Abstract
The role of cytochrome b5 (b5) in the α-naphthoflavone (α-NF)-mediated inhibition of H2O2-supported 7-benzyloxyquinoline (7-BQ) debenzylation by heterologously expressed and purified cytochrome P450 3A4 (CYP3A4) was studied. Although α-NF showed negligible effect in an NADPH-dependent reconstituted system, inhibition of 7-BQ oxidation was observed in the H2O2 system. Analysis of the effect of various constituents of a standard reconstituted system on H2O2-supported activity showed that b5 alone resulted in a 2.5-fold increase in the kcat value and reversed the inhibitory effect of α-NF. In addition, titration with b5 suggested that only 65% of the CYP3A4 participated in the interaction with b5, consistent with cytochrome P450 (P450) heterogeneity. Study of the influence of b5 on the kinetics of H2O2-dependent destruction of the P450 heme moiety suggested two distinct conformers of CYP3A4 with different sensitivity to heme loss. In the absence of b5, 66% of the wild-type enzyme was bleached in the fast phase, whereas the addition of b5 decreased the fraction of the fast phase to 16%. Finally, to locate amino acid residues that might influence b5 action, several active site mutants were tested. Substitution of Ser-119, Ile-301, Ala-305, Ile-369, or Ala-370 with the larger Phe or Trp decreased or even abolished the activation by b5. Ser-119 is in the B′-C loop, a predicted b5-P450 interaction site, and Ile-301 and Ala-305 are closest to the heme. In conclusion, the interaction of b5 with P450 apparently leads to a conformational transition, which results in redistribution of the CYP3A4 pool.
Cytochrome P450 3A4 (CYP3A4) is the most abundant P450 enzyme in human liver and plays a significant role in the metabolism of a wide variety of drugs (Guengerich, 1999; Nebert and Russell, 2002). CYP3A4 demonstrates homotropic cooperativity (sigmoidal v versus S plot) with a number of substrates (Ueng et al., 1997; Harlow and Halpert, 1998; Hutzler and Tracy, 2002). The enzyme is also known to exhibit heterotropic cooperativity, which is characterized by increased oxidation of one substrate in the presence of an effector, such as α-naphthoflavone (α-NF), which may also serve as a substrate (Harlow and Halpert, 1998; Domanski et al., 2000, 2001; He at al., 2003). Interestingly, activation by α-NF of CYP3A4 activity is reversed when the reaction is carried out using cumene hydroperoxide (CuOOH), H2O2, or iodosobenzene (Ueng et al., 1997). In addition, CYP3A4 cooperativity may be influenced by the levels of the redox partners CPR and b5 relative to the P450. Therefore, it is likely that the different effects of α-NF on CYP3A4 in the NADPH versus peroxide-supported reactions reflect the reconstitution conditions, although in one study the removal of b5 from the standard reconstituted system did not prevent stimulation by α-NF of carbamazepine 10,11-epoxidation at a single concentration of substrate, α-NF, and b5 (Ueng et al., 1997).
The role of CPR as a supplier of electron(s) for P450-mediated oxidations is well studied. However, the mode of action of b5 remains controversial (reviewed in Schenkman and Jansson, 2003). Although the role of b5 as a source of electrons for P450 is well known (Schenkman and Jansson, 1999; Guryev et al., 2001; Yamazaki et al., 2001), increasing evidence points to an allosteric effect of b5 mediated in part by an effect on the P450 spin state (Tamburini and Gibson, 1983; Jansson et al., 1985; Reed and Hollenberg, 2003a,b). The modulatory effect of b5 is further supported by the fact that b5 not only increases P450 activity but in some cases also inhibits activity (Reed and Hollenberg, 2003a,b; Yamaori et al., 2003). For instance, b5 activates oxidation of several drugs by CYP3A4 and 3A5, while inhibiting the activity of CYP3A7 (Yamaori et al., 2003). In addition, the interactions with cytochrome b5 may affect the degree of oligomerization of cytochrome P450 in the membranes, as demonstrated for P450 1A2 by rotational diffusion measurements (Yamada et al., 1995). High pressure spectroscopic studies with CYP3A4 and CYP2B4 in solution and in the microsomes have revealed the presence of two conformers with different positions of spin equilibrium and different barotropic properties (Davydov et al., 1995, 2003). It is likely that the heterogeneity of the P450 pool arises from protein-protein interactions between different P450 molecules and between P450 and its redox partners.
More recently, an allosteric effect of redox partners on CYP3A4 was demonstrated by changes in N-protoporphyrin regioisomer formation from phenyldiazene (Yamaguchi et al., 2003) Substitution of certain active site residues, as in S119W, T309A, and A370F, also caused a significant change in regioisomer formation and amplified the effects of the redox partners. This study strongly suggested that a conformational change occurs due to protein-protein interactions between CYP3A4 and b5 and modulates active site topology. The present study investigates the role of b5 in H2O2-supported 7-benzyloxyquinoline (7-BQ) debenzylation by CYP3A4. The results suggest an important mechanism by which b5 modulates CYP3A4 activity and provide further evidence for P450 heterogeneity.
Materials and Methods
Materials. 7-Benzyloxy-4-(trifluoromethyl)coumarin (7-BFC) and 7-BQ were purchased from BD Gentest (Woburn, MA), and 7-ethoxy-4-(trifluoromethyl)coumarin (7-EFC) was purchased from Molecular Probes (Eugene, OR). α-NF was obtained from Sigma-Aldrich (St. Louis, MO). Recombinant CPR and b5 from rat liver were prepared as described previously (Harlow and Halpert, 1997). All other chemicals were of the highest grade available from standard commercial sources.
Expression and Purification of CYP3A4 and Mutants. CYP3A4 wild type and mutants were expressed as His-tagged proteins in Escherichia coli TOPP3 and purified using Talon metal affinity resin (BD Biosciences Clontech, Palo Alto, CA), as described previously (Harlow and Halpert, 1998; Domanski et al., 2001). P450 contents were determined by measuring reduced carbon monoxide difference spectra. Protein concentration was determined with the bicinchoninic acid protein assay kit (Pierce Chemical, Rockford, IL) and bovine serum albumin as a standard. The specific content of CYP3A4 wild type was 15 nmol of P450 per mg protein and that of mutants, 3 to 12 nmol/mg protein (see Table 2).
Preparation of CYP3A4 Apo-Protein. Removal of the heme from CYP3A4 and preparation of the apo-protein was performed by treatment with H2O2 (Uvarov et al., 1990; Pikuleva et al., 1992). We implemented very mild conditions of treatment to prevent any peroxidative damage of the protein. In brief, CYP3A4 was diluted to 50 μM in 100 mM sodium phosphate buffer, pH 7.2, 20% glycerol (buffer A). Catalase was added directly to this solution to a final concentration of 1.7 units/ml. The sample was then dialyzed in a Spectra/Por Type I molecular weight cut-off 6000-8000, 10-cm flat width dialysis bag (Spectrum Laboratories, Rancho Dominguez, CA) against 50 volumes of buffer A containing 100 mM H2O2. Dialysis was performed at 4°C with continuous stirring for 24 h. The sample was removed and was further dialyzed against 50 volumes of 100 mM 4-morpholinepropanesulfonic acid, pH 7.4, 10% glycerol, 1 mM EDTA, 0.2 mM dithiothreitol. Protein concentration was determined using the Bradford protein assay kit (Bio-Rad, Hercules, CA).
7-BQ Oxidation Assay. The enzyme assay was essentially done as described previously with P450eryF (Khan and Halpert, 2002) except for the use of a 300-μl 96-well microtiter plate and a Biomek 2000 robotic sample handling system (Beckman Coulter, Fullerton, CA), as described previously (Kumar et al., 2005). In brief, a substrate mixture of different concentrations (0-200 μM) in the absence and presence of α-NF (as defined in the individual experiments) was prepared in 100 mM Hepes buffer, pH 7.4, with 2% as the final concentration of methanol. The substrate mixture was preincubated with 25 pmol CYP3A4 (in the absence and presence of various redox partners as defined in the individual experiments) for 5 min at room temperature in a 300-μl 96-well microplate before initiation of the reaction with H2O2 (10 mM final). The total reaction volume of the assay was 100 μl. Thus, the concentration of CYP3A4 in the reaction mixture was equal to 0.25 μM, if not otherwise indicated. After 5 min of incubation, the reactions were stopped by adding 340 units (50 μl) of catalase. Subsequently, 50 μl of 100 mM Hepes buffer, pH 7.4, was added before recording the fluorescence intensity at λex = 405 and λem = 510 nm using a fluorescence plate reader (Fluoroskan Ascent; Thermo Electron Corporation, Waltham, MA). In all the cases, the final activity was calculated by comparison to a standard curve. The kcat, S50, and Hill coefficient (n) were determined by fitting the data to the Hill equation using the SpectraLab software package (Davydov et al., 1995).
P450 Heme-Depletion Assay. Determination of the kinetics of CYP3A4 heme depletion in the presence of H2O2 was done using conditions similar to those previously described (Uvarov et al., 1990; Pikuleva et al., 1992). The reaction was carried out at 25°C in 100 mM Hepes buffer, pH 7.4, in a 1-ml semi-microspectrophotometric cell with constant stirring. The reaction mixture contained 1 μM protein and 60 mM H2O2. Bleaching of the hemoprotein was followed by measuring a series of absorbance spectra in the 340- to 700-nm range at 2- to 5-min intervals over a 30- to 60-min incubation time. Each series contained at least 10 spectra. The measurements were made with a Shimadzu-2600 spectrophotometer (Shimadzu, Kyoto, Japan). Determination of the total concentration of the heme protein was done by linear least-squares approximation of the spectra by a linear combination of spectral standards of CYP3A4 low-spin, high-spin, and P420 states (Davydov et al., 2003). Fluctuations of the baseline due to turbidity changes were compensated by polynomial correction in combination with the principal component analysis technique as previously described (Davydov et al., 1995, 2003; Renaud et al., 1996). All data treatment and fitting of the titration curves were performed with our SpectraLab software package (Davydov et al., 1995).
Results and Discussion
Hydrogen Peroxide-Supported 7-BQ Debenzylation by CYP3A4. Previous studies showed that the activity of CYP3A4 and P450eryF in peroxide-supported substrate oxidations is inhibited by α-NF in contrast to the activation observed in NADPH-dependent reactions (Ueng et al., 1997; Khan et al., 2003). To elucidate the mechanistic basis of this apparent paradox, we established a sensitive high-throughput assay of peroxide-supported CYP3A4 activity. The rates of H2O2- and CuOOH-dependent oxidation of a series of fluorogenic substrates, such as 7-BQ, 7-BFC, and 7-EFC, by CYP3A4 were measured in several buffer systems as described previously (Khan and Halpert, 2002). The enzyme showed maximal activity in H2O2-supported oxidation of 7-BQ, whereas the rates of oxidation of 7-BFC and 7-EFC were negligible (data not shown). Furthermore, incubation with 10 mM H2O2 in 100 mM Hepes buffer, pH 7.4, gave a linear rate of reaction for up to 5 min with CYP3A4 wild-type and the mutants in the absence and in the presence of b5 (data not shown). Under these conditions, the turnover in H2O2-supported debenzylation of 7-BQ by purified CYP3A4 reached about 1/15 of the value obtained for the NADPH-dependent reaction in a standard reconstituted system (data not shown).
Effect of the Constituents of the Reconstituted System on H2O2-Supported Oxidation of 7-BQ and Inhibition by α-NF.Figure 1 illustrates the effect of α-NF on CYP3A4-catalyzed H2O2-dependent debenzylation of 7-BQ. In the absence of α-NF, the dependence of v on [S] was given by an S-shaped curve obeying the Hill equation, with kcat, S50, and n values of 1.4 min-1, 72 μM, and 1.8, respectively. Addition of 25 μM α-NF decreased the value of kcat to 0.76 min-1, with no effect on S50 (79 μM) or n (1.9). Noncompetitive inhibition by effectors of CYP3A4 in an NADPH-supported system has previously been suggested to result from competition for protein-bound reactive oxygen, leading to a decrease in the rate of substrate oxidation. (Shou et al., 1994; Ueng et al., 1997; Tracy, 2003). A similar explanation may be valid in the case of the H2O2 system.
To test the hypothesis that the inhibition by α-NF in the H2O2-dependent reaction reflects the lack of interactions of CYP3A4 with other components of a standard reconstituted system, namely CPR, b5, or phospholipids, the effect of these constituents on H2O2-supported oxidation of 7-BQ was assessed (Fig. 1). Addition of CPR, b5, and phospholipid considerably increased the values of kcat, S50, and Hill coefficient to 1.9 min-1, 134 μM, and 2.4, respectively. Importantly, in this case, the values of kcat, S50, and n observed in the presence of α-NF (1.8 min-1, 129 μM, and 2.4, respectively) revealed no inhibition by the compound (Fig. 1). Examination of the individual constituents (Table 1) revealed that the most dramatic effect was exerted by addition of b5, which increased the value of kcat >2.5-fold, with a significant decrease in S50 (Table 1). The increase in kcat caused by b5 was also observed in the presence of α-NF, although in this case, an increase in S50 and a decrease in the n value were also observed (Table 1). Overall, the results strongly support an allosteric effect of b5 on CYP3A4. In contrast, α-NF showed negligible effect in an NADPH-dependent reconstituted system, and removal of b5 caused no significant change in the action of α-NF (data not shown).
Effect of b5 on H2O2-Supported Oxidation of 7-BQ. The role of b5 in the stimulation of CYP3A4 activity was studied in more detail. The changes in the parameters of H2O2-supported oxidation of 7-BQ caused by the addition of increasing concentrations of b5 to 1 μM CYP3A4 is illustrated in Fig. 2. The initial addition of b5 resulted in a sharp decrease in the n value and an increase in S50. Both changes were partially reversed upon further addition of b5. Due to the phenomenological nature of the Hill equation, these dependencies are hard to interpret in mechanistic terms. In contrast, the increases in kcat observed upon addition of b5 appear to be proportional to the amount of CYP3A4-b5 complex formed (Fig. 2). [Addition of apo-CYP3A4 at different P450/b5 molar ratios did not decrease the stimulation by b5, suggesting that any apo-protein present in the original preparation does not interfere with the results (data not shown).] Consequently, we attempted to fit the dependencies of kcat on [b5] with the following equation, which is derived from the canonical relationship for the equilibrium of binary complex formation at comparable concentrations of the interacting species (Segel, 1993, eq. II-53b). Here and represent kcat values characteristic of b5-free and b5-bound CYP3A4, respectively, whereas stays for the effective value of kcat exerted at given concentrations of CYP3A4 ([P450]) and b5, whose complex is characterized by a dissociation constant Kd. Although the square correlation coefficient (ρ2) for these fits was higher than 0.97, they showed a considerable systematic deviation from the experimental data (Fig. 2, solid lines). Based on prior findings of heterogeneity of the CYP3A4 pool (Davydov et al., 2003; Davydov and Halpert, 2004), we hypothesized that the concentration of P450 participating in the interactions with b5 was lower than the total concentration of the enzyme. This hypothesis is consistent with an earlier report, in which the maximal stimulation of testosterone 6β-hydroxylation and nifedipine oxidation was observed at a b5/CYP3A4 M ratio of 0.5 (Yamazaki et al., 1996). To probe this suggestion, the effective concentration of CYP3A4 was treated as one of the parameters of the optimization procedure, together with the values of Kd, , and . Such optimization showed that the best quality of the fitting (ρ2 > 0.996) was obtained by assuming that only 65% of CYP3A4 participates in the interaction with b5 (Fig. 2, dashed lines). This observation strongly supports the suggestion of functional heterogeneity of the CYP3A4 pool. The fitting of the dependencies of on the concentration of b5 yields = 1.4 ± 0.1 min-1, = 5.1 ± 0.2 min-1, and Kd = 0.042 ± 0.02 μM.
Effect of b5 on H2O2-Dependent Heme Depletion. The increased rate of H2O2-dependent substrate oxidation by CYP3A4 in the presence of b5 may be explained if the interactions of the two proteins result in increased accessibility of the heme moiety of CYP3A4 for H2O2. To probe this hypothesis, we studied the effect of b5 on H2O2-dependent destruction of the P450 heme moiety of CYP3A4 (Fig. 3). Interestingly, the kinetics did not fit the equation of a simple pseudofirst order reaction but, rather, the equation of the sum of two exponents, as given below in eq. 2. Here [P450]t is the concentration of P450 at the time t, [P450]0 is the total concentration of the enzyme, kf and ks are the kinetic constants of the fast and slow phases, respectively, and Ff is the fraction of the fast phase. The biphasicity of the kinetics of heme destruction was consistent with conformational heterogeneity of CYP3A4 and suggested the presence of two conformers of the enzyme with different accessibility of the heme moiety to H2O2. Sixty-six percent of the wild-type enzyme in the absence of b5 was represented by the conformer bleached in the fast phase, whereas the addition of b5 decreased the fraction of the fast phase to 16% (Fig. 3). Hence, b5 increases the fraction of the slow phase from 34 to 84%, i.e., approximately 2.5-fold, which is similar to the increase in CYP3A4 activity in the presence of b5. Thus, the interaction of CYP3A4 with b5 does not increase the accessibility of P450 heme moiety for H2O2 but, rather, displaces the partitioning of the enzyme toward the conformer with a slower rate of heme destruction. The results are reminiscent of the findings of Koley et al. (1995), who first proposed conformational heterogeneity of CYP3A4. That study showed that substrates have markedly variable effects on the CO binding kinetics of their target P450 3A4 conformers.
Additional evidence of conformational heterogeneity of CYP3A4 in solution and in microsomal membranes has been obtained recently in studies on pressure-induced transitions (Davydov et al., 2003). Only about 70% of CYP3A4 in solution and about 50% of the enzyme in microsomes of recombinant Saccharomyces cerevisiae was susceptible to a pressure-induced P450 → P420 transition. The results suggested that both in solution and in the membrane, CYP3A4 is represented by two conformers with different positions of spin equilibrium and different barotropic properties. Interestingly, the estimate of 70% for the fraction of pressure-susceptible P450 is close to the value of 66% found for the rapid fraction of H2O2-dependent bleaching of CYP3A4. Importantly, the barotropic heterogeneity of CYP3A4 disappears upon monomerization of P450 oligomers (Davydov et al., 2003). We suggest, therefore, that the interactions of CYP3A4 with b5 affect the partitioning of CYP3A4 conformers by modulation of the architecture of P450 oligomers.
Effect of Substitutions of Active Site Residues on b5-Activation. To probe the structural basis of b5-mediated allosteric modulation of H2O2-supported 7-BQ debenzylation, mutants in SRSs 1, 4, and 5 were targeted because substitutions at these regions (S119W, T309A, and A370F) caused a significant change in regioisomer formation from phenyldiazene and amplified the effect of the redox partners (Yamaguchi et al., 2003). Residues at these sites were replaced by smaller and larger side chains, and kinetic analysis of H2O2-supported 7-BQ oxidation was performed as presented in Table 2. S119A, I301A, I301F, and A370F exhibited an ≥2-fold decrease in the kcat. T309F showed very low activity, and kinetics could not be measured (data not shown). S119W demonstrated a complete loss of enzyme cooperativity (Table 2), suggesting an important site for allosteric modulation.
The most interesting observations, however, were the effect of the active site mutations on b5-stimulated 7-BQ debenzylation (Table 2). Whereas b5 stimulated kcat by ∼2.5-fold in wild-type CYP3A4, the stimulation by b5 was more or less abolished in S119W and I301F, and A305F was inhibited by b5. (The lack of correlation between the specific content of the mutants and their stimulation by b5 suggests that any interactions of b5 with apo-protein may be neglected.) Another important observation was that I369F and A370F demonstrated a loss of cooperativity upon b5 addition. The changes in the profiles of regioisomer formation from phenyldiazene suggest that in S119W (B′-C loop), active site topology is altered upon b5-interaction (Yamaguchi et al., 2003). In this context it is important to note that some basic amino acid residues in the C and C′ helix are thought to be involved in binding between P450 2B4 and b5 (Bridges et al., 1998; Clarke et al., 2004). Likewise, the substitutions at Ile-301 and Ala-305 in the SRS-4 region, which is closest to the heme (Williams et al., 2004; Yano et al., 2004), with larger side chains may cause a structural transition leading to impaired b5-interaction. Recently resolved X-ray crystal structures of CYP3A4 have indicated a Phe-cluster that includes SRS 1 residue Phe-108 and SRS 4 residue Phe-304 (Williams et al., 2004; Yano et al., 2004). It has been suggested that conformational movements in the Phe-cluster could be triggered by interaction with redox partners, which is consistent with our results with active site mutants. Significantly reduced activation by b5 along with abolished enzyme cooperativity upon b5-interaction in A370F suggests an important role of SRS-5, as predicted earlier based on the altered regioisomer formation upon interaction with redox partners (Yamaguchi et al., 2003).
Conclusions
Stimulation of CYP3A4 activity by b5 and its ability to reverse α-NF inhibition of H2O2-supported 7-BQ debenzylation suggests an important role of b5 as an allosteric modulator. The results of titration with b5 on CYP3A4 activity and of heme-depletion measurements in the absence and presence of b5 suggest that the interaction with b5 results in redistribution of the CYP3A4 pool between two distinct populations with different susceptibility to H2O2-dependent bleaching. Furthermore, the active site mutants (S119W, I301F, and A305F), which show abolished b5-activation, suggest an impaired P450-b5 interaction. To our knowledge this is the first report on the effect of b5 in the peroxide-supported system, and illustrates one possible mechanism by which b5 modulates CYP3A4 heterotropic cooperativity. We are now actively engaged in studying the oligomerization-related heterogeneity of P450 as a putative mechanistic basis of CYP3A4 allosteric modulation.
Acknowledgments
We thank You-Ai He and You-Qun He for proving purified CYP3A4 active site mutant proteins. We also thank Drs. Kenneth Johnson and Cheng Wang, Pharmacology and Toxicology, University of Texas Medical Branch, for allowing us to use their fluorescence plate reader (Fluoroskan Ascent).
Footnotes
-
This work was supported by National Institutes of Health Grant GM54995 and Center Grant ES06676.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.004606.
-
ABBREVIATIONS: α-NF, α-naphthoflavone; P450, cytochrome P450; 7-BQ, 7-benzyloxyquinoline; 7-BFC, 7-benzyloxy-4-(trifluoromethyl)coumarin; 7-EFC, 7-ethoxy-4-(trifluoromethyl)coumarin; SRS, substrate recognition site; CPR, NADPH-cytochrome P450 reductase; b5, cytochrome b5; H2O2, hydrogen peroxide; CuOOH, cumene hydroperoxide; P450eryF, 6-deoxyerythronolide B hydroxylase; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; DOPC, dioleoylphosphatidylcholine.
- Received March 2, 2005.
- Accepted April 29, 2005.
- The American Society for Pharmacology and Experimental Therapeutics