Cytochrome P450 3A4
Created by David Nyenhuis
Cytochrome P450 3A4, also referred to as CYP3A4 (2J0D), from Homo sapiens is involved in the oxidation of a wide array of substrates, including xenobiotics. The most abundant of the Cytochrome P450 superfamily proteins present in the human liver, it is involved in the mono-oxidation of more than half of current drugs to facilitate their subsequent export from the body (1,2). Adverse drug-drug interactions may arise, however, as a result of a secondary drug binding to the enzyme as an allosteric inhibitor (3). The prevalence of these adverse reactions makes CYP3A4 a common target of pharmaceutical and biomedical research. CYP3A4 has been characterized in the presence of multiple ligands, including the macrolide antibiotic erythromycin considered here. CYP3A4 has a molecular weight of 554,009.64 Da, and an isoelectric point of 8.60 (4). These data points were determined via the Compute pI/Mw tool present in ExPASy, a hub for a variety of bioinformatics based tools and databases.
Structurally, CYP450 3A4 is a homodimer with identical subunits divided into a small, β-Sheet rich N-terminal domain and a larger C-terminal domain comprised primarily of α-helices, and which contains the active site (5). The secondary structure of CYP3A4 has 268, or 55%, of its residues participating in α-helices and only 35, or 7%, of its residues participating in β-sheets (6). Figure 1 illustrates the overall fold of CYP3A4 and widely accepted naming scheme for its preeminent α-helices and β-sheets (5). CYP3A4 has a globular form in vivo and is attached to the microsomal membrane via its N-terminal domain and parts of the F’G’ loop (7). The high degree of alpha helices in the catalytic domain leads to substantial enzyme flexibility and numerous substrate egress channels (7).
There are three principal structural deviations between CYP3A4 and other eukaryotic CYP450 superfamily enzymes. First, the presence of a hydrophobic region, residues 36-50, in the aforementioned N-terminal region aids in interactions with the microsomal membrane (2). Second, the F and G helices, which form the active site’s upper bound in many CYP450 enzymes, are truncated in CYP3A4. This leads the roof of CYP3A4’s active site to be composed primarily of residues 209-217 and 237-242, which bridge the F and G helices to the F’ and G’ helices (8). Third, the unique cluster of 6 phenylalanine residues: Phe-213, Phe-215, Phe-219, Phe-220, Phe-241, and Phe-304 (1). This Phe cluster is localized to the upper region of the enzyme’s active site, and constitutes the primary residues in the peripheral binding region, which is involved in allosteric regulation of the oxidation pathway (3).
The two ligand components for Cytochrome P450 3A4 are protoporphyrin IX containing Fe (HEME) and erythromycin. The HEME prosthetic group serves as the site of substrate oxidation, and is thus catalytically essential. The central iron in the HEME group is ligated via the thiolate side chain of Cys-442, and the HEME propionates are hydrogen bonded to the side chains of Arg-105, Trp-126, Arg-130, Arg-375, and Arg-440 (5). The second ligand, erythromycin, is a macrolide antibiotic, and a common xenobiotic substrate encountered by CYP3A4 in vivo. Erythromycin is also one of the largest substrates for CYP3A4, leading to greater conformational changes in the enzyme and thus frequent employment in structural studies (1). In the case of CYP3A4, binding of erythromycin induces an 80% increase in active site area (1).
The mono-oxidation of a wide array of substrates is carried out at the CYP3A4 internal active site, which the substrate accesses via one of 5 distinct channels (7). The lower bound of CYP3A4’s active site is defined by the HEME prosthetic group while the upper bound is constituted by the Phe cluster and regions bridging the F and G helices to the F’ and G’ helices (8). Initiation of the reaction pathway involves binding of the first ligand at the peripheral binding site around the Phe cluster, with Phe-304 playing a pivotal role in creating π stacking interactions for ligand stabilization (1,3). Ketoconazole, an inhibitor of CYP3A4, was the first species shown to bind at this peripheral ligand binding site (1). This initial binding event induces conformational changes in CYP3A4 that alter the hydration state and water accessibility of the HEME prosthetic group and thus induce a spin state shift in its central iron (3). This shift in spin state permits the introduction of NADPH-cytochrome P450 reductase, which catalyzes the reduction of the HEME group’s ferric center to eventually produce the active species subsequently involved in the majority of substrate oxidations (7).
Creation of this active species proceeds through 2 reactive intermediates. These intermediates carry out the oxidation of a small set of CYP3A4’s ligands, and in these cases the intermediate conversion pathway must be hindered to permit the intermediates adequate time to carry out the requisite oxidation (8). This is accomplished by Arg-212, an active site residue which is hydrogen bonded to Phe-304 in the Phe cluster by its backbone carbonyl, and Glu-308 by its backbone amide (8). This linkage causes a change in conformation around Phe-304 mediated via Arg-212 to also induce a shift in orientation of the Glu-308 and Thr-309 peptide bond (8). This bond shift then disrupts the proton transfer pathway involved in intermediate conversion, allowing the increased intermediate lifetimes necessary to oxidize some substrates (8). In ligands where this is not required, such as ketoconazole, Arg-212 is reoriented outside of the active site (1)
The erythromycin ligand is then oxidized at the dimethylamino group present in the D-dedosamine sugar (1) . Mutagenesis studies have shown several residues surrounding bound erythromycin in the CYP3A4 binding site, including Arg-119, Leu-301, Phe-304, Ala-305, Leu-369, Ala-370, and Glu-374, to be critical active site residues (8). With the exception of Leu-369, these residues were also implicated for active site interaction with erythromycin via molecular modeling (9). The active site of CYP3A4 is capable of binding multiple substrates, and the first erythromycin molecule binds near the peripheral binding region at a distance of 17 Å from the HEME group, which is too far to undergo oxidation (1). The initial binding event is sufficient, however, to induce opposing movement in the F and G helices, producing the increase in active site volume which is shown in figure 2 (B) (1). Binding of a subsequent erythromycin molecule occurs much closer to the HEME prosthetic group, allowing mono-oxidation to proceed and the substrate to exit through one of the enzyme’s five egress channels (7)
Binding of allosteric inhibitors, including ketoconazole, to CYP3A4 also induces a marked effect on the enzyme’s conformation. Binding of two ketoconazole molecules occurs in an antiparallel orientation, and involves π stacking between the side chain of Phe-304 and the chlorobenzyl ring in concert with interactions provided by Arg-372, Arg-106, and Glu-374 in an adjacent polar pocket (1). Binding of the ketoconazole inhibitors at this location disrupts the Phe cluster and also causes residues 210-213 to adopt a helical structure, placing the critical Arg-212 residue outside of the active site and extending the F helix by an additional turn (1). Figure 2 (A) shows the reorientation of the active site in CYP3A4 following ketoconazole binding caused by the movement of the F and G helices in similar directions.
Another member of the cytochrome P450 superfamily, CYP2C9 (1OG2), was selected for comparison purposes with CYP3A4 (10). This determination was made based on results from two databases, Dali and PSI-BLAST. The Dali web server aids in finding proteins with tertiary structures similar to that of the query protein. The Z-score generated for CYP2C9 via a Dali search against CYP3A4 was 36.4 (11). Protein BLAST, or PSI-BLAST, is an algorithm that aids in finding proteins with primary structures similar to the query protein based on sequence homology. BLAST scores are given in terms of E-values, which define the likelihood of an observed overlap in protein structure occurring by chance. The E value for CYP2C9 relative to CYP3A4 was 3e-146 (12). Despite the strength of these statistical scores and the inclusion of both enzymes in the same superfamily, the overall sequence identity between CYP2C9 and CYP3A4 is only 24% (5).
Structurally, CYP2C9 is markedly similar to CYP3A4. It has the same general globular shape and overall fold indicative of the CYP450 superfamily, and is also bound to the microsomal membrane primarily via its N-terminal domain (13). Secondary structure is also highly conserved between the two enzymes, with CYP2C9 exhibiting 34 residues participating in β-sheets, a deviation of only a single residue from CYP3A4 (10). CYP2C9 also has 278 residues participating in α-helices, which makes the protein 58% helical, again very similar to the 55% helical character observed in CYP3A4 (10). The major helices identified in CYP3A4 are all still present in CYP2C9, although the F and G helices present in CYP2C9 are substantially longer and form the upper boundary of the CYP2C9 active site instead of the Phe cluster and interhelical loops observed in CYP3A4 (13).
CYP2C9 serves the same general function as CYP3A4; it catalyzes the mono-oxidation of substrates for further metabolism and subsequent export. CYP2C9 has a more narrow range of target substrates, but still targets a wide array of drugs including the anti-coagulant warfarin, which has been implicated in multiple ligand binding events similar to CYP3A4 (13). To accomplish substrate oxidation, CYP2C9 also relies on a HEME prosthetic group that forms the lower bound of its active site, although activation of the HEME group via conformational changes induced by binding at a peripheral binding domain has not been observed in CYP2C9 (13). The active site of CYP2C9 is substantially larger than that of CYP3A4, and while it lacks the distinctive Phe cluster, a phenylalanine rich hydrophobic pocket is still present (13). Binding of lipophilic substrates, including warfarin, in CYP2C9 occurs via a binding mode similar to ketoconazole binding in CYP3A4. The aforementioned hydrophobic pocket, together with Phe-110 and Phe-114, and π stacking interactions provided by the phenyl group of Phe-476 form the principal hydrophobic interactions with the substrate (13). Further stabilization is then provided by hydrogen bonding of warfarin carbonyl oxygens to Phe-100 and Ala-103 via their backbone amide-nitrogens (13).