- Protein–Ligand Docking: From Basic Principles to Advanced Applications

Abstract

Given the three-dimensional (3D) structure of a protein and the molecular constitution of a ligand, what is the binding mode of the ligand, that is, its position, orientation, and conformation when bound to the protein? This is the fundamental question of the “docking problem” (Sotriffer et al. 2003b). Although at rst sight it reminds of a 3D jigsaw puzzle, where two complementary shapes have to be matched, the complexity of the docking problem goes far beyond. In fact, the binding partners are not rigid bodies, but exible molecules surrounded by water and governed in their energetics not simply by hard-sphere potentials, but rather by complex enthalpic and entropic contributions (Bissantz et al. 2010; Sotriffer et al. 2003a,b). Moreover, to deal with the problem a workow for appropriate setup and handling of the underlying molecular structures is required (Cole et al. 2011; ten Brink and ExnerCONTENTS6.1 The Docking Problem 155 6.2 Components of Docking 1576.2.1 Molecular Representations 157 6.2.2 Search Algorithms 158 6.2.3 Scoring Methods 1636.2.3.1 Knowledge-Based Scoring Functions 163 6.2.3.2 Force Field-Based Scoring Functions 164 6.2.3.3 Empirical Scoring Functions 165 6.2.3.4 Further Approaches to Scoring 1666.2.4 Docking Under Special “Scenarios” 166 6.3 Challenges for Docking Methods 168 6.4 Recent Advances and Selected New Approaches 171 6.5 Conclusion 176 References 1762009). Accordingly, the constitutive elements of any docking approach are (1) a suitable representation of the interacting molecular structures; (2) a search algorithm to generate meaningful binding poses of the ligand at the protein’s binding site; and (3) a scoring method to evaluate the generated congurations and identify the best one, ideally corresponding to the true, experimentally veriable binding mode. These three components are tightly related to each other, as not every type of molecular representation or scoring function can be combined with every possible search algorithm. In the following, commonly used approaches for “representing,” “searching,” and “scoring” in the context of docking are briey outlined (cf. Figure 6.1), without attempting to fully cover the large variety of methods grown over the past years. As docking is a well and frequently reviewed subject in the literature, the interested reader is referred to more extensive reviews for further details and additional approaches (Brooijmans and Kuntz 2003; Halperin et al. 2002; Moitessier et al. 2008; Muegge and Rarey 2001; Sotriffer 2011; Sotriffer et al. 2003a,b; Sousa et al. 2006, 2013; Yuriev and Ramsland 2013; Yuriev et al. 2011). For links to the distributor websites of docking programs, the “Click2Drug” Directory ofcomputer-aided Drug Design of the Swiss Institute of Bioinformatics should be consulted (http://www.click2drug. org/index.html#Docking).