Macroscopic models for studies of electrostatic interactions in proteins: limitations and applicability.

The validity of macroscopic models for calculations of electrostatic energies in proteins is examined. The Tanford-Kirkwood (TK) model is extended to include the self energy of the ionized groups. It is shown that ionized groups cannot exist inside nonpolar regions of proteins and argued that the experimental finding of ions inside proteins proves that the corresponding local environment is polar. The modified TK model (MTK model), which adjusts charge-charge interactions by the corresponding solvent accessibilities, is found to be inconsistent with the TK model, on which it is based. The MTK model corresponds to a polar interior whereas the TK model assumes a nonpolar interior. It is shown that models that assume a high dielectric constant for proteins give reasonable results for interactions between charged groups at equilibrium. It is then explained why, in contradiction to common belief, protein interiors are polar around charged groups. It is argued that in focusing on charge-charge interactions one overlooks the key contribution of the protein dipoles in determining the self energy of charges in the interior of proteins.