Cytochrome P450 side-chain cleavage: Insights gained from homology modeling

Molecular and Cellular Endocrinology 265–266 (2007) 65–70 Review Cytochrome P450 side-chain cleavage: Insights gained from homology modeling Karl-Heinz Storbeck, Pieter Swart, Amanda C. Swart ∗ Department of Biochemistry, University of Stellenbosch, Stellenbosch 7602, South Africa Abstract Cytochrome P450 side-chain cleavage (CYP11A1) catalyzes the conversion of cholesterol to pregnenolone, the first step in steroidogenesis. The absence of a solved crystal structure has complicated deductions pertaining to the structure/function relationships of this key enzyme. Although a number of techniques have been employed to identify domains and specific amino acid residues important for catalytic activity, these methods have been unsuccessful in predicting three-dimensional orientations in space and thus the mechanism by which they exert their kinetic effect. This review aims to demonstrate the significant contribution homology modelling, when employed as a tool in combination with other standard biochemical techniques, has made towards our understanding of CYP11A1. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Homology modeling; Cytochrome P450 side-chain cleavage; Baboon CYP11A1; P450scc; Structure/function Contents 1. 2. 3. 4. 5. 6. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interaction with adrenodoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Substrate recognition and access channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Active pocket topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction Cytochrome P450 side-chain cleavage (CYP11A1) catalyzes the conversion of cholesterol to pregnenolone, the first step in adrenal steroidogenesis. This reaction consists of three consecutive monooxygenations; a 22-hydroxylation, 20hydroxylation and the cleavage of the C20–C22 bond, yielding pregnenolone and isocaproic aldehyde. Each monooxygenation reaction requires two electrons to activate molecular oxygen, which are provided by NADPH via a mitochondrial electron transfer system. This system consists of adrenodoxin reductase (AdxR), a FAD-containing reductase, and adrenodoxin (Adx), a soluble, low molecular weight iron sulphur (Fe2 –S2 ) ferredoxin∗ Corresponding author. Tel.: +27 21 8085862; fax: +27 21 8085863. E-mail address: acswart@sun.ac.za (A.C. Swart). 0303-7207/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mce.2006.12.005 65 66 66 67 68 69 69 69 type electron transfer protein. AdxR is reduced by NADPH, which in turn reduces Adx. Adx acts as an electron shuttle between AdxR and CYP11A1 as well as other mitochondrial cytochromes P450. CYP11A1 has been the subject of numerous studies due to its physiological importance. The enzyme is a hydrophobic membrane bound protein not amenable to standard crystallization techniques and as such its three-dimensional structure has, to date, not been determined. While limited structural information is available in the absence of a crystal structure, methods predicting the three-dimensional structure of CYP11A1 have proved useful in deducing structure/function relationships. Homology or comparative modeling is a technique used to predict the three dimensional structure of proteins based on the observation that proteins with similar amino acid sequences have a tendency to adopt similar three-dimensional structures. 66 K.-H. Storbeck et al. / Molecular and Cellular Endocrinology 265–266 (2007) 65–70 Homology modeling predicts the three-dimensional structure of a protein based only on its amino acid sequence and its alignment with the solved crystal structures of proteins with similar sequences. Although such structures obtained by modeling are less accurate than those derived experimentally, they are invaluable since they provide a testable hypothesis in the absence of experimental data (Kemp et al., 2005). Until recently, homology models of mammalian cytochromes P450 have been based solely on the structurally determined bacterial cytochromes P450, which share less than 20% sequence identity with the mammalian enzymes. The recent availability of the crystal structures of more closely related mammalian cytochromes P450 have contributed towards the reliability of homology models of CYP11A1, which has not been crystallised. Here we review the contribution that homology modeling has made to our understanding of CYP11A1. homology model previously constructed by Usanov et al. (2002). In this study the CYP11A1 mutant R411Q exhibited decreased binding affinity to Adx, while charge neutralization of K267 and K270 had no significant effect on complex formation. Furthermore, at least two different binding sites on the proximal surface CYP11A1, which correspond to the two negatively charged sites of Adx, were identified. A homology model of CYP11A1, based on the crystal structure of CYP2B4, was subsequently constructed in an effort to identify residues that contribute to the redox potential of CYP11A1. This model was originally used to predict mutations increasing the redox potential of CYP11A1, which would facilitate its use as a biosensor (a topic beyond the scope of this review), and has subsequently been used to predict mutations that may facilitate crystallization of CYP11A1 (Sivozhelezova et al., 2006). 2. Interaction with adrenodoxin 3. Membrane interaction Numerous techniques, including physical chemistry, sitedirected mutagenesis and cross-linking experiments, have been employed to investigate the nature of the interaction between CYP11A1 and Adx. Studies have revealed that the Adx-binding site of CYP11A1 is located close to the heme-containing active pocket of the enzyme (Chashchin et al., 1985) and that binding results form the interaction of negatively charged acidic residues in Adx with positively charged basic residues in CYP11A1 (Usanov et al., 2002). However, the size of the interacting surfaces and the number of amino acid residues involved in the formation of the CYP11A1/Adx complex remained unclear. Usanov et al. (2002) subsequently investigated the role of specific residues involved in bovine CYP11A1 and Adx complex formation. Selected CYP11A1 lysine and arginine residues were mutated to glutamine. Each mutant was evaluated in terms of Adx binding, its ability to be reduced and enzymatic activity. A homology model of bovine CYP11A1 was constructed, using the bacterial enzymes CYP102 and CYP108 as templates, and subsequently docked to the solved crystal structure of bovine Adx (Muller et al., 1998), independently of the biochemical studies. Four putative salt bridges were identified in the CYP11A1/Adx complex: K267, K403, K405 and R426 of CYP11A1 with E47, D76 D72 and E73 of Adx, respectively. The model showed that these residues were located in the L helix, with R425 and R426, lying close to the heme moiety and K403 and K405 lying between the heme binding region and a region known as the meander on the proximal surface of the enzyme. Interestingly, prior to this study, both K403 and K405 had been mapped by chemical modification (Chashchin et al., 1985). Furthermore, their model revealed that the loop domain around the iron–sulphur complex (FeS) of Adx is associated with the heme binding domain in CYP11A1, readily allowing electron transfer from Adx to CYP11A1. These findings are in agreement with results obtained by Zollner et al. (2002), who showed that deletions in this loop domain affect the interaction with CYP11A1. More recently Strushkevich et al. (2005) selected target residues for mutagenesis based on the bovine Adx-CYP11A1 CYP11A1 is intimately associated with the inner mitochondrial membrane. The enzyme is expressed as a higher molecular weight precursor, containing an amphipathic N-terminal presequence that targets CYP11A1 to the mitochondrial surface. The preprotein is subsequently translocated to the inner mitochondrial membrane, the 39 amino acid residue leader sequence cleaved by specific peptidases found in the mitochondrial matrix, yielding the mature form of CYP11A1. Only the mitochondria of steroidogenic tissues and those tissues expressing other mitochondrial cytochromes P450 have the necessary machinery required to process and include CYP11A1 into the mitochondrial membrane (Matocha and Waterman, 1984). Furthermore, in addition to CYP11A1 requiring a mitochondrion-specific electron transfer system, the enzyme also requires the physiological milieu of the mitochondrion for activity. The catalytic activity of the enzyme is thus dependent on the interaction of CYP11A1 with the inner mitochondrial membrane since cholesterol is not water-soluble and is transferred directly from the inner mitochondrial membrane to the active pocket of the enzyme (Seybert et al., 1979). Investigations subsequently confirmed that CYP11A1 is tightly associated, but not integrated into the inner mitochondrial membrane (Schwarz et al., 1994). Electron spin resonance studies showed the heme group of CYP11A1 being roughly parallel to the plane of the inner mitochondrial membrane and not embedded within the hydrophobic membrane environment (Fig. 1) (Blum et al., 1978). Partial tryptic digestion and sodium carbonate extraction revealed that the N-terminal half of CYP11A1 remains associated with the inner mitochondrial membrane, while the C-terminal half is soluble (Ou et al., 1986). However, hydrophobicity profiles revealed that CYP11A1 lacks an N-terminal transmembrane anchor, a structural feature identified for microsomal cytochromes P450 (Nelson and Strobel, 1988). The removal of the N-terminal transmembrane anchor from the microsomal enzyme CYP2C5 produced a membrane associated protein which can be dissociated in high salt buffers, suggesting that the residual membrane binding interactions are likely monofacial (Williams et al., 2000). The absence K.-H. Storbeck et al. / Molecular and Cellular Endocrinology 265–266 (2007) 65–70 67 4. Substrate recognition and access channel Fig. 1. Homology model of baboon CYP11A1 associated with the inner mitochondrial membrane in the orientation proposed by Headlam et al. (2003). The position of the putative access channel as well as that of I98 is shown. of an N-terminal transmembrane domain in the mitochondrial cytochromes P450 therefore argues for the presence of other membrane-associated regions (von Wachenfeldt and Johnson, 1995). These findings prompted an investigation into the membrane binding potential of the F–G loop domain by Headlam et al. (2003) who constructed the first homology model of CYP11A1 using a mammalian cytochrome P450 as a structural template. Although the model was based on the crystal structure of CYP2C5, the F–G loop of CYP2C5 was not incorporated as it was shorter than the loop predicted for CYP11A1. A loop of similar length was therefore selected from the database in the Swiss Model automated server and used as a template. The model revealed that the region most likely to interact with the membrane was the distal face of CYP11A1, which contains an aromatic region surrounding a hydrophobic patch in the vicinity of the F–G loop. The proximal face had an overall positive charge, which included the residues K267, K403, K405 and R426, implicated by Usanov et al. (2002) to be involved in Adx binding. Site-directed mutagenesis and fluorescent labelling of residues appearing on the distal surface of the CYP11A1 model revealed that V212C and L219C have enhanced fluorescence and a blue shift following association of the mutant CYP11A1 with phospholipid vesicles. This finding indicated that these residues, which are located in the F–G loop, are localized in a hydrophobic environment subsequent to membrane binding (Fig. 1). Furthermore, analysis of the quenching of tryptophan residues in CYP11A1 by acrylamide, indicated that at least two of four tryptophan residues located on the distal face (W21 and W28 of the A′ helix, and W225 and W232 of the F–G loop and G helix, respectively) are involved in membrane binding. This is consistent with the important function of aromatic amino acid residues in membrane interfacial binding (Killian and von Heijne, 2000). A subsequent study confirmed that mutations to residues in the F–G loop of CYP11A1 result in a decreased association with membranes (Pikuleva, 2004). These data demonstrate that CYP11A1 has a monotopic association with the membrane that is mediated, at least in part, by the F–G loop region (Headlam et al., 2003). As the active pockets of the cytochromes P450 are buried within the enzyme, it has been proposed that substrates are first recognized on the surface of these enzymes by a domain known as the docking region, prior to movement into the active pocket via an access channel (Tsujita and lchikawa, 1993). Tsujita and lchikawa (1993) digested [14 C] methoxychlor-bound CYP11A1 with trypsin and found that the inhibitor-bound fragment was at the N-terminus, representing a likely docking region. In 1998, Lewis and Lee-Robichaud constructed a homology model of CYP11A1 based on the crystal structure of CYP102, which supported Tsujita and Ichikawa’s findings, showing the 20 amino acid residue sequence near the N-terminal of CYP11A1 to be the likely docking region for cholesterol. Furthermore, it was deduced from the model that this docking region could interact directly with the inner mitochondrial membrane and is in close proximity to the ‘mouth’ of the substrate access channel. The ‘mouth’ of the access channel is formed by the F–G loop together with ␤ sheets, ␤1 and ␤2, while the access channel itself is formed by the F, G, B′ helices, the B–B′ and B′ –C loops and the ␤1 sheet (Graham-Lorence and Peterson, 1996). These domains are all located on the distal face of CYP11A1 and are in direct contact with the inner mitochondrial membrane (Fig. 1) according to the model of Headlam et al. (2003). This orientation of CYP11A1 would provide easy access for membrane cholesterol to the active pocket of CYP11A1 via the hydrophobic access channel. Furthermore, the association of the ‘mouth’ of the access channel with the membrane environment has been proposed for other cytochromes P450 such as CYP19, which also contains a high concentration of aromatic residues in the F–G loop (Graham-Lorence et al., 1995; Graham-Lorence and Peterson, 1996). Conversely, the soluble bacterial cytochromes P450 have several charged residues in this region, facilitating the transport of less hydrophobic substrates to the active pocket (Graham-Lorence and Peterson, 1996). As mentioned above, part of the substrate access channel is formed by the B′ –C loop (Graham-Lorence and Peterson, 1996). The B′ –C loop is the first of six putative substrate recognition sites (SRS1), domains which, for all cytochromes P450, have direct contact with the bound substrate (Gotoh, 1992). Specific residues within the B′ –C loop have been shown to contribute towards substrate specificity and hydroxylation regioselectivity in both bacterial (Graham-Lorence and Peterson, 1996) and mammalian cytochromes P450 (Gotoh, 1992). Furthermore, the B′ –C loop together with the F–G loop, have been shown to be flexible domains, a feature important in allowing substrate access to the buried active pocket of cytochromes P450. Crystal structures of CYP2C5 (Wester et al., 2003), CYP2C8 (Schoch et al., 2004), CYP2C9 (Williams et al., 2003) and CYP2B4 bound with a phenylimidazole inhibitor (Scott et al., 2004), show closed conformations with no observable substrate access channel. However, the structure of inhibitor-free CYP2B4 has a large open substrate access channel, formed by the repositioning of helices B′ to C and F through G (Scott et al., 2003). In addition, the B′ –C loop adopts different orientations when CYP2C5 is bound with the substrates diclofenac or 4-methyl-N- 68 K.-H. Storbeck et al. / Molecular and Cellular Endocrinology 265–266 (2007) 65–70 methyl-N-benzenesulfonamide (Wester et al., 2003), underlying the importance of the flexibility of this domain in substrate recognition and binding. The adaptive positioning of the B′ helix in CYP2C5 also reflects the relatively weak interactions of this domain with the rest of the protein structure. The B′ helix is flanked by two GXG motifs which contribute to the flexibility of this domain in the CYP2 family and the regions preceding and following the GXG motifs, which include the B′ –C loop, are stabilized by interactions with adjacent domains (Johnson, 2003). Our group constructed a homology model of CYP11A1 using the crystal structures of CYP102 and CYP2C5 as templates to investigate the role of specific B′ –C loop residues in the catalytic activity of CYP11A1 (Storbeck et al., 2004, 2007). The model revealed that I98 interacts directly with the I helix (Fig. 1), a helix which runs straight through the active pocket of all cytochromes P450 (Graham-Lorence and Peterson, 1996). Changing the hydrophobicity of this residue by site-directed mutagenesis resulted in decreased substrate affinity. The model revealed that mutating this residue to a polar glutamine or charged lysine residue abolished the interaction of this domain with the I helix, resulting in structural changes of the B′ –C loop and a shift of the bound substrate (Storbeck et al., 2007). This data revealed that I98 plays an essential role in stabilizing this flexible domain through direct interaction with the I helix. In other crystallized mammalian cytochromes P450, a large hydrophobic residue, corresponding to I98 in the B′ –C loop of CYP11A1 is orientated towards the I helix in a similar manner – I112 in CYP2C5, V113 in CYP2B4, L110 in CYP2C8 and F110 in CYP2C9. These residues appear to stabilize the B′ through C helices by interactions with the I helix. As the flexibility of the B′ –C loop has been shown to be important in allowing substrates access to the buried active pocket of the cytochromes P450 (Wester et al., 2003), it would seem that in CYP11A1 the interaction of I98 with the I helix may act as a kind of ‘door stop’ upon substrate binding. 5. Active pocket topology In the absence of a solved crystal structure, neither the three-dimensional structure of the enzyme nor the residues that constitute its active site have, to date, been accurately identified from sequence alignments, site-directed mutagenesis or kinetic studies, thus complicating deductions pertaining to the struc- ture/function relationships. Sequence alignments of deduced amino acid residues of CYP11A1 from various species have allowed the prediction of steroid- and heme-binding domains with some success. Site-directed mutagenesis can identify amino acid residues important for catalytic activity, but cannot predict their three-dimensional orientations in space and thus the mechanism by which they exert a kinetic effect, albeit in the active pocket, on substrate binding, or via an indirect mechanism. Substrate binding studies have shown that specific amino acid residues in the active pocket of CYP11A1 interact with the 3␤hydroxyl group of cholesterol, as well as the 20␣-hydroxyl and 22(R)-hydroxyl groups of the reaction intermediates, via hydrogen bonds (Lambeth et al., 1982; Heyl et al., 1986; Tuckey et al., 1996). In bovine CYP11A1, Y93 and Y94 were hypothesised to interact with the side-chain of cholesterol (Pikuleva et al., 1995), while another study suggested that I461 lies in close proximity to the side-chain of cholesterol (Woods et al., 1998). These results are contradictory to each other as Y93 and I461 lie at opposite ends of the active pocket, with Y93 and Y94 located in the B′ –C loop and I461 located between beta sheets, ␤4-1 and ␤4-2 (Fig. 2). Docking studies were carried out by our group which, using the homology model, identified key residues interacting with cholesterol (Storbeck et al., 2007). In the most favourable orientation cholesterol was shown to bind with its side-chain orientated towards beta sheets, ␤4-1 and ␤4-2, within SRS6 and the 3␤-hydroxyl group orientated towards the B′ helix within SRS1, in accordance with results obtained by Woods et al. (1998). In this orientation the 22(R)-hydrogen of cholesterol is located in line with the heme iron, approximately perpendicular to the heme plane (3.51 Å from the iron) as was predicted by Heyl et al. (1986). The C22 atom, the site of the first hydroxylation, is located in close proximity to T291 (Fig. 2), a highly conserved threonine residue in the I Helix, believed to play a critical role in dioxygen activation (Imai et al., 1989). It has been proposed that this residue accepts a hydrogen bond from the hydroperoxy (Fe(III)–OOH) intermediate which promotes the second protonation on the distal oxygen atom, leading to O–O bond cleavage (Nagano and Poulos, 2005). Supporting this hypothesis, we found that mutating this residue to serine did not influence enzymatic activity significantly, while mutation to an alanine residue abolished enzymatic activity (Storbeck et al., 2007). These results are similar to those obtained by Imai et al. Fig. 2. Homology model of baboon CYP11A1. The magnified section shows the putative positions of the active site residues Y93, T291, S352, R357 and P464. K.-H. Storbeck et al. / Molecular and Cellular Endocrinology 265–266 (2007) 65–70 (1989) for CYP101, further demonstrating the importance of a residue with a hydroxyl group at this key position in the active pocket of the cytochromes P450. T291 and M294 are located in the I helix and interact with the side-chain of cholesterol, pushing the substrate towards the beta sheets, ␤4-1 and ␤4-2, where cholesterol interacts with L462 and P464. In substrate binding studies carried out by Woods et al. (1998), the mutation I461L resulted in a decreased catalytic rate constant for cholesterol, an increased Km for 22(R)-hydroxycholesterol, but did not affect the kinetic constants for 20␣-hydroxycholesterol. This suggested that the side-chains of cholesterol, 22(R)-hydroxycholesterol and 20␣hydroxycholesterol occupy slightly different positions in the active pocket, with I461 in close proximity to the side-chain binding site. Our model revealed that the side-chain of cholesterol interacts directly with L462 and P464 which lie adjacent to I461 within ␤4-1 and ␤4-2 (Fig. 2). Furthermore, this orientation of cholesterol offered a feasible explanation for the ability of CYP11A1 to catalyse the side-chain cleavage of cholesterol sulphate and cholesterol esters with acyl chain lengths of up to four carbon atoms (Tuckey et al., 1996). The additional carbon atoms in the fatty acid esters could be accommodated by the flexibility of the B′ –C loop, allowing the fatty acid esters to bind in the active site with Km values similar to that of cholesterol, but in an orientation not optimal for side-chain cleavage as was observed by the decreased turnover rate of CYP11A1. It has been proposed that the binding of cholesterol to CYP11A1 is facilitated by hydrogen bonding between the 3␤oxygen of cholesterol and a residue in the active pocket (Heyl et al., 1986; Tuckey et al., 1996). Our model did not identify a residue which could form a hydrogen bond with bound cholesterol. However, R357, located in the loop between the K helix and ␤1-4 in SRS5, was found to be pointed towards the docked cholesterol, within 4 Å of the 3␤-hydroxyl group (Fig. 2). We therefore hypothesised that a hydrogen bond or dipole interaction may form between R357 and the 3␤-oxygen of the reaction intermediates 22(R)-hydroxycholesterol and 22,20dihydroxycholesterol, facilitating the 20␣-hydroxylation and the cleavage of the C20 –C22 bond. Furthermore, S352, also located in SRS5, is pointed towards the heme in the active pocket, in an orientation favourable for hydrogen bonding with the 22(R)-hydroxyl group, a bond believed to be responsible for mediating the orientation of the intermediate in the subsequent 20␣-hydroxylation (Heyl et al., 1986). The proposed interaction of the intermediates with R357 and S352 is further supported by the observation that the derivatives of cholesterol, 22(R)-hydroxyl- and/or a 20␣-hydroxycholesterol, bind significantly tighter to CYP11A1 than do cholesterol or 25hydroxycholesterol. In an attempt to predict the orientation of cholesterol in the active pocket of CYP11A1, Pikuleva et al. (1995) showed that the mutation of Y93 affected only the binding of cholesterol, but not that of the reaction intermediates. In our model, Y93 is located 3.90 Å from the nearest cholesterol atom which contradicts the above study (Fig. 2). However, our model offers a feasible explanation for the effect that Y93 has on the binding of cholesterol. The binding of 22(R)-hydroxycholesterol 69 is unaffected in that a shift in the orientation of 22(R)hydroxycholesterol towards S352 and R357 results in the intermediate being orientated further from Y93. 6. Conclusion As CYP11A1 is tightly associated with the inner mitochondrial membrane, it is not amenable to standard crystallization techniques. Furthermore, CYP11A1 has no transmembrane domain that can be modified in order to facilitate crystallization. Modification made to CYP11A1 in order to solubilize the protein may therefore impact on the structure and function of this enzyme. From the evidence presented in this paper it is clear that in the absence of a crystal structure, homology modeling has contributed significantly to our understanding of CYP11A1. Homology modeling of CYP11A1 has yielded valuable data in the absence of a crystal structure when used as a tool in combination with standard biochemical techniques and has, furthermore, been used to equal effect as a predictive tool. Acknowledgements The authors wish to thank Marina Rautenbach and Patricia Storbeck for their assistance with the preparation of this manuscript. This work was financially supported by the NRF, the University of Stellenbosch and the Wilhelm Frank Bursary Fund. References Blum, H., Leigh, J.S., Salerno, J.C., Ohnishi, T., 1978. The orientation of bovine adrenal cortex cytochrome P-450 in submitochondrial particle multilayers. Arch. Biochem. Biophys. 187, 153–157. Chashchin, V.L., Turko, I.V., Akhrem, A.A., Usanov, S.A., 1985. Cross-linking studies of adrenocortical cytochrome P-450scc. Evidence for a covalent complex with adrenodoxin and localization of the adrenodoxin-binding domain. Biochim. Biophys. Acta 828, 313–324. Gotoh, O., 1992. Substrate recognition sites in cytochrome P450 family 2 (CYP2). Proteins inferred from comparative analysis of amino acid and coding nucleotide sequences. J. Biol. Chem. 267, 83–90. Graham-Lorence, S., Peterson, J.A., 1996. P450s: structural similarities and functional differences. FASEB J. 10, 206–214. Graham-Lorence, S., Amarneh, B., White, R.E., Peterson, J.A., Simpson, E.R., 1995. A three-dimensional model of aromatase cytochrome P450. Protein Sci. 4, 1065–1080. Headlam, M.J., Wilce, M.C., Tuckey, R.C., 2003. The F–G loop region of cytochrome P450scc (CYP11A1) interacts with the phospholipid membrane. Biochim. Biophys. Acta 1617, 96–108. Heyl, B.L., Tyrrell, D.J., Lambeth, J.D., 1986. Cytochrome P-450scc-substrate interactions. Role of the 3␤- and side-chain hydroxyls in binding to oxidized and reduced forms of the enzyme. J. Biol. Chem. 261, 2743–2749. Imai, M., Shimada, H., Watanabe, Y., Matsushima-Hibiya, Y., Makino, R., Koga, H., Horiuchi, T., Ishimura, Y., 1989. Uncoupling of the cytochrome P450cam monooxygenase reaction by a single mutation, threonine 252 to alanine or valine: a possible role of the hydroxy amino acid in oxygen activation. Proc. Natl. Acad. Sci. USA 86, 7823–7827. Johnson, E.F., 2003. Deciphering substrate recognition by drug-metabolizing cytochromes P450. Drug Metab. Dispos. 31, 1532–1540. Kemp, C.A., Marechala, J.D., Sutcliffe, M.J., 2005. Progress in cytochrome P450 active site modeling. Arch. Biochem. Biophys. 433, 361–368. Killian, J.A., von Heijne, G., 2000. How proteins adapt to a membrane–water interface. TIBS 25, 429–434. 70 K.-H. Storbeck et al. / Molecular and Cellular Endocrinology 265–266 (2007) 65–70 Lambeth, J.D., Kitchen, S.E., Farooqui, A.A., Tuckey, R.C., Kamin, H., 1982. Cytochrome P450scc substrate interactions. J. Biol. Chem. 257, 1876–1884. Lewis, D.F.V., Lee-Robichaud, P., 1998. Molecular modelling of steroidogenic cytochromes P450 from families CYP11, CYP17, CYP19 and CYP21 based on the CYP102 crystal structure. J. Steroid Biochem. Mol. Biol. 66, 217–233. Matocha, M.F., Waterman, M.R., 1984. Discriminatory processing of the precursor forms of cytochrome P-450scc, and adrenodoxin by adrenocortical heart mitochondria. J. Biol. Chem. 259, 8672–8678. Muller, A., Muller, J.J., Muller, Y.A., Uhlmann, H., Bernhardt, R., Heinemann, U., 1998. New aspects of electron transfer revealed by the crystal structure of a truncated bovine adrenodoxin. Adx. Struct. 6, 269–280. Nagano, S., Poulos, T.L., 2005. Crystallographic study on the dioxygen complex of wild-type and mutant cytochrome P450cam: implications for the dioxygen activation mechanism. J. Biol. Chem. 280, 31659–31663. Nelson, D.R., Strobel, H.W., 1988. On the membrane topology of vertebrate cytochrome P-450 proteins. J. Biol. Chem. 263, 6038–6050. Ou, W.J., Ito, A., Morohashi, K., Fujii-Kuriyama, Y., Omura, T., 1986. Processing-independent in vitro translocation of cytochrome P-450(SCC) precursor across mitochondrial membranes. J. Biochem. (Tokyo) 100, 1287–1296. Pikuleva, I.A., 2004. Putative F–G loop is involved in association with the membrane in P450scc (P450 11A1). Mol. Cell. Endrocrinol. 215, 161–164. Pikuleva, I.A., Mackman, R.L., Kagawa, N., Waterman, M.R., Ortiz de Montellano, P.R., 1995. Active-site topology of bovine cholesterol side-chain cleavage cytochrome P450 (P450scc) and evidence for interaction of tyrosine 94 with the side-chain of cholesterol. Arch. Biochem. Biophys. 322, 189–197. Schoch, G.A., Yano, J.K., Wester, M.R., Griffin, K.J., Stout, C.D., Johnson, E.F., 2004. Structure of human microsomal cytochrome P450 2C8: evidence for a peripheral fatty acid binding site. J. Biol. Chem. 279, 9497–9503. Schwarz, D., Richter, W., Kruger, V., Chernogolov, A., Usanov, S., Stier, A., 1994. Direct visualization of a cardiolipin-dependent cytochrome P450sccinduced vesicle aggregation. J. Struct. Biol. 113, 207–215. Scott, E.E., He, Y.A., Wester, M.R., White, M.A., Chin, C.C., Halpert, J.R., Johnson, E.F., Stout, C.D., 2003. An open conformation of mammalian cytochrome P4502B4 at 1.6 Å resolution. Proc. Natl. Acad. Sci. USA 100, 13196–13201. Scott, E.E., White, M.A., He, Y.A., Johnson, E.F., Scout, C.D., Halbert, J.R., 2004. Structure of mammalian cytochrome P450 2B4 complexed with 4-(4chlorophenyl) imidazole at 1.9 Å resolution: insight into the range of P450 conformations and the coordination of redox partner binding. J. Biol. Chem. 279, 27294–27301. Seybert, D.W., Lancaster, J.J., Lambeth, J.D., Kamin, H., 1979. Participation of the membrane in the side chain cleavage of cholesterol. Reconstitution of cytochrome P-450scc into phospholipid vesicles. J. Biol. Chem. 254, 12088–12098. Sivozhelezova, V., Pechkova, E., Nicolinia, C., 2006. Mapping electrostatic potential of a protein on its hydrophobic surface: implications for the crystallization of cytochrome P450scc. J. Theor. Biol. 241, 73–80. Storbeck, K., Swart, P., Graham, S., Swart, A.C., 2004. The influence of the amino acid substitution I98K on the catalytic activity of baboon cytochrome P450 side-chain cleavage (CYP11A1). Endocr. Res. 30, 761–767. Storbeck, K., Swart, P., Graham, S., Swart, A.C., 2007. Evidence for the functional role of residues in the B′ -C loop of baboon cytochrome P450 side-chain cleavage (CYP11A1) obtained by site-directed mutagenesis, kinetic analysis and homology modelling. J. Steroid Biochem. Mol. Biol. 103, 65–75. Strushkevich, N.V., Azeva, T.N., Lepesheva, G.I., Usanov, S.A., 2005. Role of positively charged residues lys267, lys270, and arg411 of cytochrome p450scc (CYP11A1) in interaction with adrenodoxin. Biochemistry (Mosc) 70, 664–671. Tsujita, M., lchikawa, Y., 1993. Substrate-binding region of cytochrome P450scc (P-450X1A1). Identifcation and primary structure of the cholesterol binding region in cytochrome P-450scc. Biochim. Biophys. Acta 1161, 124–130. Tuckey, R.C., Lawrence, J., Cameron, K.J., 1996. Side-chain cleavage of cholesterol esters by human P450scc. J. Steroid Biochem. Mol. Biol. 58, 605–610. Usanov, S.A., Graham, S.E., Lepesheva, G.I., Azeva, T.N., Strushkevich, N.V., Gilep, A.A., Estabrook, R.W., Peterson, J.A., 2002. Probing the interaction of bovine cytochrome P450scc (CYP11A1) with adrenodoxin: evaluating sitedirected mutations by molecular modelling. Biochemistry 41, 8310–8320. von Wachenfeldt, C., Johnson, E.F., 1995. Structures of eukaryotic cytochrome P450 enzymes (Cytochrome P450: Structure, Mechanism, and Biochemistry, 2nd ed), pp. 183–244, Plenum Press, New York. Wester, M.R., Johnson, E.F., Marques-Soares, C., Dijols, S., Dansette, P.M., Mansuy, D., Stout, C.D., 2003. Structure of mammalian cytochrome P450 2C5 complexed with diclofenac at 2.1 Å resolution: evidence for an induced fit model of substrate binding. Biochemistry 42, 9335–9345. Williams, P.A., Cosme, J., Sridhar, V., Johnson, E.F., McRee, D.E., 2000. Mammalian microsomal cytochrome P450 monooxygenase: structural adaptations for membrane binding and functional diversity. Mol. Cell. 5, 121– 131. Williams, P.A., Cosme, J., Ward, A., Angove, H.C., Matak, V.D., Jhoti, H., 2003. Crystal structure of human cytochrome P450 2C9 with bound warfarin. Nature 424, 464–468. Woods, S.T., Sadleir, J., Downs, T., Triantopoulos, T., Headlam, M., Tuckey, R.C., 1998. Expression of catalytically active human cytochrome P450scc in Escherichia coli and mutagenesis of isoleucine-462. Arch. Biochem. Biophys. 353, 109–115. Zollner, A., Hannemann, F., Lisurek, M., Bernhardt, R., 2002. Deletions in the loop surrounding the iron-sulphur cluster of adrenodoxin severely affect the interactions with its native redox partners adrenodoxin reductase and cytochrome P450scc (CYP11A1). J. Inorg. Biochem. 91, 644–654.