Cytochromes P450 (CYPs) constitute the most important family of drugs metabolizing enzymes. They play an important role in the disposition of drugs, and their pharmacological and toxicological effects. Early consideration of ADME(T)-properties (Absorption (A), Disposition (D), Metabolism (M), Excretion (E), and Toxicology (T)) is increasingly seen as essential for efficient discovery of drug candidates and the development of new drugs. Apart from the application of experimental in vitro tools, this necessitates the application of novel in silico tools that can accurately predict ADME properties of drug candidates already in early stages of the lead finding and optimization process.
The primary aim of this paper is to present in an integrative manner computational approaches used to understand, rationalize and predict the activity and substrate selectivity of CYPs, as well as the possibilities and limitations of these approaches now and in the future. The computational models and methods are divided into ligand based (QM calculations, classical and 3D-QSAR pharmacophore models), protein based (crystallographical and homology-based protein models), and ligand-protein binding (automated docking, molecular dynamics) methods. General and more specific conclusions are presented for the different computational methods applied to CYPs, and for each of the most important CYP isoforms. Expected future developments of novel methodologies and of the application of existing methods not yet applied to CYPs are discussed.
This review shows that a variety of in silico modeling approaches have been applied to CYP enzymes, thereby successfully adding to our understanding of CYP structure and function in a way that is complementary to experimental studies. It is concluded that combining existing computational methodologies into an integrative in silico approach for modeling of CYP structure, function, and dynamics is necessary to arrive at successful and meaningful rationalization of experimental data, and interpretation and prediction of CYP function, notably in terms of substrate binding, catalytic regiospecificity, and turnover. In contrast to product formation, accurate rationalization and prediction of rates of product formation is still beyond reach. Finally, it is concluded that the combination of new computational methods and solid experimental data will prove to be critical for the elucidation of as yet unknown functional and structural properties and mechanisms of CYPs.