Expression of CYP2D6 in developing human liver

Transcription

1 Eur. J. Biochem. 202, (1991) 63 FEBS 1991 Expression of CYP2D6 in developing human liver Jean-Marc TRELUYER ', Evelyne JACQZ-AIGRAIN ', Fernando ALVAREZ and Thierry CRESTEIL Institut National de la Santi: et de la Recherche MCdicale U75, Paris, France Laboratoire de Pharmacologie Clinique, Institut National de la Santi: et de la Recherche MCdicale U120, HBpital Robert Debrk, Paris, France Institut National de la SantC et de la Recherche MCdicale U56, HBpital de BicEtre, Le Kremlin BicEtre, France (Received July 25, 1991) - EJB The CYP2D6 protein is a polymorphic isoenzyme involved in the biotransformation of several drugs including the probe drug dextromethorphan. The rise in the protein concentration, immunochemically determined with a specific antibody, was shown to occur within the first week following birth, whatever the gestational age at birth. In fetuses, the concentration of hepatic CYP2D6 protein was very low or undetectable in 70% of samples tested. In the remaining 30%, its concentration was comparable to that of newborns aged 1-7 days. This early rise was associated with spontaneous abortion in 70% of positive samples, whereas in fetuses with an intermediate CYP2D6 protein concentration, 80% were from induced abortions. The rise in CYP2D6 protein was associated with the developmental onset of dextromethorphan O-demethylation, but not N-demethylation, even if activity was lower in fetal than in neonatal and in adult liver microsomes. Lastly, the CYP2D6 RNA is detectable earlier than the protein and exhibits a peak of hepatic accumulation in newborns, before declining in adulthood. A positive correlation between RNA accumulation and protein concentration can be demonstrated only in the adult. This suggests that regulation is primarily at the transcriptional level, but cannot rule out the participation of post-transcriptional events in the regulation process throughout ontogenesis. Monooxygenase activities involved in the biotransformation of endogenous and exogenous hydrophobic molecules are mainly supported by cytochrome P-450 isoenzymes. To date, more than 30 isoenzymes of cytochrome P-450 have been identified, purified and/or cloned from the human liver (for review see [l]). These P-450s exhibit different profiles for their developmental evolution. In fetal liver, only one P-450 has been repeatedly detected: a CYP3A protein is present at a level roughly similar to that of the adult liver when determined by Western blotting [2] and is able to catalyze enzymatic activities like benzphetamine demethylation, aldrin epoxidation and dehydroepiandrosterone 16P-hydroxylation as the adult isoenzyme does. CYP2C8, one of the major P-450 isozymes immunologically detected in adult liver, is totally absent from fetal liver. The absence of this P-450 is responsible for the very low or negligible activity towards various substrates including mephenytoin, benzopyrene, p-nitroanisole [2, 31. To date, little or no data are available for other P-450 isoenzymes. We have focused our attention on CYP2D6, a gene coding for a polymorphic isoenzyme involved in the biotransformation of many drugs like P-blockers, tricyclic antidepressants, antitussives drugs, etc. In the adult Caucasian population, approximately 7% of individuals (poor vs extensive metabolizers) are defective in the active enzyme able to carry Correspondence to T. Cresteil, Laboratoire de Biochimie, TN- SERM U 75, CHU Necker, 156 rue de Vaugirard, F Paris cedex 15, France Enzymes. P-450, cytochrome P-450 (EC ); isocitrate dehydrogenase (EC 1.I.I.42). out the reaction [4]. This deficiency leads to an altered profile of urinary metabolites and consequently the ratio metabolite/ parent drug is markedly lowered [5]. Debrisoquine is the prototype of drugs subjected to CYP2D6 polymorphism: later, this definition was extended to spartein, bufuralol and dextromethorphan after both in vivo and in vitro investigations [6, 71. In rat, the CYP2D6 subfamily is expressed only after birth [8]. Since several of these drugs biotransformed by the CYP2D6 are widely administered to newborns and neonates, we were concerned by the ontogenesis of this enzyme during the fetal and neonatal period in the human liver, and secondarily by the influence of the gestational age upon drug metabolism in premature newborns. Determinations were carried out at the protein level by immunological quantitation and by assaying a CYP2D6-dependent activity to ascertain its biological function, and at the RNA level to determine whether or not a transcriptional regulation is involved. Results clearly indicate that the surge in CYP2D6 protein is mostly restricted to the postnatal period, whatever the postconceptual age of newborns. EXPERIMENTAL PROCEDURES Tissue collection Liver samples were collected from fetuses aged weeks aborted spontaneously or for therapeutic purposes (Down syndrome, severe hypotrophy or malformations). Postnatal samples were obtained from children aged 0.5 h to 5 years : they suffered from respiratory distress resulting from

2 584 immature pulmonary function, non-reversible hypotrophy, etc. Adult liver samples were obtained from donors for kidney transplantation. The protocol was submitted to and carried out following the recommendations of the Ethical Committee of Znstitut National de la SantC et de la Recherche Mkdicale. To facilitate the analysis of results, samples were divided into seven groups: group 1 refers to fetuses less than 30 weeks of gestational age (n = 60, weeks), group 2 to fetuses aged more than 30 weeks (n = 15, weeks), group 3 to newborns aged less than 24 h (n = 11, gestational age 29.3 f 3.3 weeks, 0.5 k 0.3 days postnatal), group 4 to newborns aged 1-7 days (n= 13, gestational age 33.8 f4.5 weeks, 4.3 f 1.6 days postnatal), group 5 to newborns aged 8-28 days (n = 8, gestational age 31.5 f 5.9 weeks, 13.9 k 9.5 days postnatal), group 6 to infants aged 4 weeks to 5 years (n = 12, gestational age weeks, 355 rt_ 555 days postnatal) and group 7 are adult samples (n = 8). Usually, hepatic samples were removed within the first hour after death. Immediately after removal samples were frozen in liquid nitrogen and kept at - 80 C until use within a few days. The quality of tissue sampling was determined by examination of RNA extracts and microsomal preparations : any trace of degradation (smear in denaturing gels for RNA, presence of P-420 and absence of P-450) was taken into consideration of discarding the sample. A crucial element affecting the integrity of cellular functions was the oxygen delivery to the liver: hepatocytes could be damaged leading to an irreversible degradation of proteins and RNA when the labor is very long after induction of abortion with prostaglandin or in newborns suffering from respiratory distress. RNA preparation and analysis Total RNA was extracted with hot phenol from about 1 g liver according to Moroy et al. [9]. RNA was subjected to electrophoresis in agarose gels containing 2.2 M formaldehyde: the presence of 28s and 18s ribosomal RNA with a very limited smear was required for further analysis. The amount of CYP2D6 RNA was examined in slot blots: 5-10 pg total RNA were applied to nitrocellulose (Hybond C extra, Amersham) and processed as previously described [lo]. Membranes were hybridized with CYP2D6 cdna and actin probes labeled with [32P]dCTP. High specific radioactivity probes (lo9 dpm/pg) were synthesized by random priming using a kit purchased from Amersham. The amount of hybridizable material was estimated after scanning the autoradiograms [lo] and standardized using actin as reference: results are routinely expressed as the ratio CYP2D6 RNA/ actin RNA. We have checked that the expression of the reference actin was not significantly modified (when expressed per mass total RNA extracted) throughout the whole period studied. Probes consisted in a genomic clone pac.h8h supplied by Dr Battula (NCI, Bethesda) for P-actin [l 11 and an EcoRI fragment, 1.2 kbp long, inserted in pgem for CYP2D6. This clone isolated from a human liver cdna library is strictly identical to the cdna sequenced by Gonzalez et al. [12]. Microsome preparation and determination of CYP2D6 protein content Microsomes were prepared as described elsewhere [2]. Total P-450 content was determined according to Greim [13] and protein by the procedure of Lowry et al. [14]. Microsomal proteins were subjected to electrophoresis on a SDS/9'h absorbancemg protein n- " fetuses fetuses newborns newborns newborns newborns adults GOW,30w t24h 1-7days 7-28days,28dByS Fig. 1. Age-related variation of CYP2D6protein in human liver. Results (expressed as absorbance. mg protein-') are the mean _+ SEM of individual values. Group 1 refers to fetuses less than 30 weeks of gestational age (n = 60, 23.2 f 3.8 weeks) group 2 fetuses aged more than 30 weeks (n = 15,34.6 _+ 2.4 weeks), group 3 newborns aged less than 24 h (n = 11, gestational age 29.3 f 3.3 weeks, 0.5 _+ 0.3 days postnatal), group 4 newborns aged 1-7 days (n = 13, gestational age 33.8 & 4.5 weeks, days postnatal), group 5 newborns aged 8-28 days (n = 8, gestational age 31.5 f 5.9 weeks, 13.9 f 9.5 days postnatal), group 6 infants aged 4 weeks to 5 years (n = 12, gestational age 31.2 f 5.2 weeks, days postnatal) and group 7 are samples from adult extensive metabolizers (n = 7) polyacrylamide gel [15]. After migration, proteins were transferred to nitrocellulose [16] and the CYP2D6 protein was immunochemically detected with the serum from a patient with chronic active autoimmune hepatitis containing anti- LKMl antibodies. The serum was diluted l/looo and the antigen-antibody complex was visualized after addition of peroxidase-conjugated anti-(human IgG) antibodies (Dako, Denmark), using 4-chloronaphthol as dye. The amount of reacting material was measured by scanning the nitrocellulose. Reference adult samples were routinely incorporated in each experiment to calibrate determinations. Dextromethorphan demethylase activity Microsomal protein (0.3 nmol total P-450) was preincubated for 10 min at 37 C with the NADPH-generating system (1 mm NADP, 5 mm isocitrate, isocitrate dehydrogenase 10 units/ml and 5mM MgC12) in 0.1 M sodium phosphate ph 7.4. The reaction was started by addition of 500 pm racemic dextromethorphan. The reaction was stopped after 30 min by addition of 0.1 M citrate ph 2.6. Denaturated proteins were precipitated by centrifugation. The production of dextrorphan and 3-methoxymorphinan was determined by HPLC with fluorescence detection. The extraction procedure, chromatographic conditions and equipment have been reported previously [I 71. Dextromethorphan hydrobromide and dextrorphan tartrate were generously provided by Hoffmann-La Roche company (Basle, Switzerland). RESULTS Ontogenesis of CYP2D6 protein The amount of material present in hepatic microsomes and cross-reacting with anti-lkm1 serum is shown in Fig. 1 :

3 585 c I.- al c e a 3? aj C m n CYP2D6 protein 1 * t *..3 J A1 1 A negative 47% positive 47% 33 0% intermediate negative positive fetuses < 30 weeks fetuses 30 weeks CYP2D6 RNA Gestational Age (weeks) Fig. 2. Age-related variation of CYP2D6 protein and RNA in human fetal liver. (A) Protein (expressed as absorbance mg protein-') and (B) RNA (expressed as the ratio CYP2D6 RNA/actin RNA) were plotted versus gestational age of fetuses (expressed in weeks) results are expressed as absorbance/mass microsomal protein and given as mg protein-'. In adult livers, the mean value was approximately 7 mg protein-' in most samples which are considered as extensive metabolizers. In an unique sample, the concentration was very low (< 0.01 mg protein-') and likely corresponded to a poor metabolizer. Fetal, neonatal and infant samples have been divided into six groups based upon gestational or postnatal age. Groups 1 and 2 included fetuses aged less and more than 30 weeks of gestation, respectively: mean values accounted for about 5% of the adult average value and were only marginally affected by age (Figs 1 and 2A); this was also true in neonates less than 24 h old (group 3). In addition, the CYP2D6 protein content was independent of gestational age at delivery (between weeks). The protein level began to rise in newborns aged 1-7 days (group 4) and steadily increased in neonates of 7-28 days and in infants over 28 days of age (groups 5 and 6). In each group, interindividual variations were very important. Several explanations could account for the dispersion observed: first, a large variability is known to exist in adult samples and is expected in fetal and neonatal ones. Second, based on a 70/0 polymorphism ratio within the Caucasian population, a few poor metabolizers may be present in fetal and neonatal groups lowering the average values. Third, a rapid degradation of the CYP2D6 protein could occur in some samples in relation to the delay between death and tissue freezing. This is unlikely because the CYP3A protein level (index of a limited degradation of P-450s; data not shown), and the N-demethylation rate of dextromethorphan remain positive in all samples (see below). Lastly, the very large variation occurring in fetuses resulted from the association of two clearly different populations: most of the fetuses were devoid BI B intermediate 51% negative spontaneous abortions therapeutic abortions Fig. 3. CYP2D6protein content in human,fetal liver. C'YPZD6 content was estimated in fetuses in relation to gestational age (A) or to the cause of abortion (B). Samples were classified as negative (< 0.02 mg protein-'), intermediate (0.02 < 0.35 mg protein-') or positive (> 0.35 mg protein-') of any detectable CYP2D6 protein, whereas in part of the population, the protein was readily detected and quantified. When data were carefully examined, three subgroups could be defined: a first group, called negative, where the protein was totally absent; a second group of fetuses where the protein reached a mean value > 0.35 mg protein-' (approximately 5% of adult value) and called positive; a third group including all intermediate but detectable values. Among fetuses aged less than 30 weeks (n = 60), only 30% were positive (Fig. 3A). This proportion rose to 47% in fetuses over 30 weeks (n = 15). In these positive samples, the mean concentrationwas and 1.10 f 0.34 mgprotein-',e.g. the value observed in group 4 (1-7 days newborns) where all the samples were positive. This observation suggested that the gestational age had only a moderate influence by increasing the number of positive samples associated with a limited augmentation of the CYP2D6 protein concentration (Fig. 2A). The cause of abortion was also considered: using the same classification as negative, intermediate and positive values, samples have been divided into spontaneous (n = 27) versus therapeutic (n = 38) abortions. In fetuses from spontaneous abortions (all were aged less than 30 weeks) 65% were positive (Fig. 3B) compared with 18% in therapeutic abortions (e.g. induced by prostaglandin), and 70% of positive were spontaneously delivered fetuses compared to only 20% with intermediate values. From a qualitative point of view, no difference in the apparent molecular mass of the protein was noticed in SDS/polyacrylamide gels (data not shown).

4 CYP2D6 RNA/Actin RNA dextror phan 8 nmo1.h-1.mg protein T " fetuses fetuses newborns newborns newborns newborns adults '30w '3OW <24h l-7days 7-28days,28days Fig. 4. Age-reluted variation of CYP2D6 RNA in human liver. Slotblots of RNA were probed with CYPZD6 cdna and actin DNA. Results are the mean +. SEM of individual values calculated as the ratio P-450/actin RNA. Groups are defined in Fig. 1 methoxymorphinan nmo1.h-l.mg proteln-1 7- Ontogenesis of'cyp2d6 RNA On the same samples, total RNA was extracted and subjected to slot-blot and Northern blot analysis. First, only one band was detected in samples (data not shown) and the size of the CYP2D6 RNA (1.8 kb) was identical in adult and neonatal livers. The evolution of the CYP2D6 RNA content, examined by slot-blot, was clearly different from the onset of the corresponding protein (Fig. 4): the concentration gradually rose in fetal and early neonatal samples to reach a value exceeding adult ones in groups 4-6, before declining in adulthood. Like the protein content, the CYP2D6 RNA concentration in fetal samples was distributed into two distinct populations: one negative (even if other RNAs could be detected) and one positive where the RNA concentration steadily increased with gestational age (Fig. 2B). When the RNA concentration was plotted versus the protein concentration, a close relationship was noticed in adult samples (Y = 0.95, P < 0.005) in contrast with developing livers where no positive correlation could be clearly reported. Dextromethorphan metabolism Dextromethorphan undergoes two pathways for its metabolism leading to the formation of three metabolites: 0- and N-monodemethylated and the didemethylated derivatives [18]. In vitro, only the two monodemethylated metabolites are formed in sufficient amount to be accurately assayed. In order to determine whether or not the protein detected in immunoblots was active, we have examined the in vitro metabolism of dextromethorphan in microsomes from fetal, neonatal and adult livers. To facilitate the comparison with previous data, results were analyzed using the same classification as for the CYP2D6 protein. In adult liver samples, the formation of 0-demethylated dextromethorphan (dextrorphan) was higher than the formation of the N-demethylated metabolite (3-methoxymorphinan). In a sample from a suspected poor metabolizer, the absence of CYP2D6 protein was associated with an extremely low rate of 0-demethylation: versus 10.5 nmol. h-'. mg protein-' in extensive metabolizers n- fetuses fetuses newborns newborns newborns adults (30w )30W (24h 1-7 days )8 days Fig. 5. Age-rekutedvariution of dextromethorphan metabolism in humun 1ivc.r. Incubations were carried out with 500 pm racemic dextromethorphan. Results (expressed as nmol metabolites formed in 1 h by 1 mg protein) are the mean SEM of individual values. Groups are defined in Fig. 1, except that groups 5 and 6 were pooled In fetal liver samples, the average 0-demethylation activity accounted for only 1% of the adult value: f nmol. h-'. mg protein-' (Fig. 5). Once again, activity was present in only 5 out of the 17 investigated samples (30%): activity in these 5 positive samples was f nmol. h-'. mg protein-' in fetuses aged less than 30 weeks. Activity rose within the first week after birth but did not exceed 25% of adult activity in groups 5 and 6: activities were identical in these two groups and thus were pooled. The N-demethylation reaction exhibited a different pattern of evolution: activity was present in all fetal samples and remained stable in fetal and neonatal samples but accounted for 20-30% of adult activity. When the 0-demethylation activity was plotted versus the CYP2D6 protein content in microsomes, no positive relationship was observed when all groups were investigated together. Because this was unexpected, results were examined in each group: as shown in Table 1, a positive correlation existed between the two parameters in fetal and neonatal samples. As expected, there was no correlation between the N-demethylation rate and the CYP2D6 protein concentration, and between

5 Table 1. Correlution studies between dextromethorphan metabolism and CYP2D6 protein concentration Activities (expressed as nmol metabolites formed in 1 h by 1 mg protein) were plotted against CYP2D6 protein concentrations (expressed as absorbance mg protein- ') and the equations (slope in boldface type) and correlations of the curves are tabulated Sample Compound Slope Correlation, Y Fetuses Dextrorphan/CY P2D6 y = 0.467~ (n = 24) Dextrorphan/methoxymorphinan Methoxymorphinan/CYP2D Newborns Dextrorphan/CYP2D6 y = 2.87~ + (-0.13) (n = 13) Dextrorphan/methoxymorphinan Methoxymorphinan/CYP2D Adults Dextrorphan/CY P2D6 y = 14.03~ (n = 7) Dextrorphan/methoxymorphinan Methoxymorphinan/CYP2D and N-demethylation. The slope of the regression equation was clearly different: this could indicate that the efficiency of the protein to carry out the 0-demethylation reaction was much higher in adult than in neonatal samples, and in neonatal than in fetal samples. DISCUSSION The bioavailability of a drug and consequently its therapeutic efficiency depends on several factors in the human newborn and neonate. Among those factors are the volume of distribution, the binding to proteins, etc. Another crucial element is the maturity of the enzymatic system implicated in its metabolism. To date, only few data have reported the onset of P-450 isoenzymes in the human liver: these studies concluded on the presence of CYP3A protein in the fetal liver as early as the 14th week of gestation [2, 19-23] and the absence of proteins of the CYP2C subfamily, explaining the very low activity towards benzopyrene and mephenytoin [2, 31. The CYP2D subfamily is a third group of P-450 genes, composed of three genes in the human species, but only one is functional We have accumulated evidence indicating that, during ontogenesis, the surge in CYP2D6 protein is mainly postnatal since its concentration rises in newborns 24 h after birth. This increase of the protein becomes evident in 1- week-old newborns and reaches about two-thirds of the adult concentration in infants aged 1 month to 5 years. The level of protein immunochemically detected in Western blots is associated with the rise in dextromethorphan O-demethylation, one of the probe reactions for the CYP2D6 protein in humans. The rise of CYP2D6 protein in the early newborn seems to be independent of the gestational age at birth (between 25 and 35 weeks). We can thus hypothesize that the event triggering CYP2D6 expression is related to birth itself rather than to a programmed age-dependent evolution. This hypothesis is strengthened by a careful analysis of data collected from fetuses: in newborns and infants, as in adults, all the samples are positive except a few samples from putative poor metabolizers. In fetuses and in early newborns, the picture is clearly different: a fraction of the population is totally deficient in the protein, whereas some samples were definitively positive. The proportion of positive samples increases slightly from group 1 (30%) to groups 2 and 3 (500/,), but remains much lower than the expected proportion of extensive metabolizers (93% in the adult Caucasian population). In these positive samples, the concentration of CYP2D6 protein reaches a value comparable to that observed in newborns aged 1-7 days: this supports the assumption that the onset of CYP2D6 expression was identical whatever the age, e. g. independent of gestational age, but could rather be related to the delay between the initiation of the activation process and sampling. From its developmental pattern in the neonatal liver, the rise in CYP2D6 protein can be attributed either to the loss of a maternal repressing factor or to stimulation by a neonatal intrinsic factor expressed immediately after birth. However, this cannot account for the precocious surge observed in some fetuses. If we hypothesize that the plasma concentration in hormones involved in the maintenance of pregnancy negatively regulates the expression of CYP2D6 in the fetus, spontaneous abortion as well as delivery suddenly cut down the impregnation of the newborn liver by maternal or placental hormones, and in turn switch on the expression of CYP2D6. This could explain why 65% of spontaneously aborted fetuses exhibit a positive concentration of CYP2D6 protein compared to only 18% in therapeutically aborted fetuses. Only speculations can be drawn from these clinical observations and they require experimental data to be conclusively accepted. In the human liver, a single functional gene codes for the CYP2D6 protein [24]. Data presented in this study clearly establish that gene transcription could be activated in the fetal and neonatal liver and precedes the rise in CYP2D6 protein. When individually examined, no correlation was found between RNA accumulation and protein concentration, although such a correlation is found in adult samples. In addition, RNA concentration is markedly lower in the adult liver than in the liver of newborns or infants. Such results were unexpected and might suggest that the CYP2D6 protein synthesis is less efficient in fetal and neonatal liver than in adult liver. Another unexpected result is the different catalytic activity of dextromethorphan 0-demethylase during development: within a defined age group, activity is quite constant, but varies from the fetal to the neonatal and to the adult age. No gross modification in the molecular mass of the gene product has been observed by Western blotting. Since only one gene is expressed in the human genome, modifications of enzymatic activity might be due to minor changes in protein conformation and/or affinity for the substrate. It has been repeatedly reported that activity of membrane-bound enzymes is dependent on the phospholipid composition of the membrane: interestingly, the lipid apparent viscosity of the microsomal membrane is much higher in fetal than in adult rat liver, in

6 588 relation to phospholipid concentration [25]. The same researchers have reported similar conclusions for human liver microsomes (Weil, E., et al., cited in [26]). Thus, by increasing the phospholipid concentration in the endoplasmic reticulum membrane, a relative fluidization might occur during development which modifies the catalytic efficiency of P-450 either directly changing the conformation of CYP2D6 protein or indirectly via a better substrate accessibility to the active site or a higher rate of electron transfer from NADPHcytochrome P-450 reductase to the P-450 moiety. Such a difference in the catalytic properties of P-450 has already been described in fetal compared to adult liver microsomes for the CYP3A subfamily, but in that case a modification of the protein itself cannot be ruled out [2, 20,211. In conclusion, this extensive study conducted with fetal and neonatal human liver demonstrates that ontogenesis of CYP2D6 protein is essentially postnatal, but could also occur in fetuses and might be associated with spontaneous abortion. This surge in the protein level is associated with the onset of dextromethorphan 0-demethylation. The evolution of the CYP2D6 RNA concentration shows an earlier rise reaching a maximum in neonatal samples. Thus, the regulation of CYP2D6 is primarily transcriptional and could be secondarily modulated by post-transcriptional events. The authors wish to thank Prof. M. Forest and Dr F. Narcy (Dept Anatomie Pathologique, Hbpital Cochin, Paris), Dr P. Mussat (Dcpt Neonatalogie, Hbpital Cochin, Paris) and Dr A. Lemoine (IN- SERM U75) for providing fetal, neonatal and adult liver samples. REFERENCES I. Nebert, 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, Cresteil, T., Beaune, P., Kremers, P., Celier, C., Guengerich, F. P. & Leroux, J. P. (1985) Eur. J. Biochem. 151, Shimada, T., Misono, K. S. & Guengerich, F. P. (1986) J. Biol. Chem. 261, Alvan, G., Bechtel, P., Iselius, L. & Gundert-Remy, U. (1990) Eur. J. Clin. Pharmacol. 39, Mahgoub, A,, Idle, J. R., Dring, L. G., Lancaster, R. & Smith, R. L. (1977) Lancet 2, Dayer, P., Kubli, A., Kupfer, A,, Courvoisier, F., Balant, L. & Fabre, J. (1982) Br. J. Clin. Pharmacol. 13, Schmid, B., Bircher, J., Preisig, R. & Kupfer, A. (1985) Clin. Phurmacol. Ther. 38, Gonzalez, F. J., Matsunaga, T., Nagata, K., Meyer, U. A., Nebert, D. W., Pastewka, J., Kozak, C. A,, Gillette, J., Gelboin, H. V. & Hardwick, J. P. (1987) DNA 6, Moroy, T., Etiernble, J., Trepo, C., Tiollais, P. & Buendia, M. A. (1985) EMBO J. 4, Cresteil, T., Jaiswal, A. K. & Eisen, H. J. (1987) Arch. Biochem. Biophys. 253, Hamada, H., Petrino, M. G. & Kakunaga, T. (1982) Proc. Natl Acad. Sci. USA 79, Gonzalez, F. J., Vilbois, F., Hardwick, J. P., McBride, 0. W., Nebert, D. W., Gelboin, H. V. & Meyer, U. A. (1 988) Genomics 2, Greim, H. (1970) Arch. Pharmacol. 266, Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, Laemmli, U. K. (1970) Nature 227, Towbin, H., Staehlin, T. & Gordin, J. (1979) Proc. Natl Acad. Sci. USA 76, Jacqz-Aigrain, E., Menard, Y., Popon, M. & Mathieu, H. (1989) J. Chromatogr. 495, Mortimer, O., Lindstrom, B., Laurell, H., Bergman, U. & Rane, A. (1989) Br. J. Clin. Pharmacol. 27, Kitada, M., Kamataki, T., Itahashi, K., Rikihisa, T., Sato, R. & Kanakubo, Y. (1985) Arch. Biochem. Biophys. 241, Wrighton, S. A. & Vandenbranden, M. (1989) Arch. Biochem. Biophys. 268, Ladona, M. G., Park, S. S., Gelboin, H. V., Hammar, L. & Ram, A. (1988) Biochem. Pharmacol. 37, Komori, M., Nishio, K., Fujitani, T., Ohi, H., Kitada, M., Mima, S., Itahashi, K. & Kamataki, T. (1989) Arch. Biochem. Bioph-vs. 272, Komori, M., Kanako, N., Kitada, M., Shimarazu, K., Muroya, K., Soma, M., Nagashima, K. & Kamataki, T. (1990) Biochemistry 29, Kimura, S., Umeno, M., Skoda, R. C., Meyer, U. A. & Gonzalez, F. J. (1989) Am. J. Hum. Genet. 45, Kapitulnik, J., Tshershedsky, M. & Barenholz, Y. (1980) in Microsomes, drug oxidations und chemical carcinogenesis (Coon, M. J., Conney, A. H., Estabrook, R. W., Gelboin, H. V., Gillette, J. R. & O Brien, P. J., eds) pp , Academic Press, New York. 26. Kapitulnik, J., Weil, E. & Rabinowitz, R. (1986) Biochem. J. 239,