Type III CRISPR–Cas systems provide immunity to foreign DNA by targeting its transcripts. Target recognition activates RNases and DNases that may either destroy foreign DNA directly or elicit collateral damage inducing death of infected cells. While some Type III systems encode a reverse transcriptase to acquire spacers from foreign transcripts, most contain conventional spacer acquisition machinery found in DNA-targeting systems. We studied Type III spacer acquisition in phage-infected Thermus thermophilus, a bacterium that lacks either a standalone reverse transcriptase or its fusion to spacer integrase Cas1. Cells with spacers targeting a subset of phage transcripts survived the infection, indicating that Type III immunity does not operate through altruistic suicide. In the absence of selection spacers were acquired from both strands of phage DNA, indicating that no mechanism ensuring acquisition of RNA-targeting spacers exists. Spacers that protect the host from the phage demonstrate a very strong strand bias due to positive selection during infection. Phages that escaped Type III interference accumulated deletions of integral number of codons in an essential gene and much longer deletions in a non-essential gene. This and the fact that Type III immunity can be provided by plasmid-borne mini-arrays open ways for genomic manipulation of Thermus phages.

CrAss-like phages are a recently described family-level group of viruses that includes the most abundant virus in the human gut 1,2 . Genomes of all crAss-like phages encode a large virionpackaged protein 2,3 that contains a DFDxD sequence motif, which forms the catalytic site in cellular multisubunit RNA polymerases (RNAPs) 4 . Using Cellulophaga baltica crAss-like phage phi14:2 as a model system, we show that this protein is a novel DNA-dependent RNAP that is translocated into the host cell along with the phage DNA and transcribes early phage genes. We determined the crystal structure of this 2,180-residue enzyme in a self-inhibited, likely pre-virionpackaged state. This conformation is attained with the help of a Cleft-blocking domain that interacts with the active site motif and occupies the RNA-DNA hybrid binding grove. Structurally, phi14:2 RNAP is most similar to eukaryotic RNAPs involved in RNA interference 5,6 , although most of phi14:2 RNAP structure (nearly 1,600 residues) maps to a new region of protein folding space.Considering the structural similarity, we propose that eukaryal RNA interference polymerases take their origin in a phage, which parallels the emergence of the mitochondrial transcription apparatus 7 .

We investigated the diversity of CRISPR spacers of Thermus communities from two locations in Italy, two in Chile and one location in Russia. Among the five sampling sites, a total of more than 7200 unique spacers belonging to different CRISPR-Cas systems types and subtypes were identified. Most of these spacers are not found in CRISPR arrays of sequenced Thermus strains. Comparison of spacer sets revealed that samples within the same area (separated by few to hundreds of metres) have similar spacer sets, which appear to be largely stable at least over the course of several years. While at further distances (hundreds of kilometres and more) the similarity of spacer sets is decreased, there are still multiple common spacers in Thermus communities from different continents. The common spacers can be reconstructed in identical or similar CRISPR arrays, excluding their independent appearance and suggesting an extensive migration of thermophilic bacteria over long distances. Several new Thermus phages were isolated in the sampling sites. Mapping of spacers to bacteriophage sequences revealed examples of local acquisition of spacers from some phages and distinct patterns of targeting of phage genomes by different CRISPR-Cas systems. This article is part of a discussion meeting issue ‘The ecology and evolution of prokaryotic CRISPR-Cas adaptive immune systems’.

25CrAss-like phages are a recently described family-level group of viruses that includes the most 26 abundant virus in the human gut 1,2 . Genomes of all crAss-like phages encode a large virion-27 packaged protein 2,3 that contains a DFDxD sequence motif, which forms the catalytic site in 28 cellular multisubunit RNA polymerases (RNAPs) 4 . Using Cellulophaga baltica crAss-like phage 29 phi14:2 as a model system, we show that this protein is a novel DNA-dependent RNAP that is 30 translocated into the host cell along with the phage DNA and transcribes early phage genes. We 31 determined the crystal structure of this 2,180-residue enzyme in a self-inhibited, likely pre-virion-32 packaged state. This conformation is attained with the help of a Cleft-blocking domain that 33 interacts with the active site motif and occupies the RNA-DNA hybrid binding grove. Structurally, 34 phi14:2 RNAP is most similar to eukaryotic RNAPs involved in RNA interference 5,6 , although most 35 of phi14:2 RNAP structure (nearly 1,600 residues) maps to a new region of protein folding space. 36Considering the structural similarity, we propose that eukaryal RNA interference polymerases 37 take their origin in a phage, which parallels the emergence of the mitochondrial transcription 38 apparatus 7 . 39 40 Transcription of bacterial, archaeal, and nuclear eukaryal genes is performed by multisubunit 41 DNA-dependent RNA polymerases (RNAPs), complex molecular machines that have a common 42 ancestor 4,8-10 . Their active site is located at the interface of two double-psi β-barrel (DPBB) 43 domains that belong to two different polypeptide chains. One of the DPBB domains carries the 44 universally conserved amino acid motif DFDGD, where the three aspartates coordinate Mg 2+ ions 45 required for catalysis 11,12 . Gene g066 of Cellulophaga baltica crAss-like phage phi14:2 encodes 46 a 2,180-residue protein that shows a limited sequence similarity to one of the two DPBB domains 47 of cellular RNAPs and contains a motif ( 1361 DFDID 1365 ) that is conserved in orthologs of this protein 48 across the crAss-like phage family 2 . Gp66 protein has been identified as a component of the 49 phage particle 3 . We hypothesized that gp66 is an evolutionarily divergent virion-packaged RNAP 50 of phi14:2 that is delivered into the host cell early in the infection process where it transcribes the 51 early phi14:2 genes. To test this hypothesis, we examined the in vitro and in vivo activity of gp66 52 and solved its crystal structure. 53 RNAP gp66 transcribes single-stranded and denatured double-stranded DNA in vitro 54We expressed recombinant gp66 in Escherichia coli, purified it (Extended data Fig. 1), and tested 55 its RNA synthesis activity in a diverse set of assays. 56First, we tested whether gp66 could extend the RNA primer of an 8-nucleotide long RNA-DNA 57 hybrid in the presence of ribonucleoside triphosphates (rNTPs). This hybrid molecule mimics the 58 nucleic acid structure in the transcription elongation complex 13 . Gp66 was inactive in this assay 59whe...

The emergence and persistence of selfish genetic elements is an intrinsic feature of all living systems. Cellular organisms have evolved a plethora of elaborate defense systems that limit the spread of such genetic parasites. CRISPR-Cas are RNA-guided defense systems used by prokaryotes to recognize and destroy foreign nucleic acids. These systems acquire and store fragments of foreign nucleic acids and utilize the stored sequences as guides to recognize and destroy genetic invaders. CRISPR-Cas systems have been extensively studied, as some of them are used in various genome editing technologies. Although Type III CRISPR-Cas systems are among the most common CRISPR-Cas systems, they are also some of the least investigated ones, mostly due to the complexity of their action compared to other CRISPR-Cas system types. Type III effector complexes specifically recognize and cleave RNA molecules. The recognition of the target RNA activates the effector large subunit – the so-called CRISPR polymerase – which cleaves DNA and produces small cyclic oligonucleotides that act as signaling molecules to activate auxiliary effectors, notably non-specific RNases. In this review, we provide a historical overview of the sometimes meandering pathway of the Type III CRISPR research. We also review the current data on the structures and activities of Type III CRISPR-Cas systems components, their biological roles, and evolutionary history. Finally, using structural modeling with AlphaFold2, we show that the archaeal HRAMP signature protein, which heretofore has had no assigned function, is a degenerate relative of Type III CRISPR-Cas signature protein Cas10, suggesting that HRAMP systems have descended from Type III CRISPR-Cas systems or their ancestors.