These are three major challenges in using computational methods to simulate protein and protein-ligand systems: (1) the need for improvement in the accuracy of the representation of the energy of the system; (2) the need for improvement in the sampling of both comformational and chemical space and (3) the need for better ways to combine quantum mechanical electronic structure and molecular mechanical methods to study enzyme catalyzed chemical reactions. We have made some advance in meeting these challenges in the last period of support and propose studies to further address them. We have developed a new force field for protein systems which, combined with a new approach to include long-range electrostatic effects, is a step forward in meeting challenge (1). But these is a clear need for more accurate short range electrostatic interactions and the inclusion of many-body effects - our proposed studies describe the development of models which contain these. We have shown that locally enhanced sampling (LES) to be very promising in calculation of loop structures and free energies. To address the conformational sampling challenge (2), we propose the continued development of LES methods and their application to antibody hypervariable loops, among others. Free energy calculations have been powerful in addressing the chemical sampling challenge (2), but they have been too computational intensive for many applications. We have developed a new qualitative free energy approach, PROFEC, which shows promise and we propose to develop a new method (MD/MC) that should broaden the applicability of free energy approaches to ligand and protein design problems. We have begun the development of a new, more powerful method to study enzyme catalysis, challenge (3), using a combination of high level ab initio calculation and classical free energy calculations. We propose to apply it to a variety of enzyme catalyzed reactions.
Showing the most recent 10 out of 81 publications
