(71e) Ground and Transition State Computations On Cytochrome P450 BM-3

In this work, we introduce combined use of ground state and transition state computations to understand how specific mutations present in engineered variants of cytochrome P450 BM-3 confer improved reactivity or specificity. The cytochrome P450 BM-3 monooxygenase has been the target of extensive directed evolution by Arnold and coworkers. The fatty acid hydroxylase is functionally expressed at high levels in E. coli and has been engineered to convert small alkanes to their corresponding alcohols, with an emphasis in biofuel production. We first computationally assessed the effects of 14 different experimentally isolated mutations in P450 mutant 535-h (3 mutations lie in the active site) on ethane binding utilizing CHARMM binding energy calculations. In more than half of the positions mutated, the experimentally isolated mutation corresponds to either the maximum or near maximum improvement in binding energy between the enzyme and ethane, compared to all other possible amino acid substitutions. The remaining mutations caused only minor changes in binding energy, suggesting other factors such as transition state stabilization may explain the beneficial role of these selected mutations. We assessed this possibility by integrating quantum mechanical (QM) calculations and force-field methods to assess transition state stabilization of the enzyme. The structure and atomic charges of the transition state were determined and used to parameterize force-field calculations of transition state stabilization energies for each mutant. For the wild-type enzyme, our results indicate that the activation barrier is significantly reduced over the isolated porphyrin. The transition state stabilization energies offered by each single mutation demonstrated that some mutations that have little to no effect on binding at the ground state have a substantial effect on the oxidation activity by reducing the energy barrier. A comparison of the QM-parameterized force-field calculations and QM/MM embedding calculations was done to assess the reliability of the former, faster approach. Molecular dynamics simulations of the 535-h mutant and wild-type enzyme were performed to assess the importance of conformational sampling in examining mutations. Overall perspective on the reliability of these approaches in evaluating ground state and transition state stability upon mutation will be discussed.