A novel immunity system for bacterial nucleic acid degrading toxins and its recruitment in various eukaryotic and DNA viral systems

Zhang, Dapeng; Iyer, Lakshminarayan M.; Aravind, L.

Relationship of toxin systems to genetic conflicts in the bacterial world. Classical colicins and earlier characterized CDIs act primarily on related bacterial strains of the same 'species'. Although the systems identified in our studies are abundantly represented in extracellular pathogenic bacteria, they are rare in intracellular symbionts or pathogens. This might be because intracellular bacteria are much less likely to encounter a heavy load of competing cells in the same niche. The bacterial toxin systems which we uncovered in this study and the related CDIs are also different in certain features from the classical colicin-like systems. Classical colicins are in large part encoded on plasmids, which might be either single copy, medium-sized conjugative plasmids or small multi-copy small plasmids that depend on the conjugative plasmids for their transmission (8) . Such bacteriocins are relatively rare on chromosomes. In contrast, 99.25% of the systems recovered in our study are chromosomally encoded. Majority of the plasmid-encoded classical colicin-like toxins are accompanied by a gene encoding a lysis protein and their release is concomitant with the lysis of the host cell. However, none of the systems identified in this study or the CDIs have lysis genes in their neighborhoods (25) . This difference suggests that, while both the plasmid-borne bacteriocins and these systems might be directed at close relatives, they appear to be geared toward distinct genetic conflicts. The lysis of the cell nullifies the fitness of the chromosome; hence, it would be largely deleterious for the chromosome to encode systems that require lysis. The plasmid being a selfish element is not completely affected by loss of fitness of the host as long as it can offset it by holding on to, or spreading in the host population (i.e. the plasmid's own fitness is enhanced or maintained). Cells of the host type without the bacteriocinogenic plasmid are competitors that affect the plasmid fitness, especially under stationary phase or starvation conditions. Hence, the plasmid-borne colicin would be primarily selected to act against host cells that have lost the plasmid or lack it by default under these stress conditions. Further, the plasmid toxins are unlikely to have ready access to trafficking by the host because, given the large amounts in which the colicins are produced (8) , their export is likely to impair host fitness. Further, it has been shown that under starvation only $3% of the cells produce colicin (91) . Although the loss of the cells producing the colicin would endanger the resident plasmid, a relatively small fraction of the host population is affected. By the principle of inclusive fitness of kin (92), the plasmid could still have an enhancement of fitness from the copies in the surviving cell along with the elimination of competitors by the released toxin. On the other hand, the toxin domains of many of the chromosomal versions like the CDIs and those identified in this study appear to be borne on filamentous structures that are primarily geared toward to elimination of competitors that come in physical contact with the cell-surface (25, 93) . Therefore, these systems are likely to be critical in the context of the formation and organization of biofilms and solid substrate colonies. When bacterial cells are aggregating in the above contexts it would benefit to eliminate resource sharing with non-kin competitors. Hence, presence of a chromosomall