A robust suite of tools will be developed and validated for the accurate prediction of catalytic proficiencies of enzymes through combined quantum mechanical-molecular mechanical computations of unprecedented accuracy. Accurate QM potentials will be computed using local correlation methods [LMP2, L-CCSD, L- CCSD(TO)] and through the use of a new composite ab initio approach based on the extrapolation of local coupled cluster energies to the complete basis set limit. This extrapolation procedure will be calibrated using a set of small molecules, to allow for the explicit comparison of extrapolated local correlation energies with results from established extrapolation approaches applied to conventional ab initio methods. The proposed QM/MM approaches will be validated and refined through comparison with an extensive set of experimentally determined kinetic data for wild type and mutated chorismate mutase and subtilisin proteins. DFT/MM methods will also be benchmarked against this experimental test set, utilizing new DFT functions that should offer substantial improvements to the venerable B3LYP. Once validated and fine- tuned through extensive comparisons with experimental data, these computational tools will be utilized in the ranking of novel enzyme designs to catalyze synthetically useful aldol condensations. Further, the QM/MM methods developed will allow for definitive quantitative studies of enzyme catalysis, including definitive studies of ODCase and other enzymes for which the catalytic mechanism is still unsettled. Computational tools to accurately model the catalytic efficacy of enzymes will be developed by combining quantum mechanical and classical mechanical methods. The immediate applications of these powerful computational tools will be in the design of new enzymes that catalyze reactions not found in nature. Such enzymes will be instrumental in novel synthetic pathways as well as for the development of enzyme-based therapeutics.
Showing the most recent 10 out of 11 publications
