Toward accurate transferable electrostatic models for polypeptides: A distributed multipole study of blocked amino acid residue charge distributions

We have developed a method for building up accurate electrostatic models for polypeptides, based on a distributed multipole representation of the SCF charge densities for the dipeptides (CH3.CO.NH.CHR.CO.NH.CH3) of the naturally occurring amino acids. It is based on the observation that each peptide residue has almost the correct formal charge (0, ±1). We find that the variations in the backbone charge distributions (excluding proline) with sidechain have a negligible effect on the predicted electrostatic potential around the residue. However changes in the atomic multipoles with the torsion angles (ϕ, ψ, χ) are more significant, and may need to be taken into account if electrostatic potential close to the residue is required to high accuracy. This type of DMA peptide library provides more accurate, more theoretically based, estimates of the electrostatic potential around polypeptides than current models.