A large, distinct family of RNase H fold nucleases associated with numerous defense systems in prokaryotes

During the last decade, systematic efforts aimed at the characterization of prokaryotic immunity resulted in the discovery of several hundred distinct defense systems in bacteria and archaea. This suggests that the common defense systems were already known and only rare ones remained to be identified. Here we present a genomic census and analysis of a vast family of bacterial and archaeal proteins (known as Domain of Unknown Function, DUF3800) that contain a distinct variant of the RNase H fold and are encoded in a striking 15% of the available complete prokaryotic genomes. Most of these proteins are encoded in defense islands, often in a close association with known defense systems, hence the designation WARD ( W eaponized A ntiphage R Nase-dependent D efense). The WARDs display a broad variety of protein domain architectures and genomic contexts, including a strong link to restriction-modification, CRISPR-Cas, retrons, and non-coding RNAs, suggestive of a substantial diversity in defense mechanisms. We experimentally tested 10 WARDs, mostly, encoded by standalone genes, and found that 9 of these conferred robust protection against diverse phages in Escherichia coli . Therefore, WARDs appear to represent the third most abundant defense system type in prokaryotes.