Metabolic Model-Based Analysis of the Emergence of Bacterial Cross-Feeding through Extensive Gene Loss

Metabolic dependencies between microbial species are common and have a significant impact on the assembly and resilience of microbial communities. However, the origins of such metabolic dependencies, the evolutionary forces that drive metabolic cross-feeding, and the impact of metabolic and genomic architecture on their emergence are not clear. To address these questions, we developed a novel simulation-based framework coupling a model of reductive evolution with a multi-species genome-scale model of microbial metabolism. We used this framework to model the evolution of a two-species microbial community, simulating thousands of independent evolutionary trajectories and investigating the link between genome reductive evolution and the emergence of metabolic interactions. Surprisingly, even though our model does not impose explicit selection for cooperation, metabolic dependencies emerged in nearly half of all evolutionary runs. Evolved dependencies involved cross-feeding of a diverse set of metabolites at varying frequencies, reflecting various constraints imposed by the metabolic network architecture. We additionally found metabolic ‘missed opportunities’, wherein species failed to capitalize on metabolites made available by their partners. When cross-feeding did evolve, it generally emerged immediately after a metabolite became available, but a complete dependence on such cross-fed metabolites often evolved relatively slowly. Examining the genes deleted and retained in each evolutionary trajectory and the timing of gene deletion events along these trajectories, we were further able to identify both genome-wide properties and specific gene retentions that were associated with metabolic phenotypes. Our findings provide insight into the evolution of cooperative metabolic interaction among microbial species, offering a unique view into the way such relationships could emerge in natural settings.