Explicit treatment of water molecules in data-driven protein-protein docking: the solvated HADDOCKing approach

Water molecules are active components in, literally, every biochemical event, forming hydrogen bonds, filling cavities, and mediating interactions with other (bio)molecules. Therefore, solvent drastically affects the kinetics and thermodynamics of numerous cellular events, including protein-protein interactions. While docking techniques are becoming successful in predicting the three-dimensional structure of protein-protein complexes, they are still limited in accounting explicitly for water in the binding process. HADDOCK is one of the few programs so far that can explicitly deal with water molecules during docking. Its solvated docking protocol starts from hydrated molecules, and a fraction of the interfacial water is subsequently removed from the docked models in a biased Monte Carlo procedure. The Monte Carlo-based removal is based on interfacial amino acid-water contact propensities derived from a dataset of high-resolution crystal structures of protein-protein complexes. In this chapter, this solvated docking protocol is described and associated methodological aspects are illustrated through an application example. It is shown that, although docking results do not always improve when the solvated docking protocol is applied, scoring is improved and the positions of buried water molecules in an interface are correctly predicted. Therefore, by identifying important water molecules, solvated docking can aid the development of novel inhibitors of protein-protein complexes that might be better accommodated at an interface.