Computation of protein pK's values by an integrated density functional theory/polarizable contimuum model approach

This paper describes the extension of our computational strategy for pK predictions of small molecules to large solutes. The basic computational tool results from the coupling of quantum mechanical methods rooted in the density functional theory with the most recent version of the Polarizable Continuum Model. However, a third level is introduced, which includes solute regions far from the reactive center, which are described at a simplified level. This partition, together with the recent implementation of fast cavity generation, powerful iterative solvers, and fast multipole technology, allows us to tackle solutes of the dimension of a small protein. The problems and perspectives of this methodology are analyzed with special reference to the behavior of different Polarizable Continuum Model versions on the challenging playground represented by the pK's of the different histidine residues occurring in the human prion protein.