Abstract
The results of quantitative structure-activity relationship (QSAR) studies on inhibitors and substrates of cytochrome P450 2B (CYP2B) subfamily enzymes are reported. It was found that lipophilicity (in the form of log P) is the most important property for explaining the variations in inhibitory activity, and there are similarities between QSARs for both substrates and inhibitors for CYP2B6 (human), and also between those of other CYP2B enzymes, such as CYP2B1 (rat) and CYP2B4 (rabbit). Both linear and quadratic lipophilicity relationships are evidenced in human and other mammalian species, and the particular type of expression found is probably due to the nature of the compounds under investigation, as it is usually the homologous series which tend to show quadratic relationships in log P. The findings from QSAR studies can be rationalized by molecular modelling of the active site interactions with both P450 crystal structures and homology models of CYP2B subfamily enzymes.
| Abbreviations: | ||
| CYP, | = | cytochrome P450; |
| QSAR, | = | quantitative structure-activity relationship; |
| log P, | = | logarithm of the octanol/water partition coefficient; |
| HOMO, | = | highest occupied molecular orbital; |
| AM1, | = | Austin Model 1. For statistical results, |
| n, | = | number of observations; |
| s, | = | standard error; |
| R, | = | correlation coefficient; |
| F, | = | variance ratio. |
Introduction
The cytochromes P450 (CYP) are a superfamily of haem-thiolate enzymes for which over 8500 individual members are currently known [Citation1]. In mammalian species, P450s are generally associated with the Phase 1 oxidative metabolism of drugs and other xenobiotics, although various endogenous functionalities are known, including biosynthesis of the steroid hormones and long chain fatty acid metabolism [Citation2]. Enzymes of the CYP2B subfamily, which are inducible by phenobarbital, catalyse the metabolism of many foreign compounds in mammals, including: 7-benzyloxyresorufin, 7-pentoxyresorufin, 7-ethoxycoumarin, benzphetamine, aminopyrine and testosterone [Citation3,Citation4] many of which have been employed as diagnostic probe substrates for CYP2B isoforms in mouse, rat, rabbit, hamster, monkey and man.
Substrates of CYP2B are characterised by generally non-planar molecules which usually possess the capability for hydrogen bond formation, and most inhibitors of these enzymes are substrate analogues such as: secobarbital, orphenadrine and proadifen (SKF-525A) although chloramphenicol probably inhibits CYP2B via covalent binding to an active site lysine [Citation5]. CYP2B6, the human orthologue, accounts for some 4% involvement in the Phase 1 metabolism of drugs and other foreign compounds [Citation6,Citation7] although CYP2B isoforms play somewhat greater roles in exogenous metabolism for other mammalian species such as the rat and rabbit [Citation3].
There have been several recent QSAR studies on inhibitors of CYP1 and CYP2 family P450s [Citation8–10]. We have previously reported QSARs in selected substrates of CYP2B subfamily enzymes [Citation11–13] and the current work presents the results of related investigations on CYP2B inhibitors and some additional substrates, including those of the human orthologue, CYP2B6. As with the previous studies, we have utilised compound lipophilicity in the form of log P values, where P is the octanol/water partition coefficient, to aid in an understanding of the likely structure-activity relationships for inhibition, which also includes molecular modelling with the enzymes themselves, where crystal structures of related forms such as CYP2B4 and CYP2C5 with and without bound substrates or inhibitors are now available. However, it is thought that the inhibitor-bound CYP2B4 crystal structure [Citation14] represents the most satisfactory template for homology modelling of CYP2B6, and shows the active site of CYP2B6, modelled from CYP2B4, which contains a bound inhibitor. This inhibitor molecule (CBP) docked into the CYP2B6 haem pocket () displays the “butterfly wing” conformation which is common to many CYP2B substrates and inhibitors such as phenobarbital, proadifen (SKF525A) and DDT.
Figure 1. The selective CYP2B6 inhibitor, 4-(4-chlorobenzyl)pyridine (CBP), is shown docked into the active site of this enzyme. The mode of binding and its calculated energy are consistent with the experimental findings for this compound [Citation25].
![Figure 1. The selective CYP2B6 inhibitor, 4-(4-chlorobenzyl)pyridine (CBP), is shown docked into the active site of this enzyme. The mode of binding and its calculated energy are consistent with the experimental findings for this compound [Citation25].](/cms/asset/cb4e9ff3-bb7e-4b9c-bfd0-74003da68ab0/ienz_a_451878_f0001_b.gif)
Methods
The inhibition data was collated from the literature for several series of structurally related compounds, including: primary aliphatic amines [Citation15] and 2-alkyl-benzimidazoles [Citation16], together with a range of diverse compounds [Citation17]. In addition, CYP2B-related activities for a series of 7-alkoxycoumarins [Citation18] and of barbiturates [Citation19] were retrieved from previously reported studies. Compound lipophilicities in the form of log P values were obtained from tabulations of experimental data [Citation20, Citation21] or calculated via the ClogP (BioByte, Pomona, CA) and Pallas (CompuDrug, Budapest, Hungary) software systems. These have both been shown to give good correlations with experimental values of the order of 0.95 for the correlation coefficients with significant dataset sizes [Citation22]. However, for the homologous series of eight 7-alkoxycoumarins, it was necessary to employ calculated log P values for 50% of the compounds. Consequently, we have derived a calibration equation between the experimental log P values of 4 congeners and calculated log P produced by Clog P (BioByte). For these data, there is an excellent correlation between experimental log P and Clog P, according to Equation 1:
Using this equation, the calculated log P values (based on Clog P) are very close to the experimentally determined log P data. Furthermore, one can also employ the above equation to extrapolate calculated log P values for the four congeners where experimental log P data are unavailable. These new calculated log P values (calibrated log P) are all slightly higher than those calculated directly from ClogP. The relevant data are shown below in .
Table 1. Data set of log P values for 7-alkoxycoumarins.
where N/A signifies that the data were not available. The calibrated log P values can be utilised to see if there is any significant difference in the correlation with biological activity. In fact, the results demonstrate that there is a good agreement between QSAR equations produced from the calibrated log P values and the dataset of experimental and Clog P values.
These equations are also presented in the summary of QSARs in , 3a)i) Equation (6) and 3a)ii) Equation (7).
Results and discussion
shows the dataset for a homologous series of primary aliphatic amines which act as inhibitors towards the PB-inducible P450 isoform in the rabbit, which is assumed to be primarily composed of CYP2B4 [Citation3]. to provide the relevant dataset information for the other CYP2B inhibitors of relevance to this study. Analysis of the data in reveals that there is a clear parabolic relationship between -log Ki and logP, in accordance with the quadratic expression presented as equation 1 in . This latter table summarises all of the lipophilicity relationships investigated in this work, and thus facilitates comparison between the various series of compounds and the different types of relationship for governing variations in inhibitory potency or substrate binding affinity. provides information for a diverse group of 12 CYP2B6 inhibitors and there appears to be a clear linear relationship between inhibition and lipophilicity, as given in .
Table 2. Inhibition and log P data for primary aliphatic amines inhibiting CYP2B4 in rabbit. Reference [Citation15].
Table 3. Inhibition and log P data for diverse inhibitors of the human enzyme CYP2B6. Reference [Citation17].
Table 4. O-dealkylation rates of 7-alkoxycoumarins mediated by CYP2B4 in the rabbit: HOMO energies and log P data. Reference [Citation18].
Table 5. Induction of O-dealkylation activity of 7-alkoxycoumarins (CYP2B1-related) by phenobarbital in rat. Reference [Citation26].
Table 6. Binding and log P data for diverse substrates of the human enzyme CYP2B6. Reference [Citation27].
Table 7. Inhibition and log P data for 2-n-alkyl-benzimidazoles inhibiting CYP2B1 in rat. Reference [Citation16].
Table 8. Spectroscopic binding and log P data for barbiturates metabolised by CYP2B1. Reference [Citation19].
Table 9. Summary of QSARs in CYP2B inhibitors and substrates.
For the homologous series of 8 alkoxycoumarins shown in , there is a linear relationship between rate of dealkylation mediated by CYP2B4 and log P itself, and also when combined with HOMO energy, which is equivalent to the compound ionization potential. However a quadratic relation exists between logP and the logarithm of the dealkylation activity in phenobarbital-induced animals for substituted alkoxycoumarins (). Careful analysis of the expression involving a combination of log P and E (HOMO) indicates that this relationship is not as significant statistically as that which only involved log P and, consequently, this correlation should be viewed with caution. The relevant expressions are shown in , as equations 3a, 3b and 3c, respectively. In a set of 16 diverse substrates () of CYP2B6, the human orthologue, there is a clear linear relationship between log P and -logKm, as shown in equation 4a () and a plot of this is shown in . This expression can also be formulated as a lipophilicity relationship by conversation of log P and log Km to their respective free energies, as presented in equation 4b in . The intercept of this relation indicates that the most common type of interaction within the CYP2B6 active site is likely to be hydrogen bonding [Citation11,Citation12,Citation23,Citation24], and molecular modelling supports this viewpoint. For a series of n-alkyl-benzimidazoles, inhibition of CYP2B1 () exhibits a parabolic relationship with log P, as shown in , equation 5. In this case, the negative logarithm of the Ki value has been used as a measure of the inhibitory activity of the n-alkyl-benzimidazoles.
Figure 2. Lipophilicity relationship for 16 diverse CYP2B6 substrates showing the clustering of points about the line, based on the data presented in and .
Finally, for a series of barbiturate derivatives binding to CYP2B1 (), there is also a parabolic relationship between log P and, in this case, -log KS, where KS, is the spectroscopic binding constant. The data used to generate equation 6 in are shown in . The relevant equation in (equation 6) can be compared with that of the 7-alkoxycoumarins (equation 3a, ) where it can be appreciated that the terms are of similar value. This tends to suggest that the optimal characteristics of the CYP2B active site are translated into providing similar quadratic expressions relating compound lipophilicity to binding in homologous series, and equation 1 in governing the inhibition of primary aliphatic amines also displays similar characteristics to the other quadratic expressions mentioned previously. However, it is noted that the optimal log P values for certain series of compounds which show quadratic relationships between activity and this lipophilicity parameter tend to vary somewhat, as listed in . To some extent, the optimal log P for a homologous series depends upon the way in which the various congeners become orientated within the enzyme active site. There is also some degree of chain folding that is likely to occur as the series is ascended, together with the actual length of the substrate access channel in the CYP2B enzyme which have to be taken into account when considering the implications of the optimal log P for a series of related compounds.
Conclusions
In general, there is a trend of binding and inhibition being dependent on compound lipophilicity, although it is also important to emphasise the role of active site interactions (hydrogen bonding and π-π stacking) plus the relevance of specific parameters in certain species of compounds, which have a particular significance in terms of the essential structural features in the compounds concerned. Therefore formulating the lipophilicity relationship in terms of free energies of binding and partitioning represents a very simple and straightforward method for estimating the contributions to the overall binding affinity between substrates or inhibitors and the relevant CYP2B active site. Hydrophobicity (or lipophilicity) is, therefore, a key feature of CYP2B-substrate (or inhibitor) interactions and there is also a hydrogen bond formed in many cases, as shown for the docked CBP inhibitor in the active site of CYP2B6 (). For homologous series, where alkyl chain length varies over a significant range, the normally linear relationship of bioactivity with log P becomes quadratic in nature due to the fact that the size of the pocket will be exceeded after a certain point in the series, and/or that the lengthy hydrocarbon chain will start to adopt a more folded conformation within the active site environment. Consequently, one can regard the overall lipophilicity of a compound as being a combination of hydrophobicity and polarity, such that the optimal log P within a series of structurally related compounds binding to, for example, a P450 enzyme active site may not necessarily take the same value when compared between different series of homologues, because the log P of the parent group will be different in each case and this clearly represents a component of the overall log P value.
Declaration of interest
The financial support of the British Technology Group and ExxonMobil Biomedical Sciences is gratefully acknowledged. Yuko Ito would also like to thank the Japanese Foundation for research funding of a PhD studentship.
References
- Isin EM, Guengerich FP. Substrate binding to cytochromes P450. Anal Bioanal Chem 2008;392:1019–1030.
- Lewis DFV. Guide to Cytochromes P450 Structure and Function. London:Taylor & Francis, 2001:76–117.
- Nims RW, Lubet RA. The CYP2B subfamily. In: C Ioannides, ed. Cytochromes P450 - Metabolic and Toxicological Aspects. Boca Raton, Fl: CRC Press, 1996:135–160.
- Turpeinen M, Raunio H and Pelkonen O. The functional role of CYP2B6 in human drug metabolism: substrates and inhibitors in vitro, in vivo and in silico. Curr Drug Metab 2006;7:705–714.
- Halpert JR. Structural basis of selective cytochrome P450 inhibition. Ann Rev Pharmacol Toxicol 1995;35: 29–53.
- Guengerich FP. (2005) Human cytochrome P450 enzymes. In: P.R. Ortiz de Montellano, ed. Cytochrome P450 Kluwer/Plenum, NY, 377–530.
- Rendic S. Sumary of information on human CYP enzymes: human P450 metabolism data. Drug Metabol Rev 2002;34:83–448.
- Roy K, Roy PP. Exploring QSAR and QAAR for inhibitors of cytochromes P4502A6 and 2A5 enzymes using GFA and G/PLS techniques. Eur J Med Chem 2009;44:1941–1951.
- Roy PP, Roy K. QSAR studies of CYP2D6 inhibitor aryloxypropanolamines using 2D and 3D descriptors. Chem Biol Drug Des 2009;73:442–455.
- Roy K, Roy PP. Comparative QSAR studies of CYP1A2 inhibitor flavonoids using 2D and 3D descriptors. Chem Biol Drug Des 2008;72:370–382.
- Lewis DFV, Dickins M. Baseline lipophilicity relationships in human cytochromes P450 associated with drug metabolism. Drug Metabol Rev 2003;35:1–18.
- Lewis DFV, Dickins M, Goldfarb PS. Molecular modelling of CYP2B6 based on homology with the CYP2C5 crystal structure: analysis of enzyme-substrate interactions. Drug Metab Drug Interact 2003;19:115–135.
- Lewis DFV, Lake BG, Ito Y, Anzenbacher P. Quantitative structure-activity relationships (QSARs) with cytochromes P450 2B (CYP2B) subfamily enzymes: the importance of lipophilicity for binding and metabolism. Drug Metab Drug Interact 2006;21:213–231.
- Scott EE, White MA, He YA, Johnson EF, Stout CD, Halpert JR. Structure of mammalian cytochrome P450 2B4 complexed with 4-(4-chlorophenyl)imidazole at 1.9å´ resolution. Journal of Biol Chem 2004;279:27294–27301.
- Jefcoate CR, Gaylor JL, and Calabrese R. Ligand interactions with cytochrome P450. 1. Binding of primary amines. Biochemistry, 8, 3455–3463.
- Dickins M, Bridges JW. The relationship between the binding of 2-n-alkyl-benzimidazoles to rat heptic microsomal cytochrome P450 and the inhibition of mono-oxygenation. Biochem Pharmacol 1982;31:1315–1320.
- Turpeinen M, Nieminen R, Juntunen T, Taavitsainen P, Raunio H Pelkonen O. Selective inhibition of CYP2B6-catalyzed bupropion hydroxylation in human liver microsomes in vitro. Drug Metab Dispos 2004;32:626–631.
- Vermeir M, Boens N, Heirwegh KPM. Interaction of 7-n-alkoxycoumarins with cytochrome P4502 and their partitioning into liposomal membranes. Biochem J 1992;284:483–490.
- Yih TD, van Rossum JM. KS values of some homologous series of barbiturates and the relationship with the lipophilicity and metabolic clearance. Biochem Pharmacol 1977;26:2117–2120.
- Sangster J. Octanol-water partition coefficients of simple organic compounds. Journal of Phys Chem Ref Data 1989;18:1111–1229.
- Hansch C, Leo A, Hoekman D. Exploring QSAR: Hydrophobic, Electronic and Steric Constants. Washington, American Chemical Society, 1995
- van de Waterbeemd H, Mannhold, R. Lipophilicity descriptors for structure-property correlation studies: overview of experimental and theoretical methods and a benchmark of log P calculations. In: Pliska, B. Testa and H. van de Waterbeemd, eds. Lipophilicity in Drug Action and Toxicology V. VCH, Weinheim, 1996:401–418.
- Lewis DFV, Jacobs MN, Dickins M. Compound lipophilicity for substrate binding to human P450s in drug metabolism. Drug DiscovToday 2004;9:530–537.
- Lewis DFV, Ito Y, Goldfarb PS. (2008) Enzyme-substrate binding interaction energies and their application to the cytochrome P450 system. Current Computer-Aided Drug Design 4, 111–122.
- Korhonen LE, Turpeinen M, Rahnasto M, Wittekindt C, Poso A, Pelkonen O, Raunio, H, Juvonen RO. (2007) New potent and selective cytochrome P450 2B6 (CYP2B6) inhibitors based on three-dimensional quantitative structure-activity relationship (3D-QSAR) analysis. Br J Pharmacol 150, 932–942.
- Fry JR, Hammond AH, Jassi KL, Bass AE, Bruce G, Laughton C, Shaw PN, Bylov IE, Kovalenko SM. (2004) Selectivity of 7-alkoxycoumarins as probe substrates for rat hepatic cytochrome P450 forms is influenced by the substitution pattern on the coumarin nucleus. Xenobiotica 34, 707–722.
- Ekins S, Wrighton SA. (1999) The role of CYP2B6 in human xenobiotic metabolism. Drug Metabol Rev 31, 719–754.
- Lewis DFV, Lake BG, Dickins M, Eddershaw PJ, Tarbit MH, Goldfarb PS. (1999) Molecular modelling of the phenobarbital-inducible P450 isoforms: CYP2B1, CYP2B4 and CYP2B6 by homology with the substrate-bound CYP102 crystal structure, and evaluation of CYP2B substrate binding affinity. Xenobiotica 29, 361–393.