Abstract
Recent studies have demonstrated that the catalytic behavior of one cytochrome P450 (P450) enzyme can be influenced by the presence of a second P450. This effect has been observed using reconstituted systems containing reductase, CYP2B4, and CYP1A2, primarily at subsaturating reductase. Addition of 1A2 caused a 75% inhibition of CYP2B4-dependent 7-pentoxyresorufin-O-dealkylation (PROD). Conversely, CYP2B4-dependent benzphetamine (bzp) demethylation did not exhibit this response after CYP1A2 addition. Addition of CYP2B4 to a reconstituted system containing reductase and CYP1A2 caused synergism of CYP1A2-dependent 7-ethoxyresorufin-O-dealkylation (EROD). This behavior was consistent with the formation of heteromeric CYP1A2-CYP2B4 complexes with altered catalytic properties. Although such responses have been documented in reconstituted systems, they have not been demonstrated in microsomal preparations. The goal of the present study was to determine whether such interactions were observed in rabbit liver microsomes. In an effort to detect such changes, we took advantage of the differential effect of CYP1A2 on CYP2B4-selective PROD and bzp metabolism. Rabbits were treated with phenobarbital (PB), β-naphthoflavone (βNF), and both PB + βNF—conditions that enrich microsomes with CYP2B4, CYP1A2, or both enzymes, respectively. Benzphetamine demethylation activity was equivalently elevated in both the PB and the PB + βNF groups, consistent with the induction of CYP2B4 in both groups. In contrast, PROD activity in the PB + βNF group was less than 25% of that found in the PB-treated rabbits. These results demonstrate that the interactions observed in reconstituted systems are not an artifact of reconstitution but are observed under the more natural conditions of the microsomal membrane.
There have been numerous reports related to the organization of cytochrome P450 and its redox partners in the microsomal membrane. Many of these studies have examined the interaction between NADPH-cytochrome P450 reductase and P4501 (Bernhardt et al., 1988;Nadler and Strobel, 1988; Shimizu et al., 1991; Voznesensky and Schenkman, 1992, 1994; Shen and Strobel, 1994; Cvrk and Strobel, 2001) as well as cytochrome b5 and P450 (Yamazaki et al., 1995, 1996; Bridges et al., 1998; Schenkman and Jansson, 1999). These reports have largely focused on the points of interaction between reductase and P450 or reductase and cytochromeb5. Recently, the interactions among multiple P450 enzymes and reductase were examined. The goal of these studies was to determine whether the presence of one P450 enzyme could influence the catalytic properties of another P450 (Cawley et al., 1995). Using mixtures of CYP2B4, CYP1A2, and NADPH-cytochrome P450 reductase reconstituted into dilauroylphosphatidylcholine liposomes, the addition of a second P450 enzyme dramatically alters the activity of the original P450 enzyme in the mixed reconstituted system. This effect is dependent on the substrate present. For example, CYP2B4-dependent 7-pentoxyresorufin-O-dealkylation (PROD) is dramatically inhibited by the addition of CYP1A2. This inhibitory response of CYP2B4-dependent PROD is more pronounced at subsaturating reductase, exhibiting an 80% inhibition but only showing a 20% inhibition at saturating reductase. On the other hand, CYP2B4-dependent benzphetamine dealkylation shows a much weaker inhibitory response at subsaturating reductase and appears to be roughly additive when reductase levels are not limiting. Other substrates actually produce a synergistic stimulation of monooxygenase activities in mixed reconstituted systems. CYP1A2-selective 7-ethoxyresorufin-O-dealkylation (EROD) is stimulated by the addition of a second P450 enzyme (CYP2B4). Under these conditions, the rate is greater than the sum of the rates of the simple binary systems. Synergism of testosterone 6β-hydroxylation has also been reported in mixed reconstituted systems containing reductase, CYP3A4, and CYP1A2 (Yamazaki et al., 1997). Taken together, the interaction among these enzymes is consistent with the formation of heteromeric P450-P450 complexes, which have an altered ability to associate with reductase (Backes et al., 1998).
Although the interaction among reductase and these P450 enzymes has been established in reconstituted systems, it is possible that these enzymes do not respond in a similar manner in the more natural environment of the microsomal membrane. The purpose of this study was to provide evidence for similar interactions among P450 enzymes in microsomal preparations. This was accomplished by treating rabbits with phenobarbital (PB), β-naphthoflavone (βNF), or PB + βNF to induce CYP2B4, CYP1A2, or both CYP2B4 and CYP1A2 and comparing benzphetamine and 7-pentoxyresorufin (7-PR) metabolism with each of these microsomal preparations.
Experimental Procedures
Materials.
The substrates, 7-ethoxyresorufin and 7-PR, were purchased from Molecular Probes, Inc. (Eugene, OR). PB and βNF were obtained from Mallinckrodt (St. Louis, MO) and Sigma (St. Louis, MO), respectively. Benzphetamine was a gift from Upjohn (Kalamazoo, MI). Antibody to CYP2B4 (chicken anti-rabbit) was a gift from Dr. John B. Schenkman (University Connecticut Health Center, Farmington, Connecticut). Anti-rat CYP1A1/1A2 monoclonal antibody was purchased from Oxford Biomedical Research (Oxford, MI). Purified recombinant CYP2B4 was a gift from Dr. Lucy Waskell (University of Michigan, Ann Arbor, MI). All other chemicals were reagent quality and were purchased from Sigma.
Methods.
Rabbits, divided into three groups, were treated with PB, βNF, or both PB + βNF to enrich the microsomal preparations with CYP2B4, CYP1A2, and both CYP2B4 and CYP1A2, respectively. PB was administered by daily intraperitoneal injection for 3 days at a dose of 80 mg/kg. Rabbits were euthanized 24 h after the final injection. βNF was administered as a single injection (40 mg/kg), and the rabbits were killed 72 h afterward. Rabbits treated with both PB and βNF were given injections with both inducers on day 1, followed by PB injections on the second and third days. Rabbits in this group were euthanized 24 h after the final PB injection.
Microsomes were prepared by differential centrifugation (Sequeira et al., 1994), and the protein concentration was determined (Lowry et al., 1951). Several microsomal activities were then examined including, PROD, EROD, benzphetamine demethylation, and cytochrome creductase. Benzphetamine demethylation was measured by quantifying production of the fluorescent derivative of formaldehyde after reaction with Nash reagent (Nash, 1953; Causey et al., 1990). Dealkylation of 7-PR was measured by a direct assay following the production of the fluorescent product resorufin (Perrin et al., 1990). The final concentrations for benzphetamine demethylation were microsomes (1 mg/ml), 1 mM benzphetamine, 5 mM glucose 6-phosphate, 2 units/ml glucose-6-phosphate dehydrogenase, and 10 mM magnesium chloride in 100 mM potassium phosphate, pH 7.25. The final component concentrations for PROD were microsomes (0.09 mg/ml), 1.3 μM 7-pentoxyresorufin, 0.1 mM EDTA, and 15 mM magnesium chloride in 50 mM Hepes buffer, pH 7.5. The reaction was initiated by the addition of NADPH to a final concentration of 250 μM (500 μM for benzphetamine demethylation), and the samples were incubated at 37°C. In those instances where purified P450s were used, 0.05 μM P450 and reductase (either 0.025 μM or 0.075 μM) were incorporated into liposomes made of DLPC, as described previously (Cawley et al., 1995).
P450 levels were determined by measuring the carbon monoxy-ferrous complex (Omura and Sato, 1964). The relative levels of CYP2B4 and CYP1A2 were estimated by immune blotting. Cytochrome creductase activity was measured to estimate NADPH-cytochrome P450 reductase content using a standard curve of purified NADPH-cytochrome P450 reductase (Yasukochi and Masters, 1976). Under these assay conditions, the specific activity for cytochrome c reductase activity was 3300 nmol of cytochrome c reduced (min)−1 (nmol reductase)−1. The assay contained microsomes and 60 μM cytochrome c in 100 mM potassium phosphate buffer, pH 7.25, and was initiated by the addition of 500 μM NADPH. Reduction of cytochrome c was determined by monitoring the absorbance at 550 nm.
Significance was estimated using analysis of variance followed by Student-Newman-Keuls multiple comparison procedure withp < 0.05 being considered significant. Data are expressed as the mean ± S.E.M. for four rabbits.
Results
Reconstituted Systems Containing CYP2B4, CYP1A2, and NADPH-cytochrome P450 reductase.
In previous studies with reconstituted systems, the ability of one P450 enzyme to influence the catalytic properties of another P450 was established. In these studies, the addition of CYP1A2 to a reconstituted system containing reductase and CYP2B4 exhibited an almost stoichiometric inhibition of PROD (Cawley et al., 1995). This effect was not observed with all substrates—benzphetamine demethylation did not exhibit this response. Furthermore, the amount of inhibition was dependent on the reductase-P450 ratio—the inhibition of PROD was much greater at subsaturating reductase when compared with that found at higher reductase-P450 ratios. The inhibition of CYP2B4-dependent activities in mixed reconstituted systems can be observed in Fig. 1, using recombinant CYP2B4 rather than the CYP2B4 purified from rabbit liver used in earlier studies (Cawley et al., 1995; Backes et al., 1998). Generally, the activities observed with recombinant CYP2B4 (Fig. 1) and CYP2B4 purified from rabbit liver were similar (Backes et al., 1998). Additionally, similar interactions among reductase, CYP1A2, and CYP2B4 were obtained with both preparations of CYP2B4.
Both benzphetamine demethylation and PROD are primarily CYP2B4-dependent activities. To determine whether the function of one P450 enzyme is modulated by the presence of a second P450, benzphetamine metabolism was examined in binary reconstituted systems containing reductase and a single P450 and in a mixed reconstituted system containing reductase, CYP1A2, and CYP2B4. If the interactions between reductase and a particular P450 were unaffected by the presence of an additional P450 in the complex reconstituted system, then the rate of benzphetamine metabolism would be the sum of the rates from the binary reconstituted systems. On the other hand, if the presence of a second P450 influences the characteristics of another P450 enzyme, then either a significant synergism or inhibition of the response might be expected. The results in Fig. 1 demonstrate very different behavior when comparing bzp and PROD activities. Benzphetamine activity was synergistically stimulated at both 0.5:1 and 1.5:1 [reductase]/[P450] (Fig. 1A). In contrast, PROD was dramatically inhibited at subsaturating reductase, with the inhibition being largely relieved at higher reductase to P450 ratios (Fig. 1B). EROD (a CYP1A2-selective activity) exhibited a synergism at both 0.5:1 and 1.5:1 [reductase]/[P450] (Fig. 1C). The effect of addition of a second P450 was similar to that obtained with CYP2B4 purified from rabbit liver, indicating that the presence of CYP1A2 can significantly influence the behavior CYP2B4.
Are the Interactions among P450 Enzymes and Reductase also Observed in Microsomal Preparations?
To determine whether similar interactions can be found with microsomal preparations, groups of rabbits were treated with PB, βNF, or both PB and βNF. Table 1 shows the quantification of overall P450, NADPH-cytochrome creductase, and immunoreactive CYP1A2 and CYP2B4 levels. NADPH-cytochrome P450 reductase levels were estimated from cytochromec reductase activities. These levels were elevated in PB-treated rabbits compared with both the βNF- and the PB + βNF-treated groups. Interestingly, βNF (in the PB + βNF-treated group) appeared to block the elevation in NADPH-cytochrome P450 reductase elicited by PB alone. Overall, P450 levels in the PB + βNF group were higher than those found in βNF- and PB-treated rabbits. Although P450 levels appeared to be higher in the PB versus the βNF-treated group, the difference did not rise to the level of significance.
Table 1 also shows the effect of treatment with the inducers on immunoreactive CYP1A2 and CYP2B4. CYP2B4 was significantly higher in both groups treated with PB compared with the βNF-treated group. CYP1A2 was found to be significantly elevated only in the group treated with both PB + βNF.
Microsomal studies: examination of the behavior of CYP2B4-selective substrates.
Benzphetamine demethylation and PROD, both CYP2B4-selective substrates, were examined in each group of microsomal preparations (Fig.2A). Benzphetamine demethylation was lowest in the βNF-treated group and was more than 3-fold higher in PB-treated rabbit liver microsomes. This activity was also elevated in the PB + βNF group, being 80% that found in the PB group. These results suggest that the elevated levels of CYP1A2 in the PB + βNF group only marginally affected benzphetamine demethylation.
Although 7-pentoxyresorufin is also a CYP2B4-selective substrate, PROD activity exhibited a different pattern in the rabbits treated with different P450 inducers (Fig. 2A). This activity was 10-fold higher in PB rabbits than in the βNF-treated group, a much larger relative difference than that observed with benzphetamine as substrate. Treatment with both inducers did not elevate this CYP2B4-selective activity in a similar manner as that seen with PB treatment alone. PROD in the PB + βNF group was actually much lower than expected based on estimates of CYP2B4 and more closely matched that found in the βNF-treated rabbits. Inspection of this data suggests that CYP2B4 and CYP1A2 behave similarly in both microsomal preparations and the mixed reconstituted systems.
The catalytic activities were normalized to the amount of NADPH-cytochrome P450 reductase present in the microsomal preparations (Fig. 2B). When corrected for differences in reductase activity, the activity pattern is essentially the same as observed when expressed in terms of protein concentration. The only difference was a relative decrease in the PB-induced rates when compared with the other groups. In this figure, benzphetamine demethylation was not significantly different in either the PB or the PB + βNF group, whereas PROD was dramatically diminished in the PB + βNF group compared with that observed after PB treatment. These results further support the contention that the interactions among CYP2B4, CYP1A2, and reductase are not an artifact of the reconstitution process but can also be observed in microsomal preparations.
Microsomal studies: examination of the behavior of CYP1A2-selective substrates.
When the CYP1A2-selective substrate 7-ethoxyresorufin was examined, the results showed an elevation of EROD activity in the βNF-treated groups (Fig. 3A). There was no difference between the PB- and PB + βNF-treated groups. This was particularly surprising because of the apparent elevation of immunoreactive CYP1A2 protein in the group treated with both compounds (Table 1). The lack of a difference between these groups may be, at least in part, a reflection of the 50% higher levels of reductase in the PB-treated group. When normalized to the amount of reductase, EROD in the βNF-treated group was elevated compared with that treated with PB alone (Fig. 3B); however, this activity was smaller than anticipated based on CYP1A2 immunoreactive protein levels (Table 1).
Discussion
Despite the knowledge for more than 30 years that P450s exist in multiple forms, little is known about how these heme proteins interact and how they are organized within the endoplasmic reticulum. There have been several attempts to examine these potential interactions. Because multiple P450s exist in the presence of limiting concentrations of NADPH-cytochrome P450 reductase, the potential for different P450 enzymes to compete for the available reductase would be a logical assumption. Evidence for such interactions has been acquired. Early studies by West and Lu (1972) demonstrated that P448 and P450 would compete at limiting reductase. More recently, human CYP2A6 and CYP2E1, expressed in baculovirus-infected Sf9 cells, were shown to compete for reductase in a similar manner (Tan et al., 1997). These reports are consistent with each P450 having its own affinity for reductase and the ensuing competition between different P450s being dependent not only on the relative concentrations of the different P450 enzymes but also on the relative affinities of the P450s for reductase. These interactions can be described by the following species: 1A2 + 2B4 + OR ↔ 1A2-OR + 2B4-OR (model 1), where the amount of each complex is dependent on the relative concentration of the component proteins and the affinities of reductase for each of the P450 enzymes. As an additional complication to this model, it is possible for the affinities of specific P450 enzymes for reductase to be affected by a particular substrate or effector molecule.
Although simple competition between P450s for limiting reductase is an expected outcome that has support in the literature, several studies suggest that, in some cases, specific interactions among multiple P450s and reductase (aside from the competition described above) may occur. The nearly stoichiometric inhibition of PROD in the reductase-CYP1A2-CYP2B4 system suggests that something more than simple competition for reductase is involved (Cawley et al., 1995). Subsequent data showed that in the reductase-CYP1A2-CYP2B4 system, the results were consistent with the formation of a CYP2B4-CYP1A2 complex that was able to tightly bind reductase at the CYP1A2 binding site (Backes et al., 1998). The details of such a model include: OR + 2B4 + 1A2 ↔ OR-2B4 + OR-1A2 + OR-1A2-2B4 + 1A2-2B4-OR + OR-1A2-2B4-OR (model 2). There are five potential functional complexes to this model: two binary complexes (OR-2B4 and OR-1A2), two ternary complexes (OR-1A2-2B4 + 1A2-2B4-OR), and a quaternary complex (OR-1A2-2B4-OR). The unique feature of model 2 is that substrate is able to facilitate the formation of heteromeric P450 complexes with altered function.
The concept of the formation of heteromeric P450 complexes is supported by the results obtained with the CYP1A2-preferring substrate 7-ethoxyresorufin in which the addition of CYP2B4 synergistically stimulated this CYP1A2-dependent activity at subsaturating reductase concentrations (Backes et al., 1998). Synergism has also been reported for mixed reconstituted systems containing reductase, CYP3A4, and CYP1A2 (Yamazaki et al., 1997). In this report, there was a 2-fold stimulation of testosterone 6β-hydroxylation by the addition of CYP1A2 at saturating reductase to a reconstituted system containing reductase and CYP3A4. A similar synergistic response was observed in mixed reconstituted systems containing reductase, CYP1A2, and recombinant CYP2B4 at a 1.5:1 [reductase]/[P450] (Fig. 1, A and C). These types of interactions can be most readily explained through the formation of heteromeric complexes in the mixed reconstituted systems (Backes et al., 1998).
The major advantage to examining the interaction of P450s in reconstituted systems is that the concentrations of the components and the lipid environment can be carefully regulated. A potential drawback to studies in reconstituted systems is that the proteins may not interact in the same manner as observed in the more natural environment of the microsomal membranes where the cellular machinery is used to incorporate the proteins. On the other hand, measurement of these inhibitory responses is confounded by the complexity of the microsomal systems. Not only do these systems contain a much larger number of P450 enzymes, other proteins, such as cytochromeb5, heme oxygenase, and squalene monooxygenase, are present, each capable of interacting with the reductase. The larger number of components and the inability to carefully control their concentrations, makes it much more difficult to systematically examine the interactions of P450 proteins in microsomes. Therefore, it is important to determine whether the results obtained in microsomes are consistent with those observed in the reconstituted systems.
In an effort to detect these interactions in microsomal preparations, we took advantage of the fact that the addition of CYP1A2 to reconstituted systems containing reductase and CYP2B4 led to a large degree of inhibition of one CYP2B4 activity (PROD), whereas another CYP2B4 activity (benzphetamine demethylation) was only marginally affected. We tried to mimic the CYP2B4-, CYP1A2-, and CYP1A2 + CYP2B4-containing reconstituted systems by induction of these P450s in rabbit liver with PB, βNF, and PB + βNF, respectively. The results clearly showed that when CYP2B4 was preferentially induced (PB treatment), both PROD and bzp were elevated. However, when both CYP2B4 and CYP1A2 were induced (PB + βNF treatment), only bzp activity remained elevated—PROD was dramatically decreased. These results are consistent with the results obtained in reconstituted systems. The synergism of EROD at subsaturating reductase in the mixed reconstituted system had a much smaller effect than that of PROD. Consequently, we were not able to detect analogous changes in rabbit liver microsomes treated with both PB + βNF.
There are several potential complications to such experiments. First, in each of the treatment groups, CYP1A2 is present. Consequently, a relative decrease in PROD would be expected in each of the treatments groups. However, this CYP2B4 activity should be decreased to a greater extent in the presence of higher levels of CYP1A2—as we were able to detect in the PB + βNF-treated group (Backes et al., 1998). Second, in the microsomal studies, some of the metabolism of a substrate may be catalyzed by P450 enzymes other than CYP2B4 and CYP1A2. This would be expected to either diminish or completely mask the inhibition/synergism predicted by the defined reconstituted systems. The fact that we were able to detect inhibition of PROD in the microsomal preparations from PB + βNF rabbits is most likely due to the large magnitude of the inhibition by CYP1A2. It is possible that the relative differences in PROD and benzphetamine metabolism in the PB and PB + βNF groups could be due to differential induction of P450 enzymes other than CYP2B4 and CYP1A2. However, this condition would require the induction of a form of P450 in PB rabbits that specifically induces PROD metabolism (and not benzphetamine). Furthermore, induction of this form of P450 would then have to be inhibited in rabbits treated with both PB + βNF. Although possible, this condition would require that CYP2B4 catalyze no more than 20% of the total PROD metabolism in the microsomal preparations (the fractional difference between PB and PB + βNF microsomes). In separate experiments, antibody to CYP2B4 caused a 64% inhibition of PROD in PB-treated rabbits (not shown) and is not consistent with this alternative explanation. Finally, although the current data indicates that the CYP1A2 and CYP2B4 form a complex in the presence of 7-pentoxyresorufin, the data does not give us direct information on the organization of single P450s in the binary reconstituted systems. It is possible that homodimers (or even higher order multimers) exist both in reconstituted systems and in microsomes. The use of kinetic data in these studies allows us to focus on alterations that influence formation of functional complexes.
In conclusion, the results demonstrate that the interactions among CYP2B4, CYP1A2, and reductase observed in defined reconstituted systems are also found in microsomal preparations. These results support the idea that specific interactions among various P450 enzymes are possible. Although these studies have focused on the interaction of only two P450 enzymes, evidence from the literature suggests that the interactions described in this article are by no means unique and can have a considerable impact on the metabolism and toxicity of foreign compounds.
Footnotes
-
This work was supported by Grant ES04344 from the National Institute of Environmental Health Sciences.
- Abbreviations used are::
- P450
- cytochrome P450
- PROD
- 7-pentoxyresorufin-O-dealkylation
- EROD
- 7-ethoxyresorufin-O-dealkylation
- PB
- phenobarbital
- βNF
- β-naphthoflavone
- 7-PR
- 7-pentoxyresorufin
- DLPC
- dilauroylphosphatidylcholine
- bzp
- benzphetamine
- OR
- NADPH-cytochrome P450 reductase
- Received April 13, 2001.
- Accepted August 28, 2001.
- The American Society for Pharmacology and Experimental Therapeutics