LAUSR

A recent paper in Science by Gao et al. extended a comprehensive comparative genomics approach to identify new antiviral systems in prokaryotes for combating invading phages. These defense systems manifest a variety of host defense mechanisms by utilizing enzymatic activities including reverse transcriptases, adenosine deaminases of RNA editing, and retroms (Fig. 1). Some of these antiviral systems not only enrich our understanding the phage-bacterial interaction but also represent a versatile, powerful tool for biomedical research and biotechnological applications. Because of the arms race between viruses and prokaryotes, bacteria and archaea have evolved multiple sophisticated antiviral defense strategies to combat phages, for example, restriction modification (RM), abortive infection (Abi) systems, and CRISPRCas systems. Gao et al. described computational analysis of all bacterial and archaeal genomes, encoding over 620 million proteins to discover new antiviral systems. The authors first used known antiviral genes cassettes locating ‘defense islands’ as anchors to search for neighboring uncharacterized genes (Fig. 1), because defense islands often contain abundant antiviral genes that function together between different types of antiviral defense systems with overlap functions to maintain homeostasis. Similar computational pipeline has been published in another Science paper published in 2018 by Doron et al. to discover bacterial defense systems from 160 million genes and revealing nine new antiphage systems and one novel anti-plasmid system. Using this approach, Gao and colleagues detected a total of 7472 putative defense gene families, containing 1687 uncharacteristic function genes, proximity to known defense systems in defense islands. Further studies identified homologs without use of domain annotations showing that candidate defense gene clusters are evolutionarily conserved across multiple genomes across widespread microorganisms, indicating the existence of other unidentified defense systems. To identify these predicted defense gene families, Gao et al. selected 48 candidates from new defense systems to experimentally validate antiviral activities through heterologous reconstitution (Fig. 1). To test multiple variants of candidate defense systems, a collection of one to four homologs of each novel system were chosen to engineer into Escherichia coli. A diversity of coliphages were used to infect these engineered E. coli strains to investigate their antiviral activity. To the authors’ pleasant surprise, 29 of candidate defense systems (29/48, 60%) provided resistance against phages (Fig. 1), but phage specificity was typically narrow and varied widely across systems. Of note, whether the other 19 candidate systems contain no antiviral immunity needs future investigation, because they were only tested experimentally under specific laboratory conditions and were expressed in E. coli hosts that do not normally express these genes, or lack of appropriate phage targets: defense mechanisms are often effective only against specific phage groups, or inadvertent choice of unfunctional gene sets for testing. It would be interesting to verify that these defense systems truly exert defense functions in the environmental conditions in future. These validated defense systems demonstrate an abundance ranging from ~0.1 to ~10% of bacterial and archaeal phyla. One of these, phage restriction by an adenosine deaminase acting on RNA (RADAR), forms three subtypes: RADAR standalone containing only adenosine triphosphatase (ATPase) and adenosine deaminase (rdrAB), Csx27-associated RADAR (rdrABC), and SLATT-associated RADAR (rdrABCD). Authors suggest that RADAR represents an example of defense via adenosine-to-inosine (A-to-G) RNA editing, in which both RADAR system and phage infection are required for the occurrence of RNA editing. Broad distribution of editing sites is noted in both phage transcriptomes and engineered E. coli, resulting in host growth arrest. Therefore, RADAR is analogous to editingdependent Abi, in which prokaryotes commit altruistic cellular suicide to protect larger populations from phage infection. Another interesting system is defense-associated reverse transcriptases (DRTs) identified by enrichment of a family of RTs: DRT type 1 (UG1), DRT type 2 (UG2), DRT type 3 (UG3 and UG8), DRT type 4 (UG15), and DRT type 5 (UG16) (Fig. 1), displaying the distinct pattern of phage resistance. In DRT type 1, UG1 encodes nitrilase domains. Nitrilases are involved in natural product biosynthesis, including nucleotide metabolism. Authors reveal that nitrilase domain is key for anti-phage ability, exemplifying a non-defense domain that was apparently co-opted for a defense function. DRT type I inhibits late viral gene expression but not early/middle genes and may have no effect on early phage DNA injection. Another RTsmediated antiviral defense is dependent on retrons: Retron-TIR (Toll/ interleukin-1 receptor domain), Ec67 (Retron-TOPRIM, topoisomerase-primase domain), Ec86 (Nuc_deoxy+retron) and Ec78 (Retron+ATPase+HNH) (Fig. 1). Retron affects on producing extrachromosomal satellite DNA (msDNA). Gao et al. find that both synthesis and structure of msDNA are required for defense activity. Retron-TIR systems are associated with TIR domain for sensing pathogen and immune signal transduction, a common feature of innate immune systems in animals, plants, belonging to Thoeris system (thsAB). The motif in Retron-TIR and Thoeris system may be