Applications of Monte Carlo statistical mechanics to the analysis of protein-ligand binding affinities and to the effects of solvent on reaction rates

This work presents a variety of applications of the Monte Carlo method of statistical sampling in biological and chemical systems using a classical force-field, OPLS-AA. The key feature of this molecular mechanics description is the careful parameterization of the force field to accurately reproduce experimentally measurable physical properties. An example of this approach is presented for the specific case of the development of parameters for nitriles, esters, and nitro compounds. In all cases there was excellent agreement between experiment and thermodynamic properties computed from condensed-phase simulations. Once a basic force field has been developed and validated, more complex computations may be attempted. Protein-ligand interactions for 60 factor Xa inhibitors were computed and analyzed using the extended linear response method. The results of the simulations were in good agreement with experimentally derived binding affinities, and provided insight into the interactions important for high affinity binding. Finally, a combined QM/MM approach was used to study the dramatic solvent effects on a class of decarboxylation reactions. In the instance where internal hydrogen bonding was experimentally observed to have a small solvent effect, only our improved semi-empirical method was able to reproduce the experimental observations.