice. The anti-depressant action was evaluated using the test ofreserpine-induced hypothermia in mice (reserpine2.5 mg/kg s.c., 20 h before injection of the investigatedsubstances, and then body temperature was measured at30 min intervals for 3 h). The influence of the compoundson serotonin neurotransmission was evaluated in thehead-twitch responses of mice following 5HTP injec-tion [13, 14]. Neurotoxic properties were observed in therota-rod test [15, 16] in mice.

Statistical analysis was carried out using the Student’st-test (convulsions) and Mann-Whitney U test (catalepsy,stereotypy).

1090

References

[1] Drabczynska A., Pawłowski M., Gorczyca M., Malec D., Mod-zelewski J., Pol. J. Pharmacol. Pharm. 41 (1989) 385–394.

[2] Pawłowski M., Drabczyn´ska A., Gorczyca M., Malec D., Mod-zelewski J., Pol. J. Pharmacol. Pharm. 43 (1991) 61–70.

[3] Drabczynska A., Pawłowski M., Gorczyca M., Malec D., Mod-zelewski J., Pol. J. Pharmacol. Pharm. 44 (1992) 487–503.

[4] Pawłowski M., Drabczyn´ska A., Gorczyca M., Malec D., Mod-zelewski J., Acta Pol. Pharm. 51 (1994) 385–391.

[5] Pawłowski M., Drabczyn´ska A., Gorczyca M., Malec D., Mod-zelewski J., Pharmazie 50 (1995) 453–456.

[6] Chojnacka-Wójcik E., Kłodzin´ska A., Drabczyn´ska A., PawłowskiM., Charakchieva-Minol S., Chłon´ G., Gorczyca M., Eur. J. Med.Chem. 30 (1995) 587–592.

[7] Malec D., Modzelewski J., Drabczyn´ska A., Pawłowski M., Gorc-

zyca M., Pol. J. Pharmacol. Pharm. 47 (1995) 169–173.

[8] Cacace F., Masironi R., Ann. Chim. 46 (1956) 806–809.

[9] Cacace F., Crisera R., Zifferero M., Ann. Chim. 46 (1956) 99–102.

[10] Geis U., Grahner B., Pawłowski M., Drabczyn´ska A., Gorczyca M.,Müller C.E., Pharmazie 50 (1995) 333–336.

[11] Eckstein M., Drabczyn´ska A., Synthesis 8 (1979) 581–584.

[12] Litchfield L.T., Wilcoxon F., J. Pharmacol. Exp. Ther. 96 (1949)99–113.

[13] Corn S.J., Pickering R.W., Werner B.T., Br. J. Pharmacol. 20 (1963)106–120.

[14] Moore N.A., Tye N.C., Axton M.S., Risius F.C., J. Pharm. Exp. Ther.262 (1992) 545–551.

[15] Gross F., Tripod J., Meir R., Schweiz. Med. Wochschr. 85 (1955)305–309.

[16] Dar M.S., Wooles W.R., Life Sci. 39 (1986) 1429–1437.

1091

Short communication

Synthesis and antimycobacterial activity of some coupling productsfrom 4-aminobenzoic acid hydrazones

Sq. Güniz Küçükgüzela, Sevim Rollasa*,Ilkay Küçükgüzela, Muammer Kirazb

aUniversity of Marmara, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 81010 Haydarpasa, Istanbul, TurkeybUniversity of Istanbul, Faculty of Medicine, Center for Research and Application of Culture Collections of Microorganisms,

Çapa, Istanbul, Turkey

(Received 25 February 1999; revised 2 July 1999; accepted 15 July 1999)

Abstract – Various 2,3,4-pentanetrione-3-[4-[[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazinocarbonyl]phenyl]hydrazones3a–j were synthesized by the reactions of acetylacetone with the diazonium salts of 4-aminobenzoic acid-[(5-nitro-2-furyl/pyridyl/substitutedphenyl)methylene]hydrazides2a–j at 0–5 °C. The structures of these compounds were determined using spectral data. All the synthesizedcompounds were evaluated for their antimycobacterial activity againstMycobacterium fortuitumATCC 6841 andMycobacterium tuberculosisH37Rv. Of the compounds screened,2e, 2i, 3eand3g were found to be active againstM. fortuitumat an MIC value of 32µg/mL. Compound3a, which exhibited> 90% inhibition in the primary screen at 12.5µg/mL againstM. tuberculosisH37Rv, was the most promising derivativefor antituberculosis activity. Results obtained from the level II screening showed that the actual MIC and IC50 values of3a were 3.13 and0.32µg/mL, respectively. The same compound was also tested againstMycobacterium avium, which was observed not to be susceptible to3a. © 1999 Éditions scientifiques et médicales Elsevier SAS

hydrazide-hydrazones / coupling products / antimycobacterial activity / cytotoxicity

1. Introduction

Antituberculosis and antibacterial effects have beenshown with various 4-aminobenzoic acid substitutedbenzalhydrazones [1]. Furthermore, the coupling prod-ucts starting from diazonium salts of 4-aminobenzoicacid hydrazide substituted benzalhydrazones with indolehave also been reported to possess promising antibact-erial and antitubercular activities [2]. These observationsprompted us to synthesize some novel coupling productsof 4-aminobenzoic acid hydrazide substituted benzalhy-drazone derivatives.

Earlier reports indicated that certain compounds bear-ing aromatic amine functions, such as sulfaguanidine [3],PABA [4], benzocaine [4], sulfanilamide [5], 4-amino-benzoic acid hydrazide [6], oxadiazole [7, 8], thiadiaz-ole [7], quinazolinone [7], 4-aminobenzoic acid hy-drazide substituted benzalhydrazones [2], substituted

anilines [9, 10], 1,2,4-triazoline-3(2H)-thiones [11, 12],1,2,4-triazoline-3(2H)-one [12], 1,3,4-oxadiazoline-2(3H)-thione [11, 12], 3,5-dimethyl-1H-pyrazole [11]and 4-aminobenzoic acid-[(4-fluorophenyl)methylene]-hydrazide [12] might be selected to couple their diazo-nium salts with compounds possessing an active hydro-gen.

2. Chemistry

4-Aminobenzoic acid hydrazide1 was prepared by thereaction of ethyl 4-aminobenzoate with hydrazine hy-drate. 4-Aminobenzoic acid-[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazides2a–j were syn-thesized by condensation of1 with appropriate aldehydes[13] as original compounds, except2a [14] and2c [15],which were previously reported elsewhere. The diazo-nium salts of 4-aminobenzoic acid-[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazide werethen coupled with acetylacetone in ethanol (50%) con-*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1093−1100 1093© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

taining sodium acetate. The resulting coupling productswere formulated as 2,3,4-pentanetrione-3-[4-[[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazino-carbonyl]phenyl]hydrazones3a–j in view of spectralfindings (figure 1). Spectral investigations also revealedthat the aromatic primary amine function present at1 wasnot affected during the synthesis of2a–j, although noprotective precautions were taken, probably because the-NH2 moiety of hydrazide function is more nucleophilicthan the Ar-NH2 group.

The novel compounds3a–j gave satisfactory elementalanalyses. Their structures were established using spectraldata (UV, IR,1H-NMR and EI-MS).

UV spectra of the synthesized compounds showedthree absorption maxima at 236–249, 265–303 and369–402 nm values. The lack of absorption bands be-tween 332–360 nm [16, 17] and above 400 nm [18],which could be attributed to an azo function, togetherwith observation of bands at around 369–402 nm, indi-cated that compounds3a–j were in a hydrazoneform [19].

The IR spectra of3a–j exhibited the N–H band ofhydrazone, the C=O band of ketone and hydrazone

arising from stretching vibrations at expected wavenum-bers according to the literature [3, 4].

In the 1H-NMR spectra, methyl protons of the ketonemoiety gave singlets at 2.52, 2.62 ppm (CDCl3, only forcompound3b) or 2.42–2.46 ppm (DMSO-d6, for com-pounds3a, 3c–j). The azomethine proton (-CH=N-) andNH proton of the hydrazide hydrazone N1–H(-CONHN=CH-) exhibited the expected singlets at8.32–8.71 ppm and 11.49–13.52 ppm, respectively. Un-like the others, which were dissolved in DMSO-d6, theN1–H proton of compound3b resonated at 9.09 ppm,probably due to the use of CDCl3 as solvent [20]. The1H-NMR spectra of3b in CDCl3 and 3c–d and 3f–j inDMSO-d6 displayed the hydrazone N2–H(-CH=N–NH-)protons at 13.71–14.58 ppm. The hydrazone N–H protonsof 3a and3e were observed to exchange with deuteriumin DMSO-d6 [12]. In addition, in the1H-NMR spectra ofcompounds3a–j, signals arising from the>CH–N=N-structure at 3.00–4.00 ppm [3, 4] were not observed. Thisfinding also supported the idea that the structures of thesecompounds might be given in hydrazone form. Thephenolic proton of compound3g displayed a singlet at9.69 ppm.

Figure 1. Synthetic route to compounds3a–j.

1094

EI-Mass spectra of three selected prototypes,3a, 3i and3j confirmed their molecular weights and displayedcharacteristic fragment ions as shown infigure 2. Themajor fragmentation pathway appeared by the cleavageof -CONH-N= bonds of a hydrazide-hydrazone moiety.Initial loss of an acetyl fragment (m/z 43) followed byexpulsion of CH3–CO–CN from the molecular ion andthe subsequent acquirement of a hydrogen radical af-forded the fragment ions at m/z 282 and 308 for com-pounds3i and3j, respectively. The same type of fragmention also appeared at m/z 274 in the EI-mass spectrum ofcompound3a by the cleavage of>C=N–NH- bonds ofhydrazone followed by protonation. This latter ion exhib-ited the expected fragmentation pattern of a hydrazide-hydrazone structure [13].

3. Results and discussion

The synthesized compounds2a–j and3a–j were evalu-ated for in vitro antimycobacterial activity againstM.fortuitum ATCC 6841 using the microdilutionmethod [21–24]. Of the screened compounds,2eand therelated coupling product3e, compounds2i and3g werefound as active as tobramycin against the examined

strain. The remaining compounds showed less activity(64 µg/mL) or no considerable effect (> 128µg/mL) asshown intable I.

Compounds2a–j and3a–j were also tested for in vitroantituberculosis activity againstM. tuberculosisH37Rvusing the BACTEC 460 radiometric system [25, 26].Rifampicin was used as the standard in the tests. Primaryantituberculosis activity screening results of these com-pounds can be seen intable II. Compounds2b, 2i, 3eand3h were inactive againstM. tuberculosisH37Rv, whereasremaining compounds showed varying degrees of inhibi-tion in the primary screen. The compounds which exhib-ited < 90% inhibition in the primary screen (MIC>12.5µg/mL) were not evaluated further. Compound3aeffecting > 90% inhibition in the primary screen at12.5µg/mL was re-tested at lower concentrations againstM. tuberculosisH37Rv to determine the actual minimuminhibitory concentration in a broth microdilution assayusing alamar blue. The MIC was defined as the lowestconcentration inhibiting 99% of the inoculum. Com-pound3a was also tested againstM. avium, a naturallydrug-resistant opportunistic pathogen, using the sametechnique. Clarithromycin was used as the standard inthis assay. However, no inhibition was observed with

Figure 2. Mass fragmentation pattern of3a, 3i and3j.

1095

compound3a at 12.5µg/mL, whereas clarithromycinexhibited 98% inhibition at 2µg/mL. Level II assayresults of3a are given intable III. Compound3a wasfound to be active againstM. tuberculosisH37Rv at3.13µg/mL. The same compound was also tested forcytotoxicity (IC50) in VERO cells at concentrations equalto and greater than the MIC forM. tuberculosisH37Rv.The IC50 value was found at a concentration level of0.32µg/mL for compound3a. The selectivity index (SI =IC50/MIC) was calculated as 0.1, showing that thiscompound not only displayed a considerable antitubercu-losis activity, but also had a remarkable cytotoxicity.

Compounds2a–j, whose synthesis was previouslyreported [13], was the starting point of our new work inorder to obtain new coupling products in which aromatic

primary amine functions were masked. The aim of thisstudy was not only to screen the antimycobacterialactivity of 2a–j and the resulting3a–j, but also to observethe influence of coupling2a–j with ethyl acetoacetate onthis activity. As a matter of fact, series3a–j was found tobe more active againstM. tuberculosisH37Rv than thecorresponding2a–j in most cases, as shown intable II.Even in the case of3a, in which activity possibly arisesfrom the presence of the 5-nitro-2-furanyl moiety, thecoupling product3a was found to be more active, with95% inhibition (MIC = 3.13µg/mL), than the corre-sponding amine2a which inhibited the growth ofM.tuberculosisH37Rv by only 70%. From the above data itcan be concluded that3a could be a leading compoundfor further development. Structural modifications, inwhich ketone functions present in3a–j are replaced withdifferent moieties in order to minimize toxicologicaleffects, will be the subject of our continuing studies.

4. Experimental protocols

4.1. Chemistry

Benzocaine, 4-dimethylamino benzaldehyde andacetylacetone were purchased from Merck. All otherchemicals were purchased from Sigma.

Melting points were determined using a Büchi-530melting point apparatus, uncorrected. IR spectra: (KBr,ν,cm–1) Perkin-Elmer FTIR 1600 spectrophotometer. UVSpectra: (ethanol)λmax (log e) Shimadzu UV 2100Sspectrophotometer.lH-NMR: (DMSO-d6, CDCl3, TMSas internal standard, chemical shifts,δ, in ppm) BrükerAC 200 L spectrometer. EI-MS: Kratos MS-9/50 spec-trometer (for compounds3i and 3j) and ZabSpec EI+

Magnet (for compound3a) at 70 eV. Elemental analyses:Carlo-Erba 1106 instrument.

4.1.1. Preparation of aromatic primary amines2a–jCompounds2a–j were prepared as described previ-

ously [13–15]. M.p.’s: for compound2a (271–274 °C,lit. [14] 273–274 °C); for compound2c (218 °C, lit. [15]218–219 °C).

4.1.2. Synthesis of the coupling products3a–jTo a cooled solution of compounds2a–j (0.01 mol), in

2 mL of hydrochloric acid (37%), an ice-cold solution of10 mL of sodium nitrite (10%) was added. The reactionmixture was then poured into a mixture of 1 mL ofacetylacetone and 50 g of sodium acetate in ethanol(50%) by vigorous stirring. This mixture was allowed tostand in a refrigerator for 24 h. Precipitated solid wascollected, washed with water, dried and washed with anappropriate solvent to give3a–j.

Table I. Antimycobacterial activity of2a–j and3a–j.

Compound MIC values (µg/mL) ArM. fortuitumATCC 6841

2a3a

6464

2b3b

6464

2c3c

6464

2d3d

64> 128

2e3e

3232

2f3f

> 128> 128

2g3g

6432

2h3h

64> 128

2i3i

32> 128

2j3j

64> 128

Tobramycin 32

1096

4.1.2.1. 2,3,4-pentanetrione-3-[4-[[(5-nitro-2-furyl)-methylene]hydrazinocarbonyl]phenyl]hydrazone3a

Yield: 65%, m.p. 230 °C (methanol). UV (ethanol):λmax (log e) = 303 (4.1611), 380 (4.5789); IR (KBr):ν =3 450–3 420 cm–1 (H2O), 3 331 (NH), 1 672 (C=O, ke-tone), 1 625 (C=O, hydrazone); EI-MS (70 eV): m/z (%)= 385 (14.15) [M+], 274 (55.00), 273 (8.41), 262 (22.41),

259 (8.64), 244 (17.14), 231 (4.33), 219 (3.00), 206(2.50), 120 (33.74), 92 (7.44), 78 (90.77), 65 (3.88), 63(100); 1H-NMR (DMSO-d6): δ (ppm) = 2.46 (s, 6H,COCH3), 7.64–7.99 (m, 6H, Ar-H), 8.39 (s, 1H, CH=N),11.90–12.11 (s, 1H, -CONHN=CH-). Analysis:C17H15N5O6.H2O (% calculated/found) 50.62/50.21 (C);4.24/3.54 (H); 17.36/18.15 (N).

Table II. Primary antituberculosis activity screen results of2a–j and3a–j.

Compound MIC values (µg/mL) Inhibition (%) ArM. tuberculosisH37Rv

2a3a

> 12.5< 12.5

7095

2b3b

> 12.5> 12.5

–16

2c3c

> 12.5> 12.5

1635

2d3d

> 12.5> 12.5

620

2e3e

> 12.5> 12.5

11–

2f3f

> 12.5> 12.5

2313

2g3g

> 12.5> 12.5

2152

2h3h

> 12.5> 12.5

26–

2i3i

> 12.5> 12.5

–25

2j3j

> 12.5> 12.5

5676

Rifampicin 0.25 98

Table III. Level II antituberculosis activity assay result of3a.

Compound MIC value (µg/mL) IC50 (µg/mL) SI (IC50/MIC)M. tuberculosisH37Rv

3a 3.13 0.32 0.1Rifampicin 0.125

1097

4.1.2.2. 2,3,4-pentanetrione-3-[4-[[(4-pyridyl)methyl-ene]hydrazinocarbonyl]phenyl]hydrazone3b

Yield: 48%, m.p. 246 °C (ethanol). UV (ethanol):λmax

(log e) = 241 (4.1912), 295 (4.3144), 369 (4.6061); IR(KBr): ν = 3 200 cm–1 (NH), 1 690 (C=O, ketone), 1 650(C=O, hydrazone);1H-NMR (CDCl3): δ (ppm) = 2.52 (s,3H, COCH3), 2.62 (s, 3H, COCH3), 7.47–8.01 (m, 8H,Ar-H), 8.67 (s, 1H, CH=N), 9.09 (s, 1H, -CONHN=CH-),14.58 (s, 1H, -CH=N–NH-). Analysis: C18H17N5O3 (%calculated/found): 61.53/61.05 (C); 4.84/4.73 (H); 19.93/19.68 (N).

4.1.2.3. 2,3,4-pentanetrione-3-[4-[[(phenyl)methylene]-hydrazinocarbonyl]phenyl]hydrazone3c

Yield: 75%, m.p. 199–204 °C (ethanol). UV (ethanol)λmax (log e) = 290 (4.2762), 369 (4.5297); IR (KBr):ν =3 260 cm–1 (NH), 1 684 (C=O, ketone), 1 642 (C=O,hydrazone);1H-NMR (DMSO-d6): δ (ppm) = 2.46 (s,6H, COCH3), 7.27–8.11 (m, 9H, Ar-H), 8.46 (s, 1H,CH=N), 11.74 (s, 1H, -CONHN=CH-), 13.71 (s, 1H,-CH=N–NH-). Analysis: C19H18N4O3 (% calculated/found): 65.13/64.30 (C); 5.18/5.20 (H); 15.99/15.41(N).

4.1.2.4. 2,3,4-pentanetrione-3-[4-[[(2-fluorophenyl)-methylene]hydrazinocarbonyl]phenyl]hydrazone3d

Yield: 67%, m.p. 214 °C (ethanol). UV (ethanol):λmax

(log e) = 233 (4.1269), 297 (4.3132), 371 (4.6158); IR(KBr): ν = 3 450–3 400 cm–1 (H2O), 3 200 (NH), 1 680(C=O, ketone), 1 625 (C=O, hydrazone);1H-NMR(DMSO-d6): δ (ppm) = 2.46 (s, 6H, COCH3), 7.28–8.00(m, 8H, Ar-H), 8.71 (s, 1H, CH=N), 11.69 (s, 1H,-CONHN=CH-), 13.76 (s, 1H, -CH=N–NH-). Analysis:C19H17FN4O3⋅H2O (% calculated/found): 59.06/59.15(C); 4.95/4.82 (H); 14.50/15.02 (N).

4.1.2.5. 2,3,4-pentanetrione-3-[4-[[(4-fluorophenyl)-methylene]hydrazinocarbonyl]phenyl]hydrazone3e

Yield: 41%, m.p. 218 °C (ethanol). UV (ethanol):λmax

(log e) = 285 (4.5454), 371 (4.8213); IR (KBr):ν =3 500–3 450 cm–1 (H2O), 3 284 (NH), 1 684 (C=O, ke-tone), 1 625 (C=O, hydrazone);1H-NMR (DMSO-d6): δ(ppm) = 2.42 (s, 6H, COCH3),7.30–7.94 (m, 8H, Ar-H),8.47 (s, 1H, CH=N), 11.81 (b, ½H, -CONHN=CH-).Analysis: C19H17FN4O3⋅H2O (% calculated/found):59.06/58.57 (C); 4.95/4.88 (H), 14.50/14.47(N).

4.1.2.6. 2,3,4-pentanetrione-3-[4-[[(4-nitrophenyl)-methylene]hydrazinocarbonyl]phenyl]hydrazone3f

Yield: 14%, m.p. 284 °C (ethanol). UV (ethanol):λmax

(log e) = 236 (4.1306), 265 (4.0161) sh, 372 (4.5728); IR(KBr): ν = 3 500–3 450 cm–1 (H2O), 3 330 (NH), 1 672(C=O, ketone), 1 637 (C=O, hydrazone);1H-NMR(DMSO-d6): δ (ppm) = 2.46 (s, 6H, COCH3), 7.70–8.31

(m, 8H, Ar-H), 8.56 (s, 1H, CH=N), 12.09 (b, 1H,-CONHN=CH-), 13.74 (b, 1H, -CH=N–NH-). Analysis:C19H17N5O5⋅½H2O (% calculated/found): 56.43/56.38(C); 4.48/4.14 (H);17.32/17.18 (N).

4.1.2.7. 2,3,4-pentanetrione-3-[4-[[(4-hydroxyphenyl)-methylene]hydrazinocarbonyl]phenyl]hydrazone3g

Yield: 38%, m.p. 239 °C (ethanol). UV (ethanol):λmax

(log e) = 240 (4.4540), 275 (4.0277) sh, 372 (4.8302); IR(KBr): ν = 3 500 cm–1 (OH), 3 295 (NH), 1 672 (C=O,ketone), 1 642 (C=O, hydrazone);1H-NMR (DMSO-d6):δ (ppm) = 2.46 (s, 6H, COCH3), 6.83–8.00 (m, 8H,Ar-H), 8.36 (s, 1H, CH=N), 9.69 (s, 1H, Ar-OH), 13.52(b, ½H, -CONHN=CH-), 14.28 (b, ½H, -CH=N–NH-).Analysis: C19H18N4O4⋅2H2O (% calculated/found):56.71/56.07 (C); 5.51/4.80 (H); 13.92/14.68 (N).

4.1.2.8. 2,3,4-pentanetrione-3-[4-[[(4-hydroxy-3-meth-oxyphenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone3h

Yield: 69%, m.p. 178 °C (ethanol). UV (ethanol):λmax

(log e) = 237 (4.1504), 293 (4.1156) sh, 372 (4.4509); IR(KBr): ν = 3 500 cm–1 (OH), 3 307 (NH), 1 678 (C=O,ketone), 1 625 (C=O, hydrazone);1H-NMR (DMSO-d6):δ (ppm) = 2.46 (s, 6H, COCH3), 3.83 (s, 3H,-OCH3),6.84–7.31 (m, 3H, Ar-H), 7.66 (d, 2H, o-NH,J = 8.6 Hz),7.96 (d, 2H, m-NH,J = 8.6 Hz), 8.34 (s, 1H, CH=N), 9.48(s, 1H, Ar-OH), 11.56 (b, 1H, -CONHN=CH-), 13.69 (b,½H, -CH=N–NH-). Analysis: C20H20N4O5⋅4H2O (%calculated/found): 51.28/51.40 (C); 5.98/4.76 (H); 11.96/11.84 (N).

4.1.2.9. 2,3,4-pentanetrione-3-[4-[[(4-dimethylamino-phenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone3i

Yield: 22%, m.p. 228 °C (ethanol). UV (ethanol):λmax

(log e) = 240 (4.0491), 384 (4.4508); IR (KBr):ν =3 450–3 400 cm–1 (H2O), 3 225 (NH), 1 678 (C=O, ke-tone), 1 625 (C=O, hydrazone); EI-MS (70 eV): m/z (%)= 393 (66.18) [M+], 351 (1.88), 350 (2.42), 308 (1.04),282 (14.09), 281 (1.71), 231 (100), 163 (29.73), 162(15.75), 146 (47.29), 120 (50.04), 92 (10.39), 43 (48.83);1H-NMR (DMSO-d6): δ (ppm) = 2.46 (s, 6H, COCH3),2.98 (s, 6H,-N(CH3)2), 6.77 (d, 2H, o-N(CH3)2, J = 8Hz), 7.55 (d, 2H, m-N(CH3)2, J = 8 Hz), 7.68 (d, 2H,o-NH, J = 8.3 Hz), 7.97 (d, 2H, m-NH,J = 8.3 Hz), 8.32(s, 1H, CH=N), 11.49 (s, 1H, -CONHN=CH-), 13.80 (s,1H, -CH=N–NH-). Analysis: C21H23N5O3⋅H2O (%calculated/found): 61.31/61.39 (C); 6.08/6.08 (H);17.03/17.45 (N).

1098

4.1.2.10. 2,3,4-pentanetrione-3-[4-[[3-(4-dimethylamino-phenyl)-2-propenylidene]hydrazinocarbonyl]phenyl]hy-drazone3j

Yield: 47%, m.p. 225 °C (ethanol). UV (ethanol):λmax

(log e) = 249 (4.4045), 402 (4.8812); IR (KBr):ν =3 177 cm–1 (NH), 1 672 (C=O, ketone), 1 649 (C=O,hydrazone); EI-MS (70 eV): m/z (%) = 419 (7.39) [M+],376 (3.98), 308 (25.37), 231 (46.63), 188 (9.57), 172(30.68), 145 (4.34), 136 (7.60), 120 (100), 92 (20.39), 65(12.30), 43 (32.25);1H-NMR (DMSO-d6): δ (ppm) =2.46 (s, 6H, COCH3), 3.23 (s, 6H, -N(CH3)2), 6.71 (d,2H, o-N(CH3)2, J = 8.6 Hz), 6.76–6.96 (m, 2H,-CH=CH-), 7.43 (d, 2H, m-N(CH3)2, J = 8.6 Hz), 7.65 (d,2H, o-NH,J = 8.6 Hz), 7.94 (d, 2H, m-NH,J = 8.6 Hz),8.33 (d, 1H, -CH=N), 11.46 (s,1H, -CONHN=CH-),13.72 (s, ½H, -CH=N–NH-). Analysis C23H25N5O3⋅H2O(% calculated/found): 63.14/63.69 (C); 6.22/5.95 (H);16.01/15.57(N).

4.2. Microbiological procedures

4.2.1. In vitro evaluation of antimycobacterial activityagainstM. fortuitum

Preparation of Mycobacterial inoculum required a fewmodifications due to the difficulty of obtaining a homog-enous suspension ofM. fortuitum in the broth used. Fouror five colonies ofM. fortuitum which were previouslygrown in tryptic soy agar (TSA) after 72 h of incubationat 30 °C were collected by means of a swab and sus-pended in 4.5 mL of Mueller-Hinton broth enriched withTween 80 (0.2%). Following the inclusion of 4–5 glassbeads, this mixture was whirled using a vortex-mixer toensure a good suspension. The density of this culture wasthen adjusted to a turbidity equivalent to that of a 0.5McFarland standard and finally, the adjusted culture wasdiluted with sterile water so that, after inoculation, eachmicroplate well had an inoculum size of 1.5× 105

cfu/mL.Antimycobacterial testing of all compounds was car-

ried out in Mueller-Hinton broth enriched with Tween 80(0.2%) at pH 7.3. Tobramycin, which is an activeantibiotic against rapidly growing mycobacteria, wasselected as the standard drug. Quality control strains usedin the present study wereE. coli and S. aureus. Tobra-mycin was used at a concentration range of 4–0.03µg/mLagainst the quality control strains. The standard drug andtest compounds were dissolved in water and DMSO,respectively, and were diluted with the broth used. Theconcentration intervals were 32–0.5µg/mL and128–1µg/mL for the standard drug and the test com-pounds, respectively. Microplate wells, containing 100µlof broth with Tobramycin or the test compounds, were

then inoculated with 10µl of M. fortuitum suspensionwhose preparation is described above. Sheep-bloodedagar was used for the purity control. After incubation for72 h at 30 °C, the last microplate well with no growth ofmicroorganism was recorded to represent the MIC ex-pressed inµg/mL [21–24].

4.2.2. In vitro evaluation of antimycobacterial activityagainstM. tuberculosisH37Rv andM. avium

A primary screen was conducted at 12.5µg/mL (ormolar equivalent of highest molecular weight compoundin a series of congeners) againstM. tuberculosisH37Rvin BACTEC 12B medium using the BACTEC 460radiometric system. Compounds effecting< 90% inhibi-tion in the primary screen (MIC> 12.5µg/mL) were notevaluated further. Compounds demonstrating at least90% inhibition in the primary screen were re-tested atlower concentration (MIC) in a broth microdilution assaywith alamar Blue. The MIC was defined as the lowestconcentration inhibiting 99% of the inoculum. Thesecompounds were also tested againstM. avium, a naturallydrug-resistant opportunistic pathogen, using the sametechnique. Concurrent with the determination of MIC’s,compounds were tested for cytotoxicity (IC50) in VEROcells at concentrations equal to and greater than the MICfor M. tuberculosisH37Rv. After 72 h exposure, viabilitywas assessed on the basis of cellular conversion of MTTinto a formazan product using the Promega CellTiter 96Non-radioactive Cell Proliferation Assay.

4.2.3. BACTEC radiometric method of susceptibilitytesting

Inocula for susceptibility testing were either from apositive BACTEC isolation vial with a growth index (GI)of 500 or more, or a suspension of organisms isolatedearlier on a conventional medium. The culture was wellmixed with a syringe and 0.1 mL of a positive BACTECculture was added to each of the vials containing the testdrugs. The drug vials contained rifampicin (0.25µg/mL).A control vial was inoculated with a 1:100 dilution of theculture. A suspension equivalent to a McFarland No. 1standard was prepared in the same manner as a BACTECpositive vial, when growth from a solid medium wasused. Each vial was tested immediately on a BACTECinstrument to provide CO2 in the headspace. The vialswere incubated at 37 °C and tested daily with a BACTECinstrument. When the GI in the control reads at least 30,the increase in GI (∆ GI) from the previous day in thecontrol was compared with that in the drug vial. Thefollowing formula was used to interpret results:

∆ GI control > ∆ GI drug = susceptible∆ GI control < ∆ GI drug = resistant

1099

If a clear susceptibility pattern (the difference of∆ GIof control and the drug bottle) was not seen at the time thecontrol GI is 30 the vials were read for 1 or 2 additionaldays to establish a definite pattern of∆ GI differences.

Acknowledgements

We thank Dr Joseph A. Maddry from the TuberculosisAntimicrobial Acquisition and Coordinating Facility(TAACF), National Institute of Allergy and InfectiousDiseases Southern Research Institute, GWL Hansen’sDisease Center, Colorado State University, Birmingham,Alabama, USA, for the in vitro evaluation of antimyco-bacterial activity usingM. tuberculosisH37Rv andM.avium.

References

[1] Patel J.M., Dave M.P., Langalia N.A., Thaker K.A., J. Indian Chem.Soc. 61 (8) (1984) 718–720.

[2] Jani M.K., Undavia N.K., Trivedi P.B., J. Indian Chem. Soc. 67 (7)(1990) 601–602.

[3] Ergenç N., Rollas S., J. Fac. Pharm. Istanbul 10 (1974) 77–86.

[4] Ergenç N., Rollas S., J. Fac. Pharm. Istanbul 11 (1975) 8–23.

[5] Rollas S., Ergenç N., Bankaogˇlu G., Sariogˇlu E., J. Fac. Pharm.Istanbul 18 (1982) 107–115.

[6] Kalsi R., Pande K., Bhalla T.N., Barthwal J.P., Gupta G.P., ParmarS.S., J. Pharm. Sci. 79 (4) (1990) 317–320.

[7] Pande K., Kalsi R., Bhalla T.N., Barthwal J.P., Indian J. Pharm. Sci.51 (1) (1989) 18–21.

[8] Kalsi R., Pande K., Barthwal J.P., Indian J. Chem. Sect. B 27 B (2)(1988) 197–199.

[9] Tiwari M.M., Agarwal M., Saxena V.K., Bajpai S.K., Joshi M.M.,Indian J. Pharm. Sci. 57 (3) (1995) 113–116.

[10] Singh V., Srivastava V.K., Palit G., Shanker K., Arzneim.-Forsch.Drug Res. 42 (II) (1992) 993–996.

[11] Pabuççuogˇlu M.V., Rollas S., J. Pharm. Univ. Mar. 7 (1) (1991)39–49.

[12] Küçükgüzel Sq.G., Rollas S., Erdeniz H., Kiraz M., Eur. J. Med.Chem. 34 (2) (1999) 153–160.

[13] Kömürcü Sq.G., Rollas S., Ülgen M., Gorrod J.W., ÇevikbasqA., Boll.Chim. Farmaceutico 134 (7) (1995) 375–379.

[14] Casagrande C., Canova M., Ferrari G., Farmaco (Pavia), Ed. Sci. 20(8) (1965) 544–556. Chem. Abstr. 64 (1966) 9658d.

[15] Grammaticakis P., Bull. Soc. Chim. France, (1957) 1242–1254.

[16] Malik W.U., Garg H.G., Arora V., J. Pharm. Sci. 60 (11) (1971)1738–1740.

[17] Garg H.G., Arora V., J. Pharm. Sci. 61 (1) (1972) 130–132.

[18] Elguero J., Jacquier R., Tarrago G., Bull. Soc. Chim. France 9(1966) 2981–2989.

[19] Ergenç N., Rollas S., Demir S., Özdemir F., J. Fac. Pharm. Istanbul11 (1975) 183–192.

[20] Ergenç N., Rollas S., Çapan G., Dogˇan N., Özger Y., Pharmazie 44(8) (1989) 573–574.

[21] Hawkins J.E., Wallace Jr. R.J., Brown B.A., in: Balows A., HauslerJr. W.J., Herrmann K.L., Isenberg H.O., Shadami H.J. (Eds.),Manual of Clinical Microbiology, 5th ed., American Society forMicrobiology, Washington, D.C., 1991, pp. 1138–1152.

[22] NCCLS (1990). Methods for Dilution Susceptibility Tests forBacteria That Grow Aerobically, 2nd ed. Approved Standard.NCCLS document M7-A2. NCCLS, Villanova, Pa.

[23] NCCLS (1991). Performance Standards for Antimicrobial Suscep-tibility Testing. 3rd Informational Supplement. M 100-S3. NCCLS,Villanova, Pa.

[24] Swenson J.M., Thornsberry C., Silcox V.A., Antimicrob. AgentsChemother. 22 (1982) 186–192.

[25] Inderleid C.B., in: Lorian V. (Ed.), Antibiotics in LaboratoryMedicine, 3rd ed. Williams and Wilkins, Baltimore, 1991, p. 134.

[26] Karali N., Terzioglu N., Gürsoy A., Arzneim.-Forsch. Drug Res. 48(II) (1998) 7758–763.

1100

Laboratory note

Synthesis and pharmacological properties of novel benzamide derivatives actingas ligands to the 5-hydroxytryptamine 4 (5-HT4) receptor

Katsuhiko Itoh*, Hideo Tomozane, Hidetoshi Hakira, Shuji Sonda, Kiyoshi Asano,Masatake Fujimura, Noriko Sato, Keiichiro Haga, Takeshi Kawakita

Research Laboratories, Yoshitomi Pharmaceutical Industries Ltd., 955 Koiwai, Yoshitomi-cho, Chikujo-gun, Fukuoka 871-8550, Japan

(Received 12 April 1999; revised 15 June 1999; accepted 16 June 1999)

Abstract – A series of 4-amino-5-chloro-2-methoxy-N-(1-substituted piperidin-4-ylmethyl)benzamides was synthesized as novel gastropro-kinetic agents. The affinity of these compounds for the 5-hydroxytryptamine 4 (5-HT4) receptor was evaluated. Among these compounds,4-amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide (3f, Y-34959) showed ahigher affinity for the 5-HT4 receptor (Ki = 0.30 nmol/L) than for other receptors, and was confirmed to be a potent 5-HT4 receptor agonisthaving contractile effects in the isolated guinea-pig ascending colon (EC50 = 1.2 nmol/L). In dogs, compound3f increased gastroprokineticmotility of both the gastric antrum and the ascending colon. In addition, this effect on the colon was inhibited by azasetron, a selective 5-HT3receptor antagonist, demonstrating that the effect of gastroprokinetic agents having 5-HT3 receptor antagonism on the colon were reducedcompared with that of selective 5-HT4 receptor agonists. © 1999 Éditions scientifiques et médicales Elsevier SAS

5-HT4 receptor agonist / 5-HT4 receptor agonism / 5-HT3 receptor antagonism / N-[1-[5-(1-methylindol-3-ylcarbonyl-amino)pentyl]piperidin-4-ylmethyl]benzamide / gastrointestinal motility

1. Introduction

Activation of the 5-hydroxytryptamine 4 (5-HT4) re-ceptor mediates a wide variety of effects in the centraland peripheral nervous systems [1]. Benzamides (meto-clopramide [2], cisapride [3], etc.) are used clinically asgastrointestinal motility stimulants. Recently, this gastro-prokinetic effect is thought to be mediated by 5-HT4

receptor agonism [4]. In addition, we also found thecontractile potency in the isolated guinea-pig ascendingcolon and the binding affinity for the 5-HT4 receptor werewell correlated among these benzamides [5]. On the otherhand, these benzamides show antagonistic activityagainst various receptors such as dopamine D2, 5-HT2

and 5-HT3 receptors [6–9], and the antagonism shouldreduce the gastroprokinetic effect of 5-HT4 receptoragonism and/or cause unfavourable side effects. Forexample, it is well known that the 5-HT3 receptor

antagonists induce constipation, and D2 receptor antago-nists cause central nervous system effects such as depres-sion and extrapyramidal syndrome [10]. Therefore, thesearch for selective 5-HT4 receptor agonists would be animportant goal to develop novel useful gastrointestinalmotility stimulants.

In the course of our synthetic studies on 5-HT4

receptor agonists, we have found that the essentialframework for selective 5-HT4 receptor agonism was4-amino-5-chloro-2-methoxy-N-(piperidin-4-ylmethyl)-benzamide with the polar side chain at the 1- position onthe piperidine ring (compound1 has a methylsulfonyl-aminoethyl group as a polar side chain as shownin figure 1) [11]. Modifications of the polar side chain ofcompound1 led to the discovery of a novel gastropro-kinetic agent, 4-amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide (3f, Y-34959). Compound3f wasconfirmed to be a selective 5-HT4 receptor agonist.Herein, we describe the synthesis and the pharmacologi-cal data of the novel selective 5-HT4 receptor agonist.*Correspondence and reprints

Eur. J. Med. Chem. 34 (1999) 1101−1108 1101© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved

2. Chemistry

The general synthetic procedure used in this study isillustrated in figure 2. 1-Substituted-4-(tert-butoxy-carbonylaminomethyl)piperidine derivatives (2a–2f) [12]were deprotected with hydrogen chloride in dioxane,affording the corresponding amines, which were coupledwith 4-amino-5-chloro-2-methoxybenzoic acid using1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydro-chloride (WSC HCl) and 1-hydroxybenzotriazole (HOBt)in the presence of triethylamine (TEA) to afford targetbenzamides (3a–3f), whose data for physicochemicalproperties and 5-HT4 receptor affinities are listed intable I.

3. Pharmacological data and discussion

The affinity of compounds3a–3f for the 5-HT4 recep-tor was determined as their ability to inhibit the binding

of [3H]GR113808 to the receptor. The agonistic activityof the candidate was evaluated as the contractile ability ofthe ascending colon in guinea-pigs. In in vivo studies, thegastrointestinal motility index of the candidate in con-scious dogs was measured.

Binding affinities for the 5-HT4 receptor of1, 3a–3f,and 5-HT are listed intable I, where 5-HT is the referencecompound. Phenylsulfonylamine derivative3a showed ahigher affinity for the 5-HT4 receptor than methylsulfo-nylamine derivative1. Next, we investigated the influ-ence of other polar side chains. A benzamide group wasselected, regarding the group of side chains, because anamide group is considered to be a bioisoster of thesulfonylamine group [13] and the introduction of theamide group is synthetically easier. Among the benza-mide derivatives (3b–3d), anN-pentylbenzamide deriva-tive 3d showed the highest affinity for the 5-HT4 receptorindicating that a bulky group is needed as a substituent onthe piperidine ring at the 1- position and that a largepocket for the substituent should exist at the receptor.This consideration is supported by the López-Rodríguezreport [14], which described a 5-HT4 receptor mappingusing an active analogue approach.

Next, the phenyl group was replaced by a 1-methylindole group.N-butyl-1-methylindole-3-carbox-amide derivative3e showed a higher affinity than thecorresponding compound3c. N-pentyl-1-methylindole-3-carboxamide derivative3f showed a subnanomolar affin-ity. These results suggest that the 1-methylindole-3-carboxamide group may interact with some amino acid

Figure 1. Chemical structure of compound1.

Figure 2. Synthesis of benzamides3a–3f. a: HCl in 1,4-dioxane. b: TEA, 4-amino-5-chloro-2-methoxybenzoic acid, WSC HCl,HOBt, DMF.

1102

residues in the pocket of the 5-HT4 receptor. Compound3f was devoid of significant affinities for other receptors,and produced concentration dependent contractions of theguinea-pig ascending colon (table II). Therefore, com-pound3f was confirmed to be a selective 5-HT4 receptoragonist.

Finally, we examined effects of compound3f ongastrointestinal motility in conscious dogs in a postpran-dial state. When compound3f was given intravenously todogs at 0.01 mg/kg, the motility of the gastric antrum andascending colon was rapidly enhanced, as shown in

figure 3,and the motility index of the gastric antrum andascending colon increased. In addition, we tested theinfluence of azasetron [15], a selective 5-HT3 receptorantagonist, on the stimulated gastrointestinal motility bya selective 5-HT4 receptor agonist compound3f. Aza-setron inhibited the increase in the colonic motility indexcaused by compound3f. In contrast, azasetron did notinfluence the increase of the antral motility index (fig-ure 4). These results demonstrated that gastroprokineticagents which have 5-HT3 receptor antagonism would beless effective in increasing the motility of the ascending

Table I. Physicochemical properties and 5-HT4 receptor affinities of compounds1 and3a–3f.

Compound R X n M.p. (°C) Mass (m/z) Formula Binding affinity,Ki (nmol/L)b

1 CH3 SO2 2 177–178 418 C17H27ClN4O4S 2 C2H2O4c 9.6

3a SO2 2 Amorphous solid 480 C22H29ClN4O4S 3/4H2O 2.6

3b CO 3 Amorphous solid 458 C24H31ClN4O3 H2O 9.4

3c CO 4 165–168 472 C25H33ClN4O3 4.0

3d CO 5 122–124 486 C26H35ClN4O3 1/2H2O 1.7

3e CO 4 Amorphous solid 525 C28H36ClN5O3 H2O 1.6

3f CO 5 161–163 539 C29H38ClN5O3 HCl 0.3

5-HT 130

aElementary analyses were performed for C, H and N and were within± 0.4% of the calculated values for formulae shown.bEach value isthe mean from triplicate assays in a single experiment.coxalate.

1103

colon than selective 5-HT4 receptor agonists and thatselective 5-HT4 receptor agonists should be able toenhance both upper and lower gastrointestinal motility.

4. Conclusion

We described the synthesis, and affinity for the 5-HT4

receptor, of a series of benzamides. Among them,4-amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-

ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide(3f, Y-34959) was found to be a selective 5-HT4 receptoragonist. Compound3f increased the gastrointestinal mo-tility of both the gastric antrum and the ascending colonin conscious dogs in a postprandial state. Azasetron, aselective 5-HT3 receptor antagonist, inhibited the in-crease of the colonic motility index caused by compound3f without influencing the increase of the gastric motilityindex. Based on our results, we proposed that the selec-

Figure 3. Typical tracings of the effect of compound3f on gastrointestinal motility in conscious dogs in a postprandial state.

Table II. Binding profiles and 5-HT4 receptor agonistic activity of compound3f.

Binding affinitya, Ki (nmol/L)5-HT4 receptor agonistic activity

D2 5-HT1A 5-HT2 5-HT3 5-HT4

rat striatum rat hippocampus rat cerebral cortex rat striatum guinea-pig striatum EC50 (nmol/L)c Maximal response (%)d

[3H]spiperone [3H]8-OH-DPAT [3H]ketanserin [3H]granisetron [3H]GR1138081.2± 0.3 6.3± 0.1

> 1 000b > 1 000b 110 > 1 000b 0.3

aEach value is the mean from triplicate assays in a single experiment.bIC50 value. cEC50 values (mean± SE) was determined by linearregression.dA percentage (mean± SE) of the contraction caused by methacholine (30µmol/L).

1104

tive 5-HT4 receptor agonists were novel gastrointestinalmotility stimulants which can enhance both upper andlower gastrointestinal motility with few side effects.

5. Experimental protocols

5.1. Chemistry

All melting points were measured in open capillariesand are uncorrected. Proton nuclear magnetic resonance(1H-NMR) spectra were recorded on JEOL JNM-EX270spectrometers and chemical shifts are expressed in ppmwith tetramethylsilane (TMS) as an internal standard.Signal multiplicities are represented by s (singlet), d(doublet), t (triplet), q (quartet), br-s (broad singlet) andm (multiplet). Mass spectra (MS) were taken on JEOLJMS-O1SG spectrometers. Elementary analysis was per-formed for C, H, N and were within± 0.4% of thecalculated values. Silica-gel plates (Merck F254) andsilica gel 60 (Merck, 70–230 mesh) were used foranalytical and preparative column chromatography, re-spectively.

5.1.1. General procedure for the preparation of3a–3f

5.1.1.1. 4-Amino-5-chloro-2-methoxy-N-[1-[5-(1-methyl-indol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide hydrochloride3f

A mixture of 2f (5.7 g, 13 mmol) and 4 mol/L hydro-gen chloride in 1,4-dioxane (Aldrich) (60 mL) was stirred

under ice-cooling for 1 h. After evaporation, the residuewas washed with CHCl3, and the resulting compound wasdissolved in 80 mL of dimethylformamide (DMF). To thesolution were added triethylamine (TEA) (5.7 mL,41 mmol), 4-amino-5-chloro-2-methoxybenzoic acid(2.6 g, 13 mmol) and 1-hydroxybenzotriazole (HOBt)(1.9 g, 14 mmol). The reaction mixture was stirred atroom temperature for 1 h and then 1-ethyl-3-[3-(di-methylamino)propyl]carbodiimide hydrochloride (WSC)(2.7 g, 14 mmol) was added under ice-cooling. Stirringwas continued overnight at room temperature. Afterevaporation, 5% aqueous sodium bicarbonate was addedto the residue and extracted with CHCl3. The extract waswashed with brine and dried over anhydrous magnesiumsulfate. After evaporation in vacuo, the residue waschromatographed on silica gel eluting withCHCl3–MeOH–NH4OH (100:10:1) and treated with analcoholic solution of hydrogen chloride. The precipitateswere collected and recrystallized from ethanol to give3f(3.1 g, 43%);1H-NMR (DMSO-d6) δ: 1.25–1.41 (2H,m), 1.47–1.62 (4H, m), 1.67–1.95 (4H, m), 2.80–2.93(2H, m), 2.96 (2H, br-s), 3.17 (2H, br-s), 3.26 (2H, dd,J= 5.9, 13 Hz), 3.40–3.46 (2H, m), 3.82 (3H, s, CH3N),3.83 (3H, s, CH3O), 5.95 (2H, s, Ar-NH2), 6.50 (1H, s,Ar-3-H), 7.13 (1H, t,J = 6.6 Hz, ind-5-H), 7.20 (1H, t,J = 6.6 Hz, ind-6-H), 7.46 (1H, d,J = 7.9 Hz, ind-7-H),7.66 (1H, s, ind-2H), 7.93 (1H, t,J = 5.3 Hz, CONHCH2),8.02 (1H, s, Ar-6-H), 8.14 (1H, d,J = 7.9 Hz, ind-4-H),10.20 (1H, br-s, NHCO).

Figure 4. Effect of azasetron on the increase in the gastrointestinal motility of gastric antrum and ascending colon caused bycompound3f.

1105

5.1.1.2. 4-Amino-5-chloro-2-methoxy-N-[1-[2-(phenyl-sulfonyl)aminoethyl]piperidin-4-yl]benzamide3a

Similarly to 3f, 3a was prepared starting from2a(1.6 g, 3.9 mmol), hydrogen chloride in 1,4-dioxane(11 mL), TEA (1.1 mL, 12 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (0.78 g, 3.9 mmol), HOBt (0.58 g,4.3 mmol), DMF (20 mL), and WSC (0.82 g, 4.3 mmol)to give3a (0.55 g, 29%);1H-NMR (CDCl3) δ: 1.20 (2H,dd,J = 11, 15 Hz), 1.53–1.67 (3H, m), 1.86 (2H, t,J = 9.9Hz), 2.36 (2H, t,J = 5.9Hz), 2.59 (2H, d,J = 11 Hz), 2.98(2H, t, J = 6.6 Hz), 3.31 (2H, t,J = 6.0 Hz), 3.92 (3H, s,CH3O), 4.39 (2H, br-s, Ar-NH2), 6.31 (1H, s, Ar-3-H),7.48–7.60 (3H, m, Ar-H), 7.60 (1H, t,J = 6.0 Hz,CONHCH2), 7.84 (2H, dd,J = 2.0, 6.6 Hz, Ar-H), 8.11(1H, s, Ar-6-H).

5.1.1.3. 4-Amino-N-[1-(3-benzoylaminopropylpiperidin)-4-ylmethyl]-5-chloro-2-methoxybenzamide3b

Similarly to 3f, 3b was prepared starting from2b(1.0 g, 2.7 mmol), hydrogen chloride in 1,4-dioxane(11 mL), TEA (1.1 mL, 8.1 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (0.54 g, 2.7 mmol), HOBt (0.41 g,3.0 mmol), DMF (15 mL), and WSC (0.58 g, 3.0 mmol)to give3b (0.10 g, 11%),1H-NMR (CDCl3) δ: 1.44 (2H,dd,J = 12, 22 Hz), 1.72–189 (5H, m), 2.13 (2H, t,J = 11Hz), 2.68 (2H, t,J = 6.0 Hz), 3.15 (2H, d,J = 12 Hz), 3.32(2H, t, J = 6.6 Hz), 3.56 (2H, dd,J = 3.4, 12 Hz), 3.86(3H, s, CH3O), 4.42 (2H, s, Ar-NH2), 6.29 (1H, s,Ar-3-H), 7.37–7.44 (3H, m, Ar-H), 7.74 (1H, t,J = 6.0Hz, CONHCH2), 7.84 (2H, dd,J = 2.0, 8.0 Hz, Ar-H),8.10 (1H, s, Ar-6-H), 8.28 (1H, br-s, CONH).

5.1.1.4. 4-Amino-N-[1-(4-benzoylaminobutyl)piperidin-4-ylmethyl]-5-chloro-2-methoxybenzamide3c

Similarly to 3f, 3c was prepared starting from2c(0.86 g, 2.2 mmol), hydrogen chloride in 1,4-dioxane(9 mL), TEA (0.90 mL, 6.6 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (0.44 g, 2.2 mmol), HOBt (0.32 g,2.4 mmol), DMF (15 mL), and WSC (0.46 g, 2.4 mmol).The resulting solid was recrystallized from ethanol–iso-propanol to give3c (0.20 g, 19%);1H-NMR (CDCl3-CD3OD) δ: 1.32–1.56 (4H, m), 1.64–1.89 (5H, m), 2.13(2H, t, J = 11 Hz), 2.68 (2H, t,J = 6.0 Hz), 3.21 (2H, d,J = 12 Hz), 3.39 (2H, t,J = 6.6 Hz), 3.43 (2H, t,J = 6.6Hz), 3.91 (3H, s, CH3O), 4.45 (2H, br-s, Ar-NH2), 6.35(1H, s, Ar-3-H), 7.33–7.42 (3H, m, Ar-H), 7.74 (1H, t,J= 6.0 Hz, CONHCH2), 7.78 (2H, dd,J = 2.0, 8.0 Hz,Ar-H), 8.01 (1H, s, Ar-6-H).

5.1.1.5. 4-Amino-N-[1-(5-benzoylaminopentyl)piperidin-4-ylmethyl]-5-chloro-2-methoxybenzamide3d

Similarly to 3f, 3d was prepared starting from2d(1.7 g, 4.2 mmol), hydrogen chloride in 1,4-dioxane

(17 mL), TEA (1.7 mL, 13 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (0.84 g, 4.2 mmol), HOBt (0.62 g,4.6 mmol), DMF (20 mL), and WSC (0.88 g, 4.6 mmol).The resulting solid was recrystallized from ethyl acetateto give 3d (1.4 g, 72%);1H-NMR (CDCl3) δ: 1.31–1.49(4H, m), 1.50–1.85 (7H, m), 1.91 (2H, t,J = 12 Hz), 2.32(2H, t, J = 7.8 Hz), 2.92 (2H, d,J = 12 Hz), 3.30 (2H, t,J = 6.0 Hz), 3.34 (2H, dd,J = 7.3, 13 Hz), 3.88 (3H, s,CH3O), 4.46 (2H, s, Ar-NH2), 6.30 (1H, s, Ar-3-H), 6.36,(1H, br-s, CONH), 7.33–7.42 (3H, m, Ar-H), 7.70 (1H,br-s, CONHCH2), 7.77 (2H, dd,J = 7.3 Hz, Ar-H), 8.09(1H, s, Ar-6-H).

5.1.1.6. 4-Amino-5-chloro-2-methoxy-N-[1-[4-(1-methyl-indol-3-ylcarbonylamino)butyl]piperidin-4-ylmethyl]benzamide3e

Similarly to 3f, 3e was prepared starting from2e(1.4 g, 3.2 mmol), hydrogen chloride in 1,4-dioxane(14 mL), TEA (1.3 mL, 9.6 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (0.66 g, 3.2 mmol), HOBt (0.47 g,3.5 mmol), DMF (30 mL), and WSC (0.67 g, 3.5 mmol)to give3e (0.63 g, 38%);1H-NMR (CDCl3) δ: 1.24 (2H,dd,J = 11, 15 Hz), 1.30–1.72 (9H, m), 1.91 (2H, t,J = 9.9Hz), 2.38 (2H, t,J = 6.6 Hz), 2.93 (2H, d,J = 11 Hz), 3.29(2H, t, J = 6.6 Hz), 3.50 (2H, dd,J = 6.0, 12 Hz), 3.81(3H, s, CH3N), 3.87 (3H, s, CH3O), 4.35 (2H, br-s,Ar-NH2), 6.20 (1H, br-s, CONH), 6.26 (1H, s, Ar-3-H),7.21–7.34 (2H, m, ind-5,6-H), 7.63 (1H, s, ind-2-H), 7.70(1H, t, J = 3.2 Hz, CONHCH2), 7.89 (1H, d,J = 8.0 Hz,ind-7-H), 7.92 (1H, d,J = 8.0 Hz, ind-4-H), 8.10 (1H, s,Ar-6-H).

5.2. Pharmacology

5.2.1. 5-HT4 receptor bindingMale Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka,

Japan) were killed by cervical dislocation and the stria-tum was separated from each brain. The striatum washomogenized in 15 volumes of 50 mmol/L ice-coldHEPES buffer (pH 7.4) with Polytron PT-10 and thencentrifuged at 35 000g for 20 min. The resulting pelletwas resuspended in the HEPES buffer and finally dilutedto the appropriate concentration for assay (6 mg wetweight per assay tube). This suspension was used as thetissue preparation. Assay tubes contained 50 mL ofHEPES buffer or a solution of the test agents, 50 mLsolution of [3H]GR113808 (Amersham International,UK) to give a final concentration of 0.1 nmol/L and900 mL of tissue preparation. Each tube was incubatedfor 30 min at 37 °C and the reaction was terminated byrapid filtration through a Whatmann GF/B filter (pre-soaked in 0.01% v/v polyethyleneimine) followed bywashing with 1× 4 mL of ice-cold HEPES buffer. Then

1106

the filter was placed in 3 mL of scintillant and theradioactivity was determined by scintillation counting ina Beckman model LS3801 scintillation counter. Nonspecific binding was defined in the presence of unlabelledGR113808 to give a final concentration of 1µmol/L. TheIC50 value was determined by non-linear regression ofthe displacement curve, and theKi value was calculatedaccording to the formula (Ki = IC50/(1 + L/Kd)), where Lis the concentration of radioligand andKd is the disso-ciation constant of the radioligand.

5.2.2. Binding to other receptorsThe binding studies of D2 [16], 5-HT1A [17],

5-HT2 [18], and 5-HT3 receptor [19] were carried outaccording to the previously published methods.

5.2.3. 5-HT4 receptor agonismFour male Hartley guinea-pigs (Japan SLC, Ltd.,

Shizuoka, Japan) were killed by cervical dislocation andthe ascending colon was removed. The longitudinalmuscle layer was separated from the underlying circularmuscle. Each muscle strip preparation of about 2.5 cmwas individually mounted vertically for isotonic measure-ment into a tissue bath containing 10 mL Tyrode solution.This solution was kept at 37 °C and gassed with 95% O2,5% CO2. The strips were subjected to a preload of 1 g andallowed to stabilize for 20 min. After stabilization, theresponse of the longitudinal muscle to 30µmol/L metha-choline was measured. Agonist concentration-effectcurves were constructed using sequential dosing, leaving15 min between doses. A 15 min dosing cycle wasrequired to prevent desensitization. The agonist was leftin contact with a preparation until the response hadreached a maximum, the preparation was then washed.Forty minutes was left between the determination ofconcentration-effect curves. GR113808 (10 nmol/L) wasincubated for 10 min before repeating the agonistconcentration-effect curves. After each determination of aconcentration-effect curve, 30µmol/L of methacholinewas added to the tissue bath again. All responses wereexpressed as a percentage of the mean of the twocontractions induced by 30µmol/L methacholine. TheEC50 value, the concentration causing 50% of the maxi-mal response, was determined by linear regression analy-sis.

5.2.4. Gastrointestinal motilityThree mongrel dogs were used. Under pentobarbital

anaesthesia, the abdomen was opened, and strain gaugetransducers (F-12SSH, Star Medical, Tokyo, Japan) weresutured onto the serosa of the stomach and colon in amanner to detect circular muscle contraction. The trans-ducers were placed at the greater curvature of the gastric

antrum, 5 cm proximal to the pyloric ring, and at thecolon, 10 cm distal to the ileo-colic junction. The animalswere then allowed more than 2 weeks to recover from thesurgery. To measure the gastrointestinal motor activity,the signals from the gastric antrum and colon wererecorded on a recording system (ESC-2000, Star medical,Tokyo, Japan) every 100 ms by a telemetry system(DAS-800T, Star medical, Tokyo, Japan). The motilityindex of contractile activity, shown as an area undercontraction wave, was calculated both in the gastricantrum and in the colon by means of a program (Peaks,AD Instruments, Australia). Measurements of gas-trointestinal motility were performed in a postprandialstate. Approximately 150 min after feeding of a dry typemeal (200 g), azasetron (0.1 mg/kg) or vehicle was givento the dog intravenously. Ten minutes later, test com-pound was injected intravenously. The motility indexduring 10 min before and after injection of the testcompound was determined, and data were expressed as adifference between the motility index before and thatafter the injection of the test compound. Investigationconsisted of six experiments in three dogs. Statisticalanalysis of data was performed by means of the Dunnettmethod.

Acknowledgements

We thank Mrs F. Matsugaki for some of the biologicalresults. We also thank Dr M. Terasawa and Dr K. Adachifor helpful discussion.

References

[1] Eglen R.M., Wong E.F., Dumuis A., Bockaert J., Trends Pharmacol.Sci. 16 (1995) 391–398.

[2] Harrington R.A., Hamilton C.W., Brogden R.N., Linkewich J.A.,Romankiewicz J.A., Heel R.C., Drugs 25 (1983) 451–494.

[3] McCallum R.W., Prakash C., Campoli-Richards D.M., Goa K.L.,Drugs 36 (1988) 652–681.

[4] Bockaert J., Fozard J.R., Dumuis A., Clarke D.E., Trends Pharma-col. Sci. 13 (1992) 141–145.

[5] Sato N., Asano K., Haga K., Arita M., Jpn. Pharmacol. Ther. 25(1997) 383–387.

[6] Turconi M., Schiantarelli P., Borsini F., Rizzi C.A., Ladinsky H.,Donetti A., Drugs Future 16 (1991) 1011–1026.

[7] Cushing D.J., Cohen M.L., J. Pharmacol. Exp. Ther. 264 (1993)193–199.

[8] Yoshida N., Omoya H., Furukawa K., Ito T., Karasawa T., Arch. Int.Pharmacodyn. Ther. 300 (1989) 51–67.

[9] Karasawa T., Yoshida N., Furukawa K., Omoya H., Ito T., Eur. J.Pharmacol. 183 (1990) 2181.

1107

[10] Fernandez A.G., Roberts D.J., Drugs Future 16 (1991) 885–892.

[11] Itoh K., Kanzaki K., Ikebe T., Kuroita T., Tomozane H., Sonda S.,Sato N., Haga K., Kawakita T., Eur. J. Med. Chem. 34 (1999)977–989.

[12] Kawakita T., Kuroita T., Murozono T., Hakira H., WO 95/26953Chem. Abstr. 124 (1996) 14913.

[13] Thornber C.W., Chem. Soc. Rev. 8 (1979) 563–580.

[14] López-Rodríguez M.L., Morcillo M.J., Benhaú B., Rosado M.L., J.Comput. -Aided Mol. Des. 11 (1997) 589–599.

[15] Kawakita T., Kuroita T., Yasumoto M., Sano M., Inaba K., FukudaT., Tahara T., Chem. Pharm. Bull. 40 (1992) 624–630.

[16] Creese I., Schneider R., Snyder S.H., Eur. J. Pharmacol. 46 (1977)377–381.

[17] Hall M.D., El Mestikawy S., Emerit M.B., Pichat L., Hamon M.,Gozlan H., J. Neurochem. 44 (1985) 1685–1696.

[18] Leysen J.E., Niemegeers C.E., Van Nueten J.M., Laduron P.M., Mol.Pharmacol. 21 (1982) 301–314.

[19] Nelson D.R., Thomas D.R., Biochem. Pharmacol. 38 (1989)1693–1695.

1108