Bacterial Fitness Landscapes Stratify Based On Proteome Allocation Associated With Discrete Aero-Types

Author summaryGenome-scale models enable quantitative prediction of bacterial phenotypes and a fine-grained description of the underlying optimal proteome allocation. Thus, we can now analyze the phenotypic potential of a large number of Escherichia coli genotypes grown under different conditions, which leads to the discovery of a stratified distribution of phenotypes. The observed distribution is determined by distinct ATP generation strategies, defined as "aero-types", associated with optimal proteome allocation modulated upon differential usage of electron-transport-chain enzymes. This mechanistic approach offers us a genome-scale understanding of the fitness landscape, and fundamental interpretation of bacterial physiology and evolution dynamics.The fitness landscape is a concept commonly used to describe evolution towards optimal phenotypes. It can be reduced to mechanistic detail using genome-scale models (GEMs) from systems biology. We use recently developed GEMs of Metabolism and protein Expression (ME-models) to study the distribution of Escherichia coli phenotypes on the rate-yield plane. We found that the measured phenotypes distribute non-uniformly to form a highly stratified fitness landscape. Systems analysis of ME-model simulations suggest that this stratification results from discrete ATP generation strategies. Accordingly, we define "aero-types", a phenotypic trait that characterizes how a balanced proteome can achieve a given growth rate by modulating 1) the relative utilization of oxidative phosphorylation, glycolysis, and fermentation pathways; and 2) the differential employment of electron-transport-chain enzymes. This global, quantitative, and mechanistic systems biology interpretation of fitness landscape formed upon proteome allocation offers a fundamental understanding of bacterial physiology and evolution dynamics.