Novel procedure for developing implicit solvation models for peptides and proteins

Solvation is an important factor in the structure stabilization of proteins. The free energy of solvation has been commonly approximated by summing the products of the atomic solvation parameters (ASPs) and the solvent accessible surface areas, where the ASPs were obtained from thermodynamic experiments of small molecules. This summation was usually added to the force field without calibration, We propose deriving an optimized set of ASPs only by requiring that the minimized total energy of the experimentally known structure is the global minimum. This method is applied to a cyclic hexapeptide in DMSO and can also be extended to loops in proteins in water.