THE JOURNAL OF BIOLOGICAL CHEMISTRY 8 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Val. 268, No 2, Issue of Janunaq 15, pp. 823-831,1993 Printed in U.S.A.

The Purification and Characterization of a Unique Cytochrome P-450 Enzyme from Berberis stolonifera Plant Cell Cultures*

(Received for publication, July 14, 1992)

Richard StadlerS and Meinhart H. ZenkQ From the Lehrstuhl jiir Pharmazeutische Biologie, Uniuersitht Miinchen, Karlstrasse 29, 0-8000 Miinchen 2, Germany

A new cytochrome P-450 enzyme, isolated from Ber- beris stolonifera plant cell suspension cultures, has been purified to electrophoretic homogeneity. The pu- rified hemoprotein migrated as a single band in sodium dodecyl sulfate polyacrylamide gel electrophoresis with a minimal M, = 46,000. The enzyme could be purified to a high specific content of P-450 (18.2 nmol/ mg protein) after fast protein liquid chromatofocusing, displaying an isoelectric point of 6.05. Spectral analy- sis of the homogeneous enzyme showed that it is pre- dominantly low spin in the oxidized state, with a slight red-shifted ferrous carbonyl complex that exhibits a maximum at 452 nm. The purified cytochrome P-450, successfully reconstituted with NADPH-cytochrome P-450 reductase, displayed a maximal turnover rate of 50 nmol of substrate/nmol of P-450/min. In the purified and reconstituted form, the enzyme catalyzed the oxidation of three different chiral benzyltetrahy- droisoquinoline substrates, namely (S)-coclaurine, (R)-N-methylcoclaurine, and (S)-N-methylcoclaurine, leading to the formation of three distinct dimeric prod- ucts, (R,S)-berbamunine, (R,S)-2’-norberbamunine, and (R,R)-guattegaumerine, that are also present in the plant cell cultures in vivo.

This is the first report of a P-450 enzyme that me- diates regio- and stereoselective intermolecular oxi- dative phenol coupling to furnish natural dimeric com- pounds. In this catalytic cycle cytochrome P-450 func- tions as an oxidant in a bisubstrate reaction without transfer of the activated oxygen atom to either of the two chiral substrates.

A precise concept of oxidative phenol coupling in the alka- loid field was presented by the brilliant proposals of Barton and Cohen (1) as far back as 1957. Generation of phenolate radicals by oxidative attack and subsequent radical pairing furnishes the new diphenyl ether or aryl-aryl bond either inter- or intramolecularly. In the case of the bisbenzylisoqui- noline (BBIQ)’ alkaloids, which are products of dimerization

* This work was supported by Grant SFB 145 from the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4 Present address: Nestle Research Center, P. 0. Box 44, Vers- chez-les-Blanc, CH-1000 Lausanne 26, Switzerland.

244; Fax: 089-59-02-611. To whom correspondence should be addressed. Tel.: 089-59-02-

‘ The abbreviations used are: BBIQ, hishenzylisoquinoline; BTIQ, benzyltetrahydroisoquinoline; Chaps, 3-(3-~holamidopropyl)dimeth- y1ammonio)-1-propane-sulfonate; PAGE, polyacrylamide gel electro- phoresis; HPLC, high performance liquid chromatography; FPLC, fast performance liquid chromatography.

of 6,7,4’-trioxygenated benzyltetrahydroisoquinoline (BTIQ) bases, the oxidative coupling reaction proceeds in a highly regio- and stereospecific manner.

In order to elucidate the true nature of the biocatalysts involved in oxidative phenol coupling, we decided to investi- gate the formation of the simple dimer (R,S)-berbamunine at the enzymic level. From a biogenetic viewpoint, this com- pound is an ideal model for such studies as it represents the most simple class of BBIQ alkaloids that contain only a single diphenyl ether linkage between the benzylic moieties of two enantiomeric N-methylcoclaurine molecules. Also, this large and important group of natural compounds has been subject of much study and review in the past years (2-5), reflected also by the numerous in uiuo tracer experiments carried out to investigate the biosynthetic sequence of the dimerization reactions (6-10).

An ideal source for such biosynthetic studies is callus and suspension cultures of Berberis stolonifera plant cells, with contents of individual bases in the region of up to 1% calcu- lated on a dry weight basis (11). The chief dimers isolated are all biogenetically interrelated and possess opposite configu- ration l - (R) , l’-(S), with the exception of the rare (R,R)- dimer guattegaumerine (10) (Fig. I). These cells also elaborate smaller but significant amounts of more complex dimers that portray additional intramolecular linkage via the isoquinoline or “head” portions of the molecules, exemplified by (R ,S ) - aromoline and (R,S)-obamegine (Fig. l), that represent pro- gressive oxidative modifications of their progenitor (R,S)- berbamunine.

Enzymic studies on BBIQ alkaloid biosynthesis are scarce, and those performed suggest the participation of substrate- unspecific peroxidases. Such reactions are also accompanied by extremely low yields and often furnish non-biogenetic type products (12-14). Recently, intermolecular C-0-C condensa- tion of two enantiomeric N-methylcoclaurines was demon- strated at the cell-free level for the first time (15). The enzyme mediating this highly substrate-specific process was identified as a member of the cytochrome P-450 oxidases. These he- moproteins are also functional in other secondary biosyn- thetic pathways of alkaloids, catalyzing for example site- specific hydroxylations that lead to the production of defense- related metabolites such as the fully oxidized benzophenan- thridine base sanguinarine (16-19).

A major drawback in the isolation and purification of plant cytochrome P-450 enzymes are the extremely low levels of protein and also the detergent instability when attempting to solubilize the cytochrome complex out of the membrane. To date, successful purification to homogeneity of plant-derived cytochrome P-450 monooxygenases has been reported in only five cases (20-24). Of these only three, namely the cinnamic acid 4-hydroxylase, the 3,9-dihydroxypterocarpan-6-a-hy- droxylase, and the NADPH-independent allene oxide syn-

823

824 Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids

obme*,ae

Aromollnc

FIG. 1. Structures of BTIQ substrates and dimeric BBIQ products.

thase can be correlated with a definite physiological function (22-24).

This is the first report of a P-450 enzyme that mediates highly regio- and stereoselective oxidative phenol coupling to afford natural BBIQ alkaloids. The purified terminal oxidase is identified and characterized as a unique cytochrome P-450 enzyme that functions as an oxidant of two chiral phenols without the introduction of catalytically activated oxygen into either of the substrate molecules. Additionally, this report attempts to clarify the true rationale governing the BTIQ phenol coupling reaction.

EXPERIMENTAL PROCEDURES

All isolation and enzyme fractionation procedures were carried out at 4 “C, unless otherwise stated. All pH measurements were made at 22 “C. Plant cell cultures were provided by our cell culture laboratory. The suspended cells were grown in Linsmaier-Skoog medium (25) at 23 “C on a gyratory shaker (100 rpm) under constant diffuse light, subculturing every 8 days, using an inoculum of about 10% of the volume of the culture medium. The dry weight of suspension cultures was determined by drying aliquots at 60 “C for at least 2 days.

Preparation of Microsomal Frucfions-Plant tissue (typically 400 g fresh weight) of 4-day-old B. stolonifera suspended cells was shock frozen in liquid nitrogen. A 100 mM Tricine/NaOH buffer, pH 7.4, containing 10 mM thioglycolic acid was added (2 ml/g fresh weight) to the frozen material. The tissue was ground to homogeneity, squeezed through 4-layered cheesecloth, and the filtrate centrifuged (10,000 x g) for 20 min. The clear cell-free supernatant was collected and passed through an XAD-2 (Serva) column (7 X 20 cm) pre- equilibrated with homogenization buffer. The column effluent was subsequently supplemented with MgCl, to a final concentration of 50 mM. After sedimentation at 48,000 X g for 20 min, the microsomal pellet was suspended in 100 mM Tricine/NaOH buffer, pH 8.0, containing 10 mM thioglycolic acid with 20% glycerol (v/v), and stored at -70 “C until usage.

Enzyme Assays-The standard HPLC assay employed the enan- tiomeric BTIQ hydrochlorides diluted to typically 250 mM in a final reaction volume (200 ~1) containing the following: NADPH (1 mM), the reconstituted enzymes cytochrome P-450 oxidase (1-3 pmol), NADPH-cytochrome P-450 reductase (0.5-0.7 rg), and B. stoloniferu crude lipids (20 gg). The mixture was incubated for 1 h at 37 “C and the reaction terminated by addition of 20 ~1 of trichloroacetic acid (20% w/v). Aliquots were analyzed directly by HPLC.

The radiochemical assay was carried out in a final volume of 0.3 ml and comprised (S)-N-methylcoclaurine (3.34 PM), (R)-[N-methyl- ‘?]coclaurine (0.027 &i, 0.0081 &i/FM), NADPH (167 yM), the reconstituted enzymes cytochrome P-450 oxidase (l-3 pmol), NADPH-cytochrome P-450 reductase (0.5-0.7 fig), and B. stolon$eru crude lipids (20 pg). The reaction mixture was incubated at 37 “C for 1 h and terminated by addition (60 ~1) of sodium carbonate buffer (1 M, pH 9.5) and ethyl acetate (0.2 ml). An aliquot (40 ~1) was removed from the organic phase and chromatographed together with authentic compounds on TLC sheets (0.25 mm), developed in the solvent system ethyl acetate/2-butanone/formic acid/water (5:3:3:1), revealing the following & values: N-methylcoclaurine = 0.57; berbamunine/guat- tegaumerine = 0.26. After chromatography the sheets were scanned for radioactivity and analyzed on a Berthold TLC scanner.

HPLC Conditions--Samples of the standard enzyme assay (50 ~1) were chromatographed at room temperature using a Vydac SC-201 RP (Macherey-Nagel) guard column (30 X 4.6 mm) and a 5-pm Lichrospher 60 RP-Select B column (250 x 4.0 mm, Merck). Analyses were carried out eluting (1 ml/min) with a 0.02% (v/v) phosphoric acid/acetonitrile gradient going from 4.5 to 18% acetonitrile in 20 min and then to 90% acetonitrile in 5 min, detecting the alkaloidal peaks spectrophotometrically at 282 nm. Retention times (minutes) of the BTIQ and BBIQ alkaloids were as follows: coclaurine = 10.92, N-methylcoclaurine = 12.26, guattegaumerine = 17.2, 2’-norberba- munine = 17.9, berbamunine = 18.2.

Solubilization and Purification of Cytochrome P-450-The crude microsomes were adjusted to a protein concentration of 10 mg/ml with a 100 mM Tricine/NaOH buffer, pH 8.2, containing 10 mM thioglycolic acid and 20% glycerol. For solubilization, sodium cholate was added dropwise up to a detergent/protein ratio of 0.9. The suspension was left stirring for 1 h on ice and then centrifuged (100,000 X g) for 1 h. Thereafter, a solution of polyethylene glycol 8000 (40% w/v in water) was added dropwise to the supernatant to a final polyethylene glycol concentration of 4% (w/v). The preparation was gently agitated for further 20 min and then subjected to ultracen- trifugation (100,000 x g, 1 h): The clear yellow supernatant was diluted 1:2 with a 100 mM potassium phosphate buffer, pH 7.4, containing 1 mM EDTA, 1 mM dithiothreitol, and 20% glycerol. The solubilized microsomes were then fractionated on an w-aminooctyla- garose column (2.0 X 10 cm), equilibrated with the aforementioned detergent-free dilution buffer. The column was rinsed (1 ml/min) until no further protein could be detected. A cytochrome P-450-rich preparation was washed from the column with the same equilibration buffer but containing 0.5% (w/v) Chaps. Thereafter, the column was washed with the same buffer containing 0.2% Chaps (w/v) and 0.5 M NaCl to elute the NADPH-cytochrome P-450 reductase component.

The cytochrome P-450 oxidase-rich preparation was adsorbed on a hydroxylapatite column (1 x 4.5 cm) equilibrated with a 20 mM potassium phosphate buffer, pH 7.2, containing 1 mM dithiothreitol, 0.2% Chaps, and 20% glycerol. Elution of protein (0.5 ml/min) was performed with a step-gradient going from 20 to 500 mM potassium phosphate buffer with the aforementioned ingredients. Phenol oxi- dative coupling activity was detected in the 250-300, 300-350, and 400-500 mM eluents, the highest specific P-450 content, and activity being present in the latter peak.

The 400-500-mM fraction obtained from the hydroxylapatite col- umn was concentrated with Amicon Centripreps (30,000 cut-off) to a volume of 2 ml and passed over a Sephadex G-25 column (PD-10, Pharmacia) to remove the anionic buffer. This preparation was loaded directly onto a FPLC Mono Q column (HR 5/5, Pharmacia) equili- brated with a 100 mM Tricine/NaOH buffer, pH 8.0, containing 1 mM dithiothreitol, 0.2% Emulgen 911 (v/v), and 5% glycerol. The enzyme was eluted with a NaCl gradient (O-l M) in the equilibration buffer. The fractions showing enzyme activity (0.1-0.15 M) were pooled and desalted after passage over a Sephadex G-25 gel filtration column (PD-10, Pharmacia). Electrophoretically homogeneous en- zyme was obtained after a final chromatofocusing step employing a Mono P column (HR 5/20, Pharmacia), equilibrated with a 25 mM Bis-Tris buffer, pH 6.7, containing 0.2% Emulgen 911 and 1 mM dithiothreitol. The column was rinsed with 40 ml of a mixture of

Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids 825

Polybuffer 74 and 96, ratio 6 4 (both Pharmacia), adjusted to a pH 5.5 with acetic acid, containing 0.2% Emulgen 911 and 1 mM dithio- threitol. All of the fractions showing absorbance in the 420 nm region were subjected to SDS-PAGE analysis.

Isolation of Total Lipids-Microsomal lipids were extracted directly from a crude microsomal preparation (48,000 X g) of B. stolonifera suspension cultures by addition of a mixture of chloroform/methanol (21). After exhaustive extraction, the organic layer was removed, washed with 0.9% NaC1, and concentrated under a stream of nitrogen to yield a yellowish oily residue. The lipids were suspended by sonification in a 100 mM Tricine buffer, pH 8.0, containing 1 mM dithiothreitol and 20% glycerol, to give a final concentration of 20 mg/ml.

Reconstitution-To reconstitute the cytochrome P-450 oxidase and the NADPH-cytochrome P-450 reductase, the putative cytochrome P-450 oxidase containing fractions (typically 2-4 pmol) were added to the purified NADPH-cytochrome P-450 reductase preparation (typically 0.5-0.7 fig, specific activity: 26 pmol/min/mg) together with a B. stolonifera crude lipid extract (10 fig) and bovine serum albumin (100 pg). The mixture was frozen in liquid nitrogen and then left to thaw on ice to promote reconstitution. Excess detergent was removed after passage over Extracti-gel D (Pierce), and the detergent-free effluent employed directly for enzymic tests that were initiated by the addition of NADPH (see “Enzyme Assays”).

Purification of B. stolonifera NADPH-Cytochrome P-450 Reduc- tase-The cytochrome P-450 reductase-rich preparation eluted with 0.5 M KC1 from the w-aminooctylagarose column (see “Solubilization and Purification of Cytochrome P-450”) was loaded onto a 2’,5’- ADP-Sepharose affinity column (1 X 7 cm, Pbarmacia) equilibrated with a 50 mM potassium phosphate buffer, pH 7.4, containing 1 mM dithiothreitol, 0.1% Chaps, and 10% glycerol. The column was washed (1 ml/min) thoroughly with the equilibration buffer until no more protein could be detected in the flow-through. The reductase was then displaced from the gel with the same equilibration buffer but containing in addition 5 mM 2’-AMP, 1 p M FAD, and 1 p M FMN. The reductase containing fractions were pooled and stored at -70 “C until further usage.

Isolation of Bovine Liver NADPH-Cytochrome P-450 Reductase- The liver reductase was solubilized according to standard procedures (26).

Other Methods-The concentration of cytochrome P-450 was de- termined from the CO difference spectrum of the dithionite-reduced protein using an extinction coefficient of 91 mM” for the difference between the maximum absorbance at 450 nm and that registered at 490 nm (27). The absolute spectrum of the homogeneous native enzyme was determined at 20 “C in the Mono P buffer at pH 6.05 containing 0.2% Emulgen 911. NADPH-cytochrome P-450 reductase was assayed (28) employing an extinction coefficient of 21 mM” at 550 nm.

Protein was determined by the method of Bradford (29) using bovine serum albumin as a standard.

Polyacrylamide slab gel electrophoresis was carried out in the presence of SDS with the discontinuous buffer system at 10 “C as described by Laemmli (30). Molecular weight standards for SDS- PAGE were rainbow markers (Amersham). Protein was detected by staining the gels with silver nitrate (31).

Materials and Equipment-All solvents and reagents were of the highest purity commercially available. All buffers were prepared with Millipore water and degassed before use. w-Aminooctylagarose, so- dium cholate, NADPH, Tris, Bis-Tris, dithiothreitol, thioglycolic acid, L-a-phosphatidylcholinedimyristoyl, L-a-phosphatidylcholine- dilauroyl, horse heart cytochrome c Type 111, glucose oxidase, catalase (bovine liver), 2’-AMP, FAD, and FMN were purchased from Sigma. Hydroxylapatite HTP Bio-Gel was from Bio-Rad, and Extracti-gel D was from Pierce. XAD-2 and bovine serum albumin were obtained from Serva. Chaps and Tricine were from Roth, and Emulgen 911 was a kind gift of Dr. Michael Kastner, Berlin. 2’,5’-ADP-Sepharose, Sephadex G-25 (PD-10) columns, the FPLC columns Mono Q HR 5/ 5, and Mono P HR 5/20 were obtained from Pharmacia. Amicon Centripreps were purchased from Amicon and used according to the manufacturer’s instructions. Analytical and preparative TLC was conducted on Merck G60 F254 (0.25 mm) sheets and Merck G60 (0.5 mm) plates, respectively. Alkaloid bands were located under a UV lamp (254 nm) and by spraying the edges of the chromatograms with iodoplatinate reagent. Radioactivity was detected on TLC sheets with a Berthold Automatic TLC Analyzer Tracemaster 20. HPLC was carried out on a Merck-Hitachi HPLC system. FPLC was performed with a Pharmacia FPLC system. Mass spectra (chemical and direct

chemical ionization) were obtained on a Kratos MS80 RFA spectrom- eter with NH3 as the reactant gas. 13C NMR (90.6 MHz) and ‘H NMR spectra (360 MHz) were recorded with a Bruker AM 360 instrument with MerSi as internal standard.

Substrate Synthesis-The chiral coclaurines, their N-methylated congeners, the tetraoxygenated bases norreticuline and reticuline were synthesized and resolved according to standard procedures (10, 32-34). The analogues 3’-hydroxycoclaurine and 3”hydroxy-N- methylcoclaurine were synthesized as described (351, and the 13C- labeled counterparts were prepared analogously (10,35). Radiolabeled N-metbylcoclaurines [N-14CH3] were synthesized from the corre- sponding enantiomerically pure coclaurines according to the method described (10).

Product Identification-The BBIQ alkaloid 2’-norberbamunine, synthesized in the reconstitutive enzymic system, was identified by mass spectral and NMR analyses and comparison with authentic material isolated from B. stolonifera cell suspension cultures. In order to obtain sufficient material, a large scale incubation comprising 40 individual standard HPLC assays, with (R)-[l-13C]N-methylcoclaur- ine and (S)-[l-13C]coclaurine as substrates, was incubated at 37 “C overnight (16 h). The single assays were pooled and basified with a 1 M sodium carbonate buffer to a final pH 9.5. The bases were exhaus- tively extracted into ethyl acetate and the organic layer dried with sodium sulfate. The filtered solution was then concentrated in vacuo (35 “C) and the residue chromatographed (TLC) in solvent system ethyl acetate/2-butanone/formic acid/water (5:3:1:1), revealing an R, value of 0.50 for 2’-norberbamunine. The compound was eluted with methanol, concentrated to dryness under reduced pressure (35 “C), and rechromatographed in solvent system dichloromethane/metha- nol/aqueous ammonia (909:l) affording -0.2 mg of isotopically la- beled and chromatographically pure 2’-norberbamunine (R f = 0.40). Low resolution direct chemical ionization mass spectroscopy (120 eV) yielded m/z = 585 (M+H)+ (loo%), 407 (2%), 391 (5.8%), 193 (8%), 179 (3.6%), 168 (49%), 151 (47%). 13C NMR (CD30D): 65.95 ppm (C- l ) , 57.73 ppm (‘2-1’). ‘H NMR (CD30D): 6 2.79 (s, N-CH3), 3.82 (s, 0-6-CH3 or 0-6’-CH3), 3.84 (s, 0-6-CH3 or 0-6’-CH3), 5.99 (d, 4.0 HZ, H-8’), 6.43 (d, 4.0 HZ, H-8), 6.48 (d, 2 HZ, H-lo), 6.68 (s, H-5), 6.77 (s, H-5’), 6.89 (d, 8.5 HZ, H-14), 6.90 (d, 8.5 HZ, H-11’/13’), 6.92 (d, 8.5 HZ, H-13), 7.22 (d, 8.5 HZ, H-10’/14’).

(R,S)-Berbamunine was synthesized in vitro with the reconstituted enzymes as outlined above, employing ( R ) - [ 1-’3C]N-methylcoclaurine and (S)-[N-’3CH3]coclaurine. The BBIQ alkaloid was purified by TLC developed in solvent system ethyl acetate/chloroform/ethanol/ diethylamine (70:15:105), revealing an R, value of 0.33 for (R,S)- berbamunine and 0.47 for N-methylcoclaurine. Approximately 0.25 mg of amorphous 13C-labeled berbamunine was isolated. Low resolu- tion direct chemical ionization mass spectroscopy (120 eV) showed the following fragments: m/z (%) = 599 (M+H) (loo), 500 (lo%), 456 (12%), 428 (21%), 407 (40%), 391 (18%), 357 (17%), 195 (60%), 193 (72%), 177 (62%). 13C NMR (CD,OD): 66.12 ppm (C-1), 41,17 ppm (N’-CH3). ’H NMR (CD30D): 6 2.84 (br.s, N-CH3), 3.82 (0-6-CH3), 3.83 (0-6”CH3), 5.72 (s, H-8’), 5.88 (d, 4.0 HZ, H-8), 6.39 (d, 2 HZ, H-lo), 6.67 (s, H-5), 6.77 (s, H-5’), 6.86 (d, 8.5 HZ, H-14), 6.88 (d, 8.5 HZ, H-l1’/13’), 6.92 (d, 8.5 HZ, H-13), 7.08 (d, 8.5 HZ, H-14’/10’).

The isomeric dimer (R,R)-guattegaumerine was prepared in isolat- able amounts as described above, by adding (R)-[l-’3C]N-methylco- claurine as a substrate. The product was isolated exactly as described above for (R,S)-berbamunine. Due to the low turnover rate to the (R,R)-dimer, only -0.03 mg of material could be obtained, insufficient for thorough NMR analysis. Low resolution chemical ionization mass spectroscopy (120 eV) gave m/z (%): 599 (M+H)+ (18%), 527 (36%), 421 (12%), 407 (46%), 285 (49%), 257 (52%), 195 (loo%), 177 (21%).

RESULTS

Solubilization and Purification of Cytochrome P-450”Table I depicts the purification scheme for the cytochrome P-450 oxidase. In relation to the crude microsomes, the total P-450 yield after solubilization was never higher than 15%, and the specific content at this stage was always distinctly lower, suggesting conversion of cytochrome P-450 to the inactive P- 420 form. Polyethylene glycol precipitation was chosen as the first fractionation step and, at first sight, seems rather inef- ficient in terms of both recovery of total P-450 and specific P-450 content. However, fractionation of the solubilized mi- crosomes by aminooctylagarose column chromatography (Fig.

826 Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids TABLE I

Purification of R. stolonifera cvtochrome P-450 ~~ ~ ~

Fraction Volume Protein Cflochrome Specific content Purification Recovery” P-450 mi m# nmol nmol P-450/mg -fold ?6

Crude microsomesb 45 450 39.6 0.088 1.0 100 PEG supernatant 46 141 7.05 0.05 0.57 17.8 w-Aminooctylagarose, pool B 22 7.2 9.4 1.3 14.7 24 Hydroxylapatite, 0.4-0.5 M 15 0.72 2.8 3.9 44.3 7

Mono Q‘ pool

Mono PC 1 0.0044 0.08 18.2 207 0.2 6 0.05 0.3 6.0 68.2 0.75

Calculated on the basis of all forms of P-450 present in the microsomes. Approximately 1.6 kg fresh weight cells (84 g dry weight) yield -450 mg crude microsomes. Protein and P-450 content were determined after removal of Emulgen 911 (Extracti-gel D column).

B c 0 N 0 - m C

5 n a

0.40

0.30

0.20

0.10

0.00

Elution volume (ml)

- 100

.I30

60

4 0

20

- 0 0 20 4 0 60 80 100 120

Retention time (min)

P) ; E

m .- c m 0 U

-

D

0.20

0.15

0.10

0 0 5

0.0C 0

A B

4 . 4 . .

Retention time (min)

C D E F G H

4 92000

4 69000

I 46000

FIG. 2. Elution profiles and SDS-PAGE of cytochrome fractions obtained after column chromatography. Panel A, elution from w-aminooctylagarose column; A, equilibration buffer; B, equilibration buffer containing 0.5% Chaps; C, equilibration buffer containing 0.2% Chaps and 0.5 M KCI. -, A420, hemoprotein and/or other chromophores; - - -, BTIQ coupling activity. Panel R, elution from hydroxylapatite column, -, A420, hemoproteins and/or other chromophores; - - -, BTIQ coupling activity. Panel C, elution from Mono P column, -, A420, hemoprotein and/or other chromophores; - - -, BTIQ coupling activity. Panel D, SDS-PAGE of hemoprotein preparations after Mono P chromatofocusing; A-C, active fractions; D-H, non-active hemoprotein fractions eluting between PI 6.8 and 6.2; activity was strictly correlated with protein content. Chromatographic and electrophoretic conditions were as described under “Experimental Procedures.”

2, panel A ) exhibited 30-40% more total P-450 in the P-450- rich column effluent (pool B ) compared with those micro- somes treated with polyethylene glycol. The advantages of polyethylene glycol precipitation were therefore only evident after the first chromatography step that also completely re- solved the NADPH-cytochrome P-450 reductase activity. Quantification via the CO difference spectrum of the oxidase preparation revealed -15-fold enrichment and absolutely no indication of the inactive 420-nm form.

Further purification of the cytochrome P-450 oxidases on a hydroxylapatite column (Fig. 2, panel B ) ) revealed dimer- izing activity in three different fractions (200-250, 250-400, and 400-500 mM potassium phosphate), all showing varying specific P-450 contents (average values -3.3, 3.0, and 3.9 nmol of P-450/mg protein, respectively). The preparation with the highest specific P-450 content, i.e. the 400-500 mM potassium phosphate eluent, was passed over a Mono Q column equilibrated with the detergent Emulgen 911. Notably, Mono Q chromatography of the same fraction but in the

presence of Chaps (0.2%) yielded a completely different elu- tion profile. In the latter case, activity was distributed in a t least five different fractions ranging from 0.2-1.0 M NaCl. Such a chromatographic pattern led to no P-450 enrichment a t all. Similar differences in chromatographic behavior of P- 450 on Mono Q with ionic and non-ionic detergents have been reported (36, 37), suggesting that the choice of detergent is a factor not to be underestimated when applying anionic sepa- ration techniques. Phenol coupling activity was associated with a relatively broad region eluting with 0.1-0.15 M NaCl. SDS-PAGE of the Mono Q fraction revealed three major protein bands in the M , = 46,000 region, with the upper two running so extremely close together that these could only be identified as two individual proteins after the first two low molecular weight markers M, = 14,300 and 21,500 had already run out of the gel. Separation of the three major proteins was achieved with a Mono P chromatofocusing column, with the elution profile (Fig. 2, panel C) depicting a single sharp activity zone at PI = 6.05 correlating with the lower of the

Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids 827

two close bands as visualized on SDS-PAGE (Fig. 2, panel D). With this procedure the enzyme was purified 207-fold to a specific cytochrome P-450 content of 18.2 nmol/mg (Table I).

Properties of the Reconstituted Cytochrome P-450-All in- cubations were performed with the standard HPLC assay as described under “Experimental Procedures.” Conditions were chosen to measure activity within linearity and, as is evident in Fig. 3, an incubation period of 1 h seemed optimal in terms of accurate product quantification. Restoration of BTIQ cou- pling activity was achieved only after addition of NADPH- cytochrome P-450 reductase and NADPH. A further obliga- tion is the complete removal of detergent molecules, ensured by passage over Extracti-gel D columns as described under “Experimental Procedures.” As shown in Table 11, removal of molecular oxygen via an enzymic oxygen scavenging system inhibited the enzyme, whereas catalase or superoxide dismu- tase did not impede coupling activity, excluding the partici- pation of reactive oxygen species or hydrogen peroxide.

Reconstitution of oxidase activity was also investigated as a function of cytochrome P-450 oxidase, NADPH-cytochrome P-450 reductase, and lipid content, a test carried out to

5 , I

0 60 120 180 240 300

Time (mid

FIG. 3. The reconstituted oxidative BTIQ coupling reaction as a function of time. Production of the (R,S)-dimer berbamunine was monitored as described under “Experimental Procedures.”

TABLE I1 Requirements for phenol coupling activity

The complete assay mixture consisted of a cytochrome P-450 fraction containing 3.0 nmol of P-450/mg protein (2.0 pmol of P-450) and a reductase preparation (0.7 pg) having a specific activity of 26 /*mol/min/mg. The standard reconstitutive conditions were employed as described under “Experimental Procedures.” The values given represent an average of three independent measurements.

Reconstitutive conditions % activitv“

Complete 100 Only cytochrome P-450 0 Only NADPH-cytochrome P-450 reductase 0 Only crude B. stolonifera lipidsb 0 Without liuids 90 Addition of L-cu-phosphatidylcholinedimyristoylb Addition of ~-~-phosphatidylcholinedilauroyl~ Without NADPH Without oxygen‘ Additiond of catalase and superoxide dismutase Without bovine serum albumin With reductase isolated from E. californica plant

With reductase isolated from bovine liver cell cultures

24 a2 0 4

95 62 a7

0

P-450/h. 100% activity represents 1.2 nmol of berbamunine produced/pmol

Addition of 20 pg/assay. ‘Assays were carried out by adding glucose oxidase (10 units),

Assays were carried out by adding catalase (10 units) and super- catalase (10 units), and glucose (5 mM).

oxide dismutase (5 units).

ascertain optimal assay conditions. As depicted in Table 11, lipids are by no means compulsory constituents for the reac- tion to take place. Variation in crude lipid content (1-100 pg/ assay) also had no significant effect on activity. With varying NADPH-cytochrome P-450 reductase activities/assay, the BTIQ dimerizing reaction proceeded linearly starting from a minimum turnover rate required for the reductase of 3 nmol of cytochrome c/min up to 18 nmol/min, reaching a maximum at -30 nmol/min and then flattening off. The origin of the NADPH-cytochrome P-450 reductase seems rather unimpor- tant because -87% BTIQ coupling activity could be reconsti- tuted with a reductase isolated from another plant family (Eschscholtzia californica, Papaveraceae). However, reductase isolated from beef liver did not transfer electrons to the terminal oxidase, an interesting observation in view of the fact that antibodies raised against plant reductase do not interact with animal reductases (38).

At constant NADPH-cytochrome P-450 reductase activity (18 nmol/min) and increasing cytochrome P-450 content, the reaction exhibited linearity, the product being detectable with as little as 0.2 pmol of oxidase and attaining a maximum turnover rate between 10 and 12 pmol of cytochrome P-450. The purified oxidase displayed a relatively narrow tempera- ture optimum at 35-38 “C, with half-maximal activity re- corded at 22 and 46 “C. This result is a clear shift to a higher temperature optimum in the reconstituted system as com- pared to the crude microsomal preparation that had a broad temperature maximum between 25 and 30 “C (15). Similarly, a definite buffer-dependent shift in the pH maximum of the reconstituted enzymes as compared to the crude microsomes (pH optimum 8.0-8.5) was observed (15). The highest BTIQ coupling activity was recorded with a Tricine/NaOH buffer, peaking at pH 7.8. In comparison, only 57% of the maximal activity could be recorded with a potassium phosphate buffer, accompanied by a slight pH shift with a maximum in the range of pH 7.2 to 7.5.

The molecular weight of the oxidase was determined under denaturing conditions using SDS-PAGE, revealing M , = 46,000 and an isoelectric point of 6.05, the latter an average value of three runs determined by chromatofocusing on a Mono P column. Enrichment was assessed as 207-fold (based on the crude microsomes) with a specific P-450 content of 18.2 nmol/mg protein (theoretical value 21.7 nmol/mg).

The homogeneous enzyme exhibits a Soret (gamma) band at a maximum at 414 nm, characteristic of a predominantly low spin ferricytochrome P-450. The ferrocytochrome-CO complex has an absorption maximum at 452 nm, with no indication at all of the P-420 form (Fig. 4). This maximum is slightly shifted to a longer wavelength as compared to the CO difference spectra recorded during the purification procedure that were characteristically at 450 nm. This shift suggests the presence of different cytochrome P-450 forms in B. stolonifera cell suspension cultures.

Substrate Specificity-The isolation of diastereomeric di- mers of the (R,R) and (R ,S ) type as well as the 2-N-nor congeners from B. stolonifera cell suspension cultures (10,11) signifies a rather unspecific oxidative attack of the monomeric BTIQ alkaloid progenitors. In order to clarify the substrate specificity, the purified enzyme was incubated under condi- tions as described under “Experimental Procedures.” Fur- thermore, quantitative calculations were only possible with direct HPLC analyses that also enabled the separation of the isomeric and structurally related dimers (R,S)-berbamunine, (R,R)-guattegaumerine, and (R,S)-2’-norberbamunine (Fig. 1). The HPLC profiles recorded after incubation with various putative BTIQ substrates are depicted in Fig. 5. As antici-

828

1.00 -

0.50 -

0.00 -

-0.50 1

-1.00 -

Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids

I

350 400 450 500

Wavelength (nm)

FIG. 4. Absorption spectra of homogeneous cytochrome P- 450 containing 18.2 nmol of P-450/mg protein. - - -, oxidized native enzyme; -, dithionite-reduced protein after exposure to carbon monoxide. Spectra were recorded under conditions as de- scribed under “Experimental Procedures,”

-

- E, N N W

- . L. C

.I D m *jl E n F

Time lmin)

FIG. 5. Reverse-phase HPLC separation and UV detection (282 nm) of BTIQ substrates and dimeric products. Reconsti- tution was performed in the standard HPLC assay employing elec- trophoretically homogeneous oxidase as described under “Experimen- tal Procedures”. Panel A, control assay without NADPH; panel B, equimolar concentrations of (R)- and (S)-N-methylcoclaurine; panel C, ( R ) - and (S)-N-methylcoclaurine (ratio 1O:l); panel D, equimolar concentrations of (R)- and (S)-N-methylcoclaurine and (S)-coclaur- ine; panel E, equimolar concentrations of (R)-N-methylcoclaurine and (S)-coclaurine; panel F, equimolar concentrations of ( E ) - and (S)-coclaurine. The average retention times (min) of the dimeric products were (R,R)-guattegaumerine ( I ) 17.2; 2’-norberbamunine (ZI) 17.9; berbamunine (ZZZ) 18.2.

pated, incubation with equimolar amounts of both (R) - and (S)-N-methylcoclaurine lead to the formation of the major simple tail-to-tail diaryl-ether coupled dimer berbamunine (Fig. 5 B ) . However, closer examination of the products shows the presence of the bis (R)-N-methylcoclaurine dimer guat- tegaumerine. Incubation with a higher (R)-N-methylcoclaur- ine to (S)-N-methylcoclaurine ratio (1O:l) shows definite accumulation of the (R,R)-stereoisomer at the expense of berbamunine (Fig. 5 C ) . Furthermore, this oxidase is not at all

restricted to attack of the enantiomeric N-methylcoclaurines, but also generates a N-nor dimer built up of (R)-N-methyl- coclaurine and (S)-coclaurine and linked in the same tail-to- tail fashion as its N-methylated analogues (Fig. 5D).

Incubations with (5‘)-N-methylcoclaurine and (R)-coclaur- ine did not lead to a dimeric product, as well as incubations with (S)-N-methylcoclaurine alone or with either ( S ) - or ( R ) - coclaurine together or individually (Fig. 5 F ) . Other benzyli- soquinolines such as (R,S)-3’-hydroxy-N-methylcoclaurine, (R,S)-norreticuline, and (R,S)-reticuline did not act as sub- strates for the BTIQ coupling enzyme (Table 111).

Kinetic Properties-The kinetic properties of the purified and reconstituted oxidase were studied as a function of ( R ) - N-methylcoclaurine, (8)-N-methylcoclaurine, (S)-coclaur- ine, and NADPH concentrations. For the formation of the @$)-dimer berbamunine and the (R,S)-dimer 2”norberba- munine, normal Michaelis-Menten kinetics could be observed for the substrate (R)-N-methylcoclaurine. The K, values for the aforementioned substrate involved in the formation of berbamunine and 2”norberbamunine were determined as 8 and 25 p ~ , respectively. At higher (R)-N-methylcoclaurine concentrations relative to (S)-N-methylcoclaurine, the grad- ual appearance of the (R,R)-dimer could be observed (Fig. 6A) . However, slight deviation from Michaelis-Menten kinet- ics was registered when attempting to determine the K, values for the “prime-half” substrates of the enzymic phenol coupling reaction. This phenomenon is due to the formation of the (R,R)-dimer guattegaumerine a t relatively high (R)-N-meth- ylcoclaurine levels, exemplified in Fig. 6B. An increase in the concentration of (S)-N-methylcoclaurine (or (S)-coclaurine for synthesis of 2”norberbamunine) resulted in a rapid in- crease of berbamunine, with a concomitant decrease in the level of guattegaumerine, implying a competitive reaction in the “prime” substrate-binding site of the enzyme. Notably, the turnover rate to berbamunine far exceeds that of guatte- gaumerine (Table III), suggesting further that (R,R)-d‘ lmers are only formed under selective substrate pressure, i.e. defi- ciency of the (S)-isomers. Comparison of the substrate turn- over rates of the crude microsomes and the reconstituted enzymes, depicted in Table 111, not only reveals the antici- pated higher rates for the enriched enzyme, but interestingly

TABLE 111 Comparison of substrate oxidation rates

The reconstituted system contained homogeneous cytochrome P- 450 (1 pmol) and purified reductase (0.7 pg) with a specific activity of 26 pmol/min/mg. The standard reconstitutive conditions were employed as described under “Experimental Procedures.” Values given represent an average of three separate measurements.

Turnover (nmol/nmol P 450/min)

Substrate“ Microsomes Reconstituted

svstern

(R)- + (S)-N-Methylcoclaurine 3.34 50 (R)-N-Methylcoclaurine + (S)-coclaur- 0.14 50

(R)-N-Methylcoclaurine + (R)-coclaur- 0.12 3.5

(R)-N-Methylcoclaurine 0.25 9.4 (S)-N-Methylcoclaurine 0 0 (R)- + (S)-Coclaurine 0 0 (R,S)-3’-hydroxy-N-methylcoclaurine 0 0 (R,S)-Norreticuline 0 0 (R,S)-Reticuline 0 0 (R,S)-Berbamunine 0.07’ 0 (R,R)-Guattegaumerine 0 0

ine

ine

Equimolar concentrations of substrates were employed. Turnover to aromoline that features an additional diphenyl ether

bridge between C-7-OH and C-8’.

Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids 829

+ 2’oo r”-l *‘0° 1.50 i

0.50 l ’ o o v

1.50

1 .oo

0.50

300 0.00

(R)-N-Iothylcocl8urino tUu)

+ 2.00 B

2.00

- 1.50

- 1.00

- 0.50

- 0.00 0 100 200 300

~S)-N-Yo(hylcotl~urIno (VU)

FIG. 6. Production of the dimers (R,S)-berbamunine and (R,R)-guattegaumerine as a function of substrate concentra- tion. Panel A, varying concentrations of (R)-N-methylcoclaurine a t a set concentration of the (S)-enantiomer (250 pM); panel B, varying concentrations of (S)-N-methylcoclaurine a t a set concentration of the (R)-isomer (250 p ~ ) . Assay conditions were as described under “Experimental Procedures.”

also a marked deviation as far as substrate specificity is concerned. In the case of the crude microsomes, predomi- nantly (R,S)-berbamunine is formed, accompanied by less than 10% of the (R,R)-dimer guattegaumerine and less than 5% of the N-nor dimers 2”norberbamunine and 2”norguat- tegaumerine. This result suggests that the inherent membrane structure also plays an important role in substrate selection.

Product Identity-The products of the coupling reaction mediated by the reconstituted purified oxidase were further identified by large scale standard HPLC assays employing “C-labeled substrates and subsequent HPLC, MS, ‘H NMR, and 13C NMR analyses of the purified products. Isotopically labeled berbamunine was produced by adding the monomeric BTIQ substrates (S)-[N-’3CH3]coclaurine and (R)-[1-’3C]N- methylcoclaurine in the incubation mixture. Subsequent 13C NMR analysis (CD30D) of the chromatographically purified product clearly illustrates the two enriched resonances at 66.12 and 41.17 ppm, corresponding to C-1 and N’-CH3, respectively (Fig. 7A). Closer inspection of the C-1 signal shows a relatively small upfield shifted shoulder at the base, attributable to the (R,R) contaminant guattegaumerine that could not be separated from (R,S)-berbamunine by thin layer chromatography (10). The ‘H NMR spectrum of enzymically synthesized berbamunine was identical to that recorded with authentic material isolated previously from B. stolonifera cell cultures. Specific labeling of the dimeric product is clearly reflected by the doublet nature of the proton H-8 (5.88 ppm) due to 3J carbon-proton coupling with C-1.

Similarly, large scale incubation of (R)-[l-’3C]N-methyl- coclaurine and ( S ) - [ l-’3C]coclaurine afforded isotopically en- riched 2”norberbamunine. The I3C NMR spectrum (CDsOD) depicted in Fig. 7B revealed two enhanced signals a t 65.95 and 57.73 ppm, attributable to C-1 and C-l’, respectively.

c-I 1

I 1

il

11

c v i v r l %ww+w#w”w, 67 65 63 61 59 57

ppm

FIG. 7. Proton-decoupled 13C NMR spectra of isotopically enriched dimeric products. A , (R,S)-berbamunine enzymically synthesized from (R)-[l-’3CC]N-methylcoclaurine and (S)-[N-”CH,] coclaurine; E , 2‘-norberbamunine enzymically synthesized from ( E ) - [l-13C]N-methylcoclaurine and (S)-[1-’”C]coclaurine. The BBIQ products were synthesized with the reconstituted enzyme and purified as described under “Experimental Procedures.”

These experiments clearly substantiate that no scrambling of label takes place during the oxidative coupling reaction and that the chiral centers remain intact during dimerization. The ‘H NMR spectrum of 2”norberbamunine exhibits close sim- ilarity with that of its isomer 2-norberbamunine, previously isolated from the same source (11). As in the case of highly enriched berbamunine, the H-8 and H-8’ protons were both split into doublets, due to 3J carbon-proton coupling with C- 1 and C-l’, respectively.

DISCUSSION

This is the first report of the purification and reconstitution of a plant cytochrome P-450 enzyme catalyzing an oxidative phenol coupling reaction. This enzyme mediates highly regio- and stereospecific condensation of two chiral6,7,4’-trioxygen- ated coclaurines to generate the natural BBIQ alkaloid skel- eton. As suggested (I), simple tail-to-tail coupled dimers of the (R,S)-berbamunine, (R,R)-guattegaumerine, or (S ,R)- magnoline type are formed first in the plant. These relatively simple bases can be subjected to further intramolecular oxi- dative attack to elaborate more complex macrocycles linked between the isoquinoline moieties of the molecule. In the course of these studies we were indeed able to detect in vitro turnover of the (R,S)-dimer berbamunine to (R,S)-aromoline, incorporating an additional linkage between C-7-OH and C- 8’.

Closer inspection of the substrate specificities of the puri- fied enzyme revealed no turnover of either its immediate products to the more oxidized dimers as mentioned above (Table 111), inferring a related but different enzyme that catalyzes the generation of (R,S)-aromoline and its structural isomers.

As illustrated in Table 111, a single enzyme is responsible for the metabolism of three closely related substrates to afford three distinct dimeric products, (R,S)-2’-norberbamunine, (R,S)-berbamunine, and (R,R)-guattegaumerine. Earlier ob- servations (10) have assumed that the N-nor dimers are products of a specific N-demethylation reaction of the more

830 Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids

I I IRl-+(Sl-NMecoc~aurine 2) 07

FIG. 8. Postulated catalytic mech-

of chiral benzyltetrahydroisoquino- anism of oxidative phenol coupling

line alkaloids involving two one- electron oxidation steps..

abundant bis-N-methylcoclaurines. However, the results pre- sented here show that neither N-demethylation nor N-meth- ylation of biscoclaurines governs the elaboration of the N-nor bases. Their biogenesis is therefore solely dependent on the availability of the (S)-enantiomer, accepted as either the secondary or tertiary base in the coupling reaction. This result infers binding of a competitive nature on the prime site of the enzyme. The condensation of two (R)-N-methylcoclaurine units supports this view, even though the turnover rate to the (R,S)-dimer is clearly higher (Table 111). Restriction for only C-1’s stereochemistry is therefore not evident.

Of the total of the single diphenyl ether-coupled dimers in the crude Berberis extract, (R,S)-berbamunine represents 84- 86%, the rest being attributable to (R,R)-guattegaumerine. However, incubations with equimolar concentrations of (R) - and (S)-N-methylcoclaurine in the reconstituted system led to the formation of only 3-6% of the (R,R)-dimer. Considering two competitive routes, one directed toward construction of the protoberberine and the other toward condensation to the BBIQ alkaloids, it is not at all surprising that the metabolic pool of (S)-N-methylcoclaurine is much smaller than of its (R)-counterpart, due to drainage of the former en route to the tetraoxygenated base (S)-reticuline (35,39,40). Based on the in vivo relationship of the two enantiomorphs and the kinetics of the purified enzyme we can estimate a ratio of approxi- mately 0.5:l for the two major metabolic pools (5‘)- and ( R ) - N-methylcoclaurine, respectively.

Cytochrome P-450 enzymes are generally defined as mon- ooxygenases, in that in most cases an oxygen atom is inserted in a site-specific manner into a lipophilic substrate. The results presented here show that tail-to-tail coupling between the benzylic moieties does not involve a hydroxylation reac- tion, implicating cytochrome P-450 as a terminal electron sink rather than an oxygenase. In view of the results presented here and recent mechanistic concepts (41-43), that also at- tribute a far greater role as hitherto conceived to P-450 oxidases as catalysts of dehydrogenation reactions (44), we propose a biradical mechanism. A far-sighted hypothesis sug- gesting radical pairing was presented by Barton and Cohen as far back as 1957 (1): such brilliant schemes could correctly predict the structures of natural products in plants.

As predicted previously (1,45), benzylic or C-ring couplings are the first to commence to afford the biogenetically most simple dimers of, for example, the berbamunine type. The catalytic site of the terminal hemoprotein must therefore feature two substrate binding sites to accommodate each

chiral BTIQ substrate, the iron-protoporphyrin (1X)-chro- mophore most likely in close proximity to both substrates. The hydrophobic environment, defined also by phospholipids as inherent membrane constituents, probably interacts with the substrates and imparts a certain degree of specificity. In this context, the varying substrate turnover rates of the crude microsomal preparation and the purified reconstituted system should be kept in mind (see Table 111).

An important question arising when contemplating the mechanistic mode is the nature of the reactive species. As portrayed hypothetically (Fig. 8), the first steps of the reaction are reminiscent of the typical P-450 concept (41). Several resonance forms are possible for the oxidant, and precise electron distribution has not yet been fully clarified (41-43). However, H- abstraction probably takes place sequentially, with the oxoiron(1V) porphyrin cation radical removing one electron (hydrogen) from one of the monomers to afford a stabilized BTIQ phenoxy radical. The ferry1 oxidant then oxidizes the second substrate, presumably at the same site on the molecule, generating a second radical with concomitant irreversible radical combination and removal of the reduced hydroxy function as water. The stoichiometry of the reaction could therefore be written as follows.

2ArOH + NADPH + H+ + 0, + ArO

- ArOH + NADP+ + 2H20 (Eq. 1)

A comparable radical dehydrogenation process has already been described for P-450 xenobiotic modification of parace- tamol, leading to the formation of an imino quinone in two one-electron oxidation steps (46). As already proposed, the active oxygen species of the P-450 cycle are reminiscent (not with respect to oxidation potential) of the active intermediates in the peroxidase cycle HRP compound I (Fe(1V)O-porphyrin cation radical) and horseradish peroxidase compound I1 (neu- tral oxoiron(1V) complex). Some cytochrome P-450 oxidases do indeed feature peroxidase activity, suggesting a strong analogy between the two catalytic cycles (47).

The results presented here are best explained by a biradical coupling mechanism. However, one cannot exclude partici- pation of cationic (i.e. two electron attack of only one sub- strate) or disproportionation/radical recombination mecha- nisms.

The true nature of the radical intermediates can only be elucidated with the aid of ESR experiments, limited above all by the low P-450 content in the cells. The work presented here proposes that P-450 biradical catalyzed mechanisms are

Oxidative Phenol Coupling to Bisbenzylisoquinoline Alkaloids 831

not restricted to xenobiotic models (46), but also govern highly regio- and stereoselective processes in the construction of complex natural compounds. Finally, the relatively stereo- and substrate-unspecific prime enzyme pocket at the active site enables the oxidase to elaborate a far greater diversity of structural types. This idea should be kept in mind when considering exceptions to biogenetic and chemosystematic rules that are based largely only on the stereochemistry of the BBIQ alkaloids (48).

14.

15.

16. 17. 18.

20. 19.

21. 22.

23.

Kametani, T., Nemoto, H., Kohari, T., and Takano, S. (1970) J . Heterocycl.

Zenk, M. H., Gerardy, R., and Stadler, R. (1989) J. Chem. Soc. Chem.

Rueffer, M., and Zenk, M. H. (1987) Tetrahedron Lett. 28,5307-5310 Tanahashi, T., and Zenk, M. H. (1988) Tetrahedron Lett. 29,5625-5628 Bauer, W., and Zenk, M. H. (1989) Tetrahedron Lett. 30,5257-5260 Bauer, W., and Zenk, M. H. (1990) Phytochemistry 30,2953-2961 Hi ashi, K., Ikeuchi, K., Obara, M., Karasaki, Y., Hirano, H., and Gotoh,

O'Keefe, D. P., and Leto, K. J. (1989) Plant Physiol. 89, 1141-1149 Gabriac, B., Werck-Reichart, D., Teutsch, H., and Durst, F. (1991) Arch.

K y k , G., and Grisehach, H. (1989) Arch. Biochem. Biophys. 273 , 543-

Chem. 7 , 181-186

Commun. 22,1725-1727

8. (1985) Agric. Bid. Chem. 49,2399-2405

Biochem. BLophys. 288,302-309

and

Acknowledgments-We thank M. Wislicenus for excellent techni- 25. Linsmaier, E. M., and Skoog, F. (1965) Physiol. Plant. 18 , 100-127 24. Song, W.-C., and Brash, A. R. (1991) Science 253 , 781-783

cal assistance and L. Kammerer for the kind gift of E. californica i;: ~ ~ ~ ~ a M ~ ~ a ~ ~ t ~ ~ ~ ~ ~ ~ ~ ~ Y ~ ~ l , 5 ~ ~ ~ ~ 0 ~ ~ ~ , 2370-2378

mann for recording the chemical and direct chemical ionization mass 548 spectra.

NADPH-cYtochrome p-450 P3hctase. We also thank Dr. 0. Selig- 28. Ernster, L., Siekevitz, P., and Palade, G. E. (1962) J . Cell Biol. 15 , 541-

29. Bradford, M. M. (1976) Anal. Biochem. 72,248-254 30. Laemmli, U. K. (1970) Nature 227,680-685 31. Blum, H., Beier, H., and Gross, H. J. (1987) Electrophoresis 8,93-99

REFERENCES 32. Kratzl, K., and Billek, G. (1951) Monatsch. Chem. 82,568-571 1. Barton, D. H. R., and Cohen, T. (1957) in Festschrift Prof. Dr. Arthur Stoll, 33. Baxter, I.3 L. T., and Swan, G. A. (1965) J. 3645-3660

2. Guha, K. p.7 Mukherjee, B., and Mukherjee, R. (1979) J. Nat. Prod. 429 1- 35, Stadler, R,, and zenk, M. H, (1990) Liebigs Chem,, 555-562

3. Schiff, P. L., Jr. (1983) J. Nat. Prod. (Lloydia) 46 , 1-43 4. Schiff, P. L., Jr. (1987) J. Nat. Prod. 50,529-599

37. Roos, P. H. (1990) J. Chrom. 521,251-265

5. Schiff, P. L., Jr. (1991) J . Nat. Prod. 54,645-749 38. Benveniste, I., Lesot, A,, Hasenfratz, M., and Durst, F. (1989) Biochem. J .

6. Barton, D. H. R., Kirby, G . W., and Wiechers, A. (1966) J. Chem. Soc. 39. Stadler, R., Kutchan, T. M., Loeffler, S., Nagakura, N., Cassels, B. K., and

7. Barton, D. H. R., Kirby, G. W., and Wiechers, A. (1966) J . Chem. Soc. C 40. Stadler, R., Kutchan, T . M., and Zenk, M. H. (1989) Phytochemistry 28,

8. Bhakuni, D. S., Singh, A. N., and Jain, S. (1978) J. Chem. SOC. Perkin 41. Woggon, W. D. (1988) Nachr. Chem. Tech. Lab. 36,890-895

9. Bhakuni, D. S., Jain, S., and Singh, A. N. (1980) Phytochemistry 19,2347- 43. Guengerich, F. p. (19g1) J . 266* 1o019-10022

pp, 117-144, Birkhauser, Basel 34. Gopinath, K. W., Govindachari, T. R., and Viswanathan, N. (1959) Chem. Ber. 9 2 , 1657-1661 84 36. Kastner, M., and Schulz, T. (1987) J . Chrom. 397,153-163

259,847-853

Chem. Commun. 9 , 266-267 Zenk, M. H. (1987) Tetrahedron Lett. 28,1251-1254

Org. Chem., 2313-2319 1083-1086

Trans I, 380-383

2350

istry 27,2557-2565

1005-1008

Chem., 9-12

Chem., 2770-2773 Tetrahedron 40, 1975-1982

42. Guengerich, F. P. (1990) Crit. Reu. Biochem. Mol. Biol. 25,97-153

44. Ortiz de Montellano, P. R. (1989) Trends Pharmacol. Sci. 10,354-359 10, Stadler, R,, Loeffler, s., tassels, B, K,, and Zenk, M. H, (1988) phytochem. 45. Battersby, A. R. (1967) in Oxiddue CouPli% OfPhenob (Taylor, W. 1.. and

Battersby, A. R., eds) p 119-165, Marcel Dekker Inc., New York 11. Cassels, B. K., Breitmaier, E., and Zenk, M. H. (1987) Phytochemistry 26, 46' Dahlin' D' "' Miwa' G' ':' Lu, A. Y. H., and Nelson, S. D. (1984) Proc.

Natl. Acad. Sci. U. S. A. 81,1327-1331

12. Kametani, T., Takano, S., and Kobari, T. (1969) J. Chem. Soc. C Org. 47. Marnett, L. J., Weller, P., and Battista, J. R. (1986) in Cytochrome P-450,

Structure, Mechanism, and Biochemistry (Ortiz de Montellano, P. R., ed)

13. Kametani, T., Takano, S., and Kobari, T. (1969) J . Chem. Soc. C 0%. 48. Cuinaudeau, H., Freyer, A. J., Shamma, M., and Baser, K. H. C. (1984) pp. 29-76, Plenum Press, New York