Improvement of the self-consistent-charge density-functional-tight-binding theory by a modified Mulliken charge

Abstract

Although extensive efforts had been carried out to improve the accuracy of the self-consistent-charge density-functional-tight-binding (SCC-DFTB), the application of SCC-DFTB on biological systems (e.g., enzymes) is still limited. Our benchmark calculations show that the original SCC-DFTB/MM is not able to properly descript the human carboxylesterase 1 (CES1) catalyzed hydrolysis of d-threo-methylphenidate (dMD). In contrast to the ab initio QM/MM results, SCC-DFTB/MM underestimates the activation free energy barrier and renders the acylation process as a single-step reaction without a tetrahedral intermediate. It seems like that SCC-DFTB/MM misestimates the developing negative QM Mulliken charge of the substrate oxygen atom in the oxyanion hole. To improve the SCC-DFTB energy and the electrostatic interaction between QM (SCC-DFTB) and MM atoms, we adopt an optimization strategy for QM Mulliken charge, in which an empirical parameter is trained to fit the SCC-DFTB Mulliken charge to high-level DFT method along the reaction coordinate. Herein, the optimized Mulliken charge-based SCC-DFTB method is denoted as SCC-DFTBMR. Finally, benchmark and free energy calculations were performed to prove the applicability of SCC-DFTBMR to CES1-catalyzed reaction.