Exploring enzyme chemistry through molecular simulation with QM/MM

Abstract

Chapter 1 contains a detailed overview of the relevant theoretical background, including quantum chemical methods through density-functional theory and empirical potentials, as well as quantum-mechanics/molecular-mechanics (QM/MM) embedding methods, continuum solvation models, and methods for thermochemical corrections.

Chapter 2 describes results from high-level QM/MM calculations on a selection of heme peroxidase enzymes involved in lignin degradation. I demonstrate that existing conceptual models for their activity (and pH dependence) do not stand up to scrutiny and require substantial re-evaluation. I identify in previous studies the misattribution of some results to spurious effects from a residual system charge, which I argue is entirely artificial.

This chapter also describes protracted efforts to identify protein sites in lignin peroxidase that are potential mutation hotspots. I conclude that simple geometry screening protocols do not work and demonstrate the critical importance of first-principles modelling. However, I am able to validate a proposed mutant (LiP:D183N) from an earlier study with an increased redox potential and suggest a framework for active site design based on a more general environment model inspired by the well-developed concept of electrostatic preorganisation in adjacent literature. I also briefly explore the chemical modification of heme as an engineering strategy, and report calculations on a novel variant of lignin peroxidase incorporating ring-fluorinated heme.

Chapter 3 reports on the complete QM/MM characterisation of intra-molecular ester bond formation in the bacterial adhesin SaTIE:ED1. While I am able to identify an appropriate reaction path, I show from computed activation barriers that this highly unusual cross-link is unlikely to form in the crystalline phase following the proposed mechanism. I briefly address the implications of this for our collaborators, discuss alternative mechanism proposals, and explore several methods for presenting potential energies over multiple minima.