A family of contact-dependent nuclease effectors contain an exchangeable, species-identifying domain

In mixed-species communities, bacteria can deploy contact-dependent effectors to compete with other organisms, often directly injecting these proteins into neighboring cells. One current hypothesis is that the entire protein contains information specific for a single species; emergence of new effectors comes from acquiring genes. Here we characterize a family of DNA-degrading effectors that are nucleases which cause death. Like other families of chimeric nucleases, these effector toxins contain two domains. One is a PD-(D/E)XK-containing domain necessary for DNA cleavage. The other domain, which does not contain known DNA-binding structures encodes species-identifying information. By expanding current metagenomics methods, we capitalized on the species-identifying domain to differentiate among low-abundance species, as well as to reveal domain architectures within these proteins, in human gut and oral microbiomes. Emerging are questions about how low-abundance strains use toxins for survival and how strain-identifying toxins evolve.