This project aims to develop new computational tools that will speed structure-based drug-discovery by providing a detailed analysis of hydration structure and thermodynamics of water in targeted protein binding pockets. The fundamental concept is to discretize the equations of inhomogeneous solvation theory (IST), up to second order, onto 3D grids of energy and entropy in a targeted binding site. These grids will be populated by molecular dynamics (MD) simulations with explicit solvent. The resulting treatment of solvation, termed GIST, will be characterized through comparisons with experimental data and rigorous thermodynamic integration free energy calculations for protein binding pockets. The GIST method will be incorporated into visualization tools to highlight binding site regions where it is particularly favorable or unfavorable to displace solvent, in order to speed ligand design and evaluate the "druggability" of protein binding pockets. It will also be integrated into fast new functions for ligand docking and scoring, for which promising preliminary results are provided in this proposal. Finally, in order to maximize scientific and health impact, the software will be packaged, documented and disseminated as part of the freely available, widely used and open source AMBER Tools software suite.
Most medications are molecules that work by binding tightly to a specific pocket in the surface of a protein involved in a disease process, and thus blocking the protein's disease-related function. This project aims to develop advanced computational methods that will speed the discovery of new medications by helping scientists design molecules that will bind tightly to a targeted protein. We will develop concepts and software that will reveal the 3D patterning of water molecules in a targeted binding pocket and will use this information about water to pinpoint areas in the binding pocket to which a potential medication can stick particularly tightly.