Allopatric divergence limits cheating range and alters genetic requirements for a cooperative trait

Abstract

Social and genomic context may constrain the fates of mutations in cooperation genes. While some mechanisms limiting cheaters evolve in the presence of cheating, here we ask whether cheater resistance can evolve latently even in environments where cooperation is not expressed and cheaters are absent. The bacteriumMyxococcus xanthusundergoes cooperative multicellular development upon starvation, but developmentally defective cheaters can outcompete cooperators within mixed groups. Using natural isolates and an obligate cheater disrupted at the developmental-signaling genecsgA, we show that cheating range is narrow among natural strains due to antagonisms that do not specifically target cheaters. Further, we mixed the cheater with closely related cooperators that diverged from it allopatrically in nutrient-rich environments in which cooperative development does not occur, showing that even slight divergence under these conditions can eliminate cheating phenotypes. Our results suggest that such cooperation- and cheater-blind divergence can generate a geographic mosaic of local cheater-cooperator compatibility patches that limit cheater spread. We also ask whether genomic divergence can shape the fitness effects of disrupting a cooperation gene. Construction of the samecsgAmutation in several natural-isolate cooperators generated a wide range of pure-culture sporulation phenotypes, from a complete defect to no defect. Thus, we find that epistatic interactions limit the range of genomes within which a mutation creates a cooperation defect. Moreover, these results reveal Developmental System Drift in a microbial system because sporulation proficiency is conserved across the natural isolates despite divergence in the role ofcsgA.Significance statementSelection on cooperators exploited by obligate cheaters can induce evolution of resistance to cheating. Here we show that cooperators can also rapidly evolve immunity to cheating simply as a byproduct of evolutionary divergence in environments in which cooperation and cheating at the focal trait do not occur because the trait is not expressed. We also find that differences in the genomic context in which a cooperation-gene mutation arises can profoundly alter its phenotypic effect and determine whether the mutation generates a social defect at all - a pre-requisite for obligate cheating. These findings suggest that general divergence of social populations under a broad range of environmental conditions can restrict both the set of mutations that might generate social defectors in the first place and the host range of such defectors once they arise.