The proteome is a terminal electron acceptor

Microbial metabolism is impressively flexible, enabling growth even when available nutrients differ greatly from biomass in redox state. Escherichia coli, for example, rearranges its physiology to grow on reduced and oxidized carbon sources through several forms of fermentation and respiration. To understand the limits on and evolutionary consequences of this metabolic flexibility, we developed a coarse-grained mathematical framework coupling redox chemistry with principles of cellular resource allocation. Our models inherit key qualities from both of their antecedents: i) describing diverse metabolic chemistries and ii) enforcing the simultaneous balancing of atom (e.g., carbon), electron, and energy (adenosine triphosphate) flows, as in redox models, while iii) treating biomass as both the product and catalyst of the growth process, as in resource allocation models. Assembling integrated models of respiration, fermentation, and photosynthesis clarified key microbiological phenomena, including demonstrating that autotrophs grow more slowly than heterotrophs because of constraints imposed by the intracellular production of reduced carbon. Our model further predicted that heterotrophic growth is improved by matching the redox state of biomass to the nutrient environment. Through analysis of  ≈60,000 genomes and diverse proteomic datasets, we found evidence that proteins indeed accumulate amino acid substitutions promoting redox matching. We therefore propose an unexpected mode of genome evolution where substitutions neutral or even deleterious to the individual biochemical or structural functions of proteins can nonetheless be selected due to a redox-chemical benefit to the population.

We are grateful to E. Afik, L. Aristilde, A. Burlacot, J. Ciemniecki, A. Duarte, J. Goldford, S. Hirokawa, D. McRose, R. Murali, T. Roeschinger, and G. Salmon for useful discussions and to Y. M. Bar-On, G. Chure, S. Kuehn, D. LaRowe, and R. Milo for comments on the manuscript. Financial support from the NSF (PHY-1748958 to the Kavli Institute for Theoretical Physics to A.I.F. and A.G.), fellowships from the Jane Coffin Childs Memorial Fund and Burroughs Wellcome Fund (to A.I.F.), and the Govt of India’s Ramalingaswami Fellowship (to A.G.); Additional support from the Gordon and Betty Moore Foundation (grant GBMF4513 to AG), the Caltech Center for Evolutionary Sciences (W.W.F.), NIH (1R01AI127850-01A1 to D.K.N.), the Rosen Center at Caltech and NIH MIRA grant 1R35 GM118043 (to R.P.).