Calculations of electrostatic properties in proteins. Analysis of contributions from induced protein dipoles

The calculation of induced dipole moments and of their contribution to electrostatic effects in proteins is implemented following the approach of Warshel. Isotropic polarizabilities are assigned to individual atoms, and the resulting deviation from pairwise interactions is treated by a self-consistent iterative procedure. We give a detailed description of how the formalism is implemented in molecular mechanics and molecular dynamics simulation procedures, and report results based on calculations performed on crystal structures of crambin, liver alcohol dehydrogenase and ribonuclease T1. We focus our analysis on evaluating the contribution of polarizability of the protein matrix to electrostatic energies, local fields, to dipole moments of peptide groups and of secondary structure elements in the polypeptide chain. Our calculations confirm that induced dipole moments in proteins provide important stabilizing contributions to electrostatic energies, and that these contributions cannot be mimicked by the usual approximations where either a continuum dielectric constant, or a distance-dependent dielectric function is used. We find that induced protein dipoles appreciably affect the magnitude and direction of local electrostatic fields in a manner that is strongly influenced by the microscopic environment in the protein. Most strongly affected are fields in charged groups that are involved in close interactions with other charged groups, while the influence on local fields of aliphatic groups is marginal. We find, moreover, that induction effects from surrounding protein atoms tend on average to increase peptide dipoles and helix macro-dipoles by about 16%, again reflecting electrostatic stabilization by the protein matrix, and show that (at least in the alpha/beta domain of alcohol dehydrogenase) the contribution of side-chains to this stabilization is significant.