Selection for toxin production in spatially structured environments increases with growth rate.

Microbes adopt diverse strategies to successfully compete with coexisting strains for space and resources. One common strategy is the production of toxic compounds to inhibit competitors, but the strength and direction of selection for this strategy vary depending on the environment. Existing theoretical and experimental evidence suggests that growth in spatially structured environments makes toxin production more beneficial because competitive interactions are localized. Because higher growth rates reduce the length scale of interactions in structured environments, theory predicts that toxin production should be especially beneficial under these conditions. We tested this hypothesis by developing a genome-scale metabolic modeling approach and complementing it with comparative genomics to investigate the impact of growth rate on selection for costly toxin production. Our modeling approach expands the current abilities of the dynamic flux balance analysis platform Computation Of Microbial Ecosystems in Time and Space (COMETS) to incorporate signaling and toxin production. Using this capability, we find that our modeling framework predicts that the strength of selection for toxin production increases as growth rate increases. This finding is supported by comparative genomics analyses that include diverse microbial species. Our work emphasizes that toxin production is more likely to be maintained in rapidly growing, spatially structured communities, thus improving our ability to manage microbial communities and informing natural product discovery.