Growth-Coupled Fermentation Predicted By Next-Generation Genome-Scale Models

Fermentation is an optimal growth strategy for microbes living in the absence of oxygen and other external electron acceptors. During fermentation, microbes naturally secrete large quantities of small molecules to balance their redox states, and these secretions have long been exploited for industrial production of ethanol, butanol, acetone, glycerol, and lactate. Recently, genetic engineering has been used to modify microbial fermentation and thereby produce a more diverse range of molecules. We generated a database of 69 engineered, growth-coupled strains of Escherichia coli from the literature and simulated these engineered strains in the next-generation genome-scale model of metabolism and expression (ME-model). In the ME-model, the tradeoff between pathway-enzyme cost and thermodynamics plays a major role in determining the optimal fermentation pathways. This tradeoff is absent from models strictly comprising of metabolic reactions (M-models), and we show that the ME-model predicts fermentation pathway activity more accurately than any existing M-model of E. coli. Furthermore, we use the ME-model to show that certain kinetic parameters can determine whether a target molecule will be growth-coupled. In particular, kinetics are necessary for understanding the production of succinate by E. coli. By predicting the effects of pathway-enzyme cost and kinetics on fermentation and growth coupling, ME-models can help guide the design of new growth-coupled production strains.