Solvation Energetic Costs of Cognate Binding Site Formation

Structural fluctuations of proteins can reveal alternate binding site conformations or cryptic pockets that may be exploited to discover novel, tightly bound chemical compounds. While significant effort has been dedicated to the exploration of protein conformational space, the thermodynamic role of solvation and how it is coupled to a protein's structural fluctuations, particularly in binding site formation, has not been well characterized. In this study, we examine how binding site solvation energetics differ between unligated rigid cavities restrained about their ligand-bound conformations and the same cavities with flexible side chains free to explore conformational space in molecular dynamics simulations. We find that, on average, the solvation energy of flexible binding sites is significantly more favorable, 14.4 kcal/mol, than that of their rigid counterparts. Our analysis of the solvation reveals that this energetic discrepancy is driven by the flexible binding sites structuring themselves to form more energetically favorable protein-water hydrogen bonds than in the rigid cavities. The substantial solvation energetic cost for a flexible protein to adopt conformations that are complementary to cognate ligands (We use the term cognate ligand to refer to the ligand in the cocrystallized complex in the corresponding pdb entry. The term cognate structure refers to the experimentally determined protein-ligand complex containing this ligand.) led us to hypothesize that there may be little overlap between binding site side chain configurations of unligated proteins and those of ligated proteins that have structured their cavities to optimize protein-ligand interactions. We therefore investigate the configurations of flexible binding site side chains in unligated systems and find that in some proteins, they do not sample conformations that are complementary to their cognate ligands in molecular dynamics simulations. Notably, we identify a class of binding sites characterized by highly enclosed cavities with bidentate ligand interactions that are especially prone to this solvation-induced conformational occlusion, in which there is little to no overlap in the conformational landscapes of ligated and unligated binding cavities. We discuss how understanding the interplay between solvation energetics and protein structural fluctuations can inform the development of methods aimed at discovering alternative binding pockets, improve methods such as WaterMap and GIST that estimate the contribution to binding affinity of displacing water upon ligand binding, and can be used to inform bindability assessments of revealed cryptic pockets.