Eur. J. Biochem. 207, 109- 116 (1992)

0FEBS 1992

Optimization of yeast-expressed human liver cytochrome P450 3A4 catalytic activities by coexpressing NADPH-cytochrome P450 reductase and cytochrome b5 Marie-Anne PEY RONNEAU ’, Jean-Paul RENAUD Gilles TRUAN’, Philippe URBAN ’, Denis POMPON and Daniel MANSUY I,

I

Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, Centre National de la Recherche Scientifique, Universite Paris 5 , France Centre de Genbtique Moltculaire, Centre National de la Recherche Scientifique Associee i I’Universite Paris 6, Gif-sur-Yvette, France

(Received March 11, 1992) - EJB 92 0338

Human liver P450 NF25 (CYP3A4) had been previously expressed in Succhuromyces cerevisiue using the inducible GALIO-CYCl promoter and the phosphoglycerate kinase gene terminator [Renaud, J. p., Cullin, C., Pompon, D., Beaune, P. and Mansuy, D. (1990) Eur. J. Biochem. 194, 8898961. The use of an improved expression vector [Urban, P., Cullin, C. and Pompon, D. (1990) Biochimie 72,463 -4721 increased the amounts of P450 NF25 produced/culture medium by a factor of five, yielding up to 10 nmol/l. The availability of recently developed host cells that simultaneously overexpress yeast NADPH-P450 reductase and/or express human liver cytochrome b5, obtained through stable integration of the corresponding coding sequences into the yeast genome, led to biotechnological systems with much higher activities of yeast-expressed P450 NF25 and with much better ability to form P450 NF25 -iron-metabolite complexes. %fold, g-fold, and 30-fold rate increases were found respectively for nifedipine 1,4-oxidation, lidocaine N-deethylation and testosterone 6@-hydroxylationbetween P450 NF25-containing yeast microsomes from the basic strain and from the strain that both overexpresses yeast NADPH-P450 reductase and expresses human cytochrome b5. Even higher turnovers (15-fold, 20-fold and 50-fold rate increases) were obtained using P450 NF25-containing microsomes from the yeast just overexpressing yeast NADPH-P450 reductase in the presence of externally added, purified rabbit liver cytochrome b5. This is explained by the fact that the latter strain contained the highest level of NADPH-P450 reductase activity. It is noteworthy that for the three tested substrates, the presence of human or rabbit cytochrome b, always showed a stimulating effect on the catalytic activities and this effect was saturable. Indeed, addition of rabbit cytochrome b5 to microsomes from a strain expressing human cytochrome b5 did not further enhance the catalytic rates. The yeast expression system was also used to study the formation of a P450NF25 -iron-metabolite complex. A P450 Fe(I1)-(RNO) complex was obtained upon oxidation of N hydroxyamphetamine, catalyzed by P450-NF25-containing yeast microsomes. In microsomes from the basic strain expressing P450 NF25, 10% of the starting P450 NF25 was transformed into this metabolite complex, whereas more than 80% of the starting P450 NF25 led to complex formation in microsomes from the strain overexpressing yeast NADPH-P450 reductase. These results show that specific activities of yeast-expressed P450 NF25 may be artificially low, owing to limiting amounts of the associated microsomal redox proteins and emphasize the importance of controlling the amounts of the different components of the monooxygenase complex in order to optimize these catalytic activities, especially when the expression system is to be used for demonstrating metabolic capacities towards new substrates. Correspondence to D. Mansuy, Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, CNRS URA 400,45 rue des Saints-Pkres, F-15270 Paris Cedex 06, France Fax; 33 1 42 86 04 02. Abbreviations. Cyt. b5,cytochrome b5 ; P450 reductase, NADPHP450 reductase; V8, YeDP1/8-2 plasmid; V60, YeDP60 plasmid. Enzymes. NADPH-P450 reductase [EC 1.6.2.41; P450 [EC 1.14.14.1]. Note. The updated recommended nomenclature for P450 species [Ncbert, D. W., Nelson, D. R., Coon, M. J., Estabrook, R. W., Feyereisen, R., Fujii-Kuriyama, Y., Gonzalez, F. J., Guengerich, F. P., Gunsalus, I. C., Johnson, E. F., Loper, J. C., Sato, R., Waterman, M. R., & Waxman, D. J. (1991) DNA 10, 1 - 141 is used throughout the text. The name ‘cytochrome’ has been abandoned according to the Nomenclature Committee of the International Union of Biochemistry, Nomenclature of electron-transfer proteins, Recommendations 1989 [Eur. J . Biochem. 200, 599-612 (1991)l the appropriate name being ‘haem-thiolate protein’.

P450 form a superfamily of heme-thiolate proteins involved in the primary oxidation of numerous lipophilic compounds including endogenous substrates like fatty acids, steroids and vitamins, as well as exogenous substrates like drugs, dietary substances and environmental pollutants. The b r o a d substrate specificity is now well understood on the basis of enzyme multiplicity (Gonzalez, 1989). More than 160 cDNA species coding for P450 have been isolated so far and have been classified on the basis of primary amino acid sequence similarities (Nebert et al., 1991). Although the normal fate of xenobiotic oxidation products is excretion, directly or after further conjugation with a polar group, P450 catalysis sometimes yields highly reactive metabolites that can injure cells by altering macromolecular components, leading especially to carcinogenesis (Kadlubar and Hammons, 1987). Other deleterious effects can arise during the course of xenobiotic oxi-

110 dation by human liver P450 due to drug interactions. For example, when two or more drugs are competitively metabolized by the same form of P450 or when a drug specifically induces or inhibits the form responsible for the metabolization of other drugs, this can lead to a reduced or prolonged therapeutic action and potentially toxic side effects. In order to predict hepatic drug metabolism and possible toxic effects and drug interactions, a precise knowledge of the specific activities of individual forms of P450 from human liver toward different classes of substrates is required. For that purpose, numerous groups have studied the properties of human hepatic P450 species using either hepatocytes, liver microsomes or purified enzymes. However, working with human tissues leads to many problems from an ethical and technical point of view, as well as problems of availability, reproducibility and interindividual variability. A fruitful alternative consists of the expression of cloned cDNA, coding for individual forms of P450 enzymes, in an adequate heterologous system. Several human liver P450 from the subfamilies involved in xenobiotic oxidation have already been expressed in different systems, including recombinant simian virus 40 in COS cells (Gonzalez, 1989 and references therein; Romkes et al., 1991; Veronese et al., 1991), recombinant vaccinia viruses in human hepatoma Hep G2 cells (Aoyama et al., 1990a and 1990b, and references therein) and recombinant plasmids in the yeast Succharomyces cerevisiae (Yasumori et al., 1989; Brian et al., 1989; Brian et al., 1990; Renaud et al., 1990; Eugster et al., 1990; Ching et al., 1991; Truan, G., Cullin, C., Reisdorf, P., Urban, P. and Pompon, D., unpublished results). P450 NF25 (CYP3A4) is important in pharmacology and toxicology, not only because it is probably the major form of human liver (Guengerich and Turvy, 1991) but also because it is involved in the metabolism of numerous widely used drugs such as nifedipine (Guengerich et al., 1986a), erythromycin and troleandomycin (Renaud et al., 1990 and references therein), quinidine (Guengerich et al., 1986b), cyclosporin A (Kronbach et al., 1988; Aoyama et al., 1989; Combalbert et al., 1989), 17a-ethynylestradiol (Guengerich, 1988), midazolam (Kronbach et al., 1989), lidocaine (Bargetzi et al., 1989; Imaoka et al., 1990), and diltiazem (Pichard et al., 1990). P450 NF25 was recently functionally expressed in S. cerevisiue (Renaud et al., 1990; Brian et al., 1990). Galactoseinducible expression using GALIO-CYCl, a hybrid promoter composed of the yeast GAL10 gene upstream-activating sequence and the iso-1-cytochrome c gene transcription-initiation sequence (Guarente et al., 1982), allowed relatively good levels of P450 NF25 to be obtained in transformed yeast (approximately 2 - 3 nmol/l culture), yeast microsomes to be obtained containing NF25 as the only detectable P450 and in a catalytically active state, and an expression level high enough to allow spectrophotometric binding studies on yeast microsomes (Renaud et al., 1990). Actually, the main factor limiting the catalytic activity of transformed yeast microsomes was the endogenous NADPH-P450 reductase (P450 reductase) present in yeast but not in significant amounts. A possible way to obtain more catalytically active yeast strains expressing P450 NF25, could be to coexpress the associated electron-transfer proteins by integration of galactoseinducible expression cassettes into the yeast genome under the control of an efficient and externally controllable promoter. This kind of coexpression in yeast has been recently described and characterized elsewhere (Pompon et al., 1991; Truan et al., unpublished results). Here we report the catalytic activity of P450 NF25 expressed in yeast also overexpressing its own P450 reductase

Table 1. Strains used in this study. (0ver)expression was under the control of the artificial, galactose-inducible GAL10-CYCl promoter. Strain

Strain type

W(N) W(R) W(R,N) W(B) W(B,N) W(B,R)

Wild(W303) W (R) x W (N) W (B) x W (N) W (B) x W (R)

(over) expression phenotype -

yeast reductase yeast reductase human cyt. b5 human cyt. b5 yeast reductase human cyt. b5

[URA-, ADE-] [URA+/-, ADE-] [URA+'-, ADE-] [URA', ADE-] [URA', ADE-] [URA', ADE-]

and/or coexpressing human cytochrome bs (cyt. bs) at different levels towards different substrates to show the usefulness of such a coexpression system in determining the effects of changes in the electron-transfer chain within the monooxygenase complex on P450 NF25 activities, and in dramatically increasing the specific activities of this major human hepatic isoform in transformed-yeast microsomes. MATERIALS AND METHODS Chemicals and reagents All molecular biological reagents were of analytical grade. The ingredients for culture media were obtained from Difco (OSI, Paris, France). Nifedipine, nitrendipine, testosterone, 1dehydrotestosterone and lidocaine were purchased from Sigma (St Louis, MO), and 6P-hydroxytestosterone was obtained from Steraloids (Wilton, NH). NADPH and horse heart cytochrome c were obtained from Boehringer Mannheim France (Meylan, France). Monoethylglycine xylidide was given by Laboratoires Roger Bellon (Neuilly-sur-Seine, France) and 1SP-hydroxytestosterone was a gift from Searle (Skokie, IL). N-Hydroxyamphetamine was prepared from 2nitro-1 -phenylpropene by LiAlH4 reduction (Gilsdorf and Nord, 1952). Eschevichiu coli strains

E. coli strain DH5- 1 (F-, recAl, gyrA96, thi-1, hsdR17, supE441, A-) was used for cloning. S. cerevisiue strains (Table 1)

W303-1B (leu2, his3, trpl, ade2-1, ura3, canR, cyr') was constructed by R. Rothstein. This is the 'wild-type' strain W(N). The construction and characterization of the other strains used in this work are reported elsewhere (Pompon et al., 1991; Truan et al., unpublished results): W(R) (MATE) comes from the stable integration into the genome of W(N) (MATa) of the galactose-inducible GALlOCYCl promoter at the 5'-end of the yeast P450 reductase gene open reading frame; W(B) (MATa) comes from the stable integration into the genome of W(N) (MAT'), by disruption of the yeast P450 reductase gene, of the human cyt.-b5-coding sequence placed under the control of the GALlO-CYCI promoter and the phosphoglycerate kinase gene terminator; W(R, N) and W(B, N) are diploids [W(R) (MATol)xW(N) (MAT') and W(B) (MATa) x W(N) (MATE) respectively]; W(R, N) contains one copy of the yeast P450 reductase gene

111 placed under the control of the GAL10-CYCI promoter and W(B, N) contains one copy of the human cyt.-b5-coding sequence placed under the control of the GALIO-CYCl promoter; W(B, R) is also a diploid [W(B) (MATa)xW(R) (MATa)]which contains one copy of the human cyt.-b,-coding sequence and one copy of the yeast P450 reductase gene, both placed under the control of the GAL10-CYCI promoter. Expression vectors The construction of plasmids YeDP1/8 - 2 (V8) (Pompon, 1988; Cullin and Pompon, 1988) and YeDP60 (V60) (Urban et al., 1990) was reported earlier. Plasmid V8 contains a URA3 marker and plasmid V60 carries both the URA3 and ADE2 selection markers. Insertion of NF25 cDNA into V8 to give the expression plasmid NF25-V8 (formerly called pVNF25) w v described in a previous paper (Renaud et al., 1990). Insertion of NF25 cDNA into V60 (Truan et al., unpublished results) was achieved by using homologous recombination properties of yeast (Pompon and Nicolas, 1989). Yeast transformation and culture NF25-V8 is compatible with W(N), W(R)[URA-1, and W(R, N)[URA-1, while NF25-V60 is compatible with all the strains (Table 1). The transformed yeasts are denoted by the simple juxtaposition of the name of the heterologous P450 cDNA carried on the plasmid and the name of the starting strain. For instance, the strain obtained by transforming W(R) by NF25-V8 is called NF25-W(R). The strain obtained by transforming W(R) by V8 (empty plasmid not containing the foreign cDNA) is called CONTROL-W(R). Transformations were performed according to a modified lithium acetate method (Cullin and Pompon, 1988). Transformed cells were grown at 28 "C in solution A (minimal medium) for V8 (Cullin and Pompon, 1988) or solution B (complete medium) for V60 (Urban et al., 1990). Solution A was 2% (massjvol.) D-galactose, 0.7% (massjvol.) yeast nitrogen base without amino-acids, 0.1YO(massjvol.) bactocasaminoacids, 0.002% (massivol.) tryptophan and 0.004% (massjvol.) adenine. Solution B contained 2% (massjvol.) D-galactose, 1YO(massjvol.) yeast extract and 1YO(massivol.) bactopeptone. Microsome preparation After centrifugation, 4 g wet cells were resuspended in 50 ml 50 mM Tris/HCl, pH 7.4, 5 mM EDTA and 100 mM KCl containing 87 pl 2-mercaptoethanol and incubated for 5 min at room temperature. After centrifugation, cells were washed with 50ml of the same buffer, not containing 2mercaptoethanol, and centrifuged again. The pellet was resuspended in 4 ml cold 50 mM Tris/HCl, pH 7.4, 2 mM EDTA and 1.2 M sorbitol and the volume was adjusted to 25 ml with cold 50 mM Tris/HCl, pH 7.4, 2 mM EDTA and 0.6 M sorbitol buffer. 40 g 0.45 -0.50 mm diameter glass beads (B. Braun, Melsungen, FRG) were added to the suspension and yeast cell walls were disrupted mechanically using a MKS homogenizer (B. Braun, Melsungen, FRG; 4 x 15 s at 4000 rpm) cooled with liquid COz. The beads were removed by glass-filtration and rinsed with 25 ml cold 50 mM Trisj HC1, pH 7.4, 2 mM EDTA and 0.6 M sorbitol. The filtrate was centrifuged at 4°C for 5 min at 1000 g , then for 10 rnin at 14000 g. CaClz was added to the supernatant (15mM final concentration) and the suspension was left on ice for 15 min. Microsomes were spun down by centrifuging at 4°C for

15 min at 14000 g , resuspended in a minimum volume (approximately 2 - 3 ml) cold 50 mM TrisjHCI, pH 7.4, 1 mM EDTA and 20% glycerol and kept at - 80 "C for months.

Quantitation of the different enzymes in microsomal fractions The total microsomal protein concentration was determined according to the method of Lowry et al. (1951). Total P450 was measured according to Omura and Sat0 (1964). Cyt. b5 was quantified from the reduced versus oxidized difference spectrum of microsomes, using a differential absorption coefficient 4 8 4 2 4 - 4 0 9 of 185mM-' . cm-' (Omura and Sato, 1964). NADPH-P450 reductase activity was expressed as the rate of cytochrome c reduction under slightly modified conditions from a previously described protocol (Urban et al., 1990). A solution containing 0.1 mg cytochrome c (approximately 8 nmol) in 50 mM Tris/HCl, pH 7.4, and 1 mM EDTA (final volume 940 p1) was divided equally between both cuvettes of a Kontron 820 spectrophotometer. 20 p1 fresh NADPH (12 mM in distilled water) was added to the sample cuvette, while the same volume of distilled water was added to the reference cuvette. Reduction was initiated by the addition to both cuvettes of 10 pl of a suspension containing 10 pg microsomal protein (only 1 pg for microsomes from P450-reductase-overexpressing yeasts). The absorbance change at 550nm was monitored at 20°C and the rate of cytochrome c reduction was calculated using an absorption coefficient of 21 mM-' . cm-'.

Catalytic activity studies Rabbit liver cyt. b5 was purified according to Strittmatter et al. (1978). Nifedipine oxidation was performed as previously described (Renaud et al., 1990). The assay for lidocaine oxidation essentially followed the procedure of Oda et al. (1989). Typical incubations included lidocaine (0.5 pmol, added from a fresh 0.1 M stock solution in water/acetonitrile (99.5 :OS), yeast microsomes containing P450 NF25 (lOOpmol), NADPH (0.25 pmol) and 50mM TrisjHCl, pH 7.4 and 1 mM EDTA in a final volume of 0.5 ml. When rabbit liver cyt. b5 (1 mol/mol P450 from a 20 pM stock solution) was included, microsomes were incubated with cyt. b5 for 20 rnin on ice, then with the substrate for 3 min at 37°C before addition of NADPH. The reaction proceeded for 15 rnin at 37°C and was then stopped by the addition of 50 pl 1 M NaOH. The reaction mixture was extracted twice with CHzClz (800 pl and 500 p1 portions). The combined organic phases were evaporated to dryness under a N2 stream at 30°C. The residue was dissolved in 600 p1 of the mobile phase of HPLC and 100 pl were injected onto a Nucleosil CI8 reversephase HPLC column (5 p particle size, 4.6 m m x 2 5 cm, Socikte Franqaise de Chromatographie sur Colonne, NeuillyPlaisance, France) placed in line following a 1-cm-long octyldecylsilyl guard column, using 20 mM potassium phosphate, pH 3.0 acetonitrile (9: 1) as the eluent (flow rate 1 ml/ min). The metabolites were monitored at 219 nm. The assay for testosterone 6P-hydroxylation (Brian et al., 1990) was carried out exactly as for nifedipine oxidation except that the testosterone concentration was 50 pM and the same HPLC conditions were used for analysis (Renaud et al., 1990). 1-Dehydrotestosterone was used as a standard.

112 Formation of a P450 - NF25-Fe(lI) - nitrosoalkane complex followed by difference visible spectroscopy Yeast microsomes were suspended in 100 mM Tris/HCl, pH 7.4 and 1 mM EDTA (P450 NF25 final concentration 200 nM), rabbit cyt. b, from a 20 pM stock solution was added (final concentration 200 nM) and the suspension (1 ml) was equally divided between both cuvettes of a Kontron 820 spectrophotometer. After recording the baseline, 1 p1 fresh N-hydroxyamphetamine (0.01 M in dimethylsulfoxide) was added to the sample cuvette, the same volume of solvent being added to the reference cuvette. 10 pl fresh NADPH (12 mM in water) was added to both cuvettes and difference spectra were recorded over 380 - 500 nm at different times. Measurements were carried out at 20°C.

RESULTS

Table 2. NADPH-P450 reductase content in microsomes from transformed yeast cells. P450 NF25 content was always in the range 50150 pmol/mg microsomal protein. Cyt. b5 content was in the range 100- 300 pmol/mg microsomal protein but it was not possible to distinguish yeast cyt. b5 and human cyt. b5 in microsomes from NF25-W(B, N), NF25-W(B, R), and NF25-W(B) with the available techniques (see Materials and Methods). The reductase activity shows the mean values of a t least two independent duplicate determinations. The standard deviation was approximately 10%. Microsomes

Reductase activity nmol cytochrome c reduced . (mg protein)-' . min-' 60 1000 2000 25 690