Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The level of expression of CYP2B6 in human liver and its metabolic capabilities are still in question. The specificity of several purported antibody, inhibitor and substrate probes of CYP2B6 have been previously examined in this laboratory (Ekins et al., 1997). Both the antibodies examined and the substrate, 7-EFC, were shown to be nonselective for this enzyme. 7-EFC also demonstrated allosteric activation with microsomes derived from B-lymphoblastoid expressed CYP2B6 and CYP2E1 as well as certain human liver microsomal samples. Thus it has been questioned as to whether allosteric activation is an intrinsic characteristic of this enzyme (Ekins et al., 1997). In addition, this same study demonstrated that purported inhibitory and antibody probes for CYP2B6 cross reacted with several other CYPs (Ekins et al., 1997). This laboratory has also confirmed the observation of others (Heyn et al., 1996), suggesting that CYP2B6 may be specifically involved in S-mephenytoin N-demethylation to nirvanol (Ekins et al., 1997). These results indicate that nirvanol formation may be suitable as a marker for CYP2B6.

Numerous laboratories have indicated that hepatic CYP2B6 expression frequency is variable across a population (Shimada et al., 1994), being detected in 20-100% of the livers examined in different studies (Ekins et al., 1997). It was also suggested that CYP2B6 may not be present in all livers (Yamano et al., 1989). This is despite the fact that its mRNA was found in all livers examined (Czerwinski et al., 1994). An explanation for this disparity between measurement of mRNA and protein expression may be due to poor selectivity and sensitivity of antibodies used for detection of CYP2B6. For example, the murine monoclonal antibodies specific to rat CYP2B1 have been shown to cross react with human CYP2E1 (Wrighton et al., 1992). However, recent reports using the monoclonal antibody MAb 49-10-20 to recombinant human CYP2B6 have shown that this antibody does not cross-react with other CYPs and is also immunoinhibitory (Yang et al., 1997, 1998). Recently, liver samples from a Caucasian population demonstrated 3-fold higher levels of CYP2B6 than a Chinese population when detected immunochemically using a polyclonal anti-CYP2B1 antibody which cross reacts with CYP2B6 (Kim et al., 1997). This observation suggests a genetic, environmental or dietary factor may be responsible for these differences between the two populations. Cultured human hepatocytes have also been utilized to show that CYP2B6 is induced by the CYP3A inducers rifampicin and dexamethasone as well as the standard CYP3A/2B inducer, phenobarbital (Strom et al., 1996), indicative that CYP2B6 may be co-regulated with CYP3A. However, the specificity of the polyclonal anti-CYP2B1 antibody used in both of these cases (Kim et al., 1997; Strom et al., 1996) is questionable due to our previous finding of its cross reactivity with other CYPs (Ekins et al., 1997).

Although CYP2B6 is suggested to represent less than 0.2% of total human hepatic P450 (Shimada et al., 1994), it has been shown to be capable of catalyzing the oxidation of a number of structurally diverse xenobiotics that may be clinically significant. The number of literature examples of xenobiotics and the reactions catalyzed by CYP2B6 continues to amass (Ekins et al., 1997; Rendic and Di Carlo, 1997), although there have been no definitive characterization studies describing absolute requirements of substrates or inhibitors for CYP2B6.

The current study describes the further evaluation of the monoclonal antibody MAb 49-10-20 for CYP2B6 (Yang et al., 1997) and confirms its selective nature. Using this antibody, the expression levels of CYP2B6 in a phenotyped liver bank were determined and correlated with various selective CYP probes for CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4. In addition, motivated by the prior observation of the atypical Michaelis-Menten kinetics of 7-EFC metabolism by CYP2B6, we have analyzed six further commercially available substrates of this enzyme to assess whether atypical kinetics could be an important characteristic potentially enabling the differentiation of CYP2B6 activity from that of other CYPs.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. 3CN-7-EC, 3CN-7-HC, 4Cl-7-EC, 4Cl-7-HC, resorufin and benzyloxyresorufin were obtained from Molecular Probes (Eugene, OR). Testosterone, hydroxytestosterone standards, NADPH, benzphetamine, flunitrazepam and acetylacetone were purchased from Sigma Chemical Co. (St. Louis, MO). TBA was purchased from Aldrich Chemical Co. (Milwaukee, WI). Formaldehyde was obtained from EM Science (Gibbstown, NJ). Midazolam and 1'-hydroxy midazolam were gifts from Hoffman La Roche (Nutley, NJ). Methanol and acetonitrile were purchased from Burdick and Jackson (Muskegan, MI). Microsomes prepared from control cells and human B-lymphoblastoid cell lines expressing CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1 or CYP3A4 and a vector control were purchased from Gentest Corp. (Woburn, MA).

Liver specimens. Human liver specimens were obtained from the liver transplant unit at the Medical College of Wisconsin, Medical College of Virginia and the Pathology Department of the Indiana School of Medicine, under protocols approved by the appropriate committee for the conduct of human research. Microsomes were prepared from these specimens using differential centrifugation (van der Hoeven and Coon, 1974). Liver specimens were characterized for various P450 activities and expression as described previously (Ekins et al., 1997).

Western blots of CYP2B6. Microsomes prepared from human livers, purified P450s and P450s expressed in B-lymphoblastoid cells were resolved by SDS-PAGE (Laemmli, 1970) and transferred to nitrocellulose membranes (Towbin et al., 1979) as described previously (Ekins et al., 1997). The monoclonal antibody to CYP2B6, MAb 49-10-20 (Yang et al., 1997, 1998), was used as the primary antibody. The immunoreactive proteins were detected and quantitated relative to the levels observed in microsomes from HL-A as previously described (Ekins et al., 1997).

Testosterone 16-hydroxylase assay. Initial rate conditions for the formation of 16-hydroxytestosterone were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (1 mg/ml) were preincubated for 3 min at 37°C with 100 mM sodium phosphate buffer (pH 7.4) and 1 mM NADPH in a 250 µl incubation volume. Reactions were initiated by addition of testosterone (2-1000 µM in methanol) and terminated after 3 hr by addition of dichloromethane (6 ml) and progesterone internal standard (10 µl of 0.2 mM stock). After shaking for 10 min followed by centrifugation, the aqueous layer was removed and 5 ml of the organic layer was evaporated at 42°C under nitrogen. Hydroxytestosterone standards (50-4000 pmol) were treated identically.

Benzyloxyresorufin O-debenzylase assay. Initial rate conditions for the formation of resorufin were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (0.5 mg/ml) were preincubated as described above. Reactions were initiated by addition of benzyloxyresorufin (0.25-333 µM in N'N'dimethylformamide) and terminated after 20 min by addition of 125 µl zinc sulfate (5% w/v) and 125 µl saturated barium hydroxide (Lake, 1987). After centrifugation, 300 µl of supernatant were combined with 400 µl of 0.5 M glycine (pH 8.5) and the fluorescence measured by direct fluorimetry at excitation  = 530 nm and emission  = 582 nm and 5 nm slit widths using a Shimadzu RF5000U spectrophotometer (Columbia, MD).

Benzphetamine N-demethylase assay. Benzphetamine N-demethylation was determined by the method of Prough and Ziegler (1977). Initial rate conditions for the formation of formaldehyde were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (0.5 mg/ml) were preincubated as described above. Reactions were initiated by addition of benzphetamine (1-1000 µM in methanol) and terminated after 60 min by addition of 125 µl trichloroacetic acid (20% v/v). After centrifugation, 300 µl of supernatant were combined with 150 µl of Nash reagent (Nash, 1953), heated at 60°C for 30 min then cooled on ice and the absorbance measured at 412 nm using a Beckman DU65 spectrophotometer (Beckman Instruments Inc., Fullerton, CA).

3-Cyano-7-ethoxycoumarin O-deethylase assay. Initial rate conditions for the formation of 3CN-7-HC were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (0.5 mg/ml) were preincubated as described above. Reactions were initiated by addition of 3CN-7-EC (1.25-500 µM in DMSO) and terminated after 20 min by addition of 250 µl acetonitrile. After centrifugation, the supernatant fluorescence was measured by direct fluorimetry at excitation  = 408 nm and emission  = 460 nm and slit widths of 5 nm using a Shimadzu RF5000U spectrophotometer (Columbia, MD). The ability of other P450s to metabolize 3CN-7-EC (50 µM) was assessed using expressed CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4 and a vector control (0.5 mg/ml) incubated as described above for 20 min.

Midazolam 1'-hydroxylase assay. Initial rate conditions for the formation of 1'-hydroxy midazolam were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (0.5 mg/ml) were preincubated as described above but in 200 µl final volume. Reactions were initiated by addition of midazolam (2.5-500 µM in methanol) and terminated after 60 min by addition of 200 µl methanol and addition of 10 µl flunitrazepam (0.01 mg/ml). After centrifugation a 200 µl aliquot of supernatant was removed, loaded into autosampler vials and 50 µl was injected for analysis by HPLC.

4-chlormethyl-7-ethoxycoumarin O-deethylase assay. Initial rate conditions for the formation of 4Cl-7-HC were determined in preliminary studies with microsomes prepared from cell lines expressing CYP2B6. Microsomes (0.5 mg/ml) were preincubated as described above but in a 250 µl final volume. Reactions were initiated by addition of 4Cl-7-EC (1.25-500 µM in DMSO) and terminated after 2 hr by addition of 250 µl acetonitrile containing 50 µM 7-ethoxy-4-methylcoumarin. After centrifugation a 40-µl aliquot of supernatant was analyzed by HPLC. The ability of other P450s to metabolize 4Cl-7-EC (50 µM) was assessed using cDNA expressed CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4 and a vector control (0.5 mg/ml) incubated as described above for 2 hr.

Kinetic evaluation, interpretation and statistical analyses. Kinetic analyses of 16-hydroxytestosterone, resorufin, formaldehyde, 3CN-7-HC, 1'-hydroxymidazolam and 4Cl-7-HC formation were initially assessed by visual examination of Eadie-Hofstee plots to determine whether an allosteric activation mechanism was apparent (Enzyme Kinetics 1.6, Window Chem Software Inc., Fairfield, CA). The estimates for kinetic parameters from these analyses were utilized as initial estimates for non-linear regression analyses (NONLIN, version VO2-G-VAX, Statistical Consultants, Incorp., Lexington, KY) for K_{m (apparent)}, V_{max (apparent)} and, when appropriate, n [number of apparent sites bound by activator (Segel, 1993)] calculations. The best fit to a particular model was determined by examination of (in order of importance) the randomness of the residuals, the sum of squares of residuals and the size of the S.E. of the parameter estimates (first within models, then between models), comparing the best fit weighted model when the difference between models was not obvious. The apparent difference between sum of squares was analyzed using the F-test (Boxenbaum et al., 1974). Correlation analysis, both univariate and multivariate, were performed using JMP version 3.1 (SAS Institute Inc., Cary, NC).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Evaluation MAb 49-10-20. Initial experiments demonstrated that as previously reported (Yang et al., 1997) the CYP2B6 MAb 49-10-20 did not cross react with other human CYPs (data not shown). The variability of expression of immunoreactive CYP2B6 across microsomes from 19 human livers in our study was 100-fold (the lowest, HL-L was 0.7 pmol CYP2B6/mg protein and the highest, HL-O was 71.1 pmol CYP2B6/mg protein) and is displayed as immunoquantified levels normalized to those found with microsomes from HL-A (table 1).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Correlation of immunoquantified levels of CYP2B6 with midazolam 1'-hydroxylation in a bank of 19 human liver samples as described in "Materials and Methods."

Enzyme kinetic evaluation of substrates of CYP2B6 in vitro. After preliminary studies to determine the linearity of metabolite formation with respect to amount of protein and incubation time, the kinetic values for benzyloxyresorufin O-deethylation, benzphetamine N-demethylation, 3-cyano 7-ethoxycoumarin O-deethylation, midazolam 1'-hydroxylation, 4-chloromethyl-7-ethoxycoumarin O-deethylation and testosterone 16-hydroxylation were determined using expressed CYP2B6. Most of these substrates demonstrated classic Michaelis-Menten kinetics (table 2). However, examination of the kinetics of testosterone 16-hydroxylation with expressed CYP2B6 by Eadie-Hofstee plots (not shown) suggested that substrate activation was occurring. The data were then modeled using the Michaelis-Menton equation and the Hill equation. The data best fit to the Hill equation yielding an apparent K_m value of 50.5 µM and an n of 1.3 for the number of substrate binding sites. The contribution of endogenous P450 activity in the microsomes from control cells incorporating the vector, was negligible for all of these reactions.

Incubation of 7-ethoxycoumarin derivatives with cDNA expressed CYPs. Previous studies had suggested that 7-EFC was a selective substrate of CYP2B6 (Code et al., 1995) until it was shown that multiple CYPs had a role in its metabolism (Code et al., 1997; Ekins et al., 1997). Therefore the ability of cDNA expressed CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4 to form 3CN-7-HC and 4Cl-7-HC from respectively, 3CN-7-EC and 4Cl-7-EC, at a concentration of 50 µM was examined. When expressed as pmol of product/min/pmol P450, 3CN-7-HC was formed to the greatest extent by CYP1A2, followed by CYP2C19, CYP2B6, CYP2E1, CYP2A6 and CYP2D6 (fig. 2). 4Cl-7-HC was formed at the greatest rate by CYP2C19, followed by CYP1A2, CYP2C8, CYP3A4, CYP2B6 and the other CYPs except CYP2C9 (fig. 3).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   3CN-7-HC formation by human B-lymphoblastoid cell expressed CYPs. Rates of formation of 3CN-7-HC were determined in duplicate for incubations containing microsomes (0.5 mg/ml), NADPH (1 mM) and 3CN-7-EC (50 µM) as described in "Materials and Methods." con, Control microsomes from B-lymphoblastoid cells.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   4Cl-7-HC formation by human B-lymphoblastoid cell expressed CYPs. Rates of formation of 4Cl-7-HC were determined in duplicate for incubations containing microsomes (0.5 mg/ml), NADPH (1 mM) and 4Cl-7-EC (50 µM) as described in "Materials and Methods." con, Control microsomes from B-lymphoblastoid cells.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

As previously reported (Yang et al., 1997, 1998) the CYP2B6 monoclonal antibody MAb 49-10-20 did not cross react with other CYPs (data not shown). This is in contrast to our previous observations with other antibodies used to determine CYP2B6 expression in 14 human liver samples (Ekins et al., 1997). The fact that all the human liver samples showed immunodetectable CYP2B6 in this study is also in contrast to previous reports which have illustrated CYP2B6 expression frequencies between 20% (Baker et al., 1995) and 90% (Kirby et al., 1993) in human livers characterized with different rabbit anti-rat CYP2B antibodies. Our evidence for all livers possessing CYP2B6 protein is in complete agreement with studies showing all livers possess CYP2B6 mRNA (Czerwinski et al., 1994). The range of immunochemically determined CYP2B6 expression in this study is similar to that described by others (Code et al., 1997) and is larger than the ranges for expression of both inducible and constitutive CYPs (CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2D6, CYP2E1 and CYP3A) characterized in our human liver bank (Wrighton et al., 1993). The 100-fold variability of interindividual expression levels of human hepatic CYP2B6 as shown in this study could be therapeutically important and may be mediated by multiple factors such as other drugs, diet, environment and genetics. Such factors as these may be responsible for the previously reported ethnic differences in CYP2B6 expression in which Chinese and Japanese human liver samples appear to express lower levels of CYP2B6 than Caucasian liver (Kim et al., 1997; Shimada et al., 1994).

The bank of human liver microsomes used in the present study can be used for correlation analysis to indicate relationships between each characterized enzyme and the enzyme being studied (Ekins et al., 1997). Using MAb 49-10-20, the immunoquantified levels of CYP2B6 correlated to a greater extent than the previously determined levels using an anti-rat CYP2B polyclonal antibody (which recognized an immunoreactive band thought to contain CYP2B6 and CYP2C19) with both nirvanol (r² = 0.89 compared with r² = 0.77) and 7-HFC (r² = 0.81 compared with r² = 0.63) formation (Ekins et al., 1997). This strong correlation between nirvanol formation and CYP2B6 levels determined with MAb 49-10-20 provides further evidence that the formation of nirvanol from S-mephenytoin is principally via CYP2B6. The correlation results presented here combined with the work of Heyn et al. (1996) using a CYP2B1 polyclonal antibody as an inhibitor of nirvanol formation by B-lymphoblastoid expressed CYPs, clearly demonstrates that the formation of nirvanol from S-mephenytoin is a specific probe for CYP2B6 (Heyn et al., 1996).

In our study, several probes for determining expression of CYP3A in the human liver bank, namely erythromycin N-demethylation, midazolam 1'-hydroxylation and immunoquantified CYP3A4 levels (Wrighton et al., 1993), also significantly correlated with immunoquantified CYP2B6 levels that suggested a relationship in the expression of CYP2B6 and CYP3A4. However, this relationship was found to be heavily influenced by the two livers with the highest CYP2B6 levels as their removal decreased the significance of the correlation (P > .05). Therefore, it is tempting to speculate that the basal or constitutive expression of CYP2B6 and CYP3A4 is not co-regulated but they may be induced by similar agents. This is indeed the case with respect to CYP2B and CYP3A in various animal species (e.g., rat).

Previously we have shown that metabolism of the coumarin derivative 7-EFC by expressed CYP2B6 demonstrated atypical enzyme kinetics and the kinetic data best fit the Hill equation, indicative of autoactivation of the enzyme by its substrate (Ekins et al., 1997). In an effort to discover whether other CYP2B6 substrates demonstrate autoactivation, we examined a number of substrates with a focus on coumarin derivatives. For these studies we used CYP2B6 expressed in microsomes from B-lymphoblastoid cells. Both coumarin derivatives examined, 3CN-7-EC and 4Cl-7-EC, illustrated Michaelis-Menten kinetics (table 2). Thus, because 7-EFC displayed atypical enzyme kinetics (Ekins et al., 1997) although 3CN-7-EC and 4Cl-7-EC did not, not all coumarin derivatives, as well as not all substrates appear to autoactivate CYP2B6.

Initially, 7-EFC was suggested as a probe for CYP2B6 but was subsequently shown to be metabolized by multiple CYPs (Code et al., 1997; Ekins et al., 1997). This lead us to investigate whether the other 7-ethoxycoumarin analogs, 3CN-7-EC and 4Cl-7-EC, might show a higher degree of selectivity for CYP2B6. When expressed as pmol of product/min/pmol P450, 3CN-7-HC and 4Cl-7-HC were formed to the greatest extent by CYP1A2 (fig. 2) and CYP2C19 (fig. 3), respectively. It is also important to note the considerable number of CYPs capable of metabolizing these and other coumarin derivatives including 7-ethoxycoumarin (Yamazaki et al., 1996) and 7-EFC (Ekins et al., 1997; Code et al., 1997). Therefore these compounds are not suitable as selective substrate probes for the catalytic activity of CYP2B6 in human liver microsomes due to the potential for metabolism by other CYPs.

The kinetics of other suggested CYP2B6 probes were examined for the first time to identify whether alternative structural classes of compounds may yield atypical Michaelis-Menten kinetics with CYP2B6. Benzyloxyresorufin is a nonselective substrate which is metabolized by most expressed CYPs including CYP2B6 (Waxman et al., 1991). In this study benzyloxyresorufin O-demethylation demonstrated Michaelis-Menten kinetics with a K_{m (apparent)} of 1.3 µM (table 2). Benzphetamine N-demethylation has been widely used as a CYP probe (Blanck et al., 1983). In the current study, benzphetamine N-demethylation demonstrated Michaelis-Menten kinetics (table 2) with a K_{m (apparent)} of 93.4 µM.

As a significant correlation of midazolam 1'-hydroxylation with 7-EFC metabolism by the bank of human liver microsomes and with immunoquantified levels of CYP2B6 was observed, an investigation of whether CYP2B6 could form 1'-hydroxy midazolam was undertaken. This benzodiazepine tranquilizer has been extensively reported to be metabolized by human hepatic CYP3A (Kronbach et al., 1989) and is used as an in vivo probe for this enzyme (Thummel et al., 1994). The K_{m (apparent)} for midazolam 1'-hydroxylation in human liver microsomes varies from study to study; 0.28 to 12.6 µM (Kronbach et al., 1989; Thummel et al., 1994; Gorski et al., 1994). These values are slightly different to the K_{m (apparent)} value reported for purified CYP3A4; 43.5 µM (Gorski et al., 1994) and cDNA expressed CYP3A4; 1.56 µM (Ghosal et al., 1996). Until now, no other enzyme (other than the highly related CYP3A5) has been identified as able to catalyze the 1'-hydroxylation of midazolam. However, as shown in table 2, CYP2B6 is a midazolam 1'-hydroxylase yielding typical Michaelis-Menten kinetics and a K_{m (apparent)} of 46.1 µM, comparable to that of purified CYP3A4 but higher than that obtained with expressed CYP3A4 and human liver microsomes. Therefore, it is possible that the variation in CYP2B6 expression levels may contribute to the differing K_{m (apparent)} values for 1'-hydroxylation of midazolam, described above in human liver microsomes.

CYP2B6 has been shown to be the only human CYP capable of testosterone 16- and 16-hydroxylations (Imaoka et al., 1996). In the current study the K_{m (apparent)} value for testosterone 16-hydroxylation with expressed CYP2B6 was found to be 50.5 µM (table 2) and substrate activation was observed, as the data best fit to the Hill equation (n = 1.3). Therefore, in addition to 7-EFC, a further substrate of CYP2B6 has been shown to activate its own metabolism.

Hill-type cooperative kinetics appear to be a growing aspect of many reports that describe the kinetics of the CYPs. As more data of this type are presented there has been great speculation regarding the CYP active site(s) (Ueng et al., 1997). Studies of carbon monoxide binding to P450BM-3 in the presence of a substrate demonstrated complex kinetics that were explained as due to multiple conformations of the enzyme, specifically open and closed states (McLean et al., 1996). The effects of substrates on binding kinetics of carbon monoxide in mammalian P450's have been similarly described (Koley et al., 1997) and it has been suggested that mammalian CYPs consist of multiple conformers (Koley et al., 1996). Our understanding of enzyme function, and particularly that of CYP, is evolving to incorporate a less rigid lock and key hypothesis which implies a more flexible binding (active) site that may be difficult to predict using present modeling techniques (Jorgensen, 1991).

It is postulated that with some substrates, the conformation of CYP2B6 may be altered, resulting in autoactivation. One important aspect of the current study is that it identifies six additional substrates of CYP2B6, only one, testosterone, demonstrated substrate activation, a characteristic we had previously observed with 7-EFC (Ekins et al., 1997). We also investigated two coumarin derivatives structurally similar to 7-EFC that did not behave kinetically in the same way as 7-EFC or testosterone. Our observations may be useful in defining characteristics of substrates important for autoactivation of CYP2B6 and modeling and understanding the active site(s) of this enzyme.

Physiologically, the CYPs will likely be faced with endogenous substrates as well as drug(s), dietary components and other xenobiotics. The many in vitro and in vivo studies examine the interaction of usually only two compounds with a single enzyme. This may be considered a disadvantage of in vitro studies as xenobiotics, exogenous and endogenous substrates may kinetically alter the characteristics of the enzyme and its interaction with substrates. Recently, furocoumarin derivatives present in grapefruit juice have been identified as specific inhibitors of CYP3A4 (Fukuda et al., 1997). In our study, two additional coumarin derivatives were identified which can be metabolized by CYP2B6 and several other CYPs. This suggests many other naturally occurring coumarin derivatives may also be substrates or inhibitors for CYP2B6 as well as other CYPs. However, as indicated above, it is important to realize in modeling in vitro data that the system is quite simple compared to that found in vivo.

In summary, our study further characterizes a highly specific monoclonal antibody for CYP2B6 which was used to demonstrate a 100-fold variation in the levels of expression of this protein in a bank of 19 human liver samples. In addition, our data combined with that previously reported (Heyn et al., 1996) clearly demonstrate that S-mephenytoin N-demethylation is a suitable selective probe for CYP2B6 catalytic activity. Furthermore, six additional CYP2B6 substrates were examined with one, testosterone, demonstrating autoactivation.