Modeling Reality for Optimal Docking of Small Molecules to Biological Targets

From virtual screening to understanding the binding mode of novel ligands, docking methods are being increasingly used at multiple points in the drug discovery pipeline. It is now well established that the amount ( and quality) of information provided to the docking programs greatly influences their accuracy. Ultimately, the docking programs should consider all the factors involved in the ligand/macromolecule binding process. In fact, developers have been moving towards the modeling of the dynamics involved in solvated protein/ligand complexes to improve the binding mode prediction accuracy. The problem of modeling "reality" can be broken down into several factors including the consideration of both ligand and receptor flexibility and the consideration of bulk and bridging water molecules. Additional factors such as directional metal coordination, covalent binding and charge or proton transfers should also be considered but are often disregarded due to time constrains or lower interest from the medicinal chemistry community. Each of these problems requires a separate or combined conformational search technique. In this review, we will discuss the current status in the development of search engines focusing on 1) ligand flexibility, including cyclic portions, 2) receptor flexibility, 3) bridging water molecules and finally 4) the inclusion of metal coordination geometry in docking.