Structural and energetic effects of the use of polarisable water to solvate proteins

Using a non-polarisable model (simple-point-charge (SPC)) for liquid water and two polarisable water models (COS/G2, COS/D), the effect of introducing molecular polarisability into the solvent upon protein structure and energetics is investigated for eight proteins, hen egg-white lysozyme (HEWL), major cold shock protein (CspA), protein G (GP), chorismate mutase (CM), the C-terminal domain of the ribosomal protein L7/L12 (RB), the amino terminal domain of phage 434 repressor (GRP), a 12-residue -hairpin (DNV) and the GCN trigger peptide (GTP), using MD simulation, one 50 ns simulation and four additional 20 ns simulations for each protein and each water model. The differences in overall structural and energetic properties of the proteins induced by the three different water models are small, except for the amino-terminal domain of phage 434 repressor (GRP). The polarisable COS/G2 water model induces a slightly stronger interaction with the proteins modelled using the GROMOS 54A7 force field than the non-polarisable SPC water model, while for the polarisable COS/D water model the opposite effect is observed.