Density Functional Theory Charge Delocalization Error Determines Computed Activation Barriers in Enzymatic Kemp...

The Kemp elimination is a hallmark reaction in computational enzyme design. Designed Kemp eliminases have undergone directed evolution to achieve activities approaching those of natural enzymes. Their success partly reflects the reaction's apparent simplicity: a highly exergonic, single-step, base-catalyzed elimination that is amenable to quantum-mechanical and hybrid simulations. Consequently, extensive computational literature has examined how mutations shape Kemp eliminase fitness landscapes. However, recent findings challenge established assumptions about these enzymes and motivate reassessment of the computational strategies used to study them. Here, we systematically analyze the barrier height and transition-state geometry of Kemp elimination catalyzed by HG3.17 using static QM/MM calculations. Both quantities show large sensitivity to the fraction of Hartree-Fock exchange in the density functional, arising from charge-delocalization error in DFT. Activation barriers for the catalytic step span 1-14 kcal mol^-1, with only minor basis-set effects. Thus, appropriately adjusted numerical parameters can seemingly reproduce experimental rate constants, which vary with subtle experimental conditions and compress multiple phenomena beyond the chemical step into a single value. Moreover, the reaction may appear perfectly synchronous, concerted, or stepwise depending on the level of theory. These results caution against using apparent agreement with experiment as the primary criterion for simulation quality.