Computational Modeling of (De)-Solvation Effects and Protein Flexibility in Protein-Ligand Binding Using Molecular Dynamics Simulations

Ying Yang, Purdue University


Water is a crucial participant in virtually all cellular functions. Evidently, water molecules in the binding site contribute significantly to the strength of intermolecular interactions in the aqueous phase by mediating protein-ligand interactions, solvating and de-solvating both ligand and protein upon protein-ligand dissociation and association. Recently many published studies use water distributions in the binding site to retrospectively explain and rationalize unexpected trends in structure-activity relationships (SAR). However, traditional approaches cannot quantitatively predict the thermodynamic properties of water molecules in the binding sites and its associated contribution to the binding free energy of a ligand. We have developed and validated a computational method named WATsite to exploit high-resolution solvation maps and thermodynamic profiles to elucidate the water molecules' potential contribution to protein-ligand and protein-protein binding. We have also demonstrated the utility of the computational method WATsite to help direct medicinal chemistry efforts by using explicit water de-solvation. In addition, protein conformational change is typically involved in the ligand-binding process which may completely change the position and thermodynamic properties of the water molecules in the binding site before or upon ligand binding. We have shown the interplay between protein flexibility and solvent reorganization, and we provide a quantitative estimation of the influence of protein flexibility on de- solvation free energy and, therefore, protein-ligand binding. Different ligands binding to the same target protein can induce different conformational adaptations. In order to apply WATsite to an ensemble of different protein conformations, a more efficient implementation of WATsite based on GPU-acceleration and system truncation has been developed. Lastly, by extending the simulation protocol from pure water to mixed water-organic probes simulations, accurate modeling of halogen atom-protein interactions has been achieved.




Lill, Purdue University.

Subject Area

Computational chemistry|Biophysics|Bioinformatics

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server