To unravel enzymatic reaction mechanisms and predict reaction rates, it is essential to construct free energy profiles. In the computation of such free energy profiles, one chooses between ab inito (ai) QM/MM and semi-empirical (se) QM/MM Hamiltonians. In general, ai-QM/MM simulation would lead to accurate FE profiles but at a very steep cost; in contrast, se-QM/MM simulations are more affordable but yield much less accurate results. In this work, we adopt the reference potential method. In particular, the configurational space is sampled using a se-QM/MM Hamiltonian (PM3 re-parameterized via force matching), then the (B3LYP) ai-QM/MM-quality FE profiles are obtained through weighted thermodynamic perturbation. The accuracy and efficiency of such accelerated free energy profile computations will be demonstrated with several examples: SN2 reaction and glycine proton transfer in aqueous solutions; chorismate mutase catalysis; and Kemp elimination.

Accelerated computation of <i>ab initio</i> QM/MM-quality enzymatic reaction free energy profile using re-parameterized semi-empirical QM/MM reference potentials

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.

Accelerated computation of ab initio QM/MM-quality enzymatic reaction free energy profile using re-parameterized semi-empirical QM/MM reference potentials