The Thermodynamics of Carbon Redox Biochemistry: A Quantum Chemical Approach

Abstract

Thermodynamics plays an increasingly important role in modeling and engineering metabolism. Here we developed quantum chemistry approaches to predicting the thermodynamics of biochemical reactions in order to help fill in the gaps in the existing experimental thermodynamic data. We focus on redox biochemistry, which is ubiquitous and fundamental to all living systems. We develop calibrated quantum chemistry prediction methods, where systematic errors in ab initio estimates are corrected via calibration against available experimental data. Our approach can predict experimentally derived standard reduction potentials with considerably higher accuracy than the Group Contribution Method, the most commonly used approach. This enables us to decipher general trends between and within different oxidoreductase reaction categories. We also ask what distinguishes observed, natural redox biochemistry from the “null hypothesis” represented by the space of all possible redox reactions. Finally we adapt recent developments at the intersection of quantum chemistry and machine learning to obtain fast and accurate methods for predicting the thermodynamics of biochemical redox reactions in a high-throughput manner and predict the potentials of more than 315,000 redox reactions.