Elucidating the essential interactions in protein-solvent preferential binding

Protein-solvent interactions are crucial for the protein stability and function. In mixtures of solvents, particularly in the presence of osmolytes, the protein stability can be enhanced and become resilient in response to external perturbations, as temperature or pressure variations. The thermodynamic basis for cossolvent-induced stabilization or denaturation of proteins relies on preferential interactions: when a cosolvent preferentially binds the protein surface (relative to water), it tends to increase the surface area of the protein, causing denaturation. On the contrary, crossolvents that are preferentially excluded (and the protein is preferentially hydrated) tend to favor compact protein structures, frequently associated with their functional forms. These mechanisms can be understood from molecular simulations using the Kirkwood-Buff solvation theory and the computation of appropriate integrals of the solvent densities around the protein. On the other side, dissecting the role of each interaction in thermodynamic effects is difficult, because of the local nature of protein-solvent interactions and the complexity of the protein structure. Here, we propose a method to understand the roles of local interactions in the observed thermodynamic stability of proteins from molecular dynamics simulations. Starting with a properly sampled simulation of the unfolding of a helical peptide in mixtures of water and TFE, we characterize the solvent-shell interactions of the peptide with water and TFE using minimum-distance distribution functions computed with the ComplexMixtures.jl package. Next, we compute the numerical derivatives of the helical content of the protein as a function of the strength of each of the key protein-solvent interactions and solvent-solvent interactions, to classify their stabilizing and denaturing effects, both qualitatively and quantitatively. The analysis allows the build up of intuitive interpretations of solvent effects on protein stability, which can be used for rational solvent and protein design.