wolf and lamb

8/7/2019 wolf and lamb

1/10

7620 J . Am. Chem. SOC. 981,103, 7620-1629then boiled briefly, cooled to room tem pera ture, and acidified with 20%DCI. The amino anhydride 3 that crystallized was removed by centri-fugation. The supernatant was again made basic with NaOD; during theaddition, more anhydride crystallized and was removed by centrifugation:NMR (of supernatant) 6 2.50 (3 H, s), 6.89 (1 H, d of d, J = 2 Hz, J= 8 Hz), 7.17 (1 H, d, J = 8 Hz), 7.37 (1 H, d, J = 2 Hz). Two singlets,at 6 2.78 and 7.54, were attributed to 4-methylaniline-3,5-disulfonate(about 12 mol %). Two multiplets at 6 7.79-8.00 and 8.59-8.75, at -tributed to residual naphthalimide impurities, integrated to a total of 0.78H. 5-Amiw-2-toluenesulfoNcAcid (Eastman): NMR (D20/OD-) 6 2.49( 3 H, S, CH3), 6.86 ( 1 H, d ofd , J = 2.5 Hz, J = 8 Hz, 4-H), 7.14 (1H, d, J = 8 Hz, 3-H), 7.34 (1 H, d, J =I 2.5 H, 6-H).( 3 H, S, CHS), 6.84 ( 1 H, d , J = 8 Hz, 5-H), 7.17 (1 H, d of d, J = 2Hz , J = 8 Hz, 4-H), 7.49 (1 H, d, J = 2 Hz, 2-H).

6-Amlno-3-tolwnesulfonicAcid (Aldrich): NMR (D20/OD-) 6 2.24Reaction of 5 with Propionaldehyde and Formaldehyde. Equal volumesof 0.2 M propionaldehyde and 0.2 M 5 in D 2 0 were mixed at roomtemperature, and the N MR spectrum was recorded as soon as possible.Within 120 s the a ldehydic proton (6 5: 9.67) was greatly reduced; a newtriplet (6 7.52, J = 5 Hz) had appeared. A similar experiment withformaldehyde was complicated by the fact that the resonance from theformaldehyde protons at 6 4.82 was too close to the HDO peak foradeq uate observation. A rapid reaction, however, seemed likely, sincethe first obtainable spectrum of the naphthalimide region was complex,

indicating a mixtu re of compounds. Gradually the spectrum of thisregion became less complex; after 30 min most of the spectrum appearedto be due to a single compound. By then two new doublets (6 6.68, J =12 Hz; 6 7.12, J = 11 Hz) had appeared.Reaction of 10with H20, Ethanolamine, and Mercaptoethanol. A60-MHz H N MR spectrum of a 10% (w/w) solution of 10 in D2 0wasobtained immediately after the compound was dissolved; this spectrumwas the same as that subsequently obtained after the solution had stoodat room temperature for 6 h, 30 h, 10 days, and 75 days. Compound 10was distinctly less stable in weak base. The N MR spectrum of a freshlyprepared 10% (w/w) solution of 10 in 0.1 M N a+ /D+ carbonate bufferin D20 was very similar to the spectrum of 10 in D20 . After 1 day atr w m temperature, the pD 9.1 spectrum was not noticeably different, butafter 9 days at room temperature the intensity of the vinyl protons wasreduced, and a new peak had appeared at 6 5: 3.50. It was estimatedthat between 20% and 50% of the compound had been converted, pre-sumably to the lithium salt of 7.The NMR spectrum of 52 mg of 10 in 0.4 mL of D2 0 and 0.1 mLof 0.5 M Na+ /D + pD 9.1 carbonate buffer was obtained. To the sampletube was added 14 WLof ethanolamine, and within 120 s the NMRspectrum of the vinyl region was recorded. No signals were seen, indi-cating that at least 90% of compound 10 had reacted. The completespectrum showed the presence of two multiplets a t 6 N 3.0and 6 = 3.7.A similar experiment was performed with 2-mercaptoethanol withvirtually identical results.

Wolf and Lam b Reactions: Equilibrium and KineticEffects in Multipolymer SystemsB. J. Cohen,* M. A. Kraus, and A. PatchornikContribution fro m the Department of Organic Ch emistry, Weizmann Institute of Science,Rehovot, Israel. Received April 8, 1981

Abstract: Two reagents reac ting avidly with each other in solution are rendered mutually inactive by attaching ea ch to a sep aratebatch of insoluble polymer. Two-stage reactions in which a soluble reagent reacts first with one polymeric reagent an d theproduct with the second polymeric reagent afford advantages over analogous reactions in solution. In acylation reactions ofcarbon acids, the simultaneous use of a polymeric strong base and a polymeric acylating reagent proved to be superior to theuse of soluble reagents, both for bringing about quantitative acylations and for coping with undesirable side reactions. Newpolymeric strong bases were prepared: polymeric trityllithium, para-substituted trityllithium polymers, and polymeric lithiumdiisopropylamide. Activ e esters of polymeric o-nitrophenol and N - 1-hydroxybenzotriazole were used a s acylation reagents.The scope and limitation of these reactions and their application to general multiphase systems are discussed.

Polym eric reagents-reactive, low mole cular weight moleculesbound to a polymeric ba ck bo ne ha ve been widely used in organicchemistry during the past two decades. Th e most significantadvantages of these reagents over their soluble counterparts ar ethe ease of their separation from the reaction mixture an d theirpossible recycling. Polymeric catalysts, polymers for specificseparations, ca rriers for sequential synthesis, polymeric blockinggroups, and polymeric transfer reagents in general organic syn-thesis, all have utilized these advantages. Use has also been mad eof the fa ct tha t chain fra gment s within a crosslinked polymericbackbone have restricted mobility. Active species attache d to thepolymer can thus be effectively isolated from eac h other a t rel-atively high concentrations, providing the advantages of highdilution and specificity, along with rapid kinetics. In other cases,properties of the bac kbone itself such as polarity, pore size, an dchirality were utilized to achieve unique reactions, the polymerproviding a specific microenvironment for the reaction. Theseaspe cts of polymeric rea gents have been extensively reviewed.-14

(1) Pittman, C. U. ; Evans, G. 0. CHEMTECH. 1973,560-566.(2) Leznoff, C. C. Chem. SOC. eo . 1974, 3 , 65-85.(3) Overberger, C. G.;annes, K. N. Angew. Chem., In t . Ed. Engl. 1974,13,99-104.

A t present, few examples exist in which more than one poly-meric reagent is used in a particular reaction. Pittm an andSmith,I5 for instance, described a reaction involving the simul-taneous use of t wo polymeric organometallic catalysts for con-secutive isomerization and hydrogenation of olefins. Us e of twoinsoluble polymers or a m ixture of a soluble an d an insolublepolymeric reagents for peptide synthesis has also been(4 ) Patchornik, A. ; Kraus, M. A. Pure Appl . Chem. 1975,13, 503-526.(5) Neckers, D. C. J . Chem. Educ. 1975, 52 , 695-702.(6) Weinshenker, N. M.; Crosby,G. A. Annu. Rep. Med. Chem. 1976,11,(7 ) Mathur, N. K.: Williams, R. E. J . Mucromol.Sci., Rev. Macromol.(8) Crowley, J. I. : Rapoprt, H. Acc. Chem. Res. 1976, 9, 135-144.(9 ) Heitz, W. Adv. Polym. Sci. 1977, 23 , 1-23.(10) Leznoff, C. C. Acc. Chem. Res. 1978, 11 , 327-333.(11) Kraus, M. A.; Patchornik, A. CHEMTECH. 1979, 118-128.(12) Kraus, M. A.; Patchornik, A. Macromol. Rev. 1980, I S , 55-106.(13) Mathur, N. K.; Narang, C. K.; Williams, R. E. Polymers as Aids(14) Hodge, P., Sherrington, D. C., Eds. Polymer Supported Reactions(15) Pittman, C. U.; Smith, L. R. J . A m . Chem.Soe. 1975,97,1749-1754.

261-270.Chem. 1976,CIS, 17-142.

in Organic Chemistry: Academic Press: New York, 1980.in Organic Synthesis; Wiley: New York, 1980.

0002-7863/8 1/ 1503-7620$01.25/0 0 198 1 American Chemical Society


2/10

Equilibrium in Multipolymer SystemsScheme I

A + B - roducts@ - A + Q -B - o reaction

J . Am. Chem. SOC.,Vol. 103, No. 25, 1981 7621Scheme 111

@ - A + S + @ - ES A -!-- S A B des ired product

Scheme I1S S A S A B

Howe ver, beyond the convenie nce of facile separation of products,no additional adva ntage s over the use of the corres ponding solublereagents were presented in those reactions.In a series of studies, mainly by R ebek and his co-workers18-z2and also by our it has been discovered that two insolublepolymeric reagents in the sam e reaction mixture c an provide someunique capabilities. Suc h mixtures, for example, have been usedin mechanistic studies to establish the existence of highly reactivefree species in solution. These species which were gene rate d byone polymeric reagent reac ted with t he second to yield a productat tached to the lat ter .We have studied the possibili ty of using a mixture of twoinsoluble polymeric reagents for synthetic purposes. We hopedthat such an undertakin g would eventually provide benefits beyondthose described above and tha t this approach could even be usedto modify the usual course of reactions. Indeed, our initial resultsz4showed that these assumptions were correct. The present workprovides the theoretical basis as well as additional experimentalevidence for the adv anta ges in using a multiplicity of polymericreagents in organic synthesis.Wolf and Lam b Reactions-Theoretical Considerations

Multipolymer systems of insoluble polymeric reagents areunique in tha t reactions can only take place within the matrix ofthe polymeric reagent. This spacial localization of various reactionsteps to their respective polymeric phases can lead to advantage snot found in conventional reactions in solution, where all com-p o n en t s can i n t e rac t s im u l t an e~u s ly .~~We investigated twe sta ge reactions in which a starting materialis modified successively by two polymeric transfer reagents. Theanalogous soluble reagents react with each other rapidly in so-lution, but they are rendered mutually inactive upon their at-tachm ent to the respective polymeric phases. These reactions wereaccordingly termed Wolf and Lamb reactions (after Isaiah11:6).A general scheme for these reactions is as shown in SchemeI. A and B react w ith each other in solution. @-A an d @-B a rethe corresponding insoluble polymeric reagents, which cannotinteract whenmixed. An indirect reaction between @-A and @-B is possible only through the mediation of a third, soluble

(16) Stern, M.; Kalir, R.; Patchornik, A. ; Warshawsky, A. ; Fridkin, M.(17) Heusel, G .; Bovermann, G.; Gohring, W.; Jung, G. Angew. Cbem.,(18) Rebek, J. ; Brown, D.; Zimmerman, S . J . Am . Chem.SOC. 975,97,(19) Rebek, J.; Gavina, F. J . Am. Cbem.SOC. 975, 97, 3453-3456.(20) Rebek, J.; Gavina, F. J . Am. Chem. SOC. 975, 97, 3221-3222.(21) Rebek, J.; Gavina, F.; Navarro, C. J . A m . Cbem. Soc. 1978, 100,(22) Wolf, S.; Foote, C. S . ; Rebek, J. J . Am. Cbem. SOC. 978, 100,(23) Warshawsky, A.; Kalir, R. ; Patchornik, A. J . A m . Cbem.SOC. 978,(24) Cohen, B. J. ; Kraus, M. A, ; Patchornik,A. J . Am . Cbem.SOC. 977,(25) Kraus, M. A . Ph.D. Thesis, Weizmann Institute of Science, Israel,

J . Solid-Phase Biochem. 1977, 2, 131-139.In t . Ed. Engl. 1977, 16 , 642-643.454-455.

8 113-8 117.7770-7771,100,4544-4550.99,4165-4167.1975.

S 2 A* SA B\ ;!de product;Scheme IV

C 6 H 5 C H z C N 4- C 6 H 5 ) 3 C C s H 5 C H C NC B H ~ C H C NC ~ H S C H C N

O C O C 6 H 5IC O C 6 H 5messenger reagent S: it will react first with @ -A leading toa soluble product SA , the latter moving to @-B and reacting withit to give the end p roduct SAB. It is impossible to carry out th ecorresponding reaction with soluble reactants because A and Bwill interact, a nd the overall reaction m ust be separa ted into twosteps: first, a reaction of S with A, followed by a reaction of SAwith B.We examined the advantages of the polymeric reagent approachin reactions, both at equilibrium conditions and under kineticcontrol. In either case, the two-polymer system proved to besuperior to a two-stage conventional reaction carried out withsoluble reagents.At equilibrium, the system in solution may be described asshown Scheme 11.The reaction of 1 equiv of A with 1 equiv of S will not leadquantitatively to SA. Equilibrium concentrations of the threecomponentsS,A, an d SA will depend on the equilibrium constant.The same is true for the reaction of SA with B. If an excess ofA is used to promote SA formation, this excess should be removedprior to the introduction of B, otherwise an ensuing reactionbetween A and B will lower the yields and complicate the sepa-ration process. With polymeric reagents, on the other hand,excesses of @-A and @-B may be used simultaneously, withoutpolymer interaction. S will react with the excess of @-A and willlead to quantitative form ation of SA , which will further react withthe excess of @-B, yielding the desired final product. Excess @-A and @ -B ar e easily separa ted out by filtratio n, eliminatinglaborious workup procedures.Und er kinetic control the following scheme may be presented(Scheme 111).The reaction of S an d A leads to an active species SA * whosereaction with B gives the desired product SAB. However, SA *may also react with unchanged starting material S or undergospontaneous decomposition, to give side products. In solution,a certain reaction time is needed to complete the transform ationof S to SA*. If B is introduced before quantita tive formation ofSA * has taken place, an undesired reaction will occur betweenB and unreacted A . During this time, however, side reactionsinvolving SA * molecules already formed may occur. The extentof these will depend on the various reaction ra te constants. Her eagain , utilization of a two-polym er system with excesses of @-Aand @- B may be helpful: any SA * molecule formed can im-mediately undergo a reaction with @-B, already available in thereaction mixture, and unlike the reaction with soluble reagents,there is no need to wait for a quantitative formation of SA*.In the following sections, these two sche mes are illustrated withacylations of carbanions, and th e advantages in the simultaneou suse of two polymeric reagents are confirmed. In hese examples,th e As are strong bases, Bs are acylation reagents of th eactive ester type, and Ss are ketones, nitriles, esters, etc.,possessing a-hydrogens abstractable by A.Results and Discussion

Polymeric Trityllithium la. When acylating carbon acids suchas nitriles, esters, or ketones, having two or three a-hydrogens,the monoacylated product has a-hydrogens more acidic than thosein the starting material. In he be nzoylation of phenylacetonitrile,


3/10

1622 J. Am. Chem. SOC.,Vol. 103, N o. 25, 1981Scheme V

Cohen, Kraus , an d Patchornik

Scheme VI-

I IaX = C1, OH; R = CH,, n-C,H,

for example, treating 1 equiv of it (S) with 1 equiv of a strongbase such as tr i tyl l i thium (A ) will produce 1 equiv of the anion(SA). The anion will be formed quanti tat ively due to the largedifference between t he pKa values of t r iphenylm ethane andphenylaceto nitrile. However, addition of 1 equiv of an acylationreagent such as an active ester of benzoic acid (B) will not leadto the quanti tat ive formation of the acylat ion product (SAB)because of proton exchange between the product and anionmolecules not yet acylated (Scheme IV).The result is an incomplete reaction and formation of mixtures.The concentrations of th e various species at equil ibrium dependon the pKa values of A, SA, and SAB. The increased acidi ty ofthe acylation product vs. that of the starting material m akes protoninterchange a com mon problem in acylations. It is impossible toadd an excess of t r i tyl l i thium to promote the formation of theproduc t since this excess will reect competitively with the a cylationreagent .Various ways to overcome this problem have been proposed:inverse addition of a dilute anio n solution to a con centrated solutionof the acylat ion reagent , use of certain base-acylat ing reagentcombinations such as ethoxide-ethyl ester, sodium hydride-esteror use of lithium diisopropylamide as the stron g base, at very lowtemperatures. The results are not always satisfactory. Accordingto our suggested two-polymer approach, a strong polymeric baseand a polymeric acylat ion reagent could be used to obtain thedesired product in high yield.No te tha t the acylat ion of carbanions does not exactly fol lowSche me 11, and a more acc urate description is according to SchemeV.Nevertheless, the use of polymeric reagents for this particularexamp le does not l imit th e validi ty of th e considerat ions givenabove for the use of polymeric reagents in reactions a t equilibrium.To accomplish the wolf and la mb acylat ions we developeda strong polymeric base, polymeric trityllithium Ia (Sc heme VI).Polymer I , ca rrying tr iph enylmeth ane moiet ies, was preparedby reacting polystyrene with benzhydrol or benzhydryl chloridein a Friedel-Crafts reactio n. Wi th benzhyd rol, lower loadings(0.7 mmol/g) of trityl groups than with benzhydryl chloride (2.3mmo l/g) were obtained. Th e latter reaction also required lesscatalyst . Therefore, polymer I obtained from t he chloride wasfurther used as our polymeric reagent .Th e blood-red polymeric anion Ia was produced by reactingpolymer I with excess methyll i thium or n-butylli thium in T H For in 1,2-dimethox yethane. Th e extent of ionization was deter-mined by washing excess reagents from the polymer, addingethanol, elut ing the result ing ethoxide, and t i t rat ing with acid.A typical loading vs. t ime grap h is shown in Figure 1 .In ordinary use, polymer la with a loading of ca. 1.5 mmol/gof trityllithium groups was prepared by applying a 100% excessof n-butyll i thium in T H F for 2 h at 0 OC.

0 2 4 6 8 IOTIME ( h r )

Figure 1. Loading of polymer I with anionic groups as a function of time.Table I

alkylation %starting material reagent produ ct yield(CH,),CHCOC,H,(CH,),CHCOC,H,(CH,),CHCNC,H,C=CH

CH,I (CH,),CCOC,H, 96C,H,CH,Br C,H,CH,C(CH,),COC,H, 98C,H,CH,Br C,H,CH,C(CH,),CN 97CHJ C,H,CsCH, 94

IC 2 C 0 2EI

I C,H,CH,Br I

96

98

ro 2 JNo2I1 IIa, R = C,H,b, R = OE t

The struct ure of polymer I will be discussed later when para-substi tuted tr i tylated polystyrenes are described.Polymer Ia is very sensitive to moisture, oxygen, and carbondioxide, as is its soluble analogue. There fore, all manipulationsinvolving this polymer were carried out under an inert atmosphere.Polymeric tr i tyl l i thium was shown to ab stract acidic protonsquantitatively from appropriate compounds. Excess of the polymerwas reacted with these compounds, followed by alkylation withan excess of soluble alkylat ing reagents in the same vessel.The alkylation products were obtained in excellent yields asTable I shows.Although the excess of polymer Ia is also alkylated under theseconditions, this does not adversely affect the isolation of the desiredalkylation product, which is in solution. Wo rkup of the reactionmixture was simple: the polymer was f il tered and washed andthe excess of alkylating reagent was reacted with pyridine. Th esalts and excess pyridine were washed with di lute acid, and th eorganic phase was evaporated to yield the pure product . Th eadvantage of polymeric vs. soluble trityllithium was clearly d em-onstrated by the fact that with the lat ter , chromatography wasrequired to separate tr iphenylmethane from the product .Using polymeric trityllithium, it was impossible to alkylate estershaving a-hydrogens. By the time proton abstraction was complete,side reactions, probably conden sations, had already o ccurred. Thiswill be discussed later.Wolf and Lamb React ions. A special apparatus, describedin the Expe rimenta l Section, was designed for these reactions. In


4/10

Equilibrium in Multipolymer SystemsScheme MI1

J . Am. Chem. SOC.,Vol. 103, No. 25, 1981 7623Scheme IX

I a

this apparatus , the polymeric base Ia was prepare d from polymerI by n-butyllithium in THF under inert atmosphere, while theacylatin g polymer-an active ester of polymeric o-nitrophenol,IIa or IIb26 (Scheme VI1)-was contained in a side arm.After polymer Ia was prepared, excess n-butyllithium waswashed out and the side arm turned so as to add the acylatingpolymer into the reaction cham ber. The wolf and lamb effectwas clearly dem onstra ted, as the blood-red polymer ic trityllithiumand the white polymeric active ester did not cha nge their colorsupon m ki ig , whereas their soluble analogues reacted immediatelyin solution. Color chan ges did occur, however, when moleculescarrying an acidic hydrogen were introduced. The red polymerIa w as converted to its white protonated from I, and the acylationof the carbanion produced the orange polymeric o-nitrophenol I1as i ts o-nitrophenolate salt (Scheme VIII).In the benzoylation of phenylacetonitrile previously described,the use of the two-polymer system offers a substantial adva ntageover the reaction in solution. Molecules of phenylacetonitrile,reformed a t the acylating polymer as a result of proton interchangebetween the product and anion molecules, can move back to thepolymeric base (present in excess), become deprotonated, acylated,etc., until product form ation is complete (Schem e VIII) . Thestoichiometry of the reaction shows that 2 equiv of base arerequired per 1 equiv of starting carbon acid (both polymer I1 andthe product will be ionized). W e used a 150% excess of polymerIa over tha t quantity a nd a 100% excess of the acylating polymer.Under these conditions, the reactions were complete within lessthan 15 min.The mechanism just described implies that molecules of carbonacid may move repeatedly between polymeric reagents. Motioninside a polymer bead is diffusion-controlled and does not dependon stirring. Theref ore, in order to shorten the residence time ofmolecules within the polymer beads, an d thereby s horten reactiontimes, the 20-50 mesh commercial polystyrene (XE-305, Ro hmand H aas ) used for preparing the polymeric reagents in this workwas ground to beads of 100-150 mesh size. Considerable im-provement in reaction rates a nd yields was observed. Thus , withthe groun d polymers, benzoylation of acetophenon e was complete din less than 15 minutes, whereas only 72% of the product wereobtained in that time with the polymers Ia and I Ia prepared fromnonground polystyrene beads. All the wolf and lamb reactionsdescribed in this work were therefore carried out with groundpolymers.Control reactions with analogous soluble reagents were per-formed in solution. An equivalent of trityllithium w as first reactedwith 1 equiv of the c arbon a cid, followed by addition of a n excessof the acylating reagent. With out exception, the resulting reactionmixtures had to be separated by chromatography. The work-upof the two-polymer reaction was much simpler: after the reactionmixture was neutralized, the polymers were filtered and washedseveral t imes. The inorganic salts were washed with water andthe organic phase dried an d evaporated to give the product in highyield.

(26) Kalir, R.; Fridkin, M.; Patchornik, A. Eur. J . Eiochem. 1974, 4 2,151-156.

H

Scheme X, - \

x111, X = CH,IV, X = OCH,

TH F n-B uLI1

IIIa, X = CH,IVa, X = OCH,Table I1 demonstra tes the results of the wolf and lamb re-actions vs. control reactions in solution.The advantage of using polymers instead of soluble reagentsis impressive. The rea ction solution has a light color and lacksthe highly colored side products typical of m any carbanion re-actions. The w orkup is very easy, mainly because of the insolubilityof the polymers. The high yields and purity of products are,though, unique to the two-polymer system. They stem from thefact that the two polymers constitute separate, noninteractingphases within the reaction system a nd th e ensuing possibility ofusing excesses thereof.The above described reactions result in an ionized product. Nosecond acylation was evident, although excesses of the basic a ndacylating polymers were still present. This was further checkedwith two of the products. They were reacted in a sepa rate ex-periment with 1 equiv of base and then with an excess of thesoluble acylation reagent. No further acylation was detected.Polymers Carrying para-Substituted Trityl Groups. Polymer

I was obtained with loadings of trityl groups as high as 2.3mmol/g. If the only modification occurring in its prepara tion isat th e phenyl rings of the polystyrene, this would mea n th at, onthe average, one out of every 5.9 phenyl rings is substituted.However, after examining molecular m odels, it seemed doubtfulthat such a high loading of the bulky trityl groups could beachieve d solely by modification of the phenyl rings of polystyrene.I t seems much m ore likely tha t benzhyd ryl substitution also occurson the trityl groups generated in the initial reaction, the ultim ateresult being grafting of the polystyrene (Scheme IX).This bushlike structure explains the low rate of hydrogen ab-straction (see Figure 1) even when using the relatively unhindered,highly reactive methyllithium reagen t. It is therefo re assumedthat under the reaction conditions used routinely which lead to


5/10

7624 J . Am. Chem. SOC..Vol. 103, No. 25 , 1981Table I1

Cohen, Kraus, and P atchornik

9 yield ofacylation % m c t i o n instarting material reagent prod uct yield solutionC,H,CH,CN C,H,CH(CN)COC,H, 94 4 5

C,H,CH,CN

CH,CNJNo2 91 37,H,CH(CN)CO,Et

C,H,COCH,CNd O C O C 6 H 5

90 27

C,H,COCH, C,H,COCH,COC,H, 96 4 8

C,H,COCH, (9-@-0c02E, C,H,COCH,CO,Et 9 2 4 0CH,CON(CH,), C,H,COCH,CON(CH,), 9 2 4 2

d O C O C 6 H 5

ca. 1.5 mmol/g of trityllithium groups, only the less stericallyhindered sites undergo ionization.Since grafting of this type is l ikely to occur only at the paraposition of the phenyl rings, blocking these positions should preventsubstitution. In order to test this point we prepared polymers I11and IV carrying para-substituted tri tyl groups (Scheme X).Loadings of only 1.1 mmol/g for polymer I11 and 1.Ommol/gfor polymer IV were obtained. These figures were determinedboth by increa se of weight, an d by conversion of the trityl polymersto their respective anions, IIIa and IVa, followed by titration, aspreviously described for polymer Ia.These loadings correspond to an average modification of oneout of every 7.5 phenyl rings of the pa rent polystyrene for polymerI11 and on e out of every 7. 4 rings for polymer IV. Wh ile thesevalues are almost identical, they differ from the value of 1/5.9which was calculated for polymer I, assuming no grafting. Thedifferences in molecular sizes of the various polymer-bound tritylgroups cannot alone account for these loading values so t h a tgraftin g probably occurs in the cas e of polymer I.The polymeric anions IIIa and IVa a re expected to be strongerbases than the polymer Ia.* N o nucleophilic characte r of th epara positions in these anions is to be expected, as observed fortrityllithium.28 Because of these features, polymers I11 an d IVmay find specific uses in the future.Reactions under Kinetic Control. Wolf and Lamb Reactionswith Esters. Ester enolates can undergo self-condensation, areaction which competes with the desired acylation. Whe n acy-lating a n ester in solution using a strong ba se (such as trityllithium)and an acylation reagent, quantitative enolate formation must beattain ed (as evidenced by the disappeara nce of the red color ofthe base) before the acylation reagent can be introduced, otherwiseit will react with th e remaining base. During this waiting period,some self-condensation ma y occur, depending on the reactivityof the enolate, i ts concentration, etc. For example, when y-bu-tyrolactone is benzoylated, two products, a-benzoyl-y-butyro-lactone and a- ( 1-benzoyloxy- -tetrahydrofury1)-y-butyrolactone,ar e produced, the latte r resulting from self-condensation (Scheme

When the sam e reaction is carried out as a wolf and lambreaction, using polymeric reagents, no waiting for complete enolateformation is necessary. The ac ylating polymer, present fro m the(27) Streitwieser,A. Prepr., Diu. Pet. Chem., Am . Chem.SOC. 967, 12 ,(28) Tomboulian,P.; Stehower, K. J . Org. Chem. 1968, 33 , 1509-1512.(29) Spencer, E. Y.;Wright, G. F. J . Am. Chem. SOC.1941, 6 3,

x1).29

C-21, C-22 (Chem. Absrr. 1967,67, 57703~) .1281-1285.

Scheme X Ido O l

Scheme XI1

start in excess, can immediately react with enolate moleculesformed a t the basic polymer, without reaction between the twopolymers. Thus, the concentration of enolate molecules can bekept low, and side products are not detected (Scheme XII).The results of several wolf a nd lam b acylations of esters an dyields of the control reactions in solution a re shown in Ta ble 111.Th e solution reactions produced various side products attribute dto self-condensation. These impurities were avoided with thepolymeric approach. In addition to the kinetic advantages, thewolf and lamb reactions also proved to be superior by providinghigh yields of acylation, a result of overcoming the proton -exchang ephenomenon in the particular esters being reacted, which had atleast two a-hydrogens.The r at e of enolization and th e lifetime of the enolate have a nimpo rtan t role in determining the outcome of these reactions. Itwas already mentioned tha t no satisfactory alkylations of esterswere possible by using polymeric trityllithium and soluble al-kylating reagents. The relatively long time (up to a few minutes)needed for complete enolization of the ester when using thepolymeric base enabled self-condensation to occur. Th e compa-rable reaction with a soluble base is much more rapid and yet m aygive rise to side products. In a wolf and lamb rea ction, on heother hand, the electrophilic polymer is present from the sta rt an d


6/10

Equilibrium in Multipolym er SystemsTable 111

J . Am. Chem. SOC., ol. 103, No. 25, 1981 1625~~% yield of% reaction inproduct yield solutionpolymericstarting material acylating agent

CH,CO,Et C,H,COCH,CO,Et 92 38

CH,CO,Et 88 35

C,H,CH,CO,Et C B H ~ C H C O Z E IICOCgHg 98 47C,H,CH,CO,Et C,H,CH(CO,Et), 92 35

C,H,CH,CH,CO,Et 92 40

b 0 95 31C 6 H 5 C 0 f 7Scheme XI11

J N 0 2

SchemeXIV

C6H5\ yCC6H5 COCeH,

/c=c

reacts with e nolate molecules as soon as they are formed.Wolf and Lamb Reactions with Short-Lived Species. Poly -meric Lithium Diisopropylamide. High - yield wolf and l amb acylations of esters can be carried out at room temperature,showing that ester enolates are relatively long-lived in the two-polymer system. Because their side reactions ar e basically sec-ond-order reactions, they can be controlled by reducing the ef-fective enolate concentration with an excess of electrophilicpolymer. However, anions can also undergo first-order intra-molecular changes, which are a function of temperature, but notof concentration. As described above, a two-polymer system maybe used for reactions under kinetic control, whereby a short-livedspecies generated at one polymer is reacted with a second polymerbefore its decomposition can occur. W e chose to study the acy-lation of l , -diphenyl-2-chloroethylene. he carbanion derivedfrom this olefin is stable in solution a t -70 OC, and m ay be acylatedat that temperature (Scheme XIII) . At h igher temperatures i tr ea rr an ge s t o d i p h e n y l a ~ e t y l e n e . ~ ~We attempted to acylate this anion at higher temperatures,using a two-polymer reaction. Trityllithiu m was found to beinsufficiently basic to abstract the ethylenic proton from 1,l-diphenyl-2-chloroethylene. herefore, a stronger polymeric base,polymeric lithium diisopropylamide Va, was prepared accordingto Scheme XIV.8-Phenyldiisopropylamine was prepared by modification of ageneral synthet ic route for secondary h indered amines.Chlorome thylation was carried out by using more than 1 equivof stannic chloride in order to complex the free amine a nd preventit from being alkylated by chloromethyl methyl ether. TheFriedel-Crafts attach ment to the polymer was accomplished byusing the hydrochloride of the chloromethyla tionproduct. PolymerV was obtaine d as white beads, with a loading (determ ined by

Va VIa

increase of weight) of 1.7 mmol of diisopropylamine groups/ g ofpolymer. It was ionized with n-butyllithium in TH F to give thepolymeric anion Va, a dar k brown polymer. Titration of theanionic polymer as described for polymeric trityllithium gave aloading value of 1.6mmollg .Tha t polymer Va is a stronger base than polymer Ia wa s visuallyverified by reacting the former with triphenylmethane. Th eblood-red color of trityllithium gradually developed, the reactionbeing slow because of the bulkiness of the two components.For the second component of the wolf and lamb experiment,the benzoyl derivative of polymeric N - 1-hydroxybenzotriazole(30) Kobrich, G. Angew. Chem., Int. Ed. Engl. 1972 , I I , 473-485.(31) Stowell, J. C.; Padegimas, S. . Synthesis 1974, 127-128.


7/10

1 6 2 6 J . Am. Chem. SOC.,Vol. 103, No. 25 , 1981 Cohen, Krau s, and PatchornikAIa32 was used. Thi s polymer is the fastest known polym ericacylatin g agent. An improved preparatio n of this polymer isdescribed in the Experimental Section.The wolf and lamb experiments in this system, however, werenot successful. No acylation product was found when the reactionswere carried out at temperatures down to -15 OC. Only therearranged product, diphenylacetylene, could be isolated (Sche meXV) .Apparen tly, the need to mig rate between the polymers l imitsthese reactions to species with relatively longer lifetimes. Alth ougha carban ion molecule has to travel less than 0.1 mm to m ake thistrip (when a thick slurry of finely ground polymers is stirred),the solid beads are surround ed by unstirred layers through whichthe species mus t diffuse. App aren tly, the diffusion is slow relativeto the lifetimes of the species we studied.While these experiments were being carried out , a paper byRebek and co-workers was published.22 In that work, singletoxygen was generated under various conditions at one polymericreagent an d trapped by an olefin supported on another polymerin the same system. Th e result of Rebeks experiments was tha tsinglet oxygen, with a half-lifetime of abo ut 0.1 s, has little chanceof being trapped at the second polymer before decomposit ion.Single oxygen is a small an d mobile molecule and Rebeks reactionwas carried out in the gaseous state. Changing to organic reactivespecies in organic solvents will m ake such reactions even less likely,as our work verif ies. Thu s, enolates were found to be suitablefor wolf a nd lam b experimen ts whereas very short-lived speciesare not .Wolf and Lamb Reactiom with SodiumHydride as the StrongBase. So far, only polymeric reag ents consisting of soluble reagentsattache d to insoluble polymers have been described. The generalconsiderations indicating that two noninteracting solid phases anda surrounding l iquid phase may bring about dist inct syntheticadvantag es in two-stage reactions a re not necessarily l imited topolymeric solids. Any insoluble, well-dispersed material ca pableof carrying out chemical t ransformations should be suitable.Polymers, however, are unique in that any soluble reagent , inprinciple, may be at tached to them, impart ing a general appli-cability. In addition, porous polymeric beads provide a largesurface area enabling easy access to solute components. Solidchemicals are usually reactive only at their surfaces and may beless efficient.W e substi tuted sodium hyd ride for polymeric tr i tyl li thium Iain two wolf an d lamb reactions. Otherw ise, these reactions wererun un der the sa me condit ions as the two-polymer reactions. Asan ex ample for an eq uilibrium-type reaction, the benzoylation ofacetoph enone was carried out with solid sodium hyd ride andpolymeric mit rop he nyl benzoate IIa. The reaction was completedwithin 1. 5 h d ue to th e slower rate of reaction of the hydride, ascompa red with the polymeric base Ia. Th e yield was however,quanti tat ive.Whe n trying to benzoylate y-butyrolactone with sodium hydrideand the polymeric ester IIa, a mixture of products resulted.Appare ntly, slow deprotonation with sodium hydride is incom-patible with the lifetimes of ester enolates, and side reactionsbecome more pronounced. This implies that the success of thewolf and lamb two-polymer reaction is also due to the rapidreaction between polymeric tr i tyll i thium and the soluble ester .

Recycling of the Polymeric Reagents. The recycling of poly-meric reagents is an important goal, as their preparation is usuallylonger and more cost ly than that of their soluble analogs. Theease of separat ion of the polymeric reagents simplif ies their re-cycling, provided that no i rreversible changes occur du ring theirreactions.The recycling of the polymer I, which was extensively usedthrough out this work, was studied by subject ing a sam ple of i tto th ree rep etitive ionizations and titrations, a s previously described.Loadings of 2.20, 2.15, and 2.25 mmol/g, respectively, were re-corded, proving that no changes occur in th e polymer with respect(32) Kalir, R.; Warshawsky, A.; Fridkin, M.; Patchornik, A. Eur. J .Biochem. 1975, 59 , 55-61.

Figure 2. Appara tus for wolf and lamb reactions: a, main reactionchamber; b, bath; c, sintered glass disk; d, side arm; e, solvent reservoir;f, injection inlet.to its basicity. Consequ ently, samples of this polymer wereroutinely reused for the wolf and lamb reactions. Othe rpolymeric bases described in this work which undergo ioniza-tion-protonation cycles are also assumed to be recyclable, althoughthis has not been checked.At the end of a wolf and lamb reaction, the mixture ofpolymeric reagents may be separated by selective flotation, utilizingdifferences in their densities. Benzene ( d = 0.88) and chloroform( d = 1.5) can be mixed in such proport ions as to cause f lotationof one polymeric reagent and precipi tation of the other. Eachmay then be collected a nd purif ied.Experimental Section

General. The polystyrene used throughou t this work was XE-305(Rohm & Haas ). The polymer beads were ground for use in wolf andlamb reactions as follows: the beads were suspended in the minimumamount of chloroform necessary for complete swelling. The slurry wastransferred to a blender operated at maximum speed for a few minutes(the blender blades must be very sharp to prevent the formation of finedust). The polymer was dried and the 100-150 mesh fraction was col-lected.App aratu s for wolf and lamb reactions: Figure 2 describes thesystem used. It enables the preparation and washing of the polymericbase under inert atmosphere in the main reaction chamber a, while thesecond polymer is stored in the side arm d. By turning it 180 the secondpolymer is introduced into chamber a. Thus, all operations can be con-ducted conveniently in the same apparatus.Tetrahydrofu ran and 1,2-dimethoxyethane were directly distilled intothe solvent reservoir e from trityllithium as follows: after a preliminarydrying and peroxide removal, half a liter of the solvent was treated withLiAIH4 until no hydrogen evolution was visible. Then 2-4 g of benzo-pinacolone was added. The solvent was distilled from the deep redsolution under an inert atmosphere. It should be noted that the polymericstrong bases described in this work are washed several times with thesolvent prior to reaction. Even the smallest amo unt of water or oxygenwill produce a color change on the surface of the polymer beads. Sucha sensitivity to solvent impurities is impossible with the soluble bases.NMR spectra were recorded on a Bruker 90-MHz NM R spectrom-eter. Melting points are uncorrected.Loadings of the polymeric reagents were measured either by increaseof weight or by titration of the relevant functional group.Polymer I Carrying Triphenylmethane Groups. a. Reaction of Poly-styrene with Benzhydrol. Benz hydr ol(l8.4 g) was dissolved in 100 mLof chloroform and impregnated into 10 .4 g of polystyrene by evaporation.A solution of 16 g of AlC13 in 50 mL of dry nitrobenzene was added. The


8/10

Equilibrium in Multipolymer Systemsresulting dark mixture was heated at 60 C for 24 h. Workup was bya mixture of 50 mL of concentrated HC I, 50 mL of DM F, and 100 g ofice. Afte r the beads became light-colored they were washed with warm(60 OC) DM F until the washings were colorless and then with six portionsof methylene chloride-methanol (2: 1). After being dried, the whitepolymer weighed 11.8 g (loading 0 .7 mmol/g).b. Reac tion of Polystyrene with Benzhydryl Chloride. Distilled ben-zhydryl chloride (20.2 g) was absorbed into 10.4 g of polystyrene. Asolution of 2 g of AICIJ in 50 mL of dry nitrobenzene was added. HC1immediately evolved. After being stirred for 24 h at room temperature,the reaction mixture was decomposed as in a. Washing and drying ofthe polymer afforded 16.9 g of white beads (loading 2.3 mmol/g).

Polymeric Trityllithium Ia. Although polymer I is indefinitely stable,the polymeric anion Ia is very sensitive to moisture, oxygen, and COZ.It was therefore prepared individually for each experiment. Polymer I(2 g) was introduced into the reaction chamber a (Figure 2). Dry TH Fwas directly distilled into the solvent reservoir e. The system was closed,the walls were heated, and the apparatus was flushed with purified ni-trogen several times. Solvent (20 mL) was introduced and an ice-watermixture was placed in bath b. An 8-mmol sample of a solution of n-butyllithium in hexane or methyllithium in ether was injected slowly.The polymer turned red, the color deepening with time. After the desiredperiod, the polymer was washed with 20 portions of solvent until thewashings were free of base. The appropriate reaction (titration, alkyla-tion, wolf and lamb reaction, etc.) was then performed.Titration of Polymer Ia. After prepa ration, polymer Ia was treatedwith excess ethanol. The resulting ethoxide was washed into water withTHF-e thanol (1 :l) solution and then titrat ed with dilute hydrochloricacid. Th e dependence of loadings on time is shown in Figure 1, wheremethyllithium in 1,2-dimethoxyethane was used for ionization of I. Mostreactions, however, were done with polymer Ia prepared with a 100%excess of n-butyllithium in T H F for 2 h a t 0 O C . Titration of the re-sulting polymer showed ca. 1.5 mmol/g of trityllithium groups.Recycling of Polymer I. Samples of polymer Ia that participated inproton abstrac tion reactions (not in alkylations) were recycled as follows:the polymer was washed with several portions of warm (60 C ) DMF,then with methylene chloride-methanol (2: ) , and dried thoroughly. Asample of the polymer was subjected to three cycles of ionization witha 100% excess of methyllithium in 1,2-dimethoxyethane for 9 h a t 0 O C ,then titr ated , washed, and recycled. The loadings of trityllithium groupswere 2.20, 2.15, and 2.25, respectively. Thus, it was demonstrated tha tpolymer I is recyclable. Sample s of the polymer were therefore usedseveral times in the various experiments described here.Alkylation Reactions Using Polymer Ia. Th e reactions were carr iedout in the apparatus shown in Figure 2. Polymer I (2 g) was ionized byn-butyllithium in TH F as previously described. To the resulting polymerIa (about 3 mmol) was injected 150 mg (1 mmol) of isobutyrophenone.Afte r stirring for a few minutes, 500 mg of benzyl bromide was injected.The reaction mixture was neutralized with hydrochloric acid, and thepolymer was washed with 8 portions of 20 mL of methylene chloride.Pyridine was added to the combined washings to react with the excessof benzyl bromide. Th e salts and excess pyridine were washed with diluteHC1. Th e organic phase was dried and evaporated to yield 235 mg (98%)of a,&-dimethyl-P-phenylpropiophenone. The results of other similarlyperformed alkylations ar e summarized in Table I.Attempts to Alkylate an Ester with Polymer Ia. A ca. 3-mmol sampleof polymer Ia was prepared, a s described above. Ethyl isobutyrate (1 17mg) was injected, and the mixture was stirred for a few minutes. Benzylbromide (500 mg) was injected and the workup accomplished as de-scribed for the other alkylation reactions. TL C analysis of the reactionmixture revealed several spots, some of which became colored whensprayed with 2,4-dinitrophenylhydrazine olution. These spots, therefore,might be a ttributed to self-condensation products of the starting ester.Prep arat ion of the Polymeric o-Nitrophenyl Active Esters IIa and IIb.Polymeric o-nitrophenol I1 was prepared from polystyrene and 4-hydroxy-3-nitrobenzyl chloride in a Fr ied elC raf ts reaction.26 A polymerloaded with 3.5 mmo l/g of o-nitrophenyl groups was obtained. Thepolymeric active esters were prepared by reacting the polymer with atwofold excess of either benzoyl chloride and pyridine (to obtain IIa) orethyl chloroformate and pyridine (to obtain IIb) in dry chloroform at 0OC for a few hours. The polymeric active ester was washed with drychloroform until no nonvolatile compounds were detected in the washings(abo ut 6-8 washings) and then dried. The polymers IIa and IIb wereobtained in loadings of 2.5 and 2.8 mmol/g, respectively, as evidencedby increase of weight or by titration with benzylamine.For control reactions in solution, o-nitrophenyl benzoate and o-nitro-phenyl ethyl carbonate were prepared according to the l i t e r a t ~ r e . ~ ~ ~ ~ ~

J . Am. Chem. SOC.,Vol. 103, N o. 25 , 1981 1621Wolf and L a m b Reactions. These reactions were carried out in theappara tus described in Figure 2. Dry TH F was distilled directly into thesolvent reservoir before each reaction. Th e acylating polymer was in-troduced into the side arm . The system was sealed and flushed severaltimes with pure nitrogen while the walls were heated to remove tracesof moisture. The bath was filled with an ice-water mixture, and thepolymeric base Ia was prepared and washed as described above. Th ebath was then filled with water at room temperature, and the acylatingpolymer was introduced by rotating the side arm. No reaction wasapparent between the two polymers. The solvent level was adjuste d soas to just cover the polymer mixture. The stirrer was rotated a t moderateto high speed, and the subs trate to be acylated was injected afte r being

dissolved in a small amount of dry TH F. Th e colors of the polymerschanged immediately. Th e reaction mixture was neutralized with hy-drochloric acid, and the polymers were washed 6-8 times with methylenechloride-methanol (2: 1) solution. The solvents were evaporated, and theresidue was taken up in ether, which after drying and evaporation af-forded the desired product. Sometimes, small polymer particles, eitherformed by the abrasive ac tion of the stirring or present due to imperfectsieving during the grinding of the polystyrene, were present. These fineparticles could not be removed by using filter paper, and a short chro-matography column was required.Thus, 120 mg of acetophenone were benzoylated by a mixture of ca.5 mmol of polymer Ia and 2 mmol of polymer IIa, to yield dibenzoyl-methane in 94% yield. Less than 15 min were required for completionof the reaction. Th e results of other acylations are shown in Table 11,along with the results of control reactions in solution.Control Reactions in Solution. Th e reactions were carried out underinert atmosphere. A 2-mmol sample of trityllithium in 30 mL of TH Fwas prepared from triphenylmethane and n-butyllithium. To the deep-red solution was injected 2 mmol of the same carbon acid used for thecorresponding wolf and lamb experiment. Th e red color disappearedwithin a few seconds, and the soluble acylation reagent was then intro-duced in 50 % excess. The reaction mixture was neutralized and extractedwith ether. The organic phase contained triphenylmethane, excess acy-lation reagent, o-nitrophenol, the starting carbon acid, and the product,which was separated by column chromatography.A Wolf and Lamb Reaction with Non gound Polymers. PolymersIa and IIa were prepared from polystyrene beads of the original, com-mercial size (20-50 m esh). A wolf and lamb benzoylation of aceto-phenone was carried out in exactly the sa me way as described for theground polymers. After 15 min, the reaction mixture still contained someacetophenone. After workup and chromatographic separation, di-benzoylmethane was isolated in 72% yield.Examination for Possible Second Acylation. Th e acylation productsare obtained at the end of the wolf and lamb reactions as their anions.No furthe r acylation was detected. This was sepa rately confirmed insolution. The benzoylation produc t of acetophenone, dibenzoylmethane,having two hydrogens, and the benzoylation product of phenylacetonitrile,phenylbenzoylacetonitrile, aving one a-hydrogen, were both reacted with1 equivalent of trityllithium, followed by 1.5 equiv of o-nitrophenylbenzoate a t room temperature. After 15 min, the reaction mixtures wereneutralized and worked up, showing no further acylation products.Separation of the Polymeric Reagents by Flotation. At th e end of thewolf and lamb reactions, polymeric triphenylmethane I and the poly-meric nitrophenol I1 (partially as its ester IIa or IIb) were mixed to-gethe r. Their efficient separa tion for recycling purposes was accom-plished by selective flotation in a mixture of 55% benzene and 45%chloroform (v:v). Polymer I floated, the other precipitated.Preparation of Polymer 11, Carrying Triphenylmethane Groups withMethyl Substituents at Par a Position. 4,4-Dimethylbenzhydryl chlo-rideg5 34.5 g) was absorbed into 15.6 g of polystyrene beads. A solutionof 3 g of A1C13 in 80 mL of dry nitrobenzene was adde d. Th e mixtureturned black and HCI evolved. After being stirred overnight at 60 C,the reaction was decomposed with a m ixture of 100 mL of DMF, 50 mLof concentrated HCI, and 100 g of ice. The polymer beads becamelight-colored and were then washed with warm (60 C ) DM F until thewashings were colorless. The white polymer was further washed with 6-8portions of methylene chloride-methanol (2 :l ) and dried. Th e yield was19.8 g (loading of 1.1 mmol/g). A sample of the polymer was convertedto the anion IIIa by means of 100%excess of n-butyllithium in T H F at0 O C for 2 h. Adding ethanol and titrating the ethoxide formed asdescribed for polymer Ia showed a loading of 1.1 mmol/g .Preparation of Polymer IV Carrying Triphenylmethane Groups withMethoxy Substituents at Pa ra Positions. Polystyrene (10.4 g) was im-pregnated with 26.2 g of 4,4-dimethoxybenzhydryl chlorideJ6 by dis-solving the chloride in chloroform, adding to the polymer, and evapo-

(33) Hubner, H. Justus Liebigs Ann. Chem. 1882, 210, 328-396.(34) Bender, G. Ber. Deutsch. Chem. Ges. 1886, 19, 226552271, (35) Norris, J. F. ; Banta, C. J . A m . Chem. SOC. 928, 50 , 1804-1808.(36) Straus, F. ; Dutzmann, A. J . Prakt. Chem. 1921, 103, 1-68.


9/10

7628 J . Am. Chem. SOC.,Vol. 103, No. 25, 1981rating to dryness. A solution of 2 g of AlCl, in 50 mL of dry nitro-benzene was added, and the reaction was carried out and worked up asdescribed for polymer 111. The polymer obtained weighed 13.5 g (loadingof 1.0 mmol/g). Th e loading obtained by titration of the anion IVa was1.Ommol/g.Benzoylation of y-Butyrolactone in Solution. y-Butyrolactone (0.17g) was treated with 2 mmol of trityllithium and then with o-nitrophenylbenzoate. Two benzoylation products were obtained and isolated bycolumn chromatography. One (1 18 mg, 31%) was a-benzoyl-y-butyro-l a~ to n e . ~The other (90 mg), according to the ratio of aromatic toaliphatic protons in its N M R spectrum, was suspected to be a productof self-condensation, followed by benzoylation, namely, a-( -benzoyl-oxy- 1-tetrahydrofury1)-y-butyrolactone. Further evidence for thisstructure was its hydrolysis and water elimination to give dibutyrolactone:200 mg of the substance was hydrolyzed overnight at room temperatureby a mixture of 1 mL of a solution of 40% benzyltrimethylammoniumhydroxide (triton B) in water, 1 mL of water, and 2 mL of dioxane. Thebasic solution was extracted with ethyl acetate to remove unreactedstart ing material. The aqueous solution was acidified and extracted withethyl ace tate, in which the benzoic acid formed was also taken up. Thiswas eliminated by aqueous sodium bicarbonate. Th e organic phase wasevaporated to yield 105 mg of dibutyrolactone.28Anal. Calcd for a- ( -benzoyloxy-1-tetrahydrofury1)-y-butyrolactonespectrum (CDC13): 6 1.96-3.40 (m, H) , 3.61-3.79 (m, H), 4.27-4.41(m, 4 H), .3C-7.55 (m, H), 7.95-8.06 (m, H) .Wolf and L a m b Reactions with Esters. These reactions were carriedout under conditions identical with those described above for the reactionsa t equilibrium. Workup procedures were the same as well. The reactionswere carried out at room tempera ture with fast stirring. Control reac-tions in solution were carried out as described for the reactions inequilibrium. Th e reaction mixtures of the control reactions were sepa-rated on TLC plates and showed triphenylmethane, o-nitrophenol, un-reacted nitrophenyl ester, starting material, and product, as well asvariable amounts of othe r products. Some of these gave colored spotswhen sprayed with 2,4-dinitrophenylhydrazine olution, revealing theexistence of carbonyl groups. Acylation products were separat ed bycolumn chromatograp hy. The ir yields ar e recorded in Table 111.Preparationof the Polymer V Carrying Diisopropylammonium Groups.a. B-Phenyldiisopropylamine. To a solution of 6.7 g of benzyl methylketone, 12.8 mL of isopropylamine, and 100 mL of dry benzene a t 0 Cwas added dropwise a solution of 3.5 mL T IC 4 n 15 mL of dry benzene.The mixture was stirred for 15 additional min and then hydrogenated at60 psi with 25 mg of platinum oxide (Edman catalyst). After hydro-genation was completed (ca. 0 min), 20 mL of 10% aqueous N aO H wasadded. Two phases formed, with solid titanium dioxide precipitating out.Th e organic phase was separated, dried, and evaporated. The amine wasdistilled at 74 OC (ImmHg). Yield: 8.1 g (91%).Anal. Calcd for CI 2H L9 N: , 81.30; H, 10.80; N , 7.90. Found: C,81.10, H, 10.88, N , 7.95. N M R spectrum (CDCl): 6 0.93-1.08 (m,H), 1.32 (s, 1 H) , 2.4-3.4 (m, 4 H), 7.05-7.52 (m, H).b. B- (Chloromethy1phenyl)diisopropylammonium Hydrochloride. 0-Phenyldiisopropylamine (10 g) was dissolved in 50 mL of dry methylenechloride and cooled by an icesalt bath to 0 OC. SnCI, (10 mL) wasadded dropwise. Chloromethyl methyl ether (2 0 mL) was added drop-wise, maintaining the temperature at 0 C or below. The solution waskept overnight at 5 C and then poured over ice. Aqueous Na OH (20%)was added dropwise until the reaction mixture was basic, and the re-sulting oil was taken up with ether. The organic phase was separated anddried, and dry HCI was bubbled through it. The ether was decanted fromthe precipitating oil, which solidified upon scratching the vessel walls.The product was triturated with acetone to give a white powder (mp

NM R spectrum (Me2SO-d6): 6 1.10-1.38 (m, 9 H), 2.53-3.60 (m,c. Binding to Polystyrene. Polystyrene (10.4 g) was mixed with 13.0g of 0-(chloromethy1phenyl)diisopropylammonium hydrochloride. Asolution of 2 g of AIClp in 50 mL of dry nitrobenzene was added, andthe mixture was stirred at 60 C for 24 h. The decomposition of themixture and washing of the polymer were carried out a s for polymer I.Polymer V was obtained a s white beads and weighed 17.3 g (loading of1.7 mm /g): nitrogen analysis, 2.40% (loading of 1.7 mmol/g) ; chlorineanalysis, 5.65% (loading of 1.6 mmol/g).Preparationof Polymeric Lithium Diisopropylamide Va. This polymerwas prepared according to th e procedure adopted for the other polymericbases described in this work, using a twofold excess of n-butyllithium inTH F. Th e loading of the resulting dark brown anion Va was determined(37) Ebel, F.; Weissbarth, 0. German Patent 801 276, Dec 28, 1950

(C16H1605): C, 65.21; H, 5.84. Found: C, 65.11; H, 6.01. NMR

158-160 C).4 H), 4.75 (s, 2 H) , 7.22-7.46 (g, 4 H) , 9.24 (s, 2 H).

(Chem. Abstr. 1950, 45 , 2972).

Cohen, Kraus, and Patchornikby titration and found to be 1.6 mmol/g.Improved Synthesis of Polymer VI Carrying N-l-Hydroxybenzo-triazole Groups. a. 4-Chloro-3-nitrobenzyl Alcohol. This substance wasprepared by in situ diborane reduction of 4-chloro-3-nitrobenzoic acid.NaB H, (8.5 g) was dissolved in 250 mL of dry diglyme and cooled inan ice bath until the solution became very viscous. 4-Chloro-3-nitro-benzoic acid (50 g) was then added in small portions. Cooling wascontinued, and 40 mL of distilled boron trifluoride ethera te was addeddropwise. Stirri ng was continued for 30 additional minutes. The mixturewas poured over 500 g of ice. Th e solid was filtered and dissolved in etherand the unreacted acid washed with 10% aqueous Na OH . The etherphase was dried and evaporated to give 40 g (yield 85%) of the alcohol.32

b. Binding to Polystyrene. The original preparation of polymer VIcalls for the binding of the alcohol to polystyrene.32 Here, however, the4-chloro-3-nitrobenzyl chloride, obtained from the alcohol by reactionwith thionyl chloride,j8 was used, the chloride affording a higher reactionra te and requiring only cata lytic amounts of aluminum chloride. Poly-styrene (20.8 mg) was impregnated with 17.3 g of the chloride, and asolution of 4 g of A1C13 in 100 mL of dry nitrobenzene was add ed. Th emixture was stirred a t 60 C for 24 h. It was hydrolyzed with a mixtureof 100 mL of DM F, 100 mL of concentrated HC1, and 200 g of ice. Th epolymer was washed with warm (60 C )DM F until the washings w erecolorless and then with 6-8 portions of methylene chloride-methanol(2 :l) . Th e dried polymer weighed 32.1 g (loading of 2.1 mmol/g):nitrogen analysis, 2.97% (2.1 mmol/g); chlorine analysis, 7.46% (2.1mmol/g).c. Preparation of Polymer V. Th e polymer obtained above was re-acted with hydrazine and then with hydrochloric acid according to theorig inal prep ar at i~ n . ~~he resulting polymer contained 1.7 mmol/g ofO H groups.Preparationof thePolymeric Active Ester VIa. Polymer VI (10 g) wassuspended in 60 mL of dry chloroform. Benzoyl chloride (5.8 mL) an d4.0 mL of pyridine were added at 0 C. After the solution was stirredfor 30 min, the polymer was washed with dry chloroform under a ni-trogen atmosphere. Drying at high vacuum afforded 11.7 g of polymerVIa (loading of 1.4 mmol/g). Titra tion with benzylamine gave 1.3mmol/g. Th e polymer is sensitive to moisture and was kept in a dessi-cator.Benzoylation of l,l-Diphenyl-2-chloroethylenen Solution. 1,l Di-phenyl-2-chl0roethylene~~430 mg) was dissolved in 20 mL of dry T HF .Th e mixture was cooled to -70 C , and 2 mmol of n-butyllithium wasadded dropwise, under a purified nitrogen atmosphere. After th e solutionwas stirred for 1 h at th at temperature, 2 mmol of o-nitrophenyl benzoatewas added. Stirri ng and cooling were continued for 1 5 addit ional min-utes, and the reaction was then neutralized with hydrochloric acid. Themixture was extracted with ether, dried, and evaporated. Th e residuewas chromatographed to give 402 mg (63% yield) of l,l-diphenyl-2-

chloro-2-benzoylethylene. t was identical with the target compoundmade according to the literature by a similar route.40Wolf and Lamb eaction with l,l-Diphenyl-2-chloroethylene nd thePolymers Va and VIa. The reaction was carried out in the apparatusdescribed in Figure 2. Dry TH F was distilled into the solvent reservoirbefore the reaction. Polymer VIa (1.5 g, 2.1 mmol)was introduced intothe side arm. Polymer V (2 g, 2.6 mmol) was introduced into the mainchamber and ionized as described above by a twofold excess of n-bu-tyllithium. After excess n-butyllithium was washed from the polymer Va,the bath was filled with water or a cooling mixture according to thedesired reaction conditions. Polymer VIa was then introduced. Noreaction was apparent between the two polymers. The amount of solventwas adjusted to just cover the polymers. With fa st stirring 215 mg (1mmol) of l,l-diphenyl-2-chloroethyleneissolved in a small amount ofdry TH F was injected. Polymer Va was immediately bleached. Aftera few minutes, the reaction mixture was neutralized, filtered, and workedup. Only the rearranged product, diphenylacetylene, was found. Iden-tical results were obtained in reactions carried out at room temperature,0 C, and -15 C.Wolf and LambReactions with Sodium Hydride as he StrongBase.These reactions were carried out exactly as were the wolf and lambexperiments with polymeric trityllithium Ia, except that ca. 250 mg ofNaH dispersed (5040%) in oil was substituted for the polymer Ia in themain reaction chamber. The oil was washed off the hydride with solvent,prior to reaction. No reaction between the sodium hydride and polymer

(38) Mckay, A. F.;Garmaise, D. L.; Gaudry, R.; Baker, H. A.; Paris, G.Y.; ay, R. W. ; Just, G. E.; Schwartz, R. J . Am. Chem. SOC.1959, 81,4328-4335.(39) Korshak, V. V.; Samplavskaya, K. K.; Andreeva, M. A. Z h . Obshch.Khim. 1949, 19 , 690-695.(40) Boehme, H.; Stammberger, W. Arch. Pharm. Ber. Deutsch. Pharm.Ges. 1912, 305, 392-395.


10/10

J. Am . Chem. SOC.1981,103, 7629-1635 7629When y-butyrolactone was introduced into the mixture of the hydrideand polymer IIa, TLC analysis of a sample of the reaction mixturerevealed several spots. No attempt was made to isolate them.Acknowledgment. W e wish to thank the Camil le and HenryDreyfuss Foundation for partial support for this research.

IIa used in these reactions was evident.When 1 mmol of acetophenone was introduced into the mixture ofsodium hydride and polymer IIa, the reaction was slower than withpolymeric trityllithium Ia. Afte r 1.5 h, excess hydride was carefullydestroyed with water, and after workup 212 mg (yield 95%) of di-benzoylmethane was obtained.

The Structure of CC-1065, a Potent Antitumor Agent, andIts Binding to D N AC. C. Chidester,* W. C. Krueger, S. A. Mizsak, D. J. Duchamp, and D. C. MartinContribution fro m The Upjohn Company, Kalamazoo, Michigan 49001, Received Ma y I I , 1981.Revised Manuscript Received July 20, 1981

Abstract: The crystal and molecular s tructure of a potent antitumor agent, CC - 1065 (C37H 33N70 8),as been determined,and circular dichroism studies have demonstrated a strong interaction between CC-1065 and DNA. The triclinic unit cell,space gro up P1, of dimensions a = 11.063 (3) A , b = 13.311 (2) A, c = 13.405 (2) A, a = 85.18 (1)O, @ = 99.52 (l)', andy = 103.36 (2)O, contains two CC-1065 molecules and disordered solvent molecules (two methanols and t hree w aters). Intensityda ta for 628 4 reflections were collected at -155 (2) OC; the final R value was 0.08 for 241 1 significant high-angle reflec tions.N M R spectroscopy was very impor tant in assigning kinds of atoms. The CC -1065 molecule consists of three substitutedbenzodipyrrole moieties linked by ami de bonds. On e terminal moiety has a cyclopropyl ring; the only asymmetric carbonsin the molecule are a t the cyclopropyl ring junctures. The two symmetry-independent molecules have the same configuration,but th e absolute configuration is not known. The CC -1065 molecules ar e curved and somewhat twisted with the outer peripheryhydrophilic and th e inner surface lipophilic. The interaction between C C-1065 and a number of D N A polymers was studiedby circular dichroism. The C D effect requires double-stranded D N A and is strongest with poly(dA-dT):poly(dA-dT).

A number of a n t i tumor agents are known to b ind to deoxy-ribonucleic acid by nonintercalative means.'-s Netro psin is anoligopeptide that binds preferentially to poly(dA):poly(dT) andis thought to intera ct through hydrogen bonds in the minor grooveof DNA.'s2 Anthr amyc in is a pyrrolo[ 1,4] benzodiazepine tha tbinds covalently to g ua ni ne in t he m in or g r o ~ v e . ~ , ~raithwaitean d Baguley have investigated netropsin, distamyc in, and severalbisamidine an d bisquaternary amm onium heterocyclic drugs byviscometric, spectrophotometric, and fluorometric methods, andpropose tha t they all bind in the minor groove of the D N A doublehelix.5A new antitumor agent, CC-1065, also binds to D N A withoutintercalation.6 CC- 1065 is produced by a soil culture, strepto-m yc es ~ e l e n s i s , ~nd w as isolated in pure form8 and found to besignificantly more cytotoxic in vitro tha n actinomycin, vinblastine,or maytansine? In vivo, CC- 1065 is highly active against a varietyof mouse tumors,* and, as a result , has been selected by the

(1) (1) Berman, H. M.; Neidle, S.; Zimmer, C.; Thrum, H. Biochim.(2) Wartell, R. M.; Larson, J. E.; an d Wells, R. D. J . Bio!. Chem. 1974,Biophys. Acta 1979, 561, 124-131.249, 6719-6731.1977. 475, 521-535.(3 ) Hurley, L. H.; Gairola, C.; Zmijewski, M. Biochim. Biophys. Acta(4 ) Kohn, K. W.; Glaubiger, D.; Spears, C. L. Biochim. Biophys. Acta

( 5 ) Braithwaite, A. W.; Baguley, B. C.Biochemistry 1980,19, 1101-1 106.(6) Swenson, D. H.; Krueger, W. C.; Lin, A . H.; Schpok, S. L.; Li, L. H.Proc. Am. Assoc. Cancer Res. 1981, 22 , 857.(7 ) Hanka, L. J.;Dietz, A ,; Gerpheide,S. A, ; Kuentzel, S. L. ; Martin, D.G. J . Anribiot. 1978, 31 , 1211-1217.(8 ) Martin, D. G.; iles, C.; Gerpheide,S.A.; Hanka, L. J.; Krueger, W.C.; McGovern, J. P.; Mizsak, S . A ,;Neil, G. L.; Stewart , J. C.; Visser, J . J .Antibiot. 1981, 34 , 1119-1125.(9) Martin, D. G . ;Hanka, L. J.;Neil, G. L. Proc. Am. Assoc. Cancer Res.1978, 19 , 99.

1974, 361, 288-302.

0002-7863 /8 1/ 1503-7629SO1.25 /O

Table I. "C NMR Shi ftschemical multi- tentativ e chemical multi- tentativeshiftu plicityb assignmentlS shift plicity assignment

9.5 9 C3M 121.3 s C2320.9 d c11 123.5 d c 221.6 t c1 0 127.2 s C1626.6 t C24 127.5 s C 2921.6 t c 37 128.9 s c 2 231.5 S c 9 129.1 s c 3 549.4 t c 1 2 129.5 s C453.3 t C25 130.4 s C1 954.8 t C38 130.7 s C3 260.0 q C20M 132.4 s c 2 160.3 q C33M 133.1 s c 3 4105.9 d C17 138.0 s c 2 0106.3 d C30 138.4 s c 3 3110.6 d c 7 1 57 .5 s C4 0113.0 S c 3 157.5 s cs117.3 S C18 160.2 s C27117.7 S C3 1 160.7 s C8118.2 S C36 161.2 s C14176.4 s C6(I Chemical shifts in ppm relative to internal Me,Si. s =singlet, d = doublet, t = triplet, q = quartet.

Nati onal Canc er Insti tute (as NSC2 98223) for preclinical tox-icology studie s.Th e structure of a fragment of CC-1065, the base degradationproduct, was reported previously.'0 A preliminary report of th estructur e of CC-1 065 has also been published." W e report here(10) Chidester,C. G.; Duchamp, D. J .; Martin, D. G. bslr. Am. Cryst.(1 l! Martin, D. G.; hidester, C. G.; uchamp, D. J .; an d Mizsak, S.A .Assoc., Winter Meeting 1979, 6 (Series 2, No. 2) , 67 .J . Antt610f . 980, 33 , 902-903.

0 1981 American Chemical Societv, I -

wolf and lamb

Documents