27 26 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 Macromolecular Crystallography Group Beamline I24 Viruses that defeatmicrobe immunity could help fight drug-resistant bacteria Related publication: Athukoralage J. S., McMahon S. A., Zhang C., Grüschow S., GrahamS., Krupovic M.,Whitaker R. J., GlosterT. M. &WhiteM. F. An anti-CRISPR viral ring nuclease subverts type III CRISPR immunity. Nature 577 , 572–575 (2020). DOI: 10.1038/s41586-019-1909-5 Publication keywords: CRISPR;Viral infection; Cyclic nucleotide; Nuclease V iruses can infect all cellular organisms, from bacteria to humans. Once they have tricked their way into the cell, they exploit the host’s cellular machinery to make more copies of themselves, which can be released and go on to infect other cells. Organisms have developed strategies to recognise and defend against any virus infection, called adaptive immunity. However, viruses have evolved a way to combat the host defence mechanism, meaning they ultimately win the infection battle. Microbes use an adaptive immunity mechanism called CRISPR against virus infection. An international team of researchers focussed on one type of CRISPR (type III) and investigated how viruses overcome themicrobial CRISPR defence system. Data they collected on the Microfocus Macromolecular Crystallography (MX) beamline I24 at Diamond Light Source provided vital insights into the mechanism used by viruses to overcome the microbial defence. Microbes with the type III CRISPR defence system produce a cyclic molecule in response to virus detection. It signals that there is an infection and kick-starts cellular processes to combat the attack. However, viruses have evolved an enzyme that binds to and destroys the cyclic molecule, neutralising the defence mechanism. This research used the structure of the viral enzyme, bound to the cyclic molecule, to reveal key details of how the viral enzyme recognises this molecule and how this translates into function. The results show the fundamental mechanism underlying how viruses out-manoeuvre microbe defences against infection. We may be able to use this understanding to engineer viruses to target drug-resistant bacteria that infect humans. The CRISPR system is a prokaryotic adaptive immune system, providing defence against viruses and other mobile genetic elements in bacteria and archaea 1 . CRISPR systems are mechanistically and structurally diverse and have been classified into six main types based on their signature proteins, with type II (Cas9) undoubtedly the best known due to its application in genome engineering. Type III CRISPR systems use a large multi-subunit ribonucleoprotein complex to detect RNA from invading genetic elements. Binding of this target RNA activates a specialised polymerase domain in the Cas10 subunit, which polymerises ATP into a set of cyclic oligoadenylate (cOA) signalling molecules 2 . These act as second messengers of infection, analogous to the role of cGAMP in the cGAS-STING innate immunity pathway of eukaryotic cells. cOA potentiates type III CRISPR defence by binding to a receptor CARF (CRISPR associated Rossman Fold) domain found in a wide range of ancillary effector proteins. This causes an allosteric activation of degradative enzymatic domains such as HEPN family ribonucleases, which are licensed to degrade RNA or DNA. These defence enzymes in turn degrade nucleic acids non-specifically, which can destroy the invading virus, but also lead to cell death if not regulated 3 .While altruistic suicide is often an appropriate response for virally-infected microbes, there are mechanisms to switch off the defence system by degrading the cOA second messenger using enzymes known as ring nucleases 4 . Viruses can combat CRISPR defence by expressing anti-CRISPR (Acr) proteins that inactivate the cellular effectors, often by binding directly to them. In this study, the hypothesis that viruses might use ring nucleases to neutralise type III CRISPR defence in the host was explored. A gene of unknown function present in many viruses was identified and experiments conducted to establish whether the gene expressed an Acr protein 5 . This was tested against the archaeon Sulfolobus islandicus , which was engineered to have only a type III CRISPR system. S. islandicus was grown on plates, andwhen challengedwith a lytic virus which lacked the Acr predicted to combat cellular defence there was no viral plaque formation. However, when the experiment was repeated in the presence of a plasmid expressing a candidate Acr, subsequently named AcrIII-1, plaque formation was evident (Fig. 1a). AcrIII-1 was recombinantly over-expressed and tested for activity on several (cOA) signalling molecules to establish if it was a ring nuclease. Experiments showed AcrIII-1 rapidly degraded cyclic tetra-adenylate (cA 4 ), at a rate 100-fold higher than the host cellular ring nuclease (Fig. 1b), demonstrating it is a highly specific and efficient enzyme, and thereby acts as an effective mechanism for blocking type III CRISPR defence mounted by the host. To understand the mechanism for AcrIII-1 activity, a mutant enzyme, lacking a highly conserved histidine residue (H47A), was crystallised in the presence of cA 4 and X-ray diffraction data to 1.55 Å resolution collected using beamline I24. The structure was solved using molecular replacement with the apo structure of the same enzyme which had been reported as part of a structural genomics program a decade earlier (PDB 2X4I). The structure shows AcrIII-1 is a dimer with cA 4 bound at the interface (Fig. 2a). Strikingly cA 4 is completely buried in the structure (Fig. 2b), suggesting movement of residues in AcrIII-1 following binding. While the apo structure (Fig. 2c,d) essentially shows the cA 4 binding site is pre-formed, a large loop encompassing residues 82-94 is flexible and was difficult to model. Comparison of the apo structure with the structure with cA 4 bound (Fig. 2e) shows there is a significant movement of this loop and subsequent a -helix following cA 4 binding, causing the cA 4 to become completely buried. This is no mean feat, given AcrIII-1 is just 16 kDa in size, and cA 4 is >1 kDa. The interactions made between AcrIII-1 and cA 4 are symmetrical, and hydrogen bonds are formed with ten residues in the active site (Fig. 3a). The residues that form interactions with cA 4 are fully or highly conserved in homologues, suggesting a consistent mode of binding. Arg85 is located on the large loop that folds over the cA 4 binding site, and interacts with the phosphate group on the opposite side of the cA 4 molecule (relative to the side of the cA 4 where the rest of the residues in the monomer interact) suggesting it acts as a ‘lock’ to keep the loop in position and binding site closed. Insights were also gleaned into the likely catalytic residues important for degradation of cA 4 . His47, which is fully conserved in homologues, and is likely to act as the general acid to protonate the leaving group, following in-line nucleophilic attack of the 2’ hydroxyl group in the ribose on the phosphodiester bond, on both sides of cA 4 (Fig. 3b). The H47A mutant showed a 2,500-fold decrease in activity, and enabled capture of the structure of AcrIII-1 in complex with cA 4 . Glu88 was also shown to be important for activity and may be involved in the modulation of the p K a of His47. The elucidation of the ring nuclease activity of AcrIII-1 demonstrates a powerful way in which viruses can neutralise the host defence system. Given cA 4 is a key signalling molecule in initiation of the host CRISPR type III defence system for a range of bacteria and archaea, the evolution of this ingenious systemby viruses means it is difficult for a host to respond, without a fundamental change to the signalling molecule required to activate their own defence upon infection. The study has provided highlights into the structure, function and mechanism of AcrIII-1. Such an understanding may provide opportunities to engineer viruses as a way to kill drug resistant bacteria that infect humans in the future. References: 1. Makarova K. S. et al. Evolutionary classification of CRISPR–Cas systems: a burst of class 2 and derived variants. Nat. Rev. Microbiol. 18 , 67–83 (2020). DOI: 10.1038/s41579-019-0299-x 2. Kazlauskiene M. et al. A cyclic oligonucleotide signaling pathway in type III CRISPR-Cas systems. Science. 357 , 605–609 (2017). DOI: 10.1126/science.aao0100 3. Rostøl J. T. et al. Non-specific degradation of transcripts promotes plasmid clearance during type III-A CRISPR–Cas immunity. Nat. Microbiol. 4 , 656–662 (2019). DOI: 10.1038/s41564-018-0353-x 4. Athukoralage J. S. et al. Ring nucleases deactivate type III CRISPR ribonucleases by degrading cyclic oligoadenylate. Nature 562 , 277–280 (2018). DOI: 10.1038/s41586-018-0557-5 5. Athukoralage J. S. et al. An anti-CRISPR viral ring nuclease subverts type III CRISPR immunity. Nature 577 , 572–575 (2020). DOI: 10.1038/ s41586-019-1909-5 Funding acknowledgement: This work was supported by grants from the Biotechnology and Biological Sciences Research Council (BB/S000313/1 to M.F.W. and BB/R008035/1 to T.M.G.). Corresponding authors: Dr Tracey Gloster, University of St Andrews,
[email protected]; Prof. MalcolmWhite, University of St Andrews,
[email protected] Figure 1: AcrIII-1 acts as an effective Acr, by utilising highly efficient ring nuclease activity with specificity for cA 4 . (a) Viral challenge assay on S. islandicus in the presence (left) and absence (right) of a plasmid expressing AcrIII-1. Viral plaques were observed in the presence of AcrIII-1, but not in the control, demonstrating the enzyme can neutralise the host CRISPR defence system; (b) Comparison of in vitro activity for cA 4 degradation by two different viral AcrIII-1 enzymes (green, purple) and a host ring nuclease (Crn1; blue). There is ~100-fold higher activity for the viral enzymes compared to the host, meaning the cellular cA 4 levels are rapidly reduced. Figure 2: Structure of AcrIII-1 in complex with cA 4 . (a) AcrIII-1 (dimer formed frommonomers shown in blue and yellow) in complex with cA 4 (magenta); (b) AcrIII-1 in complex with cA 4 , with surface representation; (c) Apo AcrIII-1 (dimer formed frommonomers in orange and cyan); (d) Apo AcrIII-1, with surface representation; (e) Divergent stereo figure of superimposition of apo AcrIII-1 (cyan) and AcrIII-1 (blue) in complex with cA 4 (magenta); the red squares indicate the a -helices where most movement is observed. Apo AcrIII-1, PDB code 2X4I; AcrIII-1 + cA 4 , PDB code 6SCF. Figure 3: Binding site of AcrIII-1 with cA 4 . (a) Residues from the AcrIII-1 dimer (different monomers shown in blue and yellow) that interact with cA 4 (magenta); (b) His47 of AcrIII-1 (position taken from the apo structure; different monomers shown in orange and cyan) is predicted to act as the general acid in catalysis for degradation of cA 4 (magenta). Apo AcrIII-1, PDB code 2X4I; AcrIII-1 + cA 4 , PDB code 6SCF.