Computer Simulations Reveal an Entirely Entropic Activation Barrier for the Chemical Step in a Designer Enzyme
Johan Åqvist
Abstract
Despite advances in computational design of protein structures, it has proven very difficult to design efficient enzyme catalysts by such approaches. One of the challenges in the field has been to computationally design enzymes that catalyze Kemp elimination, a reaction not observed in nature. Among several such designs, there is a series for which the catalytic rate constant could be improved by several thousand-fold by laboratory evolution, although it is still modest compared to natural enzymes catalyzing similar chemistry. These evolved designer enzymes also showed unusual temperature optima that were not related to thermal unfolding. Here, we report extensive computer simulations of both the catalyzed reaction and conformational thermodynamics of one these enzymes to analyze the underlying reasons for low catalytic activity and the anomalous temperature behavior. The results reveal that there exists a lower energy state of the enzyme-substrate complex, not seen in crystal structures with transition state analogues, which explains the low activity. Computational Arrhenius and van't Hoff plots for the chemical step and the transition between the two reactant states are both linear, and the resulting reaction thermodynamics is found to render the catalytic barrier entirely entropic. Kinetic modeling based on our calculated thermodynamic parameters gives two possible quantitative explanations for the temperature optimum: a change of rate-limiting step at 308 K or a heat capacity change of -0.3 kcal/mol/K upon substrate binding, where experimental data appear most consistent with the former.