Zhan Chen, Shan Zhao, Jaehun Chun, Dennis G. Thomas, Nathan A. Baker, Peter W. Bates, and G. W. Wei Journal of Chemical Physics 2012, 137, 084101 (Paywall)
As I discussed in a previous highlight, work by Ryde and co-workers suggest that one of the main obstacles to predicting accurate protein-ligand binding constants using continuum solvation models is the non-polar contributions to the change solvation energy.
I was therefore intrigued when ResearchGate alerted me to a new publication co-authored by Nathan Baker on a variational approach to non-polar solvation. In this approach the solute-solvent boundary surface is varied to minimize the non-polar solvation energy. The non-polar solvation energy is a function of the area and volume of the boundary surface and the solute-solvent van der Waals interactions.
The current implementation of the method contains three adjustable parameters: the surface tension and hydrodynamic pressure of the solvent and the well-depth of the Lennard-Jones potential describing the solute-solvent interaction. These parameters were obtained by fitting to experimental solvation energies of 11 alkanes, with an RMSD of 0.12 kcal/mol. These parameters were then used to predict the solvation energy of 19 new alkanes, with a resulting RMSD of 0.31 kcal/mol. The solvation energies of the 11 alkanes used for calibration are shown to be significantly more accurate than those obtained by explicit solvation.
It will be very interesting to see whether this approach will lead to significantly better protein-ligand binding energies, but several issues will have to be addressed first: How will the be necessary parameters be obtained? Is the approach efficient enough to handle protein-sized systems? Does the variationally determined solute-solvent surface lead to good polar solvation energies? It will be interesting to find out.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.