Sealy Center for Structural Biology, University of Texas Medical Branch at Galveston, Galveston, TX 77555-1157, U.S.A.

Incorporation of solvation energy contributions for energy refinement and folding of proteins.

Computer Simulation of Biomolecular Systems. 3:270-283; van Gunsteren, W.F., Weiner, P.K., Wilkinson, A.J., Eds., KLUWER/ESCOM, Dordrecht, The Netherlands.

Protein structures determined from X-ray or NMR data are generally refined by including empirical energy terms for intramolecular interactions. A good stereochemical quality of an experimentally determined protein structure is a necessary requirement for a high-resolution structure. Several force fields for intramolecular interactions in proteins are nowadays in widespread use. However, a low intramolecular energy of an experimentally determined structure does not prove that this structure is correct, as the analysis of incorrectly determined experimental and deliberately misfolded protein structures shows. It is necessary to include the protein-solvent interaction in the refinement process.

Various models for treating the protein-solvent interaction have been suggested and their strengths and limitations were critically evaluated in several recent reviews. Protein-solvent interaction can be computed with either explicit moving solvent particles, or in continuum models based on the Poisson-Boltzmann equation or on the solvent accessible surface area. Treating protein-solvent interactions in a continuum approximation is an order-of-magnitude faster than computations with explicit molecules. This efficiency makes this approach attractive for refinement calculations and studies of protein docking and folding.

This chapter describes basic mathematical features for calculating the accessible surface areas and their derivatives analytically. The strengths and limitations of each refinement method are evaluated and compared to the use of explicit water molecules in unconstrained and constrained molecular dynamics calculation.