Theory and empirical studies in metazoans predict that apex predators should shape the behavior and ecology of mesopredators and prey at lower trophic levels. Despite the ecological importance of microbial communities, few studies of predatory microbes examine such behavioral res-ponses and the multiplicity of trophic interactions. Here, we sought to assemble a three-level microbial food chain and to test for behavioral interactions between the predatory nematode Caenorhabditis elegans and the predatory social bacterium Myxococcus xanthus when cultured together with two basal prey bacteria that both predators can eat—Escherichia coli and Flavobacterium johnsoniae. We found that >90% of C. elegans worms failed to interact with M. xanthus even when it was the only potential prey species available, whereas most worms were attracted to pure patches of E. coli and F. johnsoniae. In addition, M. xanthus altered nematode predatory behavior on basal prey, repelling C. elegans from two-species patches that would be attractive without M. xanthus, an effect similar to that of C. elegans pathogens. The nematode also influenced the behavior of the bacterial predator: M. xanthus increased its predatory swarming rate in response to C. elegans in a manner dependent both on basal-prey identity and on worm density. Our results suggest that M. xanthus is an unattractive prey for some soil nematodes and is actively avoided when other prey are available. Most broadly, we found that nematode and bacterial predators mutually influence one another’s predatory behavior, with likely consequences for coevolution within complex microbial food webs.

Theory and empirical studies in metazoans predict that apex predators should shape the behavior and ecology of mesopredators and prey at lower trophic levels. Despite the ecological importance of microbial communities, few studies of predatory microbes examine such behavioral responses and the multiplicity of trophic interactions. Here, we sought to assemble a three-level microbial food chain and to test for behavioral interactions between the predatory nematode Caenorhabditis elegans and the predatory social bacterium Myxococcus xanthus when cultured together with two basal prey bacteria that both predators can eat - Escherichia coli and Flavobacterium johnsoniae. We find that >90% of C. elegans worms failed to interact with M. xanthus even when it was the only potential prey species available, whereas most worms were attracted to pure patches of E. coli and F. johnsoniae. In addition, M. xanthus altered nematode predatory behavior on basal prey, repelling C. elegans from two-species patches that would be attractive without M. xanthus, an effect similar to that of C. elegans pathogens. The nematode also influenced the behavior of the bacterial predator: M. xanthus increased its predatory swarming rate in response to C. elegans in a manner dependent both on basal-prey identity and on worm density. Our results suggest that M. xanthus is an unattractive prey for some soil nematodes and is actively avoided when more lucrative prey are available. Most broadly, we find that nematode and bacterial predators mutually influence one another’s predatory behavior, with likely consequences for coevolution within complex microbial food webs.