Emergent sub-population behavior uncovered with a community dynamic metabolic model of Escherichia coli diauxic growth
Microbial communities have adapted to greatly variable environments in order to survive both short-term perturbations and permanent changes. The diauxic shift of Escherichia coli, growing first on glucose and, after it is exhausted, on acetate, is still today actively studied. As a fundamental example of metabolic adaptation, we are interested in understanding if diauxie in monocultures of E. coli is a coordinated and uniform metabolic shift, or rather the observable emergent result of individual sub-populations behavior. To do so we first develop a modeling framework that integrates dynamic models (ordinary differential equation systems) with structural models (metabolic networks), providing an open source modeling framework that is suitable to investigate the dynamics of microbial communities. We apply our methods to model E. coli either as having an average, unique metabolic state or as being the combination of two E. coli populations adapted to one of the two carbon sources. Our results are in strong agreement with previously published data and suggest that rather than a coordinated metabolic shift, diauxie could be the emergent pattern resulting from a survival strategy where individual cells differentiate for optimal growth on different substrates in view of environmental fluctuations. This work offers a new perspective on how to use dynamic metabolic modeling to investigate population dynamics, as the proposed approach can be easily transfered to studies on other multi-species communities as well as single cells.\n\nImportanceEscherichia coli diauxie is a fundamental example of metabolic adaptation that has not yet been completely understood, and to this aim approaches integrating experimental and theoretical biology are needed. We present a novel dynamic metabolic modeling approach that captures diauxie as an emergent property of sub-population dynamics rather than a homogeneous metabolic shift in E. coli monocultures. Without fine tuning of the parameters of E. coli core genome-scale model we obtain good agreement with published data. Our results suggest a change of paradigm in using single organism metabolic models, which can only to a certain approximation represent the average population metabolic state. We finally provide an open source modeling framework that can be applied to model multi-organism dynamics in variable environments.