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.

Nuclease toxins of the HNH/ENDOVII fold. The HNH or the ENDOVII fold is a version of the treble-clef fold. The treble-clef fold is one of the most prevalent Zn-binding motifs across the three superkingdoms of life (56) . Classical HNH nucleases, like the restriction endonuclease (REase) McrA and the T4 endonuclease VII, contain the four conserved, Zn-chelating cysteines of the treble-clef fold (52) . However, these cysteines are lost in several forms, such as the REase MboII, colicin E8 and the NucA family, but these domains still retain the characteristic structural geometry of the treble-clef (52, 56) . The active site of these enzymes is formed at the interface of the characteristic helix and b-hairpin and contains a divalent cation, which is chelated by three polar residues usually from the first strand of the b-hairpin and the C-terminal helix of the treble-clef fold. The residues chelating the metal are typically histidine, aspartate and asparagine but their exact configuration can greatly vary between different members of this fold making them difficult targets for identification through sequence analysis (52) . Among the nucleases of this fold occurring in the neighborhood of the SUKH superfamily we observed eight distinct families spanning the entire gamut ranging from conventional HNH nucleases to certain highly derived forms that have not be identified before. The conventional HNH versions (e.g.AM1_0143, gi: 158333371 from the cyanobacterium A. marina) retain all the four cysteines of the treble-clef fold and a typical arrangement of residues chelating the catalytic metal. Others, like the nuclease domains of the PSPTO_3229 protein from P. syringae (gi: 28870395) and some CDI proteins, belong to the colicin E7/E8/E9 family (Figure 2) . A highly derived version is represented by the NucA family (57) , where structural analysis reveals that a treble-clef domain which has lost the characteristic cysteines is inserted between two copies of a three-stranded domain with distinct loop-like C-terminal extensions (Figure 3 ). We uncovered several divergent, earlier unrecognized NucA family nuclease domains in both the SUKH superfamily neighborhoods and CDI systems, such as those typified by the B. subtilis protein YeeF (gi: 251757354). The structural organization of the NucA domain suggests that it arose from an ancestral HNH/EndoVII domain, which 'carried' these duplicated three-stranded units along with it to form a more complex domain. Consistent with this proposal, we discovered a family of novel HNH fold nucleases in our gene-neighborhoods, which contain an active site similar to the NucA nucleases, but are standalone versions without the two flanking three-stranded units. We called this family GH-E after the three conserved residues associated with the active site typical of these domains. Interestingly, a subset of the GH-E family preserves the conserved cysteines of the treble-clef suggesting that they indeed represent the potential evolutionary intermediate from a classical HNH domain to the derived NucA-like forms (Figure 3) .