Flavonoid 6-Hydroxylase from Soybean (Glycine maxL.), a Novel Plant P-450 Monooxygenase
2001; Elsevier BV; Volume: 276; Issue: 3 Linguagem: Inglês
10.1074/jbc.m006277200
ISSN1083-351X
AutoresA. O. Latunde‐Dada, Francisco Cabello‐Hurtado, Nikola Czittrich, Luc Didierjean, Christel R Schopfer, Norbert Hertkorn, Danièle Werck‐Reichhart, Jürgen Ebel,
Tópico(s)Phytochemicals and Antioxidant Activities
ResumoCytochrome P-450-dependent hydroxylases are typical enzymes for the modification of basic flavonoid skeletons. We show in this study that CYP71D9cDNA, previously isolated from elicitor-induced soybean (Glycine max L.) cells, codes for a protein with a novel hydroxylase activity. When heterologously expressed in yeast, this protein bound various flavonoids with high affinity (1.6 to 52 μm) and showed typical type I absorption spectra. These flavonoids were hydroxylated at position 6 of both resorcinol- and phloroglucinol-based A-rings. Flavonoid 6-hydroxylase (CYP71D9) catalyzed the conversion of flavanones more efficiently than flavones. Isoflavones were hardly hydroxylated. As soybean produces isoflavonoid constituents possessing 6,7-dihydroxy substitution patterns on ring A, the biosynthetic relationship of flavonoid 6-hydroxylase to isoflavonoid biosynthesis was investigated. Recombinant 2-hydroxyisoflavanone synthase (CYP93C1v2) efficiently used 6,7,4′-trihydroxyflavanone as substrate. For its structural identification, the chemically labile reaction product was converted to 6,7,4′-trihydroxyisoflavone by acid treatment. The structures of the final reaction products for both enzymes were confirmed by NMR and mass spectrometry. Our results strongly support the conclusion that, in soybean, the 6-hydroxylation of the A-ring occurs before the 1,2-aryl migration of the flavonoid B-ring during isoflavanone formation. This is the first identification of a flavonoid 6-hydroxylase cDNA from any plant species. Cytochrome P-450-dependent hydroxylases are typical enzymes for the modification of basic flavonoid skeletons. We show in this study that CYP71D9cDNA, previously isolated from elicitor-induced soybean (Glycine max L.) cells, codes for a protein with a novel hydroxylase activity. When heterologously expressed in yeast, this protein bound various flavonoids with high affinity (1.6 to 52 μm) and showed typical type I absorption spectra. These flavonoids were hydroxylated at position 6 of both resorcinol- and phloroglucinol-based A-rings. Flavonoid 6-hydroxylase (CYP71D9) catalyzed the conversion of flavanones more efficiently than flavones. Isoflavones were hardly hydroxylated. As soybean produces isoflavonoid constituents possessing 6,7-dihydroxy substitution patterns on ring A, the biosynthetic relationship of flavonoid 6-hydroxylase to isoflavonoid biosynthesis was investigated. Recombinant 2-hydroxyisoflavanone synthase (CYP93C1v2) efficiently used 6,7,4′-trihydroxyflavanone as substrate. For its structural identification, the chemically labile reaction product was converted to 6,7,4′-trihydroxyisoflavone by acid treatment. The structures of the final reaction products for both enzymes were confirmed by NMR and mass spectrometry. Our results strongly support the conclusion that, in soybean, the 6-hydroxylation of the A-ring occurs before the 1,2-aryl migration of the flavonoid B-ring during isoflavanone formation. This is the first identification of a flavonoid 6-hydroxylase cDNA from any plant species. 2-hydroxyisoflavanone synthase high performance liquid chromatography reverse-phase flavonoid 6-hydroxylase N-[2-hydroxy-1,1-bis (hydroxymethyl)ethyl]glycine Flavonoids are a diverse group of natural products that serve important roles in plants during growth, during development, and in defense against microorganisms and pests (1Harborne J.B. Grayer R.J. Harborne J.B. The Flavonoids: Advances in Research since 1986. Chapman and Hall, London1994: 589-618Google Scholar, 2Dixon R.A. Steele C.L. Trends Plant Sci. 1999; 4: 394-400Abstract Full Text Full Text PDF PubMed Scopus (593) Google Scholar). These compounds are synthesized from phenylpropanoid- and acetate-derived precursors through central pathways furnishing basic C6-C3-C6 flavonoid skeletons and, in addition, through a variety of reactions leading to a range of modified aglycones and subsequently to their glycosylated derivatives within each flavonoid class. Many of the enzymes of flavonoid biosynthesis have been extensively studied (3Heller W. Forkmann G,. Harborne J.B. The Flavonoids: Advances in Research since 1986. Chapman and Hall, London1994: 499-535Google Scholar), and recent molecular biological approaches have complemented biochemical methods in elucidating the mechanism and regulation of flavonoid biosynthesis (4Forkmann G. Harborne J.B. The Flavonoids: Advances in Research since 1986. Chapman and Hall, London1994: 537-564Google Scholar). Typical enzymes belonging to the complex branch pathways for the elaboration of flavonoid skeletons are cytochrome P-450-dependent hydroxylases (3Heller W. Forkmann G,. Harborne J.B. The Flavonoids: Advances in Research since 1986. Chapman and Hall, London1994: 499-535Google Scholar), such as flavonoid 3′-hydroxylase (5Brugliera F. Barri-Rewell G. Holton T.A. Mason J.G. Plant J. 1999; 19: 441-451Crossref PubMed Scopus (152) Google Scholar), flavonoid 3′,5′-hydroxylase (6Holton T.A. Brugliera F. Lester D.R. Tanaka Y. Hyland C.D. Menting J.G.T. Lu C.-Y. Farcy E. Stevenson T.W. Cornish E.C. Nature. 1993; 366: 276-279Crossref PubMed Scopus (292) Google Scholar), isoflavone 2′-hydroxylase (7Akashi T. Aoki T. Ayabe S. Biochem. Biophys. Res. Commun. 1998; 251: 67-70Crossref PubMed Scopus (73) Google Scholar), flavanone 2-hydroxylase (8Akashi T. Aoki T. Ayabe S. FEBS Lett. 1998; 431: 287-290Crossref PubMed Scopus (91) Google Scholar), flavone synthase II (9Martens S. Forkmann G. Plant J. 1999; 20: 611-618Crossref PubMed Scopus (92) Google Scholar, 10Akashi T. Fukuchi-Mizutani M. Aoki T. Ueyama Y. Yonekura- Sakakibara K. Tanaka Y. Kusumi T. Ayabe S. Plant Cell Physiol. 1999; 40: 1182-1186Crossref PubMed Scopus (80) Google Scholar), and 2-hydroxyisoflavanone synthase (2HIS)1 (11Steele C.L. Gijzen M. Qutob D. Dixon R.A. Arch. Biochem. Biophys. 1999; 367: 146-150Crossref PubMed Scopus (167) Google Scholar, 12Akashi T. Aoki T. Ayabe S. Plant Physiol. 1999; 121: 821-828Crossref PubMed Scopus (170) Google Scholar, 13Jung W., Yu, O. Lau S.-M.C. O'Keefe D.P.O. Odell J. Fader G. McGonigle B. Nat. Biotechnol. 2000; 18: 208-212Crossref PubMed Scopus (371) Google Scholar). Whereas flavonoid 3′-hydroxylase and flavonoid 3′,5′-hydroxylase are responsible for the formation of the 3′,5′-hydroxylation pattern of the flavonoid B-ring, hydroxylation of the isoflavone B-ring at the 2′ position (isoflavone 2′-hydroxylase) is one of the key reactions leading to pterocarpan structures. The formation of flavones and isoflavones from flavanones is catalyzed by several evolutionarily related P-450s, either in a single concerted reaction leading directly to the flavone double bond (flavone synthase II) or in a two-step process, the first being a monooxygenation of the C-ring, which yields a 2-hydroxyflavanone (flavanone 2-hydroxylase) or a 2-hydroxyisoflavanone (2HIS) intermediate. An alternative route for the conversion of flavanones to flavones involves, instead, a 2-oxoglutarate-dependent dioxygenase (14Britsch L. Arch. Biochem. Biophys. 1990; 282: 152-160Crossref PubMed Scopus (77) Google Scholar). The A-ring hydroxyl group in positions 5 and/or 7 is formed during the synthesis of the flavonoid skeleton catalyzed by chalcone synthase, a member of plant polyketide synthases (15Schröder J. Trends Plant Sci. 1997; 2: 373-378Abstract Full Text PDF Google Scholar). Additional hydroxyl groups in the A-ring of some flavonoid classes are found in positions 6 and 8. Enzymes involved in the hydroxylation of the A-ring at these positions have, however, not yet been described. Isoflavone and pterocarpan derivatives play important roles in plant-microbe interactions as phytoalexins and nodulation factors and as phytoestrogens. In contrast to the constitutive production of isoflavonoids in plants, pterocarpans, such as glyceollin from soybean (Glycine max L.), are inducible and accumulate in pathogen-infected or elicitor-treated plant tissues (16Ebel J. Grisebach H. Trends Biochem. Sci. 1988; 13: 23-27Abstract Full Text PDF PubMed Scopus (102) Google Scholar). In an attempt to investigate transcriptionally regulated cytochromes P-450 activated by biotic stress in soybean, the technique of differential display of mRNA was recently employed (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar). Eight full-length cDNA clones were subsequently isolated that represented elicitor-activated cytochromes P-450. One of these, CYP73A11 cDNA, encoded cinnamate 4-hydroxylase, and a second one, CYP93A1 cDNA, coded for 3,9-dihydroxypterocarpan 6a-hydroxylase (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar, 18Schopfer C.R. Kochs G. Lottspeich F. Ebel J. FEBS Lett. 1998; 432: 182-186Crossref PubMed Scopus (57) Google Scholar). We now present the functional identification of CYP71D9 whose cDNA was also previously isolated from elicitor-induced soybean cells by using the differential display method (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar). By employing heterologous expression the CYP71D9 cDNA was demonstrated to encode a protein capable of catalyzing the hydroxylation of ring A of flavonoid substrates. Combined studies with recombinant 2-hydroxyisoflavanone synthase (2HIS; CYP93C1v2) indicated that A-ring hydroxylation occurs before the 1,2-aryl shift of the flavonoid B-ring during isoflavanone formation. Daidzein (7,4′-dihydroxyisoflavone), eriodictyol (5,7,3′ 4′-tetrahydroxyflavanone), genistein (5,7,4′-trihydroxyisoflavone), kaempferol (3,5,7,4′-tetrahydroxyflavone), luteolin (5,7,3′,4′-tetrahydroxyflavone), naringenin (5,7,4′-trihydroxyflavanone), and quercetin (3,5,7,3′,4′-pentahydroxyflavone) were purchased from Roth (Karlsruhe, Germany); factor 2 (6,7,4′-trihydroxyisoflavone) and liquiritigenin (7,4′-dihydroxyflavanone) were from Extrasynthèse (Genay, France). Apigenin (5,7,4′-trihydroxyflavone), biochanin A (5,7-dihydroxy-4′-methoxyisoflavone), dihydroquercetin (3,5,7,3′,4′-pentahydroxyflavanone), and isoliquiritigenin (2′,6′,4-trihydroxychalcone) were from Sigma. Dihydrobiochanin A (5,7-dihydroxy-4′-methoxyisoflavanone), dihydrokaempferol (3,5,7,4′-tetrahydroxyflavanone), formononetin (7-hydroxy-4′-methoxyisoflavone), 7-methoxy-4′-hydroxyflavanone, 7,4′-dimethoxyflavanone, and medicarpin (3-hydroxy-9-methoxypterocarpan) were from our laboratory collection. Soybean (G. max L. cv. Harosoy 63) cell suspension cultures were propagated in the dark as described earlier (19Ebel J. Ayers A. Albersheim P. Plant Physiol. 1976; 57: 775-779Crossref PubMed Google Scholar). For elicitation experiments, 6-day-old cultures were transferred into fresh medium 12 h prior to treatment with a β-glucan elicitor fraction (20Schmidt W.E. Ebel J. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 4117-4121Crossref PubMed Scopus (149) Google Scholar) from Phytophthora sojae(200 μg/ml glucose equivalents) for 3–30 h. Cells were harvested by filtration, frozen in liquid nitrogen, and stored at −80 °C. Microsomal fractions were isolated by a modified version of the protocol described by Diesperger et al. (21Diesperger H. Müller C.R. Sandermann Jr., H. FEBS Lett. 1974; 43: 155-158Crossref PubMed Scopus (78) Google Scholar). Frozen cells were homogenized in a mortar with a pestle and suspended in 0.2m Tris-HCl, pH 7.5, 15% sucrose, 30 mmMgCl2, 1 mm phenylmethylsulfonyl fluoride, and 1 mm dithiothreitol, in the presence of Dowex 1 × 2. After filtration through a nylon mesh and centrifugation at 12,000 × g for 20 min, the microsomal fraction was collected from the supernatant by centrifugation for 30 min at 50,000 ×g, resuspended in 0.1 mKH2PO4/K2HPO4, pH 7.4, containing 30% glycerol, frozen in liquid nitrogen, and stored at −80 °C. TheSaccharomyces cerevisiae strain W303-1B, designated W(N), and its derivatives W(R) and WAT11 (22Truan G. Cullin C. Reisdorf P. Urban P. Pompon D. Gene (Amst.). 1993; 125: 49-55Crossref PubMed Scopus (150) Google Scholar, 23Urban P. Mignotte C. Kazmaier M. Delorme F. Pompon D. J. Biol. Chem. 1997; 272: 19176-19186Abstract Full Text Full Text PDF PubMed Scopus (281) Google Scholar), as well as the expression vector pYeDP60 (24Urban P. Cullin C. Pompon D. Biochimie (Paris). 1990; 72: 463-472Crossref PubMed Scopus (110) Google Scholar), were provided by Rhône-Poulenc Agro (Lyon, France) and D. Pompon (Gif-sur-Yvette, France). The yeast strains had previously been engineered to either overexpress the NADPH-cytochrome P-450 reductase from yeast (W(R)) or the Arabidopsis thaliana isoform ATR1 (WAT11) upon galactose induction. The yeast strain WVS1 was engineered in the same way to overexpress theVicia sativa P-450 reductase VS1 (accession number Z26250);VS1 was inserted into the integrative plasmid pYeDP110 (25Pompon D. Louerat B. Bronnie A. Urban P. Methods Enzymol. 1996; 272: 51-64Crossref PubMed Google Scholar) after polymerase chain reaction amplification for addition ofBamHI and SacI restriction sites just 5′ and 3′ of the coding sequence using primers 5′-CGGGATCCATGACTTCCTCTAATTCCG (5′-end) and 5′-CGGGAGCTCTCACCAAACATCTCTTAGG (3′-end) and then integrated into W(N) with the GAL10-CYC1 promoter at the locus of the endogenous reductase by homologous recombination. The coding regions of the soybean P-450s were amplified using the primers specified earlier (26Schopfer C. Isolation of elicitor inducible CYPs of soybean (G. max L.) and identification and functional expression of cDNAs encoding cinnamate 4-hydroxylase and dihydroxypterocarpan 6a-hydroxylase.Doctoral thesis. University of Munich, Germany1998Google Scholar) and inserted into pYeDP60 according to Urban et al. (27Urban P. Werck-Reichhart D. Teutsch H.G. Durst F. Regnier S. Kazmaier M. Pompon D. Eur. J. Biochem. 1994; 222: 843-850Crossref PubMed Scopus (143) Google Scholar). CYP93C1v2 cDNA (AF135484; a gift from R. A. Dixon and M. Gijzen, Samuel Roberts Noble Foundation, Ardmore, OK) (11Steele C.L. Gijzen M. Qutob D. Dixon R.A. Arch. Biochem. Biophys. 1999; 367: 146-150Crossref PubMed Scopus (167) Google Scholar) was amplified with the polymerase chain reaction primers 5′-atatatggatccATGTTGCTTGAACTTGCAC (5′-end) and 5′-tatataggtaccTAATTAAGAAAGGAGTTTAG (3′-end) to generateBamHI and KpnI restriction sites just 5′ and 3′ of the coding sequence, before insertion into pYeDP60. Polymerase chain reaction was performed as described earlier (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar), and the resulting plasmids were confirmed for identity by restriction and sequence analyses of the CYP-coding regions. Yeast strains W(R) (for CYP93C1v2) or WAT11 and WVS1 (for the other soybean P-450s) were transformed, and microsomal fractions were isolated as described earlier (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar, 27Urban P. Werck-Reichhart D. Teutsch H.G. Durst F. Regnier S. Kazmaier M. Pompon D. Eur. J. Biochem. 1994; 222: 843-850Crossref PubMed Scopus (143) Google Scholar). CYP expression was induced using the high density procedure (25Pompon D. Louerat B. Bronnie A. Urban P. Methods Enzymol. 1996; 272: 51-64Crossref PubMed Google Scholar), and cultures of 1–5 × 108 cells/ml were used for microsome preparation. The standard assay for flavonoid 6-hydroxylase (F6H; CYP71D9, accession number: Y10490) contained in a total volume of 100 μl of 50 mm Tricine/KOH, pH 7.9, 0.5 mm reduced glutathione, 60 μg of microsomal protein from yeast, and 100 μm substrate (dissolved in either Me2SO or 2-methoxy-ethanol). After equilibration for 2 min at 18 °C, the reaction was started with the addition of 20 μl of a NADPH-regenerating system (comprising 100 μmNADPH, 160 μm glucose 6-phosphate, and 0.04 unit of glucose 6-phosphate dehydrogenase in the reaction mixture) and terminated by the addition of 100 μl of 3% acetic acid in ethyl acetate. The products were extracted once with this solvent and twice thereafter with ethyl acetate from the reaction mixture. The organic phase was pooled and evaporated, and the residue was dissolved in 200 μl of a mixture of 40% methanol, 60% water, 0.2% acetic acid (v/v) and analyzed by reverse-phase high performance liquid chromatography (RP-HPLC) (LiChrosorb RP-18, 4 × 250 mm; flow rate, 1 ml min−1; linear gradient from 35 to 65% methanol in 16 min; eluent 1). Compounds were detected at 290 nm and, when (R,S)-liquiritigenin (7,4′-dihydroxyflavanone) was used as substrate, the retention times were 10 min for the reaction product and 15 min for liquiritigenin. Amounts of products were calculated using molar extinction coefficients (ε290 nm, MeOH = 6500 m−1cm−1 for liquiritigenin; ε = 15500m−1 cm−1 for naringenin; ε = 16800 m−1 cm−1 for eriodictyol) or relative to the conversion of liquiritigenin (all other substrates tested). The pH optimum was determined as described for the standard assay using 50 mm Tricine/KOH buffer, pH 7–9. The assay for 2HIS contained in a total volume of 100 μl of 50 mm Tricine/KOH, pH 8.6 (28Kochs G. Grisebach H. Eur. J. Biochem. 1986; 155: 311-318Crossref PubMed Scopus (134) Google Scholar), 0.5 mm reduced glutathione, 60 μg of microsomal protein from yeast, and 100 μm liquiritigenin or 6,7,4′-trihydroxyflavanone. The reaction was run for 1 h at 15 °C. Products were extracted as described for the F6H assay and were analyzed by RP-HPLC using eluents comprising 1% acetic acid in a 50:50 mixture of methanol and acetonitrile (solvent A) and 1% acetic acid in H2O (solvent B). Compounds were eluted at a flow rate of 1 ml min−1, in a linear gradient of solvent A from 25 to 70% (eluent 2), within 18 min. Eluates were monitored at 290 nm. For characterization of the reaction product from the combined catalytic action of F6H and 2HIS, the incubation with (R,S)-liquiritigenin was carried out at a larger scale (240 times that described above for the F6H and 2HIS standard assays). The F6H reaction was initiated by incubation at 18 °C, pH 7.9, for 1 h. This was followed by the addition of 2HIS, adjustment of the pH to 8.6 and a switch to 15 °C for an extra hour. The reaction was terminated, and the products were extracted as described above. To achieve a positive identification of the tetrahydroxylated isoflavanone product, the ethyl acetate residue was subjected to acid treatment by stirring in 500 μl of 10% HCl (v/v) in methanol for 1 h at room temperature. The mixture was extracted thrice with ethyl acetate, and, upon evaporation, the pooled organic phase was dissolved in 300 μl of methanol and analyzed by RP-HPLC as described above for the standard 2HIS assay. Fractions of the eluate containing the isoflavone derivative (R t = ∼16 min) were collected, concentrated, reapplied on to the column to ascertain purity, reduced to dryness, and subjected to UV and NMR spectroscopy. Spectrophotometric measurements of total P-450 content and evaluation of substrate binding were performed according to Omura and Sato (29Omura T. Sato R. J. Biol. Chem. 1964; 239: 2370-2378Abstract Full Text PDF PubMed Google Scholar) and Schalk et al. (30Schalk M. Batard Y. Seyer A. Nedelkina S. Durst F. Werck- Reichhart D. Biochemistry. 1997; 36: 15253-15261Crossref PubMed Scopus (27) Google Scholar), respectively. Substrate-binding spectra were recorded using double cuvettes. K s , ΔA max, and the corresponding S.D. values were calculated from the ΔA 390–420 nm for eight to ten ligand concentrations using the nonlinear regression program DNRPEASY. Flavonoids were dissolved in Me2SO. Products of the F6H-catalyzed reaction were analyzed by mass spectrometry to determine the site of hydroxylation of the flavonoid substrates. For gas chromatography-mass spectrometry analysis, the samples were converted to trimethylsilyl ether derivatives with a mixture of bis-(trimethylsilyl)trifluoroacetamide containing 1% trimethylchlorosilane and pyridine (1:1 v/v) for 2 h at room temperature. The electronic impact analysis of 1-μl samples was performed on a Trio 2000 Micromass Quadrupole apparatus fitted with a J and W Scientific DB 5 MS (5% phenyl, 95% methyl) column (30 m × 0.320 mm inner diameter, 0.1 μm film) using He TPH 55 at 60 kPa as carrier gas. Initial column temperature was 120 °C, held for 2 min, and ramped to 250 °C at 15 °C min−1 and then from 250 °C to 280 °C at 2 °C min−1. Nuclear magnetic resonance spectra of reaction products were acquired with a Bruker DMX 500 NMR spectrometer using a 5-mm inverse geometry probehead (90°(1H) = 9.3 μs; 90°(13C) = 9.8 μs) in acetone-d 6 (2.04, 29.8 ppm) at 303 K. A phase-sensitive (echo-antiecho selection) and sensitivity enhanced1H,13C heteronuclear single quantum coherence NMR spectrum of 6,7,4′-trihydroxyisoflavone was acquired using Bruker standard software (1J(CH) = 165 Hz; acquisition time, 203 ms; spectral width, 5040 Hz (F2), 106 F1 (13C) increments with a final resolution of 83 Hz;13C GARP decoupling: 70 μs, gradient; pulse, 1 ms; recovery, 450 μs). The pK values of 6,7,4′-trihydroxyisoflavone were calculated with the ACD/Labs (Pegnitz, Germany) pK a Data Base, Version 4.5. Earlier investigations (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar) identified eight cDNA clones representing cytochromes P-450 whose expression in soybean cell cultures was activated by elicitor treatment concomitantly with the production of glyceollins, the pterocarpanoid phytoalexins of soybean: CYP73A11, CYP82A2, CYP82A3, CYP82A4, CYP93A1, CYP93A3, CYP71D8, and CYP71D9. Activated expression of another P-450 (CYP71A9) isolated in the same mRNA differential display screening was refuted when tested by Northern blot analysis. Previous analyses of the catalytic properties of recombinant proteins expressed in yeast disclosed the function of two of the clones. CYP73A11 cDNA encoded a cinnamate 4-hydroxylase (17Schopfer C.R. Ebel J. Mol. Gen. Genet. 1998; 258: 315-322Crossref PubMed Scopus (68) Google Scholar), whereas CYP93A1 cDNA coded for 3,9-dihydroxypterocarpan 6a-hydroxylase (18Schopfer C.R. Kochs G. Lottspeich F. Ebel J. FEBS Lett. 1998; 432: 182-186Crossref PubMed Scopus (57) Google Scholar). The former cytochrome P-450 thus represented a well studied enzyme of general phenylpropanoid metabolism whose action gives rise to the hydroxyl group in position 4′ of the flavonoid B-ring. More significantly, the latter cytochrome P-450 catalyzes the stereoselective and regioselective hydroxylation of position 6a of 3,9-dihydroxypterocarpan to give (S)-3,6a,9-trihydroxypterocarpan (glycinol), a biosynthetic intermediate of the glyceollins (18Schopfer C.R. Kochs G. Lottspeich F. Ebel J. FEBS Lett. 1998; 432: 182-186Crossref PubMed Scopus (57) Google Scholar). These earlier studies thus demonstrated that within the isolated group of CYP clones at least two cDNAs were related to phenylpropanoid and isoflavonoid pathways. In an attempt to identify other candidate P-450 cDNAs involved in flavonoid biosynthesis, all the other coding sequences were also expressed in yeast. Because we previously showed that the level of P-450 expression in yeast might be strongly dependent on the coexpressed P-450 reductase (31Cabello-Hurtado F. Batard Y. Salaün J.P. Durst F. Pinot F. Werck-Reichhart D. J. Biol. Chem. 1998; 273: 7260-7267Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar, 32Robineau T. Batard Y. Nedelkina S. Cabello-Hurtado F. Le Ret M. Sorokine O. Didierjean L. Werck-Reichhart D. Plant Physiol. 1998; 118: 1049-1056Crossref PubMed Scopus (106) Google Scholar), a yeast strain overexpressing a P-450 reductase (accession number Z26250), isolated from the legumeV. sativa, was constructed. The expression of the different P-450s in this strain WVS1 was compared with that in the strain WAT11, overexpressing the A. thaliana reductase ATR1 (TableI). No expression of CYP82A4 was obtained in either strain. For all other CYPs, except CYP71D8, twice as much expression was obtained in WAT11 compared with WVS1. The ratio was reversed for CYP71D8, which appeared to be more stable in the presence of the reductase from V. sativa. Microsomal fractions isolated from the most favorable yeast strains were then systematically screened with a variety of flavonoid compounds for specific binding into the CYPs active sites. Type I binding spectra, indicative of a displacement of solvent in the vicinity of heme (33Raag R. Poulos T.L. Biochemistry. 1989; 28: 917-922Crossref PubMed Scopus (193) Google Scholar, 34Loida P.J. Sligar S.G. Biochemistry. 1993; 32: 11530-11538Crossref PubMed Scopus (214) Google Scholar), were recorded upon addition of naringenin to CYP71D9 and CYP82A2, dihydrokaempferol to CYP71D8 and CYP71D9, and eriodictyol to CYP71A9, CYP71D8, CYP71D9, and CYP82A2. Largest amplitude spectra were obtained upon binding of eriodictyol to CYP71D8 and CYP71D9 (TableII). As an example, spectra for CYP71D9 are shown in Fig. 1. No interaction of formononetin, genistein, or daidzein with any of the CYPs was detected.Table IComparison of expression of the soybean CYPs in yeast strains overexpressing V. sativa (WVS1) or A. thaliana ATR1 (WAT11) P450 reductasecDNAP450WVS1WAT11pmol mg −1 proteinCYP71A914.144.9CYP71D824.713.1CYP71D972.4127.8CYP82A235.9106.5CYP82A3016.5CYP82A400CYP93A321.356.7P450 content in microsomal fractions prepared from yeast cells grown for 16 h in the presence of galactose was determined from CO-reduced versus reduced difference spectra. Open table in a new tab Table IIScreening for flavonoid binding to soybean CYPs in recombinant yeast microsomesCompounds testedAmplitude of the type I spectrumCYP93A3CYP71A9CYP71D8CYP71D9CYP82A2CYP82A3ΔA390–420 nmpmol −1 P450Naringenin0001370Dihydrokaempferol0082200Eriodictyol028565680Dihydroxypterocarpan180001.40Formononetin00000.80Genistein000000Daidzein000000The microsomal fraction was prepared from yeast cells grown for 16 h in the presence of galactose. Difference spectra were recorded in 0.1m Tris-HCl, pH 7.5, containing 30% glycerol and 1 mm EDTA after adding 100 μm of each of the compounds to the oxidized microsomes in the sample cuvette and an equal volume of solvent to the reference microsomes. Open table in a new tab P450 content in microsomal fractions prepared from yeast cells grown for 16 h in the presence of galactose was determined from CO-reduced versus reduced difference spectra. The microsomal fraction was prepared from yeast cells grown for 16 h in the presence of galactose. Difference spectra were recorded in 0.1m Tris-HCl, pH 7.5, containing 30% glycerol and 1 mm EDTA after adding 100 μm of each of the compounds to the oxidized microsomes in the sample cuvette and an equal volume of solvent to the reference microsomes. The preliminary screening thus indicated that some of the soybean CYPs were binding flavonoids in their active site. For some of them, in particular CYP71D8 and CYP71D9, displacement of solvent was effective enough so that a positioning suitable for an oxidative attack was likely to be achieved. To test such a possibility, recombinant yeast microsomes were incubated with NADPH and liquiritigenin, eriodictyol, naringenin, or dihydrokaempferol. Formation of polar metabolites was observed with all four flavonoids but only upon incubation with CYP71D9. To further characterize the metabolite formed by CYP71D9, microsomes of recombinant WAT11 yeast were incubated with liquiritigenin and NADPH, and the ethyl acetate extract of the reaction mixture was analyzed by RP-HPLC (Fig. 2). Following RP-HPLC, the reaction product was identified by three criteria: retention time during HPLC, mass spectrometry, and NMR spectroscopy. As shown in Fig. 2, the product P formed from liquiritigenin (S) had a smaller retention time (6 min) than the substrate (10 min), required NADPH for its formation, and was not formed when yeast cells were transformed with the empty vector. Similar results were recorded with naringenin as a substrate. The higher polarity of the products when compared with the substrates was fully supported by mass spectrometry of their trimethylsilyl derivatives. The metabolites of liquiritigenin and naringenin exhibited molecular ion peaks atm/z 488.2 and 576.2, respectively. The retro-Diels-Alder fragment peaks were found atm/z 296.1 (A-ring) and m/z192.1 (B-ring) for the product formed from liquiritigenin, and atm/z 384.2 (A-ring) and atm/z 192.1 (B-ring) for that formed from naringenin. These results indicated that recombinant CYP71D9 protein catalyzed the monooxygenation of ring A of both flavanones (Fig.3). Fragments hydroxylated on ring A were also observed upon gas chromatography-mass spectrometry analysis of the metabolites of eriodictyol and dihydrokaempferol. To elucidate the position of hydroxylation of the A-ring,1H NMR spectra were recorded in acetone-d 6 of the product formed from liquiritigenin (Table III). The spectra clearly showed that the proton signal of H-6 was absent in the reaction product, whereas H-5 and H-8 formed singlets. All other signals were very similar to those observed for liquiritigenin. Taken together, the chemical characterization identified 6,7,4′-trihydroxyflavanone as the product formed from liquiritigenin in the reaction catalyzed by the recombinant CYP71D9 protein (Fig. 3). We conclude that the enzyme encoded by CYP71D9 is a F6H.Table III1 H and 13 C NMR spectra of liquiritigenin, 6,7,4′-trihydroxyflavanone, and 6,7,4′-trihydroxyisoflavone measured in acetone-d6Proton7,4′-Dihydroxyflavanone (Liquiritigenin)6,7,4′-Trihydroxyflavanone6,7,4′-TrihydroxyisoflavoneδJδJδJδ1H 3-aChemical shift values taken from the 1H,13C heteronuclear single quantum coherence NMR spectrum before vacuum treatment.δ13C 3-aChemical shift values taken from the 1H,13C
Referência(s)