Dynamic Charge Distribution as a Key Driver of Catalytic Reactivity in an Artificial Metalloenzyme

Abstract

Miniature artificial metalloenzymes such as mimochromes, which consist of a heme center protected by an upper and lower helix, provide a simplified platform to extract design principles for engineering rate enhancements beyond that of natural enzymes. However, design optimizations have largely focused on structural properties, leaving the impact of the electronic environment less explored. To investigate how the electronic environment influences reactivity, we carry out both classical and ab initio molecular dynamics (MD) simulations of a series of mimochromes, MC6, MC6*, and MC6*a, with known differences in reactivity. To uncover patterns in electronic structure, we train supervised machine learning (ML) models and carry out statistical analysis. The ab initio MD simulations uncover key charge coupling interactions between Arg10’ and Glu2’ that are not identified in classical dynamics. These residues are determined to be essential in orienting Arg10’ in the active site. Furthermore, using our ML classifiers trained to distinguish mimochromes based on partial charge dynamics alone, we identify critical residues such as Arg10’, Lys11, Glu2’, and Ser6’ that distinguish the mimochrome analogues. Disruption of the computationally-identified Glu2’···Lys11 salt bridge in a new mimochrome analogue, MC6*a-LysBoc, that we synthesize and characterize experimentally supports this prediction by resulting in a seven-fold decrease in the total turnover number. Quantum mechanical calculations show that Arg10’ stabilizes the transition state for the formation of the reactive oxo from hydrogen peroxide. These results indicate that partial charge distribution dynamics are important in the reactivity series in mimochromes, which should guide the engineering of a broad array of next-generation metalloenzymes.