Structure‐Based Methods for Predicting the Sites and Products of Metabolism

Abstract

IntroductionCytochrome P450s (CYPs) are heme-containing enzymes that can be found in virtually all organisms. In humans, the CYP superfamily constitutes the most important enzymes in drug metabolism (see Chapters 4-8).Various computational approaches to study CYP metabolism have been reported over the years [1][2][3][4], and depending on the question that is being addressed, different methods may be more or less appropriate. For instance, if the actual reaction mechanism of a particular substrate is to be studied, a quantum mechanical description explicitly describing the electrons that are responsible for bonds being broken and formed is required [5]. On the other hand, a classification of a large number of compounds in terms of their likelihood to show an interaction with a CYP may rather be performed by application of cheminformatics machine learning tools [6]. In almost all cases, the versatility of the enzymes, the diversity of the substrates and inhibitors, and the malleability of the active sites pose additional challenges to the computational approach at hand. This chapter focuses on structure-based approaches to study the metabolism of substrates by CYP enzymes. Rather than addressing the actual enzymatic reactions, as in Chapters 6 and 11, we will focus here on the preferred orientations of substrates in the active site of CYP enzymes as a crucial first step in the prediction of metabolism. Using mostly examples from our own work, we will highlight some of the possibilities and challenges in structure-based predictions of sites of metabolism (SoMs). 6 Å RuleThe basic concept behind structure-based prediction of metabolism by CYP enzymes is extremely simple. The crucial cofactor of the enzymes is a heme 243 Drug Metabolism Prediction, First Edition. Edited by Johannes Kirchmair.

Computational methods and tools to predict cytochrome P450 metabolism for drug discovery

“…The output from these methods is easy to interrogate as they generate binding poses that can be inspected by the user and are rooted in the physical reality of explicitly modeling the binding event. Today, crystal structures are available for almost all CYP isoforms relevant to xenobiotic metabolism (Guengerich, ; Oostenbrink, ). Malleability of these enzymes, their complex interplay with water, and the hydrophobic character of their—in part—very large binding sites pose significant challenges to the application of structure‐based approaches, in particular docking.…”

CYP 2D6 Binding Affinity Predictions Using Multiple Ligand and Protein Conformations

“…Interestingly, it was possible to select from MD simulations only a few CYP 2D6 structures that can be used as docking templates to obtain SOM prediction accuracies of more than 80% [14]. Later, positional fluctuations observed during MD [14] of side chains of active-site residues (Glu216, Phe483) could be confirmed by comparison to newly published X-ray structures (PDB codes 3QM4 [15], 3TDA and 3TBG) [22]. …”

Alignment-Based Prediction of Sites of Metabolism

“…Current computational methods for SoM prediction range from knowledge-based approaches to machine learning, reactivity models, and structure-based approaches (mainly ligand docking). ,,, These methods for SoM prediction can be generalized into three categories: knowledge-, structure-, and ligand-based. Knowledge-based systems use relatively simple rules derived from empirical knowledge accumulated by medicinal chemists over decades of research.…”