Stereospecificity of (+)-Pinoresin01 and (+)-Lariciresinol Reductases ...

18 downloads 45 Views 958KB Size Report
Jun 7, 1993 - teniposide derivatives) is used for treating a variety of cancers. (l), such as testicular cancer and acute lymphocytic leukemia. The “mammalian” ...
THEJOURNAL OF BIOWICAL CHEMISTRY

vel. 268,No. 36,Issue of December 25,pp. 27026-27033, 1993 Printed in U S A .

0 1993 by The American Society for Biochemistry and Molecular Biolom, Inc

Stereospecificity of (+)-Pinoresin01 and (+)-Lariciresinol Reductases from Forsythia intermedia* (Received for publication, June 7, 1993, and in revised form, August 26, 1993)

Alex Chu, Albena Dinkova, Laurence B. Davin, Diana L. Bedgar, and Norman G. Lewis$ From the Institute of Biological Chemistry, Washington State University, Pullman, Washington 99163-6340

PinoresinoVlariciresinolreductase catalyzes the first known example of a highly unusual benzylic ether reduction in plants; its mechanism of hydride t r a n s f e r is described. The enzyme was found in Forsythia i n t e r n diu and catalyzes the presumed regulatory branchpoints in the p a t h w a y leading to benzylaryltetrahydrofuran, dibenzylbutane, dibenzylbutyrolactone, and aryltetrahydronaphthalene lignans. Using [7,7’-2H2]pinoresinol and [7,7’-2H3]lariciresinol as substrates, the hydride transfers ofthe highly unusual reductase were demonstrated to be completely stereospecific (>go%). The incoming hydrides were found to take up the pro-R position at C-7’ (and/or C-7) in lariciresinol and secoisolariciresinol, t h e r e b y eliminating the possibility ofrandom hydride delivery to a planar quinone methide intermediate. As might be expected, the mode of hydride abstraction &om NADPH was also stereospecific: using [ 4 W “ ] and [4S-3H]NADPH, it w a s f o u n dthat only the 4 pro-R hydrogen was abstracted for enzymatic hydride transfer.

structural subfamilies (1).Among the most common (Fig. 1)are the dibenzylbutanes (e.g. secoisolariciresinol 11, the dibenzylbutyrolactones (e.g., matairesinol 2, arctigenin 3),the furofurans (e.g. pinoresinol4, medioresinol 5), the arylnaphthalenes ( e g . chinensin 6) andthe aryltetrahydronaphthalenes (e.g. podophyllotoxin 7). In terms of biosynthesis, the enzymatic steps leading toformation of the CsC3 monomeric units (monolignols) have been firmly established andrecently reviewed (17, 18). Surprisingly, the subsequent enzymatic transformations involved in monomeric coupling and post-coupling modifications are only now being delineated (16, 19-25). Thus, it has been established thatinsoluble enzymepreparations from Forsythia sp. catalyze the formation of (+)-pinoresin014 a from two achiral molecules of E-coniferyl alcohol 8 (22). (More recent studiestargeted towardpurification of this enzymehave revealed that O2 is required as a cofactor.)l Once formed, (+)-pinoresin01 4 a undergoes highly enantiospecific NADPHdependent reductions to first afford the benzylaryltetrahydrofuran lignan, (+)-lariciresinol9a(24, 25) and then thedibenzylbutanelignan, (-)-secoisolariciresinol l b (23-25); toour knowledge, these are the first examples of benzylic ether reductions in plants. (-)-Secoisolariciresinol l b can next be steLignans are a structurally diverse familyof phenylpropanoid reospecifically oxidized by a NADP+-dependent dehydrogenase metabolites found throughout the plant kingdom, principally in to give the (-1 antipode of matairesinol 2 (20, 211, thereby woody gymnosperms and angiosperms; theyare mostfreproviding entry into the dibenzylbutyrolactone subgroup (16). quently found as dimers (l),although higher oligomers exist Matairesinol2 hasalso been proposed to serve asa precursor of (2). Based on their known properties,various physiological aryltetrahydronaphthalene lignans, such aspodophyllotoxin 7 roles in plants have been proposed. These include antioxidant (3, 41, bactericidal (5), fungicidal (6), antiviral (71, insect anti- (26). Thus, the sequential reductive fission of (+)-pinoresin014 a to feedant (81, phytotoxic (to competing species) @),and perhaps give (+)-lariciresinolga and(-)-secoisolariciresinol l b permits even cytokinin-like (10) functions. Although direct evidence is a rational and hitherto unexpected entry into the various liglacking, it has been proposed that lignans are involved in lignan subgroups and, indeed,may represent key regulatory nification (11).Many lignans exhibit pharmacologically imporpoints in lignan pathway branching. Moreover, these highly tant effects: for example, podophyllotoxin (as itsetoposide and unusual reductive steps raise intriguing questions regarding teniposide derivatives) is used for treating a variety of cancers enzyme-catalyzed hydride transfer mechanisms. Using (+)-pi(l), such as testicular cancer and acute lymphocytic leukemia. noresinol 4a as an example (Fig. 2), three possible hydride The “mammalian” lignans, enterolactone and enterodiol, aptransfer scenarios can be envisaged, resulting in either the parently reduce incidence rates of breast and prostratecancers original C-Hbond geometry of the benzylic C-7’ carbon being in humans on high fiber diets (12) by modulating steroidal retained,inverted, or undergoing“racemization” as shown. hormone synthesis(13); both lignansare presumed tobe (Note that this is the case, regardless whether a pentacoordiformed in the intestine via metabolism of ingested plant matenate SN2 or a quinone methide enzyme-bound transition state rials (14). Lastly, (-)-arctigenin and (-)-trachelogenin strongly is involved.) This report describes delineation of the stereoinhibit in vitro replication of the human immunodeficiency vichemical course of these unusual benzylic ether reductions, rus (HIV-1) ( E ) , and like the mammalian lignans, both are affording (+)-lariciresinol9a and(-1-secoisolariciresinol lb, represumed to be formed from secoisolariciresinol 1 and spectively. matairesinol 2 (14, 16). Lignans can be conveniently divided into several different EXPERIMENTAL.PROCEDURES Plant Materials

* This work

was supported by National Science Foundation Grant MCB9219586 andUnited States Department of AgricultureGrant 91371036638. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertzsement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed. Tel.: 509-335-2682;

Fax: 509-335-7643.

Forsythia intermedia (var. Lynwood Gold)plants, obtained from Bailey’s Nursery (St. Paul, MN), were maintained in Washington State

University greenhouse facilities. P. Pad, L. B. Davin, D. L. Bedgar, and N. G. Lewis, unpublished results.

27026

27027

Stereospecificity of Benzylic Ether Reductases

CH,O

Q OH

OH

OCH,

bCH,

OH

OCH,

OH

I R'

R' 10a

10b

11

R' = Ribose-diphosphate-ribose-adenine FIG.

1. common lignan structural classes, E-coniferyl alcohol, and tsHINADP'~~PH.

Inversion n

OCH3

Non-stereospecific

Ar = 4-Hydroxy-3-methoxyphenyl

@

t

C-7/C-7' protons of (+)-pinoresin01 and (+)-lariciresinol

(+)-Lariciresinol9a

FIG.2. Possible stereochemical consequences during hydride transfer by (+)-pinoresin01reductase from F. intermedia. Designations HR and Hs refer to pro-R and pro-S, respectively. Materials Sodium borodeuteride (98 atom % 2H)and MeOZH(99.5+ atom % 2H) were purchased from Aldrichand deuterium gas (99.8 atom % 2H) from Isotec Inc. ~-[l-~H]Glucose (222 GBq mmol") was obtained from New England Nuclear. Yeast Glc-6-Pdehydrogenase (Type lX,6.2 g a t m g ' protein) and yeast hexokinase (TypeF300,4.2 g a t mg" protein) were purchased from Sigma and dihydrofolate reductase (Lactobacillus casei, 9.3 pkat m g l protein) from Biopure Co. Znstrumentation a n d Chromatography Materials Silica gel thin layer and column chromatography were performed on Kieselgel 60 FZS4(0.25 mm) and Silica Gel-60 (EM Science, 230-400

mesh), respectively. DEAE-cellulose (fine mesh) was purchased from Sigma, and AfFi-Gel Blue Gel. (100-200 mesh) from Bio-Rad. Allsolvents and chemicals used were reagent or HPLC2grade; tetrahydrofuran was distilled over sodiunhenzophenone immediately prior t o use. 'H nuclear magnetic resonance spectra (300 and 500 MHz) were recorded on Briiker AMX300 and Varian VXR500S spectrometers, respectively, using C2HCI3as solvent with chemical S h i h (6 ppm) reported downfield from tetramethylsilane (internal standard). IR and W spectra The abbreviations used are: HPLC, high performance liquid chromatography; ppm, partdmillion; 2D-NOESY, two-dimensional nuclear Overhauser effect;2D-COSY, two-dimensional correlation spectroscopy.

27028

Stereospecificity of Benzylic Ether Reductases

we e obtained on Perkin-Elmer 1720-X Infra-red Fourier Transform an Lambda 6 W M S spectrometers, respectively. Mass spectra were recorded on a VG 7070E spectrometer (ionizing voltage 70eV).All melting points are uncorrected. High performance liquid chromatography was performed as described (23). HPLC separations of lignans were ca+d out using either reversed-phase (Waters, Nova-pak C18, 150 x 3.9 mm inner diameter)or chiral (Daicel, Chiralcel OC, or Chiralcel OD, 240 x 4.6 mm inner diameter) columns, with detection a t 280 nm. For HPLC separations, the EtOH was denatured with 2-PrOH (5%, Aldrich), and hexanes contained >95% n-hexane. All HPLC samples were filtered (ACRO LC3A disposable filter, 0.45 pm, Gelman Science) prior t o analysis. Reversed-phase HPLC elution conditions for enzyme reaction mixtures containing secoisolariciresinol 1, lariciresinol 9, and pinoresinol4 consisted of MeOW3%AcOH in H,O (30:70) with a flow rate of 0.4 ml min". Chiral column HPLC separations of (+)- and (-)-seeoisolariciresinols la and lb employed a Chiralcel OD column eluted with EtOWhexanes (30:70) a t a flow rate of 0.5 ml min-', whereas (+)and (-1-pinoresinols 4a and 4butilized EtOWhexanes (50:50) as eluant, a t a flow rate of 0.8 ml min-'. Separation of (+)- and (-)-lariciresinols Sa and 9b was achieved using a Chiralcel OC column eluted with EtOW hexanes (8020)a t a flow rate of 0.2 ml min". Purities of [4S3HlNADPH 10a and [4R-3HlNADPHlob were individually determined using reversed-phase HPLC with MeOWO.l M potassium phosphate buffer, pH 7.1 (3:97) a s eluant, a t a flow rate of 0.3 ml min" with simultaneous detection a t 260 and 340 nm. Radioactive samples were analyzed in Ecolume (ICN) and measured using a liquid scintillation counter (Packard, Tri Carb 2000CA).

i

by adding acetone (0.5 ml), diluted with EtOAc (50 ml), washed with water (50 ml), and saturated NaCl (50 ml), then dried (Na2S0,), filtered, and evaporated to dryness in uacuo to give a foam (1.05 g). This was reconstituted in a minimum amount of EtOAc and applied to a silica gel column (2 x 30 cm) eluted with EtOAdhexanes(12). Fractions containing the desired products were combined and evaporated to dryness in uacuo to give the diastereomeric mixture of alcohols 14d14b (673 mg), which were converted as described (25, 28) into (+[7,7'ZH21pinoresinols4 d 4 b (161.8 mg) in 37.8% overall yield. 'H N M R (500 MHz) (C2HC13):6 3.1 (ZH, m, C8H + C8,H),3.88 (ZH, dd, J8,gp= 4.0 Hz, J g u , g p = 9.0 HZ, CgHp + CyHp), 3.9 (6H, S , 2 X OCHs), 4.30 (2H, dd, Js,g, = 6.5 Hz, Jgu.gp = 9.0 Hz, CgHa + Cg.Ha), 5.64 (ZH, brs, OH), 6.82(ZH, dd, J , = 1.5 Hz, J2 = 8.0 Hz, C6H + c,.H), 6.89 (4H, m, CZH+ C2.H+ C5H + C,.H); W AZHnm: 231,280;MS m l z (%): 360 (M++ 2,51.6), 359 (MI + 1,6.6), 358 (M+,3.0), 345 (1.5),329 (3.5),328 (4.6),235(2.6),222 (2.3), 221 (1.9), 207 (10.3), 206 (13.2), 205 (14.8), 164 (27.4), 153 (42.01, 152 High resolution MS: (24.9), 151 (1001, 138 (59.11, 137 (14.6), 125 (20.8); calculated for CZOH~&2H,,360.1542; found, 360.1544. (+)-[7,7'-2H~Lariciresinols 9 d 9 6 T o a stirred solution of (2)-[7,7'2Hzlpinoresinols 4 d 4 b (100 mg, 0.294 mmol) in a mixture of EtOAc (5 ml) and Me02H (5ml) was addedPd-C (lo%, 100 mg), with the resulting suspension allowed to stir under deuterium gaswhile monitoring the reaction progress by thin layer chromatography (25). After stimng for 5 h at room temperature, the catalyst was removed by filtration, washed with MeOH (10 ml), with the MeOH solubles combined and evaporated to dryness in vacuo. The resulting foam (105 mg) was reconstituted in a minimum amount of EtOAc, applied to a silica gel column (2 x 30 cm), and eluted with EtOAdhexanes (1:l). Fractions [4S3HI and [4R-3HlNADPHI O d l O b containing the desired product were combined and evaporated to dry9d9b 132 mg, 32%) as ness in uacuo to afford ~~~-[7,7'-2H311arieiresinols These were enzymatically synthesized as described by Moran et al. an amorphous powder. 'H NMR (300 MHz) (C2HC13):6 2.39 (lH, q, 58.8. (27) with the following modifications: ~ - [ l - ~ H l G (240 l c pl, 8.75 MBq) = J8.,9.a = J8.,9.p = 7.0 Hz, C8H), 2.70 (lH, m, C8,H),3.74 (lH, dd, J8..g,a was phosphorylated with hexokinase (16.7 nkat) in the presence ofATP = 6.2HZ,J g , , , g , p = 8.7 HZ, Cg,Hp), 3.78 ( l H , dd, J 8 . g S = 7.0 HZ,JgR.93 = (1 pmol) and MgC1, (1 pmol) to give [1-3HlGlc-6-P(5.26 MBq, 60.1% 10.1 HZ, CgHS), 3.84 (3H, S , OCH,), 3.88 (3H, S , OCHs), 3.92 (1H, dd, radiochemical yield). This was oxidized with Glc-6-P dehydrogenase J 8 , g R = 7.2 HZ,J ~ R =,10.1 ~ sHZ, CgHR), 4.05 (lH, dd, J8',9'.., = 7.0 HZ, (16.7 nkat) in the presence of NADP+(50 nmol) to afford 6-phosphoglu- J9.u,9.p = 8.7 Hz, Cg.Ha), 5.71 (lH, brs, OH), 5.89 ( l H , brs, OH), 6.53conate and [4S3H1NADPH 10a (4.45 MBq, 50.9% rahochemical yield). 6.69(2H,m,ArH),6.78(1H,dd,J1=1.5Hz,Jz=8.0Hz,ArH),6.81-6.86 Formation of [4-3HlNADP+11 (4.31 MBq, 49.3% radiochemical yield) (3H, m, ArH); W A L O H ) nm: 229,281; MS m l z (%): 363 (M* + 3,34.8), was attained by incubating 10a with dihydrofolate (160 nmol) and 362 (M*+ 2,8.9), 361 (M++ 1,0.9), 360 (M*, 0.1),348 12.5),345 (6.1), 332 dihydrofolate reductase (1.67 nkat), further incubation of which with (1.5), 238 (6.81,153 (14.01, 151 (37.2), 139 (loo), 138 (39.91, 137 (16.71; Glc-6-P dehydrogenase (16.7 nkat) and Glc-6-P (10 pmol) gave 14R- High resolution MS: calculated for C Z ~ H ~ ~ 363.1761; O ~ ~ H ~ found, , ,H]NADPH 10b (3.80 MBq, 43.4% radiochemical yield). Enzymatic in- 363.1784. cubations were performed a t 30 "C for 30 min. After each incubation, (+)-Lariciresinols9 a l 9 b T h e s e were synthesized as described (25). the NADPH or NADP+ form was purified using DEAE-cellulose mini'H NMR (500 MHz) (C2HC13):6 2.41 ( l H , m, C8H), 2.55 (1H, dd, J7.R.8' columns (0.3-ml bed volume) equilibrated in distilled HZO exactly a s = 10.8 Hz, J7.S,7.R = 13.5 Hz, C7XR), 2.73 (lH, m, C8.H), 2.92 (1H, dd, described (27). [4R-3H]NADPH lob was further purified by reversed- J7,s,,.= 5.2 HZ,J7.S.7.R = 13.5 HZ, C,,HS), 3.75 (IH, dd, Jw.9.p = 6.1 HZ, phase HPLC. Js.u.wa = 8.6 HZ, C,.H@),3.78 (lH, dd, J 8 . g ~= 6.5 HZ, J g R , g s = 10.6 HZ, CgHS), 3.87 (3H, S , OCH,), 3.89 (3H, S , OCH,), 3.92 dd, J s . 9 ~= 7.1 Chemical Syntheses HZ,J9R..,S = 10.6 HZ,CgHR), 4.05 (lH, dd, J 8 ' , 9 , u = 6.6 HZ,Jg,a,g.p = 8.6 J7,8= 6.6 Hz, C7H),5.51 (lH, brs, OH),5.59 (1H, (+)-[7,7'-2Hz]Pinoresinols4 d 4 b and (-c)-[7,7'-2H311ariciresinols HZ,C,,Ha), 4.79 (lH, d, 9 d S b were prepared as described (25, 28) from the known diastereo- brs, OH), 6.69 (lH, brs, C,.H), 6.70 (lH, m, C,.H), 6.81 (1H, dd,JI = 1.5 HZ,Jz = 8.3 HZ, CsH), 6.84 (lH, d,J = 8.5 HZ,C5.H), 6.87-6.88 (ZH, m, merie alcohols 12d12b with the following modifications. (+)-0,O'-Dibenzyl-7,7"dioxomatairesinols 1 3 a t l 3 b T o a solution of C,H + C,H); 2D-NOESY(500 MHz) (C2HC13)(500 ms mixing time): C8H diastereomeric alcohols 12d12b (3 g, 5.3 mmol) in CHzClz(100 ml) was - CZH, CwH,C g W , C 7 . m- Cz.H, CgH, C,H, C7,HS;CaH - CYH,CwH, added pyridinium chlorochromate (1.25 g, 5.8mmol), with the resulting C7.HS, CEH, CyHa; C7,HS - Cz,H, CG.H,C7,HR, C8.H; Cg.HP - CTH, suspension stirred for 18 h under N2 a t room temperature. The suspen- Cs.H, Cg.Ha; CgHS - C7H, C8H, CgHR; C3,OCHS- CyH, C6.H; C3OCH3 sion was diluted with diethyl ether (200 ml), filtered through silica gel - CZH;CgHR - C7H, CgHS; CyHa - CwH, Cg,HP; C7H - CzH, CGH, (5 g) with the filtrate evaporated to dryness in vacuo (3.2 g). The whole C7,HR, CgHS,CgHR;C2.H - Cs.OCH3,C7.HS, C7,HR, Cs.H, CyHP; C6.H-C7.HS,C,.HR, Ca.H, Cg.HP; CBH-C~H; CzH-C3OCH, C,H, CaH; was reconstituted in a minimum amount of EtOAc and applied to a MS mlz (%I: 360 (M+,38.8),345 (2.7),342(5.71.329 (131,236 (6.81,151 silica gel column (5 x 20 cm) eluted with EtOAdhexanes(1:3). Fractions (34.6), 137 (100). containing the desired product 13d13b were combined and evaporated to drynessin vacuo. Recrystallization from EtOAdhexanes afforded the diketones 13d13b(2.35 g, 78.3%). Melting point, 150-151 "C; 'H NMR Partial Purificationof Pinoresinol 1Lariciresinol Reductase (500 MHz) (C2HCI,): 6 3.9 (3H, S, OCH,), 3.92 (3H, S , OCH,), 4.39 (lH, All steps were carried out a t 4 "C. F intermedia stems (20-25 cm dd, J8..9.,= 6.8 Hz, J9.s.9.b = 9.0 Hz, Cg.Ha), 4.72 (lH, t,J8..9.b = J9.a.9.b long, 100 g) were cut into 1-cm sections, frozen (liquid N2)and pulver= 9.0 HZ, CyHb), 5.04 (lH, ddd, J a . , g . - = 6.8 HZ,Js.8. = 7.0 HZ,J a , , g . b = ized (Waring Blender). The resultingpowder was further homogenized 9.0 Hz, Ca,H), 5.12 (lH, d, J8,8.= 7.0 Hz, C8H), 5.22 (4H, brs, 2 x buffer (0.1 M, pH 7.0,300 OCH,Ph), 6.91 (lH, d, J = 8.3 Hz, ArH), 6.94 (lH, d, J = 8.6 HZ, ArH), in aWaring blender with potassium phosphate 7.267.54 (12H, m,ArH), 7.61 ( l H , d , J =2.0 Hz,ArH), 7.72 (lH, dd, Jl ml)containing 10 m~ dithiothreitol. The homogenate wasfiltered = 2.1 HZ, J , = 8.6 Hz, ArH); FT-IRcm-': 2983, 1771, 1733, 1668, through four layers of cheesecloth into a beaker containing PVPP slurry 1594, 1516; W AkoH nm: 231,281,312; MS m l z (%I: 566 (M*, 4.7),522 (6 g in 100 ml of potassium phosphate buffer, 0.1 M, pH 7.0) and stirred (28.1), 506 (2.2),430 (9.3), 416 (1.8), 325(17.4), 241 (81.5), 151 (9.2), 91 for 20 min. The filtrate was centrifuged (12,000 x g, 10 min) and the (100); High resolution MS: calculated for C3,H3008, 566.1941; found, resulting supernatant fractionated with (NH,),SO,. Proteins precipitating between 40 and 60% saturation were recovered by centrifugation 566.1960. (~)-[7,7'-2H~Pinoresinols 4a146To a solution of the diketones 13d (12,000 x g, 15 min). Thepellet was reconstituted in Tris-HC1 buffer (20 pH 8.0, 10 ml) containing 5 m~ dithiothreitol (buffer A) with ali13b (1g, 1.77 mmol) in MeOH (5 ml) a t 0 "C was added sodium boro- m, quots (2.5 ml) applied to four prepacked PD-10 columns, (Pharmacia deuteride (70mg, 1.67 mmol, 98 atom % ,H) with the resultingsolution LKB Biotechnology Inc., Sephadex G-25 medium) equilibrated with The reaction wasquenched stirred for 15 min at the same temperature.

StereospecificityReductases of Ether Benzylic buffer A. The eluted fraction (14.0 ml) was next applied to an Affi-Gel Blue Gel column (1 x 15 cm) equilibrated in buffer A (29, 30). ARer rinsing the column (bufferA, 20 ml), pinoresinolflariciresinol reductase was eluted with a NaCl gradient (0-5 M in 100 ml)in buffer A at a flow rate of 0.2 ml min". Proteins were detected at 280 nmand each fraction (1ml) was assayed for pinoresinol and lariciresinol reductase activities (see below). Active fractions were combined and used as the enzyme preparation for stereospecificity studies. Protein contents were determined using the method of Bradford (31)using Bio-Rad dyereagent and bovine y-globulins as standard.

27029

resulting foam was reconstituted in a minimum amount ofEtOAc, applied to a silica gel column (0.5 x 7 cm) and eluted with EtOAd hexanes (1:l). Fractions containing the desired product were combined to (-)-[7,7'Sand evaporated to dryness in Vacuo give 2H,]secoisolariciresinol l b (1.3 mg, 37.7%)as an amorphous powder. 'H NMR (300 MHz) (C2HC13):6 1.82 (2H, m, C8H + C,.H), 2.70 (1H, d, J 7 . R , s = 7.3 HZ,C,.HR), 3.54 (2H, dd, J1 = 4.2 HZ,J z = 11.7 HZ,CgH + C9.H),3.77-3.87 (8H, m, 2 x OCH,+ C9H + C,.H), 6.56(2H, d, J = 2.0 Hz, CzH + Cz.H), 6.68(2H, dd, J1 = 2.0 HZ,J z = 8.0 HZ, C,H + Cs,H), 6.78 (2H, d, J = 8.0 Hz,C5H + C,.H); MS mlz (%): 365 (M++ 3,15.2),364 (M+ + 2,2.0), 363 (M++ 1,0.7), 362 (MI, 0.3), 347 (9.3), 346(2.3), 345 (0.4), 139 (loo), 138 (99.3),137(18.3);High resolution MS: calculated for Cz,Hz3062H3, 365.1918;found, 365.1932. Enzymatic Formation of (-)-[7,7'S-2HJSecoisolariciresinol I&(+)[7,7'-2Hz]pinoresinols4d4b (5.2 m) were incubated with the enzyme preparation under identical conditions as described above to give (-1[7S,7'S-2Hz]secoisolariciresinoll b as an amorphous powder. 'H NMR (300 MHz,C2HC1,): S 1.82 (2H, m, C8H + C,.H), 2.70 (2H, d, J 7 R . 8 = J7.R,s. = 7.5 HZ,C 7 . m + C,HR), 3.54 (2H, dd, Jz 3.7 HZ,J1 = 11.1HZ, C9H + CyH), 3.80 (8H, m, 2 x OCH, + CgH + Cg.H),6.56 (2H, d, J = 1.7 HZ,CZH + Cz,H),6.61 (2H, dd, J1=1.7 HZ,Jz = 8.0 HZ,CsH + CG,H),6.78 (2H, d, J = 8.0 Hz, CSH+ C,.H). MS m /Z (%): 364 (M++ 2,19.9), 363 (M' + 1, 2.2), 362 (M*, 0.7), 346 (21.4), 345 (2.2), 344 (1.11, 138 (100)137 (26.6); High resolution MS: calculated for CZOHz,062Hz, 364.1855; found, 364.1863.

Enzyme Assays Pinoresinol Reductase-Pinoresinol reductase activity was assayed by monitoring the formation of [3H]lariciresinol9 and secoisolariciresinol I . Each assay, conductedin quadruplicate, consisted of (a)-pinoresinols 4d4b (5m~ in MeOH, 40pl), the partially purified enzyme preparation (200 pl) and buffer (0.1 M, Tris-HC1, pH 8.0, 120 4). The enzymatic reaction was initiated by addition of either [4WH]NADPH 10a (2.9 m, 7.27 kBq mmol" in 0.1 M potassium phosphate buffer, pH 7.1, 140 p1) or t4R-3HlNADPH lob (2.9 m ~ 6.79 , kBq mmol-l in 0.1 M potassium phosphate buffer, pH 7.1, 140pl). After 1 h of incubation at 30 "C with shaking, the reaction mixture was extracted with EtOAc (400 pl) containing (d-lariciresinols 9aBb (40 pg) and (z)-secoisolariciresinols l d l b (40 pg) as radiochemical camers. After centrifugation (13,800xg, 5 min), the EtOAc solubleswere removedand theextraction For each assay, the EtOAc procedure was repeated with EtOAc (400d). RESULTSANDDISCUSSION solubles were combined, evaporated to dryness in uacuo, and reconstituted in MeOW3% AcOHin HzO (50:50, 100pl) containing ferulic acid The stereospecificity of the pinoresinolllariciresinol reduc(15 pg) as internalstandard and filtered 10.45 pm). An aliquot (60 pl) tase-catalyzed hydride transfers was investigated from two from one assay was subjected to reversed-phase HPLC analysis with perspectives, namely (i) the stereochemical consequence(s) of fractions collected every minute (t = 0-30 min) and analyzed by liquid scintillation counting. The remaining assays were divided into 6O-pl hydride addition during enzymatic product formation and (ii) aliquots and individually subjected to reversed-phase HPLC, with (2)- the mode of hydride abstraction from the NADPH cofactor. In lariciresinols 9dSb and (a)-secoisolariciresinols l d l b (retention vol- order to investigatethese stereochemicalquestions,(+)-piume 5.9 and 6.7 ml, respectively)separately collected and evaporated to noresinoY(+)-lariciresinolreductase was partially purified (50dryness in uacuo. Each lignan was reconstituted in MeOH (100 PI) and fold) by affinity chromatography (Affi-GelBlue Gel eluted with filtered (0.45 pm), with an aliquot (60 4) subjected to chiral column HPLC analysis with eluant analyzed by both UV and liquid scintillation a NaCl gradient; see "Experimental Procedures"), and used as counting. Controls wereperformed using either denatured enzyme the enzyme preparation for this study. It should be noted, however, that both pinoresinol and lariciresinol reductase activities (boiled 96 "C, 10 min) or in the absence of (2)-pinoresinols4d4b. Lariciresinol Reductase-Lariciresinol reductase activity was aseluted coincidentally during the NaCl gradient. Indeed, this sayed by monitoring the formation of [3Hlsecoisolariciresinol1. Assays was also observed to be the case even when the reductase was were carried out exactly as described above, except that (a)-laricires- purified -1,000-fold via a combination of hydroxyapatite and inols 9dSb (5 m~ in MeOH, 40 pl) were used as substrates and (a)secoisolariciresinolsl d l b (40 pg) were added as radiochemical carri- affinity yellow chromatography3 With partially purified pinoresinolllariciresinol reductase in ers. 9a-To a solu- hand, attention was first directed toward defining the stereoEnzymatic Formation of (+)-[7,7'S-2HJLariciresino1 tion of (*)-[7,7'-2Hzlpinoresinols4d4b (5.2 m~ in MeOH, 4 ml) in a specificity of hydride transfer.As can be seen inFig. 3 A , the net mixture of buffer (20 m~ Tris-HC1,pH8.0, containing 5 m~ dithio- stereochemical change during the reductive cleavage of the threitol, 22 ml) and NADPH (20 m in HzO, 4 ml) was added the furan rings of 4a and 9a is destruction of the chiral centersat partially purified enzyme preparation (20 ml). Afterincubation a t 30 "C the of hydride with shaking for 1h, the reaction mixture was extracted with EtOAc (2 C-7/C-7'. But in order to address stereospecificity transfer, it was necessary to unambiguously distinguish the x 50 ml). The EtOAc solubles were combined, washed with saturated NaCl(50 ml), dried (Na,SO,), and evaporated to dryness in uacuo. The C-7/C-7' pro-R and pro-S methylenic protons in theenzymatic resulting foam was reconstituted in a minimum amount ofEtOAc, products,(+)-lariciresinol 9a and (-)-secoisolariciresinol lb. applied to a silica gel column (0.5 x 7 cm), and eluted with EtOAcf This was achieved via a combination of 'H, ZD-COSY, and hexanes (1:2) to give 150fractions (0.25 ml each). Fractions 49-98 were 2D-NOESY NMR spectroscopic experiments aided by molecucombined and evaporated to dryness in uucuo to give (+)-[7,7'Slar modeling. Using (2)-lariciresinols 9d9b as an example, the 2HzllariciresinolSa (1.4 mg, 13%)as an amorphous powder. 'H NMR = Js..g.p= 7.0 Hz, C8H), 'H and 2D-COSY NMR spectra revealed that the(2-7' methy(300 MHz) (C'HCl,): 6 2.39 (lH, q, J8,8.= J8,,g,m 2.51 (lH, d, J 7 . R , 8 . = 10.6 Hz, C7.HR),2.71 (1H, m, C,,H), 3.73(lH, dd, lenic protons were magnetically non-equivalent, with each inJ s , , g , p = 7.0 HZ, Jg.,+.p = 8.5 HZ, Cg,Hp), 3.76 (lH, dd, J8,9s = 6.5 HZ, dividually observed as a pair of doublet of doublets at 62.55 J g R . 9 s =9.5 HZ,CgHS), 3.86 (3H, 9, OCH,), 3.88 (3H, S, OCH,), 3.92 (lH, ( J ~=c10.8 Hz, J,,, = 13.5 Hz) and2.92 ( J ~==5.2 Hz, Jgem = dd, J 8 , 9 ~= 6.0 HZ,J9R.s = 9.5 HZ,C g H R ) , 4.04 (lH, dd, Js.,gs, = 7.0 HZ, 13.5 Hz),respectively (Fig. 4A). Based on the magnitudeof the Jg,...9.a = 8.5 Hz, Cg.Hu),5.47 (lH, brs, OH), 5.55 (lH, brs, OH), 6.68coupling constants (5.2 and 10.8 Hz) between the (2-7' 6 . 7 0 ~ 2 H , m , A r H ) , 6 . 6 8 ~ 1 H , d d , J l = 1 . 8 H z , J z ~ 6 . 2 H z , A r H ) , 6 . 7 7 ~vicinal .87 (3H, m,ArH); MS mlz (%): 362 (M++ 2,58.8), 361 (M++ 1,9.1), 360 (M+, methylene protons and the adjacent C-8' proton, two possible 3.6), 347 (2.3), 344 (3.0), 331 (0.71, 237 (7.1), 153 (14.9), 151 (43.4), 138 conformers A and B (as shown by their Newman projections in (loo), 137(29.4);High resolution MS: calculated for Cz,HZz062Hz, Fig. 3 B ) bearing the requisite dihedral anglescould be envis362.1698; found, 362.1709. aged (32). But inspection of Dreiding molecular models sugEnzymatic Formation of (-)-[7, 7'S-2H~Secoisolariciresinollb-To a solution containing buffer (20 m~ Tris-HC1, pH 8.0, and 5 rn &thio- gests thatonly conformer A was favorable, since conformer B is threitol, 43ml),NADPH (20 m~ in HzO, 7.6 ml) and (+)-[7,7'- apparently more sterically hindered via unfavorable interacC-9 hydroxy2H,llariciresinols SaBb (5.0 m in MeOH, 1.9 ml), was added the par- tions between the aromatic ring and the pendant After incubating at 30 " c methylene group (Fig. 3 B ) . That this wasindeed the case was tially purified enzyme preparation (38 d). with shaking for 1h, thereaction mixture was extracted with EtOAc (2 X 25 ml). The EtOAc solubles were combined, washed with saturated A. Dinkova, D. L. Bedgor,L. B. Davin, and N. E. Lewis,unpublished NaCl(25 ml), dried (NaZSO4),and evaporated to dryness in uacuo. The results.

27030

Stereospecificity of Benzylic Ether Reductases

(+)-Lnriciresinol9a (+)-Pinoresinol4a

JTR.8.

(-)-SecoisolaricireFinoll b

= 10.8 HZ

JTs.8. = 5.2Hz

(+)-Lariciresinol(9a) Conformer A Ar = 4-Hydroxy-3-methoxyphenyl

FIG.3. Enzymatic interconversions of selected E infermedia lignans and conformational analysis of (+)-lariciresinol9a.Enzymatic reactions catalyzed by (+)-pinoresinoV(+)-lariciresinolreductase (A) and Newman projection formulaeof two possible conformers of (+I-lariciresinol 9a ( B ) ,predicted by bond angle restrictions based on observed coupling constants and MM2 force field calculations for energy minimization.

established by a 2D-NOESY experiment (500 ms mixing time), presented the new problem of differentiating theincoming hywhich reveals correlations between protons spatially located dride from that of the existingC-7 or C-7' proton present in the within 5 b; from each other (33). Thus the detection of a cross- substrates 4a and 9a. One means to differentiate both protons peak between theC-2' proton (66.67) of the aromatic ring and was to employ selective deuterium labeling, either involving the C-9'p proton (63.75) established that both protons were in deuteride transferfrom NADP'H or viaselective deuteration at close proximity to one another, as in conformer A (data not the C-7 and C-7' sites of pinoresinol 4 and lariciresinol 9; the shown). By contrast, no cross-peak was observable between the second option was employed because of readily available synC-2' aromatic proton at 66.67 and the C-9 hydroxymethylene thetic methodology (25,28). Thus,p-ketolactones 13d13b were protons at 6 3.78 and 3.92, indicating that conformer B was not obtained from the known diastereomeric alcohols 12d12b folfavored. This result agreed with those obtained from a molecu- lowing treatment with pyridiniumchlorochromate; these were lar modeling routine,usingtheprogram Macromodel with then reduced with sodium borodeuteride to give the deuterated MM2 force field calculations for energy minimizations. This diols 14d14b (Fig. 5). Subsequent LiAlH4 reduction afforded again revealedunfavorable steric interactionsbetween the aro- tetraols 15dl5b, which weretreatedwithbromotrimethyl of (~)-O,O'-dibenzyl-[7,7'give a -4:lmixture matic ring and the hydroxymethylene group (C-9) in conformer silaneto B (data not shown)which were not evident with conformer A. 'Hz1pinoresinols 16d16b and (d-0,0'-dibenzyl-[7,7'-2H21Thus, based on the results from both 2D-NOESY experiments epipinoresinols(25).Debenzylation via hydrogenation over A was the preferred con- Pd-C afforded the required (~)-[7,7'-2H21pinoresinols 4d4b. and energy minimization calculations, former (Fig. 3B). With the conformation so established, it was Subsequent comparison of the 'H NMR and mass spectra of possible to assign theC-7' pro-R andC-7' pro-S protonson the deuterated 4d4b with its unlabeled analogue unequivocally basis of diaxial interactions with the C-7 protonviaa 2D- established the positions of deuteration within the molecule. NOESY experiment. Thus,a cross-peak betweenthe C-7 proton Thus, both 'H NMR spectra were essentially identical,except 4d4b a t 64.79 and a C-7' proton resonancea t 62.55, showed that both for the absenceof a 2H doublet at 64.74 in the deuterated were inclose proximity and that the latter corresponded to the sample; this resonance had previously been assigned to the C-7'pro-R proton. No cross-peaks were observed between the C-7/C-7' protons in unlabeled (d-pinoresinols 4d4b (25, 28). C-7 proton and the C-7' pro-S proton resonance at 62.92. Fol- Analysis of the mass spectraof both deuterated andunlabeled lowing a similar line of reasoning, the pair of doublet of dou- (5)-pinoresinols 4d4b confirmed and extended the 'H NMR 'H NMR spectrum of (+secoisolariciresinols l d l b findings. With unlabeled (+pinoresinols 4d4b, a molecular blets in the ion at ( m l z ) 358 was evident together with ion fragments at at 62.66 and 2.76 were assigned to the C-7pro-S and C-7pro-R ( m l z ) 206, 152, and 137 corresponding to [M-ArCHOI', [Arprotons, respectively. Having established a NMR spectroscopic method to distin- CHO]', and [ArCH21+,respectively, where Ar = 4-hydroxy-3guish the pro-S andpro-R protonsat C-7' of (+)-lariciresinol9a methoxyphenyl(23). By contrast, the deuterated analogue had and C-7E-7' of (-)-secoisolariciresinol lb, attention was di- a molecular ion a t ( m l z )360, corresponding to thepresence of rected to defining the stereochemistryof hydride transfer. This two deuterium atoms, with ion fragments at ( m l z ) 207, 153,

Stereospecificity of Benzylic Ether Reductases

B

3!0

219

218

211

21.6

2!5

214

PPm

FIG.4. Partial 'H N M R spectra of lariciresinol9 showing spectral regions for C-T', C-B', and C-8 proton resonances. Synthetic lariciresinols 9dSb (A) and enzymatically synthesized (+)-[7,7'S2H2]lariciresinol9a ( B ) ;assignments are based ondouble quantumCOSY and OD-NOESY NMR experiments.

and 138 due to [M-ArC2H01+, [ArC2HO]+, and [ArCH'H]'. Thus, it was unambiguously established that the (d-pinoresinols 4d4b were deuterated at C-7 and C-7'. In an analogous manner, (d-[7,7'-2H3]1ariciresinols9d9b were prepared from (~)-[7,7'-2H21pinoresinols 9d9b via catalytic deuterolysis over 10% Pd-C. Comparison of the 'H NMR and mass spectra of both the deuterated and unlabeled lariciresinols 9d9b again unequivocally revealed the sitesof deuteration. Thus, the'H NMR spectra were essentially identical, except for the absence of resonances in the deuteratedproduct a t 62.55, 2.92, and 4.79 corresponding to the C-7' pro-R, C-7' pro$, and C-7 protons, respectively. Analyses of the massspectra of both provided additional support to the NMR assignments; the unlabeled (%)-lariciresinols 9d9b had a molecular ion at ( m / z )360, together with ion fragments at ( r n l z ) 236, 152,and 137corresponding to [M-ArHI', [ArCHO]', and [ M H z ] + , respectively, where Ar = 4-hydroxy-3-methoxyphenyl, as before. By comparison, the deuterated analogue had a molecular ion at ( m/ z ) of 363, establishing the introduction of three deuterium atoms. That these deuterium atoms were a t positions 7 and 7' could be deduced from the ion fragments at ( m l z ) 238, 153,139, and 138corresponding to [M-Ar'H]', [ArC2HOl+,[ArC2H2]+,and [ArCH2H]+,respectively. With the required deuterated substrates in hand, stereothe chemical basis of the enzyme-mediated hydride transfer was 4d4b (0.42 mM) investigated. Thus, (~)-[7,7'-2H21pinoresinols were incubated for 1 h at 30 "C with the partially purified pinoresinolllariciresinol reductase from F. intermedia in the presence of 1.6 mM NADPH. The resulting lariciresinol 9 so formed was purified using flashsilica gel chromatography, and its 'H NMR and mass spectra, and chiral column HPLC profiles wererecorded and compared to thatof synthetic unlabeled (2)-lariciresinols 9d9b. Fig. 4, A and B, show the pertinent

27031

regions of the lH NMR spectra corresponding to the C-8, C-7' pro-R, C-8' and C-7' pro-S proton resonances for both the synthetic (Fig. 4A) and the enzymatically formed lariciresinol 9 (Fig. a). It can immediately be noted that in the enzymatic product, only the 1Hdoublet a t 62.51 ( J = 10.6 Hz) corresponding to the C-7'pro-R proton is observed, whereas theC-7' pr0-S proton signal at 62.92 is essentially absent (>99% reduction). Note also that the C-7'pro-R proton resonance at 62.51 is only absence of geminal coupling a doublet ( J = 10.6 Hz) due to the and that itschemical shift is moved upfield by 0.04 ppm; this shift is consistent with thereported shielding effects caused by replacing a proton with deuterium in a methylene or methyl group (34). Mass spectroscopic analyses of the enzymatic product further supported this finding; the enzymatically formed deuterated lariciresinol 9 had a molecular ion at ( m l z ) 362, corresponding to the presence of two deuterium atoms. That these deuterium atoms remained at C-7 and C-7' was established from the ion fragmentation pattern; thus, theenzymatic (deuterated) product 9 gave ion fragments at ( r n l z ) 237, 153, and 138 corresponding to [M-Ar2Hl+, [ArC2HOI+, and [ArCH2H]+ where Ar = 4-hydroxy-3-methoxyphenyl.Chiral HPLC column analysis of the deuteratedproduct revealed that only (+)-lariciresinol Sa was formed, in accordance with the known enantiospecificity of this transformation (25). Thus, it was concluded that the reductase-catalyzed hydride transfer was >99% stereospecific, where the incoming hydride exclusively took up the pro-R position with the existing carbondeuterium bond geometry at C-7' undergoing inversion to assume the pro-S position, c.e. the enzymatic productis (+)-[7,7'S2H2]lariciresinol 9a. The possibility that lariciresinol9reduction proceeded in an analogous mannerwasnext investigated. Thus, (+)-[7,7'2H3]lariciresinols 9d9b (0.11 m ~ were ) incubated for 1 h at 30 "C with the partiallypurified reductases in thepresence of 1.7 m~ NADPH. As before, the secoisolariciresinol 1 formed was purified, with its lH NMR and mass spectra and chiral HPLC profiles subsequently recorded and compared to authentic unlabeled (d-secoisolariciresinolsl d l b . The 'H NMR spectroscopic findings are shown in Fig. 6, A and B,where only the pertinent regions of the 'H NMR spectra corresponding to the C-7 pro-S, C-7 pro-R, andC-9 proton resonances are shown. As can be seen, the deuteratedenzymatic product 1 (Fig. 6B) displayed only a 1H doublet centered a t 62.70 ( J = 7.5 Hz), in contrast to the pair of doublets of doublets centered a t 62.66 and 2.76 corresponding to theC-7 pro-S and C-7 pro-R protons, respectively, in unlabeled (d-secoisolariciresinols l d l b (Fig. 6.4). Thus, given the small expected upfield proton resonance shift of -0.04-0.07 ppm due to deuterium substitution(341, it can again be proposed that hydride transfer is stereospecific (>99%) resulting in replacement of the C-7 pro-S proton by deuterium. The mass spectrum of the enzymatically synthesized (deuterated) secoisolariciresinol 1 confirmed the presence of three deuterium atoms, as evidenced by the molecular ion at ( r n l z ) 365. That these deuterium atoms remained at C-7 and C-7' positions was again established from the ion fragmentation patterns with ion fragments ( m / z )at 139 and 138 corresponding to [ArC2H21+and [ArCH2HI+,respectively, where Ar = 4-hydroxy-3-methoxyphenyl.Chiral HPLC analysis of the deuterated product revealed that only (-)-secoisolariciresinol l b synthesis had occurred, again in accordance with the known enantiospecificity of the conversion (24, 25). 4a reduction, the conThus, as for (+)-[7,7'-2H2]pinoresinol version of (+)-[7,7'-2H311ariciresinol9a into (-)-[7,7'-2H3]secoisolariciresinol l b was stereospecific. But, in the latter case, a degree of uncertainty remained regarding the assignment of the 62.70 resonance to the C-7 pro-R proton, since it relied exclusively upon a 0.06 ppm upfield chemical shift from

27032

Stereospecificity of Benzylic Ether Reductases

PCC 0

OCH,Ph

12a112b

bCH,Ph

OCH2Ph

13a113b

14ai14b

1

LiAlH,

I

t

OH

16d16b : R=CH2Ph

2. Hz, lO%Pd-C

PCC = Pyridinium chlorochromate BrTMSi = Bromotrimethylsilane FIG.5. Chemical synthesis of (+)-[7,7’-aH,1pinoresinols 4d4b and (+)-[7,7’-2H,]lariciresinols 9&b. The enzymatically synthesized, deuterated, (-)-secoisolariciresinol l b formed was purified and subjected to ‘H NMR and mass spectralanalyses. As can be seen inFig. 6C, a 2H doublet at 82.70 (J = 7.5Hz) was again evident in the ‘H NMR spectrum. Since it had already been established that the hydride transferred during conversion of (+)-pinoresin01 4 a into (+Ilariciresinol9a becomes the C-7 pro-R proton, and that the‘H NMR spectrum o f (-)-deuterated l b shows only a single 2H doublet, it follows that its subsequent conversion into (-)-secoisolariciresinol l b has proceeded in ananalogous manner, i.e. in both reduction steps, the incoming hydride takes up the pro-R position. Lastly, it was also instructive to determine whether the 4 pro-R or 4 pro-S hydrogen on the nicotinamide ring of NADPH wasabstractedduring hydride transfer. [4R-3H] and [4S3HlNADPH 10b and 10a were conveniently prepared by modification of the procedure by Moran et ai. (27). Each tritiated cofactor was individually incubated for 1 h at 30 “C with the partially purified reductases in thepresence of ( a )0.5 mM (2)pinoresinols 4d4b and ( b ) 0.5 mM (+)-lariciresinols 9d9b, ref I 1 I I I I I I I I 3.6 3.4 3.2 3.0 1.S 1.6 ppm spectively. (Note that following incubation, unlabeled (*)-larFIG.6.Partial ‘H N M R spectra of secoisolariciresinol 1 show- iciresinols 9d9b (40 pg) and(2)-secoisolariciresinolsl d l b (40 ing spectral regions for C-7 and C-9 proton resonances. A, syn- pg) were added as radiochemical carriers, with each lignan thetic (+secoisolariciresinols l d l b and enzymatically synthesized(-)- individually purified and resolved by chiral column HPLC.) [7,7’SZH,]secoisolariciresinol lb, obtainedfollowingincubation of The results obtained are summarized in Table I. With (*)-pipartially purified (+)-pinoresinoV(+)-lariciresinolreductase with E , ( d [7,7’-2H,]lariciresinols9d9b,and C, (+)-[7,7‘-2H21pinoresinols 4d4b, noresinols 4 d 4 b as substrates, radiolabeled lariciresinol9 was only formed (1.37 pkat mg” protein) when [4R-3HlNADPH respectively. lob was employed as cofactor. Moreover, chiral column HPLC that of the C-7pro-Rproton resonancein theunlabeled product analysis again revealed that only the (+)-enantiomer 9a was la. In order to unambiguously establish thestereospecificity of radiolabeled. The reduction of lariciresinol 9 to (-)-secoisolarthis last reductive transformation, (r)-[7,7’-2H21-pinoresinolsiciresinol l b proceeded in an analogous manner: radiolabeled 4 d 4 b (0.42 mM) were nextincubated for 1 h withthe secoisolariciresinol 1was only synthesized (2.19pkat mg-’ propinoresinoVlariciresino1 reductase from E intermedia as before. tein) when [4R-3H]NADPH 10b was added as a cofactor; chiral

27033

Stereospecificity of Benzylic Ether Reductases

TABLEI Stereospecificity of (+)-pinoresin01I(+)-lariciresinol reductase with [4S-3H] or [4R-3H]NADPH 10dlOb as cofactors (+)-PinoresinoU(+)-lariciresinolreductase was incubated separately with (d-pinoresinols 4d4b, (+I-lariciresinolsSdSb and [4R-3Hl or [4S3H1NADPH. Enzymatically formed products were purified and analysed by chiral column HPLC coupled with radiochemical detection. Substrates

Specific reductase activity

pkat mg" protein

(d-Pinoresinols4d4b [4S-3HlNADPHloa [4R-3H]NADPHlob

(+)-Lariciresinol9a 0.00

(-)-Lariciresinols SdSb [4WHINADPH 10a [4R-3H]NADPH10b

(+)-Secoisolariciresinoll a 0.00 0.00

1.37

1-1-Lariciresinol Sb 0.00 0.00 (-1-Secoisolariciresinol l b 0.00 2.19

7. Markkanen, T.,Makinen, M. L., Maunuksela, E., and Himanen, P. (1981) column HPLC chromatographic analysis established thatonly Drugs Exptl. Clin. Res. 7, 711-718 the (-)-antipode lb was enzymatically formed. Thus, both (+I8. Harmatha, J., and Nawrot, J. (1984) Biochem. Syst. Ecol. 12,95-98 pinoresinol and (+)-lariciresinol reductases abstract the 4pro-R 9. Elakovich, S . D., and Stevens, K. L. (1985) J. Chem. Ecol. 11,27-33 10. Binns,A. N., Chen, R. H., Wood,H. N., and Lynn, D.G . (1987)Proc. Natl. Acad. hydrogen from NADPH. This mode of hydride abstraction is Sci. U.S . A . 84,98&984 comparable in action to cinnamyl alcohol dehydrogenases (EC 11. Rahman, M. M. A,, Dewick, P. M., Jackson, D. E., and Lucas, J. A. (1990) 1.1.1.195)from Forsythia (35, 36) and soybean (GZycine max) Phytochemistry 29,1841-1846 (37), which catalyze the reduction of E-coniferaldehyde to E- 12. Adlercreutz, H. (1984) Gastroentorology 88,761-764 13. Adlercreutz, H. (1991) in Nutrition, Zbxicity and Cancer (Rowland, I. R., ed), coniferyl alcohol 8. But this mode of hydride abstraction is pp. 137-195, CRC Press, B o a Raton, FL opposite to that of the preceding step in monolignol synthesis 14. Bomello, S. P., Setchell, K.D. R., Axelson, M., and Lawson, A. M. (1985) J . Appl. Bacteriol. 58, 37-43 affordingE-coniferaldehyde, i.e. p-hydroxycinnamoyl C o A 15. Schroder, H.C., Merz, H., Steffen, R..Muller, W. E. G . ,Sann, P. S., T r u m m , S., NADP' oxidoreductase (EC 1.2.1.44) which catalyzes abstracSchulz, J., and Eich, E. (1990) Z. Naturforsch. 46c, 1215-1221 tion of the 4 pro-S hydrogen in Forsythia and soybean ( G . max) 16. Ozawa, S., Davin, L. B., and Lewis, N. G . (1993)Phytochemistry 32,643-652 17. Lewis, N. G., and Yamamoto, E. (1990) Annu. Rev. Plant Physiol. Plant Mol. during E-coniferaldehyde synthesis (38). B i d . 41, 45-96 In summary, the (+)-pinoresin01 and (+)-lariciresinol reduc- 18. Davin, L. B., and Lewis, N. G . (1992)Rec. Adv. Phytochem. 28,325375 tases function in a highly stereospecific manner. Both abstract 19. Umezawa, T.,Davin, L. B., Yamamoto,E., Kingston, D. G . I., and Lewis, N. G . (1990) J. Chem. Soc. Chem. Commun. 1405-1408 the 4 pro-R hydrogen of NADPH, in such a manner that the 20. Umezawa, T., Davin, L. B., and Lewis, N. G. (1990) Biochem. Biophys. Res. incoming hydride adopts thepro-R position (>99%)at C-7fC-7' Commun. 171, 1008-1014 in the lignan product. These findings, therefore, rule out the 21. Umezawa, T.,Davin. L. B., and Lewis, N. G. (1991)J . Biol. Chem. 266.1021010217 possibility of a SN1 mechanism involving a planar transition 22. Davin, L. B., Bedgar, D. L., Katayama, T., and Lewis, N. G . (1992)Phytochemstate with random hydride delivery from either side of the istry 31,386%3874 molecule (Fig. 2). The next phases of our research will be to 23. Katayama, T.,Davin, L. B., and Lewis, N. G . (1992)Phytochemistry 31,38753881 establish whether the enzyme-bound transition state is in ei24. Lewis, N. G., Davin, L. B., Katayama, T.,and Bedgar, D. L. (1992)BuU.Liaison ther the furano or quinone methide form, and the regulatory Groupe Polyphdnols 16, g a l 0 3 25. Katayama, T.,Davin, L. B.. Chu, A,, and Lewis, N. G. (1993) Phytochemistry roles of these enzymes in lignan biosynthesis. 33,581-591 26. Kamil, W. M., and Dewick, P. M. (1986) Phytochemistry 26,2093-2102 Acknowledgments-We thank Drs. Jaroslav Zajicek(Washington State University NMR Spectroscopy Center) forobtaining the 2D- 27. Moran, R. G . , Sartori, P., and Reich, V. (1984)A d . Biochem. 138, 19G.204 H., Nakatsuko, F., and Higuchi, T.(1982) Mokuzai Gakkaishi 28, NOESY spectrumof lariciresinol,William Siems for recordingthe mass 28. Fujimoto, 555-562 spectroscopy spectra and Susan Johns (Washington State University 29. Biellmann, J.3, Samama, J.-P., Branden, C. I., and Eklund, H. (1979)Eur. J. Visualization Analysis and Design in Molecular Sciences Customer Ser- Biochem. 102,107-110 vices) for performing MM2 calculations. 30. Prochaska, H. J., and Talalay, P. (1986) J. Biol. Chem. 261, 1372-1378

REFERENCES 1. Ayres, D. C.. and b i k e , J. D. (1990) Ligmnsr Chemical, Biological and Clinical Properties, 402 pages Cambridge University Press, Cambridge, United Kingdom 2. Sakakibara, A,, Sasaya, T.,Miki, K., and Takahashi, H. (1987) Holzforschung 41, 1-11 3. Faur4, M., Lissi, E., Torres, R., and Videla, L. A. (1990) Phytochemistry 29, 3773-3775 4. Osawa, T.,Nagata, M., Namiki, M., and Fukuda, Y. (1985)Agric. Biol. Chem. 49,3351-3352 5. Hatton, M., Hada, S.,Watahiki,A., Ihara, H., Shu, Y.-Z., Kakiuchi, N., Mizuno, T., and Namba, T.(1986) Chem. Pharm. Bull. 34,3885-3893 6. Belmares, H., Barrera, A,, Castillo, E., Ramos, L. F., Hernandez, F., and Hernandez, V. (1979) Ind. Eng. Chem. Prod. Res. D e v . 18, 220-226

31. Bradford, M.M. (1976) Anal. Chem. 72,248-254 32. Slessor, K. N., and Tracey A. S. (1971) Can. J. Chem. 49, 28742884 33. Kumar, A., Emst, R. R.. and Wuthrich, K. (1980) Biochem.Biophys. Res. Commun. 96, 1 4 34. Bovey, F. A. (1969)Nuclear MagneticResonance Spectroscopy, p. 86, Academic Press, New York 35. Mansell, R. L., Gross, G . G . Stdckigt, , J., Franke, H., and Zenk, M. H. (1974) Phytochemistry 13, 2427-2435 36. Klischies, M.,Stockigt, J., and Zenk, M. H. (1978) Phytochemistry 17, 15231525 Wengenmayer, H., and Wyrambik, D. (1977)in PyridineNucleo37. Grisebach, H., tide Dependent Dehydrogenases (Sund, H., ed), pp. 458471, de Gruyter, Berlin 38. Grisebach, H. (1981) in The Biochemistry of Plants: a Comprehensive Peatise (Conn, E. E.,ed) Vol. 7, pp. 457478, Academic Press, New York