Regulation of the RNA and DNA nuclease activities required for Pyrococcus furiosus Type III-B CRISPR–Cas immunity

Regulation of the RNA and DNA nuclease activities required for Pyrococcus furiosus Type III-B... 4418–4434 Nucleic Acids Research, 2020, Vol. 48, No. 8 Published online 21 March 2020 doi: 10.1093/nar/gkaa176 Regulation of the RNA and DNA nuclease activities required for Pyrococcus furiosus Type III-B CRISPR–Cas immunity 1 2 3 2 Kawanda Foster , Sabine Grusc ¨ how , Scott Bailey , Malcolm F. White and Michael 1,4,5,* P. Terns 1 2 Department of Microbiology, University of Georgia, Athens, GA 30602, USA, Biomedical Sciences Research Complex, School of Biology, University of St Andrews, St Andrews KY16 9ST, UK, Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205, USA, Department of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, USA and Department of Genetics, University of Georgia, Athens, GA 30602, USA Received December 27, 2019; Revised February 28, 2020; Editorial Decision March 05, 2020; Accepted March 19, 2020 ABSTRACT INTRODUCTION Prokaryotes often harbor powerful CRISPR–Cas Type III CRISPR–Cas prokaryotic immune systems (clustered regularly interspaced short palindromic repeat- provide anti-viral and anti-plasmid immunity via a CRISPR associated) adaptive immune systems to protect dual mechanism of RNA and DNA destruction. Upon against infections from invading viruses and plasmids (1,2). target RNA interaction, Type III crRNP effector com- CRISPR genomic arrays are composed of short DNA plexes become activated to cleave both target RNA sequences of foreign origin (called spacers), separated by (via Cas7) and target DNA (via Cas10). Moreover, host repeat sequences. CRISPR arrays become transcribed trans-acting endoribonucleases, Csx1 or Csm6, can and the long, primary transcripts are processed into short, promote the Type III immune response by destroying mature crRNAs that assemble with Cas proteins to form both invader and host RNAs. Here, we characterize crRNP (CRISPR RNA-containing ribonucleoprotein) how the RNase and DNase activities associated with effector complexes. These effector complexes detect and Type III-B immunity in Pyrococcus furiosus (Pfu)are destroy invading nucleic acids that are complementary to their crRNAs. CRISPR–Cas systems are quite diverse regulated by target RNA features and second mes- and fall into six distinct types (Types I-VI) and over 30 senger signaling events. In vivo mutational analy- subtypes (3,4). Types I, II and V (and possibly IV) target ses reveal that either the DNase activity of Cas10 the destruction of DNA (5–7), while Type VI destroys or the RNase activity of Csx1 can effectively direct RNA (8). Type III systems are particularly noteworthy in successful anti-plasmid immunity. Biochemical anal- that they uniquely degrade both RNA and DNA of the yses confirmed that the Cas10 Palm domains con- invaders (9–19). Type III systems are further categorized vert ATP into cyclic oligoadenylate (cOA) compounds into six subtypes (III-A through III-F) with the majority that activate the ribonuclease activity of Pfu Csx1. belonging to either the Type III-A (Csm) or Type III-B Furthermore, we show that the HEPN domain of the (Cmr) systems (3). adenosine-specific endoribonuclease, Pfu Csx1, de- Types III-A (Csm) and III-B crRNP (Cmr) effector com- grades cOA signaling molecules to provide an auto- plexes exhibit an overall similar subunit organization and architecture (see Figure 1A for an example of the Cmr ef- inhibitory off-switch of Csx1 activation. Activation of fector complex). Each complex is composed of a single cr- both the DNase and cOA generation activities require RNA and vfi e (Csm 1–5 for III-A) or six (Cmr 1–6 for III-B) target RNA binding and recognition of distinct target Cas proteins (15,20–25). The mature crRNAs within these RNA 3 protospacer flanking sequences. Our results complexes contain eight nucleotides of repeat sequence at highlight the complex regulatory mechanisms con- the 5 end called the 5 tag, followed by a ∼30–40 nucleotide trolling Type III CRISPR immunity. guide sequence that base-pairs with the target RNA pro- tospacer (26–28). Multiple catalytically active Cas7 super- family proteins (Csm3 or Cmr4) that act as target RNA en- To whom correspondence should be addressed. Tel: +1 706 542 1896; Fax: +1 706 542 1752; Email: mterns@uga.edu C The Author(s) 2020. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Nucleic Acids Research, 2020, Vol. 48, No. 8 4419 Cmr5 Cmr6 Cmr5 Cmr5 5’ CARF ATP Cmr4 Cmr4 Cmr4 Cmr1 Cmr4 Cmr3 crRNA 3’ Csx1 Palm Cmr2 HEPN HD RNase (D26) cOA DNase (HD) RNase (HEPN) Plate on No Selective Media Target Plasmid No CRISPR Defense PyrF (Colony Formation) Pfu Strain Plate on Selective Media or Target Plasmid CRISPR Defense No CRISPR Defense (No/Low Colony Formation) (Colony Formation) PyrF Null Cmr Cmr2-HD Cmr2-Palm Target Plasmid Csx1 No Target Plasmid Cmr2-HD /Palm m m Cmr2-HD /Csx1 Cmr2-Palm /Csx1 /Csx1-HEPN Cmr2-HD /Csx1 Cmr2-HD m C Cmr2 /Csx1 Colonies per μg of DNA Figure 1. Csx1 is required for anti-plasmid immunity in the P. furiosus (Pfu) III-B CRISPR–Cas system. (A) Components of the Pfu Cmr defense response. The Cmr effector complex is composed of Cmr1–6 and a mature crRNA (black & orange) containing a 5 tag (black) eight nucleotides in length. Cyclic oligoadenylate compounds (orange) produced by the Palm domain of Cmr2 bind to the CARF domain of Csx1. See Supplemental Figure S1 for a list of the Cas gene family names that correspond to each of the six Cmr subunits. (B) Plasmid interference assay. Pfu strains were transformed with plasmids that contain (target) or lack (no target) a transcribed target region complementary to the 7.01 crRNA (first spacer of the Pfu CRISPR 7 array). Both plasmids contain the pyrF gene to facilitate growth in the absence of uracil. (C) Colonies produced by transforming 11 different Pfu strains. Strains were transformed with the target plasmid (blue) or no target plasmid (gray). Mean colonies obtained per #2;g of plasmid DNA are plotted for two replicates. Error bars indicate standard deviation of replicates. Pfu Strain 4420 Nucleic Acids Research, 2020, Vol. 48, No. 8 doribonucleases (16,29,30), interact along the length of the vealed a key role for the short sequence that flanks the 3 crRNA guide region and these proteins also tightly asso- end of the RNA protospacer, termed the protospacer flank- ciate with Cas11 superfamily proteins (Csm2 or Cmr5). Ad- ing sequences (PFS) in controlling Type III activities. When ditional Cas7 superfamily Cas proteins directly contact the PFS sequences are complementary to the 5 crRNA tag (as 5 crRNA tag (Csm4 or Cmr3) or 3 terminus of the guide would be the case if the cell produced antisense crRNAs), RNA segment (Csm5 or Cmr1 and Cmr6). Cas10 (Csm1 or the DNase and cOA production activities fail to become ac- Cmr2) is the signature protein of Type III complexes (4). tivated (9,13,14,53,54) while the transcripts themselves are This large, multiple domain-containing protein is situated cleaved by the Csm3/Cmr4 integral RNases (16,27). Thus, near the 5 end of the crRNA and typically contains two 5 tag/PFS pairing can negatively regulate some but not all highly conserved motifs: the HD motif capable of destroy- Type III activities. In some characterized systems, the par- ing single-stranded DNA (9,10,13) and the GGDD motif of ticular identity of nucleotides within the PFS will dictate one of two Palm domains that can convert ATP into cyclic if the effector complex is active or inactive independent of oligoadenylate (cOA) second messenger molecules (31–33) their capacity to base-pair with the 5 crRNA tag. In par- (Figure 1A). ticular, three nucleotides immediately 3 of the target RNA Interestingly, Type III systems also include trans-acting protospacer (i.e. in positions +1, +2, +3 in the target RNA ribonucleases, Csm6 (III-A) or Csx1 (III-B), that are not 5 -3 direction) have been found to be important for acti- stably associated with the effector crRNP complexes (34– vating the DNase and cOA production presumably due to 37) and appear to be capable of degrading both invading interactions of these PFS elements and subunits of the Type RNA (leading to immunity) as well as host RNAs (leading III effector complex (likely Cas10 and/or Csm4/Cmr3) to cell dormancy or cell death) when activated by cOA bind- (9,25,48,54). The DNase and cOA generation activities are ing (32,35,38). Csm6 and Csx1 proteins share two highly also switched off as a result of degradation of the trigger tar- conserved domains: the HEPN (Higher Eukaryotes and get RNA that occurs when Csm3/Cmr4 RNases of the com- Prokaryotes Nucleotide binding) domain and the CARF plex cleave the RNA protospacer (33,42,48). Finally, mech- (CRISPR Associated Rossman Fold) domain (39,40)(Fig- anisms have also recently been discovered that can reverse ure 1A). The ribonuclease activity of both Csx1 and Csm6 the effects of cOA signaling on Csm6/Csx1 HEPN ribonu- is facilitated by the HEPN motif (R-X -H) within the clease activity. Specifically, dedicated ‘ring nucleases’ have 4–6 HEPN domain (34–37). Binding of cognate cOA to the been identified that bind, cleave, and inactivate cOAs by CARF domain allosterically stimulates the single-stranded converting the compounds into short, bi-adenylate degra- RNase activity of the HEPN domain of the Csm6/Csx1 dation products with 2 -3 cyclic phosphate termini (A > p) proteins (31–33,41,42). CARF domains, either within Csm6 in certain organisms (45,46). However, other organisms ap- (43,44) or as part of unrelated proteins called ring nucleases pear not to possess distinct ring nucleases and instead have (45,46), have also recently been found to cleave and inacti- been found to rely on an intrinsic ability of the CARF do- vate cOA singling molecules to switch off the activity of the mains of Csm6/Csx1 to destroy the cOA molecules (43,44). Csm6 or Csx1 HEPN RNases (43–46). Additional layers and molecular mechanisms for control- Anti-virus or anti-plasmid immunity afforded by Type III ling the activity of each of the key nucleases of Type III ef- crRNPs, is transcription-dependent as these systems specif- fector crRNPs as well as the cOA signaling pathway likely ically recognize an RNA target having a crRNA interaction await discovery. region (i.e. the RNA protospacer) (9,10,14,19,47). Once Pyrococcus furiosus (Pfu) is a hyperthermophilic ar- bound to the target RNA protospacer, the crRNP effec- chaeon that contains three distinct functional CRISPR– tor complex changes its conformation (24,25,48–50) and the Cas effector crRNPs: two DNA-targeting systems (Types I- Type III defense response becomes activated to: (i) specif- A (Csa) and I-B (formerly known as I-G) (Cst) (55–59)) and ically cleave the target RNA at regular six-nucleotide in- a Type III-B (Cmr) system (Supplemental Figure S1) shown tervals within the protospacer region using multiple copies to destroy both DNA and RNA targets in a transcription- of the Csm3/Cmr4 integral ribonuclease via an active site coupled manner (9,11,17,27). Previous investigations on the that relies on a key aspartate residue (16,29,30), (ii) non- mechanism of action of Pfu Cmr crRNPs also revealed that specifically degrade nearby invading single-stranded DNA crRNA-mediated target RNA interaction was required for via the HD motif of Cas10 (Csm1/Cmr2) (9,10,13,14), (iii) both target RNA cleavage (by Cmr4) (11,17,29) and non- generate cyclic oligoadenylate (cOA) from ATP using the specific DNase activity (by HD domain of Cmr2) which was GGDD motif of the conserved Palm domain of Cas10 shown to cleave single-stranded DNA substrates or both (Csm1/Cmr2) and (iv) non-specifically degrade single- strands of short double-stranded DNA substrates (9). The stranded RNA via the cOA-activated and trans-acting target RNA PFS requirements for controlling immunity by Csm6/Csx1 HEPN ribonuclease (31–33,41,42,51,52). Col- Pfu Cmr crRNPs have been thoroughly investigated. Simi- lectively, these target-RNA and transcription-coupled reac- lar to other studied Type III systems, PFS sequences capa- tions provide robust Type III-mediated immunity. ble of base-pairing with the 5 tag sequence of the crRNA Type III RNase and DNase activities must be tightly prevented DNase activity but not target RNA destruction regulated so that undesirable cellular toxicity or host cell (9,11). In contrast, the identity of three nucleotides within death can be prevented before or during an immune re- the PFS immediately 3 of the target RNA protospacer sponse. The molecular details for how each of the RNases, (previously referred to as the rPAM [RNA protospacer- DNase and cOA signal generation activities of Type III sys- adjacent motif] (9)), was critical for DNase activity in vitro tems become specifically activated by target RNA binding and anti-plasmid immunity in vivo, but not target RNA is currently not fully understood. Several studies have re- cleavage (9). Whether or not Pfu crRNPs are capable of Nucleic Acids Research, 2020, Vol. 48, No. 8 4421 producing cOA second messenger compounds had not yet with 0.5 mM iso-propyl-β-D-thiogalactopyranoside (IPTG) been tested. Furthermore, deletion of the csx1 gene in Pfu at 24 C overnight. Cells were pelleted then resuspended in did not disrupt anti-plasmid immunity (9). These findings lysis buffer (40 mM Tris–HCl (pH 7.5), 500 mM NaCl, raised the key question as to whether cOA-mediated acti- 10 mM Imidazole) containing one protease inhibitor tablet vation of Csx1 RNase activity is important for conferring (Roche) and lysed via sonication. Cell lysates underwent a immunity by the Pfu Cmr crRNP effector complex. In this thermal precipitation by incubating in a 75 Cbeadbath study, we expand our understanding of the molecular mech- for 20 minutes. Insoluble material was removed by cen- anisms of action of the well-characterized Type III-B (Cmr) trifugation at 14,000 rpm for 20 minutes at 4 C and fil- system of Pyrococcus furiosus (Pfu) and further define how tered through a 0.8 #2;M syringe filter (Corning Incorpo- the Csx1 RNase and Cmr2 DNase activities are regulated rated). Proteins were purified using gravity affinity chro- by target RNA elements and second messenger signaling matography and either Ni-NTA resin (Cmr2, Cmr3, Cmr4, events. Cmr5, Cmr6, Csx1; Thermo Scientific) or Talon Cobalt resin (Cmr1–1; Clontech). Cell lysates were rotated with pre-rinsed and equilibrated resin for 1 h at 4 C. The pro- MATERIALS AND METHODS teins were then washed with the lysis buffer and wash buffer P. furiosus strains and growth conditions (40 mM Tris–HCl (pH 7.5), 500 mM NaCl, 20 mM Im- idazole). The proteins were then eluted using four differ- All P. furiosus strains utilized in this study are listed in Sup- ent elution buffers containing increasing amounts of imida- plemental Table S1. Strains were grown at 90 C under strict zole (50, 100, 200, 500 mM). Wildtype and HEPN mutant anaerobic conditions using defined medium as previously Csx1 proteins were dialyzed and underwent a second round detailed (60). Cultures were grown in 5 or 20 mL volumes of gravity affinity chromatography. Buffer exchange was and inoculated with either 1% inoculum or a single isolated performed using Slide-A-Lyzer Dialysis Cassettes (Thermo colony. Cultures were incubated for 16–24 h and plates were Scientific) in elution buffer lacking Imidazole. Protein con- incubated for 64 h. Uracil (20 #2;M) and/or 5-fluoroorotic centrations were assessed using Qubit protein concentration acid (5-FOA, 2.75 mM) were supplemented in the media assays (Invitrogen) and purity was assessed via SDS-PAGE for selection or counterselection for the pyrF marker gene. and Coomassie blue staining analysis. P. furiosus strains were produced using homologous re- combination of transformed SOE-PCR (splicing by over- lap extension-polymerase chain reaction) constructs as pre- Csx1 mutagenesis viously reported (9). Complementation strains were gen- The wildtype gene encoding Csx1 was subcloned into a erated by adding a modified wildtype cmr2 or csx1 gene modified pET24D vector. The Csx1 HEPN mutant (Csx1- back onto the P. furiosus genome that contained restriction HEPN ) contains a H436A mutation and was created as sites that enabled detection of the introduced genes from previously described (37). CARF domain mutants were cre- wildtype genes and did not affect the protein coding po- ated using the wildtype vector and either inverse PCR or tential. A BclI-HF (5 -TGATAA-3 → 5 -TGATCA-3 )re- Quikchange site-directed mutagenesis (Stratagene). Muta- striction site was introduced into cmr2 and a NruI-HF (5 - genesis primers are provided in Supplemental Table S3. TCGGGA-3 → 5 -TCGCGA-3 ) restriction site was intro- Amino acid residues 121–127 were deleted from Csx1 us- duced into csx1. Primers used to make the complementa- ing inverse PCR to create a mutant form of the protein pre- tion strains can be found in Supplemental Table S3. All dicted to not be able to organize a functional CARF motif strains underwent at least three rounds of strain purification (Csx1-CARF ). Quikchange mutagenesis was used to cre- using minus uracil selective media. Successful strain gener- ate site-specific mutations within the predicted cOA binding ation was confirmed via PCR amplification of the Pfu gene pocket of the CARF domain of Pfu Csx1. Two cOA bind- of interest and DNA sequencing (Eurofins Genomics). ing mutants were created: Csx1-INAA (I169A, N170A) and Csx1-INQQ (I169Q, N170Q). Successful mutagenesis for Recombinant protein expression and purification all mutant plasmids was confirmed via DNA sequencing The genes encoding P. furiosus Cmr1–6 and Csx1 pro- (Eurofins Genomics). teins were amplified via PCR and cloned into modified ver- Purification of Csx1-HEPN was performed as described sions of pET24-D (Cmr4, Cmr5, Csx1), pET101-D (Cmr1– above. Purification of wildtype, Csx1-CARF , Csx1-INAA 1) and pET200-D (Cmr2, Cmr3, Cmr6) as previously de- and Csx1-INQQ proteins was performed using a batch tailed (17,37). All constructs contain a 6x-histidine tag method. The batch method involved affinity purifying the on either the N-terminus (Cmr2–6, Csx1) or C-terminus proteins by incubating the soluble lysate with Ni-NTA resin (Cmr1–1) of the corresponding protein. Recombinant pro- for 1 h at 4 C then performing subsequent washes and elu- tein expressions were performed in E. coli BL21-RIPL cells tions in a 15 ml centrifuge tube and spinning at 4000 rpm for (DE3, Novagen). Expression cultures for wildtype and mu- 2 mins in between each step. Protein concentrations were as- tant proteins were grown in 1 l (Cmr1–1, Cmr4, Cmr5, sessed using Qubit assays and purity was assessed via SDS- Csx1), 2 l (Cmr2, Cmr3) or 4 l (Cmr6) cultures at 37 C. PAGE and Coomassie blue staining analysis. Luria broth (Cmr1–1, Cmr2–Cmr5, Csx1; Research Prod- ucts International (RPI)) or Terrific broth (Cmr6; RPI) Preparation of RNA and DNA substrates medium was used for cultures and supplemented with ei- ther 50 #2;g/ml kanamycin sulfate (Cmr2-Cmr6, Csx1) or Synthetic RNAs (7.01 crRNA and 7.01 Target RNA) were 100 #2;g/ml ampicillin (Cmr1–1) for plasmid selection. Cul- purchased from Integrated DNA Technologies and DNAs tures were grown to an OD of 0.7 at 37 C then induced from Eurofins Genomics. The RNA and DNA sequences 600 4422 Nucleic Acids Research, 2020, Vol. 48, No. 8 can be found in Supplemental Table S3. Synthetic target added to each reaction and incubated for 30 min at 37 C RNA was 5 -end labeled using #3;- P-ATP, gel purified, prior to gel electrophoresis. Cyclic oligoadenylate produc- eluted, extracted, and precipitated as previously described tion assays were performed by assembling the Cmr crRNPs (37). 7.01 crRNA was gel purified, eluted, extracted, and as described above using RNA assay buffer and 50 nM of precipitated prior to using in assays. 5 -end labeled, double- 7.01 crRNA. After assembly, 0.5 mM of ATP (NEB), 5 nM stranded DNA substrate was prepared by annealing com- of #4;- P-ATP (3000 Ci/mmol; Perkin Elmer), and 100 nM plementary DNA oligonucleotides as previously described of target RNA was added to the reaction and incubated (37) and gel purification. for 1 h at 70 C. Unless otherwise indicated, cOA produc- Target RNAs with the indicated PFS sequences were cre- tion assays were completed with target RNA containing a ated by in vitro transcription using T7 RNA polymerase 5 -GGG-3 PFS sequence. All Reactions were stopped by and the MEGAshortscript T7 kit (Invitrogen) as described adding Gel Loading Buffer II (Life Technologies) and vi- (9). DNA templates with a T7 phage promoter sequence sualized by using 7M urea denaturing 15% polyacrylamide were generated by amplifying a target plasmid listed in Sup- gels followed by autoradiography. Cyclic oligoadenylate re- plemental Table S2 with IVT primers listed in Supplemen- actions were also ran on 8M urea denaturing 20% poly- tal Table S3 and gel purified using the Zymoclean Gel Re- acrylamide sequencing gels. Decade Markers (Life Tech- covery Kit (Zymo Research). Following synthesis, target nologies) and partial alkaline hydrolysis ladders (Ambion) RNAs were subsequently gel purified from denaturing gels, of poly A RNA were generated as previously described eluted, extracted, and precipitated prior to adding to the as- (34,37). says. Target RNA concentrations were determined using the Qubit RNA BR Assay Kit (Invitrogen) and quality was as- Csx1 In vitro activation assays sessed using 7M urea denaturing 15% polyacrylamide gels and ethidium bromide staining (Supplemental Figure S2). Activation with the native Cmr produced cOA. Csx1 RNase activity assays were performed as previously reported (37). Plasmid interference assay Ribonuclease activity of Csx1 was assessed by incubating 500 nM of Csx1 with 0.5–1.5 nM of radiolabeled target Plasmid transformation interference assays were performed RNA in assay buffer (20 mM Tris–HCl (pH 7.5), 200 mM as previously described (9). Liquid cultures of P. furio- NaCl) for 1 h at 70 C. In order to assess activation of Csx1, sus strains were allowed to reach mid-to-late log phase 10 or 20 nM of Csx1 was incubated with radiolabeled target of growth. 100 #2;l of liquid culture was transformed with RNA and assay buffer in the presence of unlabeled cOA for 1 ng of either Target plasmid (pJE65; containing a tran- 1h70 C. Unlabeled native cOA was generated as described scribed protospacer matching the 7.01 crRNA) or No Tar- above by omitting #4;- P-ATP. The unlabeled cOA was then get (pJE47) control plasmid. The transformations were in- extracted similarly to published methods (61). Five reac- cubated for 15–45 min at room temperature prior to plat- tion volumes of phenol/chloroform/isoamyl alcohol (PCI, ing. Each transformation mixture was split between two 125:24:1 at pH 4.5; Ambion) was added to the reaction and plates and spread onto solid defined media lacking uracil. vortexed for 30 seconds. The mixture was then centrifuged The plates were incubated at 90 C in an anaerobic chamber. at 20 000 rpm at 4 C then the aqueous layer incubated with Plates were observed for colony growth and counted after vfi e reaction volumes of chloroform (Fisher Scientific) vor- 64 h of incubation. Results shown represent two replicates. texed and centrifuged. The aqueous layer was extracted, aliquoted, and stored at −80 C in single use aliquots. All Cmr crRNP in vitro activity assays reactions were stopped by adding Gel Loading Buffer II and visualized by using 7M urea denaturing 15% polyacry- RNA and DNA nuclease activity assays were performed similarly to methods previously described (9,17). Purified lamide gels followed by autoradiography. recombinant Cmr proteins were first incubated with 7.01 crRNA to form crRNPs. For RNase assays, Cmr crRNPs Activation with synthetic cA and cA . Csx1 ribonuclease 4 6 were assembled by preincubating 500 nM of each Cmr pro- activation assays with cA and cA species was performed 4 6 tein (50 nM of Cmr2) with RNA assay buffer (20 mM Tris– as described above except synthetic cOAs from BIOLOG HCl (pH 7.5), 250 mM NaCl, 1.5 mM MgCl ), and 12.5 Life Science Institute were used. Several concentrations of nM of 7.01 crRNA for 25 min at 70 C. After the prein- cA and cA were tested as indicated in the figure legend. 4 6 cubation, 0.5–1.5 nM of radiolabeled synthetic 7.01 target RNA was added to the reaction and incubated for one hour Activation with Csx1-treated cOA. Cmr crRNP-generated at 70 C. One unit of Proteinase K (NEB) was then added cOA was incubated with or without 600 nM of wildtype ◦ ◦ to each reaction and incubated for 30 min at 37 C prior Csx1 for 30 min at 70 C. Reaction products were then ex- to gel electrophoresis. For DNase assays, the Cmr crRNP tracted to deproteinize the samples (as described above) and was assembled in DNA assay buffer (20 mM Tris–HCl (pH incubated with 20 nM of Csx1, 5 -end labeled target RNA, 7.5), 250 mM NaCl, 1.5 mM MgCl , 200 #2;M NiCl )and and assay buffer for 1 h at 70 C. 2 2 50 nM of 7.01 crRNA for 25 min at 70 C. After the prein- cubation, 100 nM of 7.01 target RNA and 1 nM of radi- Mass spectrometry olabeled dsDNA was added to the reaction and incubated for one hour at 70 C. Unless otherwise indicated, DNase Unlabeled cOA production assays were performed as de- assays were completed with target RNA containing a 5 - scribed above except the reactions were incubated for 2 h. GGG-3 PFS sequence. One unit of Proteinase K was then Liquid chromatography high resolution mass spectrometry Nucleic Acids Research, 2020, Vol. 48, No. 8 4423 (LC-HRMS) analysis was performed on a Thermo Scien- ity of Csx1 (via HEPN motif and activated by Cmr2-Palm tific Velos Pro instrument equipped with HESI source and GGDD motif) are each sufficient for conferring highly ef- Dionex UltiMate 3000 chromatography system as previ- fective anti-plasmid immunity in Pfu. ously described (41). RESULTS P. furiosus Cmr crRNPs produce cyclic oligoadenylate second Effective anti-plasmid immunity is achieved by either Csx1 messengers RNase or Cas10 DNase activity Next, we addressed whether the Pfu Cmr system functioned through generating cOA signaling molecules as has been Distinct Type III-A or III-B systems have shown consider- observed for other bacterial and archaeal Type III systems able variability in the need for Cas10 (Csm1/Cmr2) DNase (31–33). In vitro reconstituted Cmr crRNPs were assayed and/or Csm6/Csx1 RNase activities for anti-plasmid and for their ability to generate cOA compounds as well as to anti-viral immunity (9,32,34,35,38,47,62). In our earlier in support previously observed RNase and DNase activities vivo work with the Pfu Cmr (III-B) system, we found that (Figure 2)(9,17). Wildtype as well as four different func- individual mutations of either the HD (H13A/D14A) or tional mutants of the Cmr crRNP complex (Cmr2-HD , Palm domain GGDD (D673A/D674A) motifs of Cmr2 Cmr2-Palm , Cmr4-D26N and Cmr2-HD /Palm ), were (Cmr2-HD or Cmr2-Palm mutants) or a single dele- m m m m m tion of the csx1 gene (Csx1 ), did not interfere with anti- assembled in vitro (Figure 2A) and tested. As expected, tar- plasmid immunity but double mutations in the Cmr2 HD get RNAse activity was observed for all Cmr crRNP com- and GGDD Palm motifs (Cmr2-HD /Palm )prevented plexes except those harboring a mutation in the Cmr4 sub- m m immunity (9) (see Figure 1A for overview of the crRNA unit (Cmr4-D26N) (Figure 2B). Moreover, DNase activity and Cas protein components). This early work was per- was only observed for wildtype, Cmr2-Palm , and Cmr4- formed prior to knowledge that the Palm domain of some D26N complexes but not Cmr2 mutants in which the HD Cas10 superfamily proteins can catalyze conversion of ATP motif was mutated (Figure 2C). To test for cOA produc- to cOA signaling molecules that activate the ribonuclease tion, the same vfi e complexes were incubated with #4;- P- activity (HEPN domain) of Csm6/Csx1 HEPN ribonucle- ATP and the products of the reactions were separated by ases (31–33,41,42,51,52). This new information motivated denaturing polyacrylamide gel electrophoresis. Conversion us to perform a more systematic in vivo mutational analy- of the #4;- P-ATP to slower migrating products indicative of cOA compounds was only observed for complexes with an ses in which specific combinations of double mutants were intact Cmr2 Palm GGDD motif. Additionally, the presence tested to more fully address if the Cmr2 DNase and Csx1 of the target RNA was required for cOA production (Fig- RNase activities were important for anti-plasmid immunity ure 2D). These findings reveal that the Pfu Type III-B Cmr (Figure 1). system possesses the highly conserved activity of cOA gen- Anti-plasmid immunity was assayed in vivo by transform- eration which is catalyzed by the Cmr2 Palm GGDD motif ing a Pfu strain containing wildtype or mutant versions of and is dependent upon interactions between crRNPs and the Cmr system with a target plasmid that harbors a tran- complementary target RNA. scribed protospacer matching an endogenous crRNA (7.01; the first crRNA from CRISPR locus 7) or empty plasmid control (Figure 1B). As we observed previously, the anti- plasmid immunity observed with wildtype Cmr was unaf- Pfu Csx1 ribonuclease activity is activated by cyclic oligoad- fected when Cmr effector complexes contained mutations enylate produced by the Cmr complex in either the Cmr2 DNase HD active site (Cmr2-HD ) or Palm GGDD motif (Cmr2-Palm ), or if the csx1 gene Next we sought to determine if the cyclic oligoadenylate was deleted from the genome (Csx1 )(Figure 1C and (9)). compounds produced by Pfu Cmr crRNPs, are capable of As expected, immunity was absent for a strain that lacked stimulating the ribonuclease activity of Pfu Csx1 in vitro the entire Cmr complex and Csx1 (null strain) or when (Figure 2E). Previously, we found that recombinant Pfu the Cmr2-HD /Palm double mutant was re-tested. Im- Csx1 is capable of cleaving RNA substrates via the HEPN m m munity was disrupted when the Pfu csx1 gene was deleted motif but only when high concentrations (e.g. 500 nM) of or contained a mutation within the RNase catalytic mo- the Csx1 enzyme is used (37). Therefore, we used levels of tif of csx1 (Csx1-HEPN ; H436A) in conjunction with a the Csx1 protein (20 nM) that showed no or low RNase ca- second mutation within the DNase catalytic site of Cmr2 pacity to test if cOA produced by Cmr crRNPs could stimu- (Cmr2-HD mutation) (Figure 1C and (9)). These same late the RNase activity of Csx1. ATP (unlabeled) was incu- csx1 mutations did not prevent immunity when combined bated with wildtype and mutant Cmr crRNPs and the prod- with a Cmr2-Palm mutation (GGDD; predicted to block ucts of the reactions were added to reactions containing cOA generation). Cmr-mediated immunity was rescued by Csx1 and radiolabeled substrate RNA (Figure 2E). Cleav- restoring either wildtype cmr2 or csx1 genes in the Cmr2- age of the radiolabeled substrate RNA by Csx1 depended HD /Csx1 double mutant strains (Cmr2c and Csx1c are upon Cmr crRNP complexes having an intact Cmr2-Palm Cmr2-HD /Csx1 strains complemented with wt cmr2 or GGDD motif (Figure 2D and E). The results indicate that wt csx1, respectively. Figure 1C). Taken together, the results Cmr2 Palm domain-mediated cOA production by the Cmr show that both the DNase activity of the Cmr effector cr- crRNP complex is a potent activator of the RNA degrada- RNPs (via HD domain of Cmr2) as well as the RNase activ- tion capacity of Pfu Csx1. 4424 Nucleic Acids Research, 2020, Vol. 48, No. 8 Cmr crRNP Cmr2 Cmr6 Cmr1-1 Cmr3 Cmr4 Cmr5 ssRNA dsDNA BC : : M Cmr crRNP M Cmr crRNP * 20 cOA Production D ssRNA : Rxn Product Cmr crRNP : Csx1 M ++ + + + + ++ + + : MM2 + Target RNA 20 20 - cOA ATP Figure 2. The ribonuclease activity of Pfu Csx1 is activated by cOA species produced by the Pfu effector complex in a Cmr2 Palm domain-dependent manner. (A) Purification of Pfu complex proteins. HD and Palm mutations are located within Cmr2. The D26N mutation is located within Cmr4. (B) RNase activity of each Pfu Cmr complex. Radiolabeled 7.01 Target RNA was incubated with Cmr complexes and reaction products were visualized by urea-PAGE. Radiolabeled RNA size standards (M) in nucleotides were used. The black arrow indicates the full-length substrate and the black asterisks indicate cleavage products. (C) DNase activity of each Pfu Cmr complex. Each complex was incubated with dsDNA (label located on DNA Target 2) and reactions were visualized as described in part B. (D) cOA production activity of the Pfu Cmr crRNP. The vfi e Cmr complexes were tested for cOA production in presence (+) and absence (–) of 7.01 Target RNA. Reaction products were run on a denaturing gel. A radiolabeled alkaline hydrolysis ladder (M2) and size standard (M) were added to the urea-PAGE analysis. (E) Activation of wildtype Csx1 by Pfu cOAs. Csx1 was tested for ribonuclease activity under activating (20 nM Csx1) conditions. Reactions were visualized as described in part B. - - - - Cmr2-HDm Cmr2-Palmm Cmr2-HDm HDm/Palmm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm WT WT Cmr4-D26N Cmr4-D26N WT WT Cmr4-D26N WT Cmr4-D26N WT Cmr4-D26N Nucleic Acids Research, 2020, Vol. 48, No. 8 4425 Pfu Csx1 is activated by cA species produced by Cmr com- served that, when paired with a Cmr2-HD mutation that plexes disrupts DNA cleavage, the Csx1-CARF and Csx1-INQQ mutants both led to a loss of anti-plasmid immunity (Figure The identity of the compounds generated after addition 4F). As expected, wild type Csx1 but not the Csx1-HEPN of ATP to wildtype Cmr complexes (Figure 2D) was de- supported anti-plasmid immunity. Collectively, the in vitro termined by mass spectrometry using established meth- and in vivo results support a key role for the CARF domain ods (41). We found that the Pfu Cmr complex primarily of Pfu Csx1 in triggering cOA-stimulated RNase activity produces cyclic-triadenylate (cA ) and cyclic-tetraadenylate and reveal important CARF domain residues responsible (cA ) species (Figure 3A). The data reveal that cA is the 4 4 for mediating cOA-triggered, Pfu Csx1 RNase activation. most abundant species and is approximately twice as abun- dant as cA ; there were traces of cA and cA which made 3 5 6 Pfu Csx1 cleaves and inactivates cOA using its adenosine- up less than 5% of the total cOA species. To address which specific HEPN RNase active site form(s) of the generated cOA leads to the activation of Csx1 ribonuclease activity, varying concentrations of commer- Once cOAs are produced by Type III crRNPs, it is not well cially available synthetic cA and cA compounds were in- understood how cOA levels are controlled to prevent un- 4 6 cubated with Pfu Csx1 in the presence of 5 -end labeled necessary destruction of vital cellular RNAs by trans-acting substrate RNA (cA and cA compounds were not com- Csx1 (or related Csm6) that could lead to host cell toxicity 3 5 mercially available and so could not be tested). cA but not or death during the immune response. Recently, a new class cA stimulated the ribonuclease activity of Csx1 (Figure 3). of CARF-domain containing proteins called ring nucleases, These results indicate that cA is both the dominant cA were discovered in Sulfolobus solfataricus and found to ex- species produced by Pfu Cmr effector complexes and a po- hibit cOA nuclease activity that halts cOA-triggered Csx1 tent activator of Pfu Csx1 RNA cleavage activity. activity (45). S. solfataricus Csx1 itself was not able to de- grade cOA molecules (45). In contrast, Csm6 of Thermo- The CARF domain of Pfu Csx1 is needed for cOA ribonucle- coccus onnurineus exhibited an intrinsic ability to utilize its ase activation CARF domain to both bind and cleave cOA which gen- erates inactive, linear di-adenylate products with 2 ,3 cyclic The X-ray structure of Pfu Csx1 has been solved (63) and re- phosphate termini (A >p) (44). Previous characterization veals two major domains that are conserved amongst Type of Pfu Csx1 revealed an adenosine specificity for endori- III-affiliated Csx1 or Csm6 RNases: the HEPN (RNase ac- bonuclease activity conferred by the HEPN ribonuclease tive site) and CARF domains (shown in other systems to motif (37). That knowledge combined with the lack of a selectively bind cOA for allosteric activation of the HEPN CARF domain-containing ring nuclease homologs in P. fu- RNase activity) (Figure 4A). Previously, we found that mu- riosus, led us determine if Pfu Csx1 was capable of control- tation to the HEPN motif (Csx1-HEPN ; H436A) dis- ling its own ribonuclease activity by recognizing and de- rupted the ribonuclease activity of Pfu Csx1 (37). To investi- grading cOA (Figure 5). Indeed, we found that wildtype gate if the CARF domain is required for the cOA-stimulated Csx1 protein efficiently converted native cOA substrates RNase activity of Pfu Csx1, we examined the effects of into products with relative mobilities consistent with inac- Csx1 mutants predicted to disrupt CARF function and tive linear di-adenylate (A > p) products (Figure 5Aand cOA binding (Figure 4A and B). Csx1-CARF contains (43,45)). The conversion of cOA into the presumed A >p a deletion of residues 121–127 predicted to be critical for products was abolished by mutations in the HEPN do- CARF domain assembly. The Csx1-INAA and Csx1-INQQ main but not mutations in the CARF domain (Figure 5A). mutants are predicted to prevent specific binding of cOA Moreover, we found that the Csx1-mediated cOA cleav- and contain a double mutation of residues Isoleucine169 age products failed to activate Csx1 ribonuclease activity and Asparagine170 to alanine or glutamine (Figure 4A (Figure 5B). The results indicate the adenosine-specific Pfu and Supplemental Figure S3). We previously showed that Csx1 endoribonuclease responsible for target RNA destruc- high (500 nM) concentrations of wildtype Csx1 cleave RNA tion (37), also utilizes its HEPN active site to recognize, without cOA activation (37). We further analyzed Pfu Csx1 cleave and inactivate cOA molecules providing an autoreg- activity and determined that lowering the concentration to ulation negative feedback control mechanism that limits 10–20 nM resulted in a loss of detectable RNase activity Csx1 RNase activity. by Csx1 in the absence of cOA (Figure 4). All three Csx1 CARF mutants disrupted the ability of low levels of Csx1 Activation of Cmr2-mediated DNA nuclease activity and (10 nM) to support RNA cleavage activity in response to cOA production is dependent upon distinct target PFS ele- cOA (Figure 4C and D) as is observed for the wildtype Csx1 ments enzyme (Figure 4C). None of the CARF mutants impaired the ability of Csx1 to cleave RNAs when high amounts of Prior investigation into regulatory mechanisms controlling enzyme (500 nM) were assayed showing that the mutations the function of Pfu Cmr complexes revealed the important per se did not negatively impact HEPN functionality (Fig- role of three nucleotides within the PFS immediately ad- ure 4C and D). In contrast, the Csx1-HEPN mutant failed jacent to the crRNA/target RNA protospacer interaction to efficiently cleave RNA at either low (10 nM) or high (500 (see Figure 6A) (9). Successful anti-plasmid immunity in nM) concentrations and with or without cOA addition (Fig- vivo and Cmr2-mediated DNA cleavage (via the HD do- ure 4E). main) in vitro required a PFS with a NGN, NNG, or NAA The impact of CARF domain mutations was further sequence (9) and Table 1. In contrast, target RNA cleav- tested in vivo on anti-plasmid immunity (Figure 4F). We ob- age by Pfu crRNPs (via Cmr4 backbone subunit) occurs 4426 Nucleic Acids Research, 2020, Vol. 48, No. 8 Figure 3. cA is the relevant activator for Pfu Csx1. (A) UV chromatogram (258 nm) and MS chromatogram (extracted ion chromatogram for m/z 494.6, 659.1, 823.6, 988.2) for cA samples from Pfu Cmr wildtype complexes. (B) Pfu Csx1 activation assays with synthetic cOA. Wildtype Csx1 (20 nM) was incubated with increasing concentrations of synthetic cA and cA . For each activator, 1.85, 18.5 and 185 pM concentrations were used. A radiolabeled 4 6 alkaline RNA size marker (M) was added to the urea-PAGE analysis. The black arrow indicates the full-length RNA substrate and the black asterisks indicate cleavage products. Dotted lines are indicative of noncontiguous data being omitted from the gel. Table 1. Summary of PFS requirements for Pfu Cmr activities independent of the PFS (9,11). Given the newly observed cOA generation activity for Pfu Cmr complexes revealed in CRISPR defense DNase cOA this study (Figure 2D), we examined whether the first three Target RNA PFS (5 -3 ) (in vivo) (in vitro) (in vitro) positions of the PFS were important for cOA production. NNN GGG + + + Specifically, we addressed whether a large panel of target AAA + + − RNAs that differed only in having distinct PFS elements CCC −− − could activate DNA cleavage (Figure 6B) or cOA produc- UUU −− − NGG GGG + + + tion (Figure 6C) in vitro in parallel reactions. As expected, AGG + + + only target RNAs containing a NGN, NNG, or NAA pro- CGG + + + tospacer flanking sequence activated DNA cleavage (Figure UGG + + + 6B). In contrast, a much smaller subset of the same target GGN GGG + + + GGA + + + RNAs activated cOA production (Figure 6C). The target GGC + + (+) RNAs eliciting strong cOA production activity contain a GGU + + − PFS consensus of NGR sequence. Weak cOA generation ac- UNG UGG + + + tivity was observed for target RNAs containing a GGC or UAG + + (+) UAG sequence within the first three positions of the PFS. A UCG + + − UUG + + − summary of the results for all PFS elements tested on either NUG GUG + + − DNase or cOA generation activities are provided in Table 1. AUG + + − The results reveal a difference in specificity for PFS elements CUG + + − needed for activating cOA generation vs. DNase activity for UUG + + − NGU GGU + + − the Pfu Cmr crRNP. AGU + + − CGU + + − UGU + + − DISCUSSION UNU UGU + + − UAU −− − The Pyrococcus furious Type III-B (Cmr) effector crRNP UCU −− − was the first example of a CRISPR system that identifies UUU −− − and pairs with RNA transcripts rather than DNA strands YUY CUU −− − of invaders (17). A decade of subsequent in vivo and in vitro UUC −− − research has revealed the detailed structure and organiza- Recap of PFS dependent CRISPR defense, DNase activity, and cOA pro- tion of the Pfu Cmr effector crRNPs (20,21,30,64,65)and duction activity results observed for target RNAs used in this study and determined that the system employs highly versatile strate- previously reported (9). + indicates the activity was observed, – denotes gies to combat invading mobile genetic elements (Figure the activity was not observed, and (+) indicates the activity was weakly 7). Previous work showed that immunity provided by the observed. Nucleic Acids Research, 2020, Vol. 48, No. 8 4427 Figure 4. Activation of Pfu Csx1 is CARF domain dependent. (A) Ribbon structure of Pfu Csx1 (PDB: 4EOG) indicating the CARF and HEPN domains. Residue H436 was mutated in the HEPN domain (green). Residues 121–127 were deleted from the CARF domain (blue), and residues I169 and N170 were mutated within the CARF domain (red). (B) SDS-PAGE analysis of Pfu Csx1 protein purifications. WT refers to protein purified using a column column method and WT refers to protein purified using a batch method. ( C–E) In vitro activation of Csx1 mutants. Wildtype Csx1 along with four batch Csx1 mutants (CARF , INAA, INQQ, HEPN ) was tested for ribonuclease activity under high (500 nM Csx1) and low (10 nM Csx1) concentration m m conditions. A radiolabeled alkaline RNA size marker (M) was added to the urea-PAGE analysis. The black arrow indicates the full length RNA substrate and the black asterisks indicate cleavage products. (F) In vivo plasmid silencing assay results for CARF domain mutants. Strains were transformed with the target plasmid (blue) or no target plasmid (gray). Mean colonies obtained per #2;g of plasmid DNA are plotted for two replicates. Error bars indicate standard deviation of replicates. 4428 Nucleic Acids Research, 2020, Vol. 48, No. 8 mune response by activating the trans-acting Csx1 ribonu- clease via their CARF domains (Figures 2-4). Furthermore, the Csx1 HEPN RNase active site, responsible for target RNA destruction via cleaving after adenosine residues (37), also degrades its cognate cyclic-tetra-AMP (cA )activator to switch off the signaling pathway and to limit the activ- ity of Csx1 through autoregulation (Figures 5 and 7). Our in vivo mutational analyses show that either the DNase ac- tivity (via HD domain of Cmr2) or RNase activity (via HEPN motif of Csx1 and triggered by cOA generation by the GGDD motif of Cmr2 Palm domain using ATP pre- cursors) leads to robust anti-plasmid immunity. Our finding that the DNase and cOA generation capacities are activated by distinct 3 protospacer flanking sequences (PFS) of the target RNA highlight that the two target-RNA stimulated enzymatic processes are differentially regulated by distinct allosteric control mechanisms. Taken together with results obtained from other investigated bacterial or archaeal Type III-B (Cmr) and III-A (Csm) systems, our results contribute to making a compelling case for Type III systems as the most complex and highly regulated CRISPR systems dis- covered thus far. Role of trans-acting Csx1 endoribonuclease in conferring anti-plasmid immunity Our previous investigation into the requirement of Pfu Csx1 in anti-plasmid immunity through single csx1 gene muta- tional analysis, failed to uncover its role (9). The more com- prehensive gene mutational analysis performed here reveals that the Csx1 RNase activity or the Cmr2 DNase activity each can independently provide robust anti-plasmid immu- nity (Figure 1). We show that immunity remains unchanged by disrupting either the DNase activity of the Pfu crRNP (via mutation of the HD motif of Cmr2), the RNase ac- tivity of Csx1 (via mutation of the HEPN motif or csx1 gene deletion), or the ability to generate cOA second mes- sengers (via the Palm domain of Cmr2) needed to acti- vate Csx1 RNase activity (Figure 1). In contrast, double mutations that block both DNase and Csx1 RNase activ- Figure 5. Degradation of cOAs by Pfu Csx1 inactivates the cOA activators. (A) Sequencing gel reactions of Csx1 proteins incubated with Pfu cOAs. ity (either directly or through interfering with cOA signal- Five different Pfu Csx1 proteins were incubated with radiolabeled cOAs ing), are required to prevent immunity (Figures 1 and 4). for two timepoints-5 and 20 mins. Urea-page analysis on sequencing gels The results reveal that Cmr2 DNase and Csx1 RNase ac- was performed. Radiolabeled RNA size marker (M) and alkaline hydroly- tivities serve a redundant function in conferring highly ef- sis ladder (M) and were included in the analysis. (B) Effect of Csx1 treat- fective anti-plasmid immunity by the Type III-B Pfu Cmr ment on cOA activation activity. Unlabeled Pfu cOAs were treated with 600 nM of wildtype Csx1 or no protein for 30 min. The plus extraction set system. In comparison, other studied Type III-A or III- of reactions were tested for activation of wildtype Csx1 by incubating 20 B systems have shown a dominant role for Csm6/Csx1 nM of wildtype Csx1 with decreasing concentrations of A >P or cOA –– 2 4 activity (32,34,41,47,62) or Cmr2/Csm1 DNase activity −1 −2 −4 roughly 1.25, 0.6, 0.6 × 10 ,0.6 × 10 , and 0.6 × 10 uM. #2; indicates (14,35,38,50,62) to execute effective immunity against plas- a putative designation for this reaction product. Native cOA without the mids (14,34,38,47,62) or phages (32,35). Type III-C, D, E 30 min incubation or extraction was utilized as a control (N). and F systems are predicted to contain Cmr2 or Csm1 (i.e. Cas10) subunits that lack functional DNase or cOA gen- Pfu Cmr crRNP effector complexes, requires that the in- erating capacities and further work on these systems is re- vasive DNA undergo transcription to produce the target quired to understand if and how they execute immunity RNA required for triggering destruction of both the target against mobile genetic elements (3,4). RNA transcript and invading genome by intrinsic Cmr cr- The exact mechanism of action of Pfu Csx1 for anti- RNP RNase (Cmr4) and DNase (Cmr2 HD domain) activi- plasmid immunity remains to be determined. For example, ties, respectively (9,11,17,27). Here we demonstrate that Pfu it is unclear if Pfu Csx1 functions through selectively de- Cmr effector crRNP complexes also produce cyclic oligoad- stroying the invading target RNA (to provide specific im- enylate second messenger compounds that amplify the im- munity) and/or by destroying host RNA transcripts (lead- Nucleic Acids Research, 2020, Vol. 48, No. 8 4429 Cmr6 Cmr5 Cmr5 Cmr5 5’ Cmr4 Cmr4 Cmr4 Cmr4 Cmr1 Cmr3 3’ crRNA Cmr2 3’ 5’ Target RNA PFS (+1,2,3) crRNA 5’-AUUGAAAGUUGUAGUAUGCGGUCCUUGCGGCUGAGAGCACUUCAG-3’ Target RNA 3’-..UAGGGAGCCAANNNAACAUCAUACGCCAGGAACGCCGACUCUCGUGAAGUCUCCUAGG..-5’ +3+2+1 PFS dsDNA PFS (5’-3’) M + ++ ++ + + + + ++ + ++ ++ + + + + ++ ++ : Cmr crRNP cOA PFS (5’-3’) + + + + + + + ++ + + + + + + + + ++ + + + + + : MM2 Cmr crRNP cOA ATP Figure 6. Pfu Cmr DNase and cOA production activities are PFS dependent. (A) Composition of the Pfu Cmr crRNP. The Cmr effector complex is composed of Cmr1–6 and a mature crRNA (black & orange) containing a 5 tag (black). The Pfu Cmr complex is activated by binding of complementary target RNA (purple) to the crRNA. The target RNA contains a PFS (red) within it that is located adjacent (3 ) to the target sequence. (B)TargetRNA dependent activation of DNase activity by the Pfu complex. Cmr crRNP complexes were incubated with radiolabeled double-stranded DNA in the presence of 23 different target RNAs. Each target RNA contains a different three nucleotide sequence in the PFS region of the target RNA. Urea-page analysis on sequencing gels was performed. Radiolabeled RNA size markers (M) were included in the analysis. The black arrow indicates the full-length RNA substrate and the black asterisks indicate cleavage products. (C) Target RNA dependent activation of cOA production activity by the Pfu complex using the same target RNAs mentioned in part B. Products were analyzed as described in part B. An alkaline hydrolysis ladder (M2) was included in the urea-PAGE analysis. ing to cellular dormancy or death) (Figure 7). We are un- the Type III-B system of S. epidermidis was found to ei- able to distinguish these two different scenarios given that ther selectively cleave target RNA (including at regions out- inducible gene expression systems have yet to be established side of the RNA protospacer segment) while sparing cel- for Pfu that would enable testing which RNAs are being de- lular RNAs (35), or to destroy both target RNA and cel- stroyed as a function of activation of the Cmr crRNP activi- lular RNAs (38). It is conceivable that low levels of target ties over time. Given that Pfu Csx1 was shown to cleave var- RNA and second messenger signaling may limit Csx1 and ious RNAs without specificity in vitro (except that it cleaves Csm6 RNase activity to target RNA destruction, while el- after adenosine residues (37)) and that Csx1 or related Csm6 evated target RNA concentrations and high cOA signaling are not normally stably associated with Type III effector cr- may lead to more indiscriminate RNA degradation. Selec- RNPs (15,17,21–23,34), to be selective for the target RNA tive targeting of invader RNAs would promote immunity would involve conditional recruitment of Csx1 to the cr- and maintain viable host cells. In contrast, indiscriminate RNP when the complex is engaged in interaction with the RNA destruction would prevent the growth or kill infected target RNA protospacer. The related Csm6 RNase from host cells, which would effectively prevent the spread of the - - - - GGG AAA CCC UUU UGG UUG UGU CGG AGG GGU GGC GGA UCG UAG CUG CUU UCU CGU AUG UUC UAU AGU GUG GGG AAA CCC UUU UGG UUG UGU CUU UCU UUC CGG AGG GGU GGC GGA UCG UAU UAG CUG CGU AUG AGU GUG 4430 Nucleic Acids Research, 2020, Vol. 48, No. 8 Dormacy or Viral/Plasmid Immunity Cell Death Type III crRNP 5’ ATP 5’ Cmr4 crRNA CARF Palm 3’ Csx1 5’ cOA PFS HEPN Cmr2 Cellular Viral HD RNA RNA Viral DNA 5’ 3’ 3’ 3’ 5’ 3’ RNAP Figure 7. Model for regulation of the RNAse, DNase, and cOA synthetase activities of the Pfu Type III-B effector crRNP in providing anti-viral immunity. Inactive Type III-B crRNPs become activated upon interaction with expressed viral RNA. crRNA-guided base-pairing to the viral RNA and Cas protein recognition of the PFS results in conformational changes that trigger three important activities: viral RNA cleavage by Cmr4, viral DNA cleavage by the HD domain of Cmr2 (a member of the Cas10 superfamily), and ATP dependent production of cyclic oligoadenylates (cOA) by the Palm domain of Cmr2. The cOA second messengers allosterically regulate the trans-acting Csx1 enzyme by binding to the CARF domain to trigger HEPN RNase activity. Clearance of the viral DNAs and RNAs provides immunity and restores the crRNPs back to the inactive state. Under conditions of controlled immunity, Csx1 degrades and inactivates cOA signaling molecules using the HEPN domain, to provide an auto-inhibitory off-switch of Csx1 RNase activation. Inefficient degradation of cOA molecules by the Csx1 or failure to efficiently destroy the viral RNAs or viral genome may cause Csx1 to destroy both viral and cellular RNA molecules, leading to cellular dormancy or death. invader to other host cells in the local environment (Figure onnurineus (Ton), was found to be capable of autoregulat- 7). ing RNase activity through binding and cleaving cA (44). However, unlike Pfu Csx1, Ton cA cleavage of cA into in- 4 4 active linearized adenosine dinucleotide products (A >P) Second messenger signaling pathway leading to Csx1 RNase depended upon the CARF rather than HEPN domain de- activation and deactivation spite evidence from high resolution structural studies show- ing that cOA binds to both CARF and HEPN domains of We show for the first time that the Pfu Cmr complex gen- Ton Csm6 (44). Interestingly, the HEPN domains of Ton erates primarily cA signaling molecules (Figure 3A) that Csm6 and Pfu Csx1 each cleave RNA with a strict adeno- significantly stimulate the RNase activity of Csx1 (Figures sine specificity ( 37,44). Furthermore, Thermus thermophilus 2-4). Either cA or cA have been found to be the rele- 4 6 Csm6 exhibits an intrinsic ability to degrade its cognate cA vant activators for various characterized bacterial or ar- using its CARF domain but not its HEPN domain (43). chaeal Csx1 and Csm6 enzymes (31–33,41,42,44,51). Inter- Studies examining RNA cleavage patterns of various estingly, the HEPN RNase active site of Pfu Csx1 (which Csx1 and Csm6 HEPN endoribonucleases revealed that is adenosine-specific) cleaves and inactivates cOA signaling many are either specific for cleaving after adenosines ( 37,44) molecules (Figure 5) in addition to executing the destruc- or purines (adenosines and guanosines) (31,34). These cat- tion of adenosine-containing, target RNA transcripts (Fig- alytic properties indicate that other Csx1 or Csm6 enzymes ures 2–5 and (37)). Previous studies have shown that the may utilize their HEPN RNase active site to degrade cOA CARF domain is the cOA sensor domain (31,32,44,46,51) molecules. Thus, Pfu Csx1 HEPN domain-mediated de- and accordingly, mutation of the CARF domain of Pfu struction of its cognate cOA activator reveals a novel and Csx1 blocks activation of Csx1 RNase activity. However, distinct auto-catalytic control mechanism that limits Csx1 Pfu Csx1 CARF domain inactivation did not abolish cleav- RNase activity that may be common to other Type III sys- age and inactivation of cOA second messengers (Figure 5). tems. Our finding that the HEPN RNase motif, rather than CARF domain of Pfu Csx1, can cleave and inactivate cOA signaling molecules expands the number of mechanisms ca- Influence of target RNA interactions in controlling DNase pable of down-regulating Type III-induced cOA signaling and RNase activities of Pfu effector crRNPs pathways. For example, Sulfolobus islandicus Csx1 appears to lack the ability to itself degrade cOA molecules and in- All characterized Type III CRISPR–Cas systems are specif- stead utilizes extrinsic CARF domain-containing factors ically activated through identifying and interacting with (called ring nucleases) to switch off Csx1 RNase activa- target RNA. Base-pairing interactions between crRNAs tion (45). Similar to our finding with Pfu Csx1, Csm6 within crRNP effector complexes and matching target from another hyperthermophlic archaeon, Thermococcus RNA protospacer elements, has been shown to trigger Nucleic Acids Research, 2020, Vol. 48, No. 8 4431 major conformational changes within Type III complexes vating the DNase or cOA synthesis. These results have pro- (25,49,50). In turn, less understood, minor target RNA- vided insight that the two important activities for immunity induced conformational changes lead to activation of sev- are differentially controlled by distinct molecular mecha- eral enzymatic activities of the complexes including RNase, nisms. Additional evidence supporting the notion that dis- DNase and cOA generation needed for downstream ac- tinct conformational changes triggered by PFS/protein in- tivation of Csx1 and Csm6 HEPN endoribonucleases teractions separately influence either DNase or cOA pro- (25,49,50,52). duction were observed with S. thermophilus Type III-A sys- Some but not all of these enzymatic activities are further tem where Csm1 mutations at residues thought to transmit regulated by a ∼eight base-pair PFS region of the target allosteric effects of the PFS were found to impair DNase RNA that is located just 3 of the target RNA protospacer activity only, cOA synthesis only, or both activities (25). Of region. For all Type III systems evaluated thus far, includ- note, the Pfu spacer acquisition machinery responsible for ing Pfu (9), if the target RNA PFS is complementary to the addition of new spacers into CRISPR arrays, recognized the 5 tag of the crRNA (as is the case if anti-sense transcrip- invading DNAs having a 5 -NGG-3 consensus PAM (pro- tion of the CRISPR locus occurs (27,66)) then base-paring tospacer adjacent motif) (70,71). In turn, our results pre- between residues of the crRNA tag and the PFS of the tar- dict that the resultant 5 -NGG-3 RNA PFS of expressed in- get RNA apparently prevents the ability of the crRNPs to vader genomes would trigger both DNA cleavage and cOA allosterically activate DNase and cOA production (9,10,12– activities of Pfu Cmr effector crRNPs (Figure 6 and Table 14,54,67,68). In contrast, cleavage of the target RNA (at 1). A functional coupling between PAM recognition during regular 6 nucleotide intervals throughout target RNA pro- CRISPR spacer acquisition and Type III interference acti- tospacer region) by Cmr4 (III-B) or Csm4 (III-A) crRNP vation may boost the specificity of certain Type III systems backbone RNases is unaffected by crRNA 5 tag/target to compensate for an observed high degree of tolerance to RNA PFS interactions (9,11)(Figure 7). mismatches in the rest of the target RNA (69). There are apparent differences in PFS requirements for activation of DNase and cOA generation activities for dis- tinct Type III systems. For example, a simple lack of base- Role for the RNAse activity of Cmr4 subunit? pairing between 5 crRNA tag and target RNA PFS, ap- pears to be sufficient for activating these enzymatic activ- We were unable to determine the possible in vivo role of the ities for the Staphylococcus epidermidis Type III-A system RNase activity of Cmr4 on anti-plasmid immunity. Cmr4 (69). In contrast, the identity of PFS sequences, especially is the backbone subunit of the III-B crRNP (see Figure 1) within the +1, +2 and/or +3 positions have been found to be that cleaves the bound target RNA at regular six nucleotide important for conferring immunity for Type III systems of intervals (27). Despite repeated attempts, we failed to gen- Pyrococcus furiosus (9), Streptococcus thermophilus (25,48) erate a strain in which the RNase activity of Cmr4 was inac- and Thermotoga maritima (54). For most studied Type III tivated (via D26N mutation that blocks target RNA cleav- systems, the PFS requirements have not been thoroughly in- age in vitro (Figure 2)). Given that we were able to read- vestigated. ily create the catalytically inactive Cmr4 mutant in a strain Based on recent Type III crRNP structural studies, the in which both the DNase and Csx1 RNase activities were PFS sequences in the +1, +2, +3 positions are predicted to also inactivated (i.e. the Cmr2-HD /Csx1 strain that can- interact with Csm1/Cmr2 (Cas10) and Csm4/Cmr3 (5 tag not execute anti-plasmid immunity (Figure 1)), we suspect interacting) subunits (25,48,49). We propose that in systems that preventing Cmr4 RNase activity results in host cell where the identity of protospacer flanking sequences mat- lethality brought about host cell DNAase activity and/or ter for activating function such as Pfu,PFS RNA/protein cOA signaling and Csx1 RNase activation as the result of contacts play a key role in initiating the conformational generation of constitutively active Type III-B crRNPs. This changes that ultimately trigger activation of the Cas10 HD would occur if there was recognition of endogenous P. fu- domains to cleave invader DNA and the Cas10 Palm do- riousus RNA(s) exhibiting complementarity to any of the main GGGDD motif to convert ATP into cOA second mes- 200 known crRNAs that are assembled into P. furious type sengers. Interestingly, our systematic analyses of the impact III-B crRNPs (55,56). Consistent with this possibility, it has of varying the PFS elements in the +1, +2 and +3 positions been shown that cleavage of the crRNA-bound target RNA of the target RNA, revealed that a given PFS can activate by Cmr4 (and related Csm3 protein of III-A crRNPs) stim- DNase activity, cOA production, both activities or neither ulates rapid release of the target RNA into solution to tem- activity (independent of the ability of these PFS nucleotides porally control (switch off) the DNase and cOA produc- to base-pair with the 5 crRNA tag element) (Figure 6 and tion activities of the complex (9,10,13). Failure to cleave Table 1). Furthermore, we found a direct correlation be- the target RNA that is required for triggering the activi- tween PFS elements that were previously found to support ties of the complex, is expected to drive the effector crRNPs anti-plasmid immunity in vivo (9) and those found to also into a long-lasting nuclease-active state. The scenario is fur- be required for in DNase activity in vitro (Figure 6 and Ta- ther supported by findings with the S. epidermidis III-A ble 1). In contrast, only a subset of PFS elements required complex that exhibits a hyperactive immunity phenotype in in vivo, resulted in cOA production in vitro (Figure 6 and vivo when target RNA cleavage was disrupted through in- Table 1). ducible expression of the comparable catalytically inactive All Pfu PFS elements that triggered activity, shared a Csm3 mutant (14). Taken together, we view it likely that the purine-rich character within position +2 and +3 of the tar- RNase activity of P. furiousus Cmr4 normally contributes get RNA PFS appears to responsible for differentially acti- to both the destruction and turnover of the target RNA and 4432 Nucleic Acids Research, 2020, Vol. 48, No. 8 plays an important role in immunity and regulation of type and structural analyses of Pfu Type III-B crRNPs in the III-B crRNP activity. presence and absence of substrate target RNAs and DNAs and with and without Csx1, would provide important in- sight into specific molecular mechanisms that control the Contribution of DNase vs. RNase activities to type III immu- timing and specificity of Type III crRNP-mediated immu- nity nity. The ability of type III CRISPR–Cas systems to act both at the level of DNase and RNase degradation for invader im- SUPPLEMENTARY DATA munity, sets them apart from all know other CRISPR types. Moreover, it is clear that there is great variation in the rel- Supplementary Data are available at NAR Online. ative importance of these two distinct nuclease activities in conferring immunity between different type III systems as ACKNOWLEDGEMENTS well as within a system for targeting specific invaders. Here, we find that anti-plasmid immunity for the P. furiosus III-B We thank members of the Terns and White laboratories for system is robust when only employing DNase activity (me- helpful discussions and guidance. We also express gratitude diated by the HD domain of Cmr2) or RNAse activity (me- to Janet Westpheling for providing critical guidance on gen- diated by the cOA-regulated, HEPN domain of Csx1) (Fig- erating the cmr2 and csx1 complementation strains. ure 1). This is in contrast to in vivo results obtained for sev- eral other type III systems including Sulfolobus islandicus (47), S. epidermidis (62,72), S. thermophilus (72), Lactococ- FUNDING cus lactis (72) shown to rely primarily on Csx1 RNase ac- National Institutes of Health (NIH) [R35GM118160 to tivity given that plasmid immunity was abolished by single M.P.T., R01GM097330 to S.B. and 1F31GM125365 to deletion or mutational inactivation of the csx1 or csm6 gene. K.F.]; Biotechnology and Biological Sciences Research Yet other systems appear to solely utilize the DNase activity Council [REF: BB/S000313/1 to M.F.W.]. Funding for as exemplified by the recently characterized Lactobacillus open access charge: NIH grant. delbrueckii found to confer anti-plasmid immunity despite Conflict of interest statement. None declared. lacking a Csm6 gene and not being capable of producing cOA (73). Anti-plasmid immunity for the S. aureus III-B system, like P. furious, appears to depend on both DNase REFERENCES and RNase targeting but a partial interference phenotype 1. Jackson,R.N., van Erp,P.B., Sternberg,S.H. and Wiedenheft,B. (2017) is observed with csx1 gene knockout, indicating relatively Conformational regulation of CRISPR-associated nucleases. Curr. weak DNase activity (74). Finally, when anti-viral immu- Opin. Microbiol., 37, 110–119. nity was examined, the DNAse activity alone was found be 2. Hille,F., Richter,H., Wong,S.P., Bratovic,M., Ressel,S. and protective except under conditions of excess viral genome Charpentier,E. (2018) The biology of CRISPR–Cas: backward and forward. Cell, 172, 1239–1259. copy numbers where Csm6 RNase activity becomes neces- 3. Makarova,K.S., Wolf,Y.I., Iranzo,J., Shmakov,S.A., Alkhnbashi,O.S., sary to provide robust immunity (75). The natural high de- Brouns,S.J.J., Charpentier,E., Cheng,D., Haft,D.H., Horvath,P. et al. gree of functional variation amongst type III systems is also (2020) Evolutionary classification of CRISPR–Cas systems: a burst highlighted by the prediction that type III-C through III- of class 2 and derived variants. Nat. Rev. Microbiol., 18, 67–83. 4. Makarova,K.S., Wolf,Y.I., Alkhnbashi,O.S., Costa,F., Shah,S.A., F systems are predicted to encode Cas10 (Cmr2 or Csm1) Saunders,S.J., Barrangou,R., Brouns,S.J., Charpentier,E., Haft,D.H. proteins that lack the ability to cut DNA or generate cOA et al. (2015) An updated evolutionary classification of CRISPR–Cas signalling molecules (3). systems. Nat. Rev. Microbiol., 13, 722–736. In addition to these functional differences, the regulatory 5. Brouns,S.J., Jore,M.M., Lundgren,M., Westra,E.R., Slijkhuis,R.J., mechanism that govern spatiotemporal control of the activ- Snijders,A.P., Dickman,M.J., Makarova,K.S., Koonin,E.V. and van der Oost,J. (2008) Small CRISPR RNAs guide antiviral defense in ities of type III systems have also shown to be quite differ- prokaryotes. Science, 321, 960–964. ent. For example, we show here clear evidence that the P. 6. Garneau,J.E., Dupuis,M.E., Villion,M., Romero,D.A., furiousus III-B system utilizes bipartite recognition of tar- Barrangou,R., Boyaval,P., Fremaux,C., Horvath,P., Magadan,A.H. get RNA (through crRNA base-pairing and likely protein- and Moineau,S. (2010) The CRISPR/Cas bacterial immune system cleaves bacteriophage and plasmid DNA. Nature, 468, 67–71. mediated PFS interaction) in stimulating the DNase and 7. Zetsche,B., Gootenberg,J.S., Abudayyeh,O.O., Slaymaker,I.M., RNase activities of the crRNP (Figure 6). In contrast, cr- Makarova,K.S., Essletzbichler,P., Volz,S.E., Joung,J., van der Oost,J., RNA pairing alone with the target RNA appears to trig- Regev,A. et al. (2015) Cpf1 is a single RNA-guided endonuclease of a ger activity for the S. epidermidis III-A system (69). Collec- class 2 CRISPR–Cas system. Cell, 163, 759–771. tively, the high degree of functional variability argues that 8. Abudayyeh,O.O., Gootenberg,J.S., Konermann,S., Joung,J., Slaymaker,I.M., Cox,D.B., Shmakov,S., Makarova,K.S., diverse type III systems must be extensively characterized to Semenova,E., Minakhin,L. et al. (2016) C2c2 is a single-component understand the full spectrum of modes of action and regu- programmable RNA-guided RNA-targeting CRISPR effector. latory mechanisms operating in type III immunity. Science, 353, aaf5573. In summary, our findings significantly contribute to our 9. Elmore,J.R., Sheppard,N.F., Ramia,N., Deighan,T., Li,H., Terns,R.M. and Terns,M.P. (2016) Bipartite recognition of target understanding of the multiple mechanisms that ensure that RNAs activates DNA cleavage by the Type III-B CRISPR–Cas Type III CRISPR systems become appropriately activated system. Genes Dev., 30, 447–459. and then rapidly deactivated to achieve maximum benefit 10. Estrella,M.A., Kuo,F.T. and Bailey,S. (2016) RNA-activated DNA to host cells experiencing infection from invading mobile cleavage by the Type III-B CRISPR–Cas effector complex. Genes genetic elements. In future studies, a detailed biochemical Dev., 30, 460–470. Nucleic Acids Research, 2020, Vol. 48, No. 8 4433 11. Hale,C.R., Cocozaki,A., Li,H., Terns,R.M. and Terns,M.P. (2014) 31. Kazlauskiene,M., Kostiuk,G., Venclovas,C., Tamulaitis,G. and Target RNA capture and cleavage by the Cmr type III-B Siksnys,V. (2017) A cyclic oligonucleotide signaling pathway in type CRISPR–Cas effector complex. Genes Dev., 28, 2432–2443. III CRISPR–Cas systems. Science, 357, 605–609. 12. Han,W., Li,Y., Deng,L., Feng,M., Peng,W., Hallstrom,S., Zhang,J., 32. Niewoehner,O., Garcia-Doval,C., Rostol,J.T., Berk,C., Schwede,F., Peng,N., Liang,Y.X., White,M.F. et al. (2017) A type III-B Bigler,L., Hall,J., Marraffini,L.A. and Jinek,M. (2017) Type III CRISPR–Cas effector complex mediating massive target DNA CRISPR–Cas systems produce cyclic oligoadenylate second destruction. Nucleic Acids Res., 45, 1983–1993. messengers. Nature, 548, 543–548. 13. Kazlauskiene,M., Tamulaitis,G., Kostiuk,G., Venclovas,C. and 33. Rouillon,C., Athukoralage,J.S., Graham,S., Gruschow,S. and Siksnys,V. (2016) Spatiotemporal control of Type III-A CRISPR–Cas White,M.F. (2018) Control of cyclic oligoadenylate synthesis in a type immunity: coupling DNA degradation with the target RNA III CRISPR system. Elife, 7, e36734. recognition. Mol. Cell, 62, 295–306. 34. Foster,K., Kalter,J., Woodside,W., Terns,R.M. and Terns,M.P. (2019) 14. Samai,P., Pyenson,N., Jiang,W., Goldberg,G.W., Hatoum-Aslan,A. The ribonuclease activity of Csm6 is required for anti-plasmid and Marraffini,L.A. (2015) Co-transcriptional DNA and RNA immunity by Type III-A CRISPR–Cas systems. RNA Biol., 16, cleavage during Type III CRISPR–Cas immunity. Cell, 161, 449–460. 1164–1174. 35. Jiang,W., Samai,P. and Marraffini,L.A. (2016) Degradation of phage 15. Staals,R.H.J., Agari,Y., Maki-Yonekura,S., Zhu,Y., Taylor,D.W., van transcripts by CRISPR-Associated RNases enables Type III Duijn,E., Barendregt,A., Vlot,M., Koehorst,J.J., Sakamoto,K. et al. CRISPR–Cas immunity. Cell, 164, 710–721. (2013) Structure and activity of the RNA-targeting Type III-B 36. Niewoehner,O. and Jinek,M. (2016) Structural basis for the CRISPR–Cas complex of Thermus thermophilus. Mol. Cell, 52, endoribonuclease activity of the type III-A CRISPR-associated 135–145. protein Csm6. RNA, 22, 318–329. 16. Tamulaitis,G., Kazlauskiene,M., Manakova,E., Venclovas,C., 37. Sheppard,N.F., Glover,C.V. 3rd, Terns,R.M. and Terns,M.P. (2016) Nwokeoji,A.O., Dickman,M.J., Horvath,P. and Siksnys,V. (2014) The CRISPR-associated Csx1 protein of Pyrococcus furiosus is an Programmable RNA shredding by the type III-A CRISPR–Cas adenosine-specific endoribonuclease. RNA, 22, 216–224. system of Streptococcus thermophilus. Mol. Cell, 56, 506–517. 38. Rostol,J.T. and Marraffini,L.A. (2019) Non-specific degradation of 17. Hale,C.R., Zhao,P., Olson,S., Duff,M.O., Graveley,B.R., Wells,L., transcripts promotes plasmid clearance during type III-A Terns,R.M. and Terns,M.P. (2009) RNA-guided RNA cleavage by a CRISPR–Cas immunity. Nat. Microbiol., 4, 656–662. CRISPR RNA-Cas protein complex. Cell, 139, 945–956. 39. Anantharaman,V., Makarova,K.S., Burroughs,A.M., Koonin,E.V. 18. Tamulaitis,G., Venclovas,C. and Siksnys,V. (2017) Type III and Aravind,L. (2013) Comprehensive analysis of the HEPN CRISPR–Cas immunity: major differences brushed aside. Trends superfamily: identification of novel roles in intra-genomic conflicts, Microbiol., 25, 49–61. defense, pathogenesis and RNA processing. Biol. Direct, 8,15. 19. Goldberg,G.W., Jiang,W., Bikard,D. and Marraffini,L.A. (2014) 40. Makarova,K.S., Anantharaman,V., Grishin,N.V., Koonin,E.V. and Conditional tolerance of temperate phages via Aravind,L. (2014) CARF and WYL domains: ligand-binding transcription-dependent CRISPR–Cas targeting. Nature, 514, regulators of prokaryotic defense systems. Front. Genet., 5, 102. 633–637. 41. Gruschow,S., Athukoralage,J.S., Graham,S., Hoogeboom,T. and 20. Spilman,M., Cocozaki,A., Hale,C., Shao,Y., Ramia,N., Terns,R., White,M.F. (2019) Cyclic oligoadenylate signalling mediates Terns,M., Li,H. and Stagg,S. (2013) Structure of an RNA silencing Mycobacterium tuberculosis CRISPR defence. Nucleic Acids Res., complex of the CRISPR–Cas immune system. Mol. Cell, 52, 146–152. 47, 9259–9270. 21. Osawa,T., Inanaga,H., Sato,C. and Numata,T. (2015) Crystal 42. Nasef,M., Mufyfl ,M.C., Beckman,A.B., Rowe,S.J., Walker,F.C., structure of the CRISPR–Cas RNA silencing Cmr complex bound to Hatoum-Aslan,A. and Dunkle,J.A. (2019) Regulation of cyclic a target analog. Mol. Cell, 58, 418–430. oligoadenylate synthesis by the Staphylococcus epidermidis 22. Rouillon,C., Zhou,M., Zhang,J., Politis,A., Beilsten-Edmands,V., Cas10-Csm complex. RNA, 25, 948–962. Cannone,G., Graham,S., Robinson,C.V., Spagnolo,L. and 43. Athukoralage,J.S., Graham,S., Gruschow,S., Rouillon,C. and White,M.F. (2013) Structure of the CRISPR interference complex White,M.F. (2019) A Type III CRISPR ancillary ribonuclease CSM reveals key similarities with cascade. Mol. Cell, 52, 124–134. degrades its cyclic oligoadenylate activator. J. Mol. Biol., 431, 23. Staals,R.H., Zhu,Y., Taylor,D.W., Kornfeld,J.E., Sharma,K., 2894–2899. Barendregt,A., Koehorst,J.J., Vlot,M., Neupane,N., Varossieau,K. 44. Jia,N., Jones,R., Yang,G., Ouerfelli,O. and Patel,D.J. (2019) et al. (2014) RNA targeting by the type III-A CRISPR–Cas Csm CRISPR–Cas III-A Csm6 CARF domain is a ring nuclease complex of Thermus thermophilus. Mol. Cell, 56, 518–530. triggering stepwise cA4 cleavage with ApA>p formation terminating 24. Taylor,D.W., Zhu,Y., Staals,R.H., Kornfeld,J.E., Shinkai,A., van der RNase activity. Mol. Cell, 75, 944–956. Oost,J., Nogales,E. and Doudna,J.A. (2015) Structural biology. 45. Athukoralage,J.S., Rouillon,C., Graham,S., Gruschow,S. and Structures of the CRISPR-Cmr complex reveal mode of RNA target White,M.F. (2018) Ring nucleases deactivate type III CRISPR positioning. Science, 348, 581–585. ribonucleases by degrading cyclic oligoadenylate. Nature, 562, 25. You,L., Ma,J., Wang,J., Artamonova,D., Wang,M., Liu,L., Xiang,H., 277–280. Severinov,K., Zhang,X. and Wang,Y. (2019) Structure studies of the 46. Molina,R., Stella,S., Feng,M., Sofos,N., Jauniskis,V., CRISPR-Csm complex reveal mechanism of co-transcriptional Pozdnyakova,I., Lopez-Mendez,B., She,Q. and Montoya,G. (2019) interference. Cell, 176, 239–253 e216. Structure of Csx1-cOA4 complex reveals the basis of RNA decay in 26. Hale,C., Kleppe,K., Terns,R.M. and Terns,M.P. (2008) Prokaryotic Type III-B CRISPR–Cas. Nat. Commun., 10, 4302. silencing (psi)RNAs in Pyrococcus furiosus. RNA, 14, 2572–2579. 47. Deng,L., Garrett,R.A., Shah,S.A., Peng,X. and She,Q. (2013) A 27. Hale,C.R., Majumdar,S., Elmore,J., Pfister,N., Compton,M., novel interference mechanism by a type IIIB CRISPR-Cmr module Olson,S., Resch,A.M., Glover,C.V. 3rd, Graveley,B.R., Terns,R.M. in Sulfolobus. Mol. Microbiol., 87, 1088–1099. et al. (2012) Essential features and rational design of CRISPR RNAs 48. Guo,M., Zhang,K., Zhu,Y., Pintilie,G.D., Guan,X., Li,S., that function with the Cas RAMP module complex to cleave RNAs. Schmid,M.F., Ma,Z., Chiu,W. and Huang,Z. (2019) Coupling of Mol. Cell, 45, 292–302. ssRNA cleavage with DNase activity in type III-A CRISPR-Csm 28. Carte,J., Wang,R., Li,H., Terns,R.M. and Terns,M.P. (2008) Cas6 is revealed by cryo-EM and biochemistry. Cell Res., 29, 305–312. an endoribonuclease that generates guide RNAs for invader defense 49. Jia,N., Mo,C.Y., Wang,C., Eng,E.T., Marraffini,L.A. and Patel,D.J. in prokaryotes. Genes Dev., 22, 3489–3496. (2019) Type III-A CRISPR–Cas Csm complexes: assembly, periodic 29. Ramia,N.F., Spilman,M., Tang,L., Shao,Y., Elmore,J., Hale,C., RNA cleavage, DNase activity regulation, and autoimmunity. Mol. Cocozaki,A., Bhattacharya,N., Terns,R.M., Terns,M.P. et al. (2014) Cell, 73, 264–277. Essential structural and functional roles of the Cmr4 subunit in RNA 50. Wang,L., Mo,C.Y., Wasserman,M.R., Rostol,J.T., Marraffini,L.A. cleavage by the Cmr CRISPR–Cas complex. Cell Rep., 9, 1610–1617. and Liu,S. (2019) Dynamics of Cas10 govern discrimination between 30. Benda,C., Ebert,J., Scheltema,R.A., Schiller,H.B., Baumgartner,M., self and Non-self in Type III CRISPR–Cas immunity. Mol. Cell, 73, Bonneau,F., Mann,M. and Conti,E. (2014) Structural model of a 278–290. CRISPR RNA-silencing complex reveals the RNA-target cleavage 51. Han,W., Stella,S., Zhang,Y., Guo,T., Sulek,K., Peng-Lundgren,L., activity in Cmr4. Mol. Cell, 56, 43–54. Montoya,G. and She,Q. (2018) A Type III-B Cmr effector complex 4434 Nucleic Acids Research, 2020, Vol. 48, No. 8 catalyzes the synthesis of cyclic oligoadenylate second messengers by 64. Cocozaki,A.I., Ramia,N.F., Shao,Y., Hale,C.R., Terns,R.M., cooperative substrate binding. Nucleic Acids Res., 46, 10319–10330. Terns,M.P. and Li,H. (2012) Structure of the Cmr2 subunit of the 52. Jia,N., Jones,R., Sukenick,G. and Patel,D.J. (2019) Second messenger CRISPR–Cas RNA silencing complex. Structure, 20, 545–553. cA4 formation within the composite Csm1 palm pocket of Type III-A 65. Shao,Y., Cocozaki,A.I., Ramia,N.F., Terns,R.M., Terns,M.P. and CRISPR–Cas Csm complex and its release path. Mol. Cell, 75, Li,H. (2013) Structure of the Cmr2-Cmr3 subcomplex of the Cmr 933–943. RNA silencing complex. Structure, 21, 376–384. 53. Marraffini,L.A. and Sontheimer,E.J. (2010) Self versus non-self 66. Lillestol,R.K., Shah,S.A., Brugger,K., Redder,P., Phan,H., discrimination during CRISPR RNA-directed immunity. Nature, Christiansen,J. and Garrett,R.A. (2009) CRISPR families of the 463, 568–571. crenarchaeal genus Sulfolobus: bidirectional transcription and 54. Johnson,K., Learn,B.A., Estrella,M.A. and Bailey,S. (2019) Target dynamic properties. Mol. Microbiol., 72, 259–272. sequence requirements of a type III-B CRISPR–Cas immune system. 67. Liu,T.Y., Iavarone,A.T. and Doudna,J.A. (2017) RNA and DNA J. Biol. Chem., 294, 10290–10299. Targeting by a Reconstituted Thermus thermophilus Type III-A 55. Majumdar,S., Zhao,P., Pfister,N.T., Compton,M., Olson,S., CRISPR–Cas System. PLoS One, 12, e0170552. Glover,C.V. 3rd, Wells,L., Graveley,B.R., Terns,R.M. and Terns,M.P. 68. Zhang,J., Graham,S., Tello,A., Liu,H. and White,M.F. (2016) (2015) Three CRISPR–Cas immune effector complexes coexist in Multiple nucleic acid cleavage modes in divergent type III CRISPR Pyrococcus furiosus. RNA, 21, 1147–1158. systems. Nucleic Acids Res., 44, 1789–1799. 56. Terns,R.M. and Terns,M.P. (2013) The RNA- and DNA-targeting 69. Pyenson,N.C., Gayvert,K., Varble,A., Elemento,O. and CRISPR–Cas immune systems of Pyrococcus furiosus. Biochem. Soc. Marraffini,L.A. (2017) Broad targeting specificity during bacterial Trans., 41, 1416–1421. Type III CRISPR–Cas immunity constrains viral escape. Cell Host 57. Majumdar,S. and Terns,M.P. (2019) CRISPR RNA-guided DNA Microbe, 22, 343–353. cleavage by reconstituted Type I-A immune effector complexes. 70. Shiimori,M., Garrett,S.C., Chambers,D.P., Glover,C.V.C. 3rd, Extremophiles, 23, 19–33. Graveley,B.R. and Terns,M.P. (2017) Role of free DNA ends and 58. Majumdar,S., Ligon,M., Skinner,W.C., Terns,R.M. and Terns,M.P. protospacer adjacent motifs for CRISPR DNA uptake in Pyrococcus (2017) Target DNA recognition and cleavage by a reconstituted Type furiosus. Nucleic. Acids. Res., 45, 11281–11294. I-G CRISPR–Cas immune effector complex. Extremophiles, 21, 71. Shiimori,M., Garrett,S.C., Graveley,B.R. and Terns,M.P. (2018) Cas4 95–107. nucleases define the PAM, length, and orientation of DNA fragments 59. Elmore,J., Deighan,T., Westpheling,J., Terns,R.M. and Terns,M.P. integrated at CRISPR Loci. Mol. Cell, 70, 814–824. (2015) DNA targeting by the type I-G and type I-A CRISPR–Cas 72. Ichikawa,H.T., Cooper,J.C., Lo,L., Potter,J., Terns,R.M. and systems of Pyrococcus furiosus. Nucleic Acids Res., 43, 10353–10363. Terns,M.P. (2017) Programmable type III-A CRISPR–Cas DNA 60. Lipscomb,G.L., Stirrett,K., Schut,G.J., Yang,F., Jenney,F.E. Jr., targeting modules. PLoS One, 12, e0176221. Scott,R.A., Adams,M.W. and Westpheling,J. (2011) Natural 73. Lin,J., Feng,M., Zhang,H. and She,Q. (2019) Characterization of a competence in the hyperthermophilic archaeon Pyrococcus furiosus novel type III effector provides new insights into the allosteric facilitates genetic manipulation: construction of markerless deletions activation and suppression of the Cas10 DNase. bioRxiv doi: of genes encoding the two cytoplasmic hydrogenases. Appl. Environ. https://doi.org/10.1101/2019.12.17.879585, 18 December 2019, Microbiol., 77, 2232–2238. preprint: not peer reviewed. 61. Rouillon,C., Athukoralage,J.S., Graham,S., Gruschow,S. and 74. Cao,L., Gao,C.H., Zhu,J., Zhao,L., Wu,Q., Li,M. and Sun,B. (2016) White,M.F. (2019) Investigation of the cyclic oligoadenylate signaling Identification and functional study of type III-A CRISPR–Cas pathway of type III CRISPR systems. Methods Enzymol., 616, systems in clinical isolates of Staphylococcus aureus. Int. J. Med. 191–218. Microbiol.: IJMM, 306, 686–696. 62. Hatoum-Aslan,A., Maniv,I., Samai,P. and Marraffini,L.A. (2014) 75. Pyenson,N.C. and Marraffini,L.A. (2017) Type III CRISPR–Cas Genetic characterization of antiplasmid immunity through a type systems: when DNA cleavage just isn’t enough. Curr. Opin. III-A CRISPR–Cas system. J. Bacteriol., 196, 310–317. Microbiol., 37, 150–154. 63. Kim,Y.K., Kim,Y.G. and Oh,B.H. (2013) Crystal structure and nucleic acid-binding activity of the CRISPR-associated protein Csx1 of Pyrococcus furiosus. Proteins, 81, 261–270. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Pubmed Central

Regulation of the RNA and DNA nuclease activities required for Pyrococcus furiosus Type III-B CRISPR–Cas immunity

Nucleic Acids Research, Volume 48 (8) – Mar 21, 2020

Loading next page...
 
/lp/pubmed-central/regulation-of-the-rna-and-dna-nuclease-activities-required-for-fBzGYQTKnG
Publisher
Pubmed Central
Copyright
© The Author(s) 2020. Published by Oxford University Press on behalf of Nucleic Acids Research.
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/gkaa176
Publisher site
See Article on Publisher Site

Abstract

4418–4434 Nucleic Acids Research, 2020, Vol. 48, No. 8 Published online 21 March 2020 doi: 10.1093/nar/gkaa176 Regulation of the RNA and DNA nuclease activities required for Pyrococcus furiosus Type III-B CRISPR–Cas immunity 1 2 3 2 Kawanda Foster , Sabine Grusc ¨ how , Scott Bailey , Malcolm F. White and Michael 1,4,5,* P. Terns 1 2 Department of Microbiology, University of Georgia, Athens, GA 30602, USA, Biomedical Sciences Research Complex, School of Biology, University of St Andrews, St Andrews KY16 9ST, UK, Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205, USA, Department of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, USA and Department of Genetics, University of Georgia, Athens, GA 30602, USA Received December 27, 2019; Revised February 28, 2020; Editorial Decision March 05, 2020; Accepted March 19, 2020 ABSTRACT INTRODUCTION Prokaryotes often harbor powerful CRISPR–Cas Type III CRISPR–Cas prokaryotic immune systems (clustered regularly interspaced short palindromic repeat- provide anti-viral and anti-plasmid immunity via a CRISPR associated) adaptive immune systems to protect dual mechanism of RNA and DNA destruction. Upon against infections from invading viruses and plasmids (1,2). target RNA interaction, Type III crRNP effector com- CRISPR genomic arrays are composed of short DNA plexes become activated to cleave both target RNA sequences of foreign origin (called spacers), separated by (via Cas7) and target DNA (via Cas10). Moreover, host repeat sequences. CRISPR arrays become transcribed trans-acting endoribonucleases, Csx1 or Csm6, can and the long, primary transcripts are processed into short, promote the Type III immune response by destroying mature crRNAs that assemble with Cas proteins to form both invader and host RNAs. Here, we characterize crRNP (CRISPR RNA-containing ribonucleoprotein) how the RNase and DNase activities associated with effector complexes. These effector complexes detect and Type III-B immunity in Pyrococcus furiosus (Pfu)are destroy invading nucleic acids that are complementary to their crRNAs. CRISPR–Cas systems are quite diverse regulated by target RNA features and second mes- and fall into six distinct types (Types I-VI) and over 30 senger signaling events. In vivo mutational analy- subtypes (3,4). Types I, II and V (and possibly IV) target ses reveal that either the DNase activity of Cas10 the destruction of DNA (5–7), while Type VI destroys or the RNase activity of Csx1 can effectively direct RNA (8). Type III systems are particularly noteworthy in successful anti-plasmid immunity. Biochemical anal- that they uniquely degrade both RNA and DNA of the yses confirmed that the Cas10 Palm domains con- invaders (9–19). Type III systems are further categorized vert ATP into cyclic oligoadenylate (cOA) compounds into six subtypes (III-A through III-F) with the majority that activate the ribonuclease activity of Pfu Csx1. belonging to either the Type III-A (Csm) or Type III-B Furthermore, we show that the HEPN domain of the (Cmr) systems (3). adenosine-specific endoribonuclease, Pfu Csx1, de- Types III-A (Csm) and III-B crRNP (Cmr) effector com- grades cOA signaling molecules to provide an auto- plexes exhibit an overall similar subunit organization and architecture (see Figure 1A for an example of the Cmr ef- inhibitory off-switch of Csx1 activation. Activation of fector complex). Each complex is composed of a single cr- both the DNase and cOA generation activities require RNA and vfi e (Csm 1–5 for III-A) or six (Cmr 1–6 for III-B) target RNA binding and recognition of distinct target Cas proteins (15,20–25). The mature crRNAs within these RNA 3 protospacer flanking sequences. Our results complexes contain eight nucleotides of repeat sequence at highlight the complex regulatory mechanisms con- the 5 end called the 5 tag, followed by a ∼30–40 nucleotide trolling Type III CRISPR immunity. guide sequence that base-pairs with the target RNA pro- tospacer (26–28). Multiple catalytically active Cas7 super- family proteins (Csm3 or Cmr4) that act as target RNA en- To whom correspondence should be addressed. Tel: +1 706 542 1896; Fax: +1 706 542 1752; Email: mterns@uga.edu C The Author(s) 2020. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Nucleic Acids Research, 2020, Vol. 48, No. 8 4419 Cmr5 Cmr6 Cmr5 Cmr5 5’ CARF ATP Cmr4 Cmr4 Cmr4 Cmr1 Cmr4 Cmr3 crRNA 3’ Csx1 Palm Cmr2 HEPN HD RNase (D26) cOA DNase (HD) RNase (HEPN) Plate on No Selective Media Target Plasmid No CRISPR Defense PyrF (Colony Formation) Pfu Strain Plate on Selective Media or Target Plasmid CRISPR Defense No CRISPR Defense (No/Low Colony Formation) (Colony Formation) PyrF Null Cmr Cmr2-HD Cmr2-Palm Target Plasmid Csx1 No Target Plasmid Cmr2-HD /Palm m m Cmr2-HD /Csx1 Cmr2-Palm /Csx1 /Csx1-HEPN Cmr2-HD /Csx1 Cmr2-HD m C Cmr2 /Csx1 Colonies per μg of DNA Figure 1. Csx1 is required for anti-plasmid immunity in the P. furiosus (Pfu) III-B CRISPR–Cas system. (A) Components of the Pfu Cmr defense response. The Cmr effector complex is composed of Cmr1–6 and a mature crRNA (black & orange) containing a 5 tag (black) eight nucleotides in length. Cyclic oligoadenylate compounds (orange) produced by the Palm domain of Cmr2 bind to the CARF domain of Csx1. See Supplemental Figure S1 for a list of the Cas gene family names that correspond to each of the six Cmr subunits. (B) Plasmid interference assay. Pfu strains were transformed with plasmids that contain (target) or lack (no target) a transcribed target region complementary to the 7.01 crRNA (first spacer of the Pfu CRISPR 7 array). Both plasmids contain the pyrF gene to facilitate growth in the absence of uracil. (C) Colonies produced by transforming 11 different Pfu strains. Strains were transformed with the target plasmid (blue) or no target plasmid (gray). Mean colonies obtained per #2;g of plasmid DNA are plotted for two replicates. Error bars indicate standard deviation of replicates. Pfu Strain 4420 Nucleic Acids Research, 2020, Vol. 48, No. 8 doribonucleases (16,29,30), interact along the length of the vealed a key role for the short sequence that flanks the 3 crRNA guide region and these proteins also tightly asso- end of the RNA protospacer, termed the protospacer flank- ciate with Cas11 superfamily proteins (Csm2 or Cmr5). Ad- ing sequences (PFS) in controlling Type III activities. When ditional Cas7 superfamily Cas proteins directly contact the PFS sequences are complementary to the 5 crRNA tag (as 5 crRNA tag (Csm4 or Cmr3) or 3 terminus of the guide would be the case if the cell produced antisense crRNAs), RNA segment (Csm5 or Cmr1 and Cmr6). Cas10 (Csm1 or the DNase and cOA production activities fail to become ac- Cmr2) is the signature protein of Type III complexes (4). tivated (9,13,14,53,54) while the transcripts themselves are This large, multiple domain-containing protein is situated cleaved by the Csm3/Cmr4 integral RNases (16,27). Thus, near the 5 end of the crRNA and typically contains two 5 tag/PFS pairing can negatively regulate some but not all highly conserved motifs: the HD motif capable of destroy- Type III activities. In some characterized systems, the par- ing single-stranded DNA (9,10,13) and the GGDD motif of ticular identity of nucleotides within the PFS will dictate one of two Palm domains that can convert ATP into cyclic if the effector complex is active or inactive independent of oligoadenylate (cOA) second messenger molecules (31–33) their capacity to base-pair with the 5 crRNA tag. In par- (Figure 1A). ticular, three nucleotides immediately 3 of the target RNA Interestingly, Type III systems also include trans-acting protospacer (i.e. in positions +1, +2, +3 in the target RNA ribonucleases, Csm6 (III-A) or Csx1 (III-B), that are not 5 -3 direction) have been found to be important for acti- stably associated with the effector crRNP complexes (34– vating the DNase and cOA production presumably due to 37) and appear to be capable of degrading both invading interactions of these PFS elements and subunits of the Type RNA (leading to immunity) as well as host RNAs (leading III effector complex (likely Cas10 and/or Csm4/Cmr3) to cell dormancy or cell death) when activated by cOA bind- (9,25,48,54). The DNase and cOA generation activities are ing (32,35,38). Csm6 and Csx1 proteins share two highly also switched off as a result of degradation of the trigger tar- conserved domains: the HEPN (Higher Eukaryotes and get RNA that occurs when Csm3/Cmr4 RNases of the com- Prokaryotes Nucleotide binding) domain and the CARF plex cleave the RNA protospacer (33,42,48). Finally, mech- (CRISPR Associated Rossman Fold) domain (39,40)(Fig- anisms have also recently been discovered that can reverse ure 1A). The ribonuclease activity of both Csx1 and Csm6 the effects of cOA signaling on Csm6/Csx1 HEPN ribonu- is facilitated by the HEPN motif (R-X -H) within the clease activity. Specifically, dedicated ‘ring nucleases’ have 4–6 HEPN domain (34–37). Binding of cognate cOA to the been identified that bind, cleave, and inactivate cOAs by CARF domain allosterically stimulates the single-stranded converting the compounds into short, bi-adenylate degra- RNase activity of the HEPN domain of the Csm6/Csx1 dation products with 2 -3 cyclic phosphate termini (A > p) proteins (31–33,41,42). CARF domains, either within Csm6 in certain organisms (45,46). However, other organisms ap- (43,44) or as part of unrelated proteins called ring nucleases pear not to possess distinct ring nucleases and instead have (45,46), have also recently been found to cleave and inacti- been found to rely on an intrinsic ability of the CARF do- vate cOA singling molecules to switch off the activity of the mains of Csm6/Csx1 to destroy the cOA molecules (43,44). Csm6 or Csx1 HEPN RNases (43–46). Additional layers and molecular mechanisms for control- Anti-virus or anti-plasmid immunity afforded by Type III ling the activity of each of the key nucleases of Type III ef- crRNPs, is transcription-dependent as these systems specif- fector crRNPs as well as the cOA signaling pathway likely ically recognize an RNA target having a crRNA interaction await discovery. region (i.e. the RNA protospacer) (9,10,14,19,47). Once Pyrococcus furiosus (Pfu) is a hyperthermophilic ar- bound to the target RNA protospacer, the crRNP effec- chaeon that contains three distinct functional CRISPR– tor complex changes its conformation (24,25,48–50) and the Cas effector crRNPs: two DNA-targeting systems (Types I- Type III defense response becomes activated to: (i) specif- A (Csa) and I-B (formerly known as I-G) (Cst) (55–59)) and ically cleave the target RNA at regular six-nucleotide in- a Type III-B (Cmr) system (Supplemental Figure S1) shown tervals within the protospacer region using multiple copies to destroy both DNA and RNA targets in a transcription- of the Csm3/Cmr4 integral ribonuclease via an active site coupled manner (9,11,17,27). Previous investigations on the that relies on a key aspartate residue (16,29,30), (ii) non- mechanism of action of Pfu Cmr crRNPs also revealed that specifically degrade nearby invading single-stranded DNA crRNA-mediated target RNA interaction was required for via the HD motif of Cas10 (Csm1/Cmr2) (9,10,13,14), (iii) both target RNA cleavage (by Cmr4) (11,17,29) and non- generate cyclic oligoadenylate (cOA) from ATP using the specific DNase activity (by HD domain of Cmr2) which was GGDD motif of the conserved Palm domain of Cas10 shown to cleave single-stranded DNA substrates or both (Csm1/Cmr2) and (iv) non-specifically degrade single- strands of short double-stranded DNA substrates (9). The stranded RNA via the cOA-activated and trans-acting target RNA PFS requirements for controlling immunity by Csm6/Csx1 HEPN ribonuclease (31–33,41,42,51,52). Col- Pfu Cmr crRNPs have been thoroughly investigated. Simi- lectively, these target-RNA and transcription-coupled reac- lar to other studied Type III systems, PFS sequences capa- tions provide robust Type III-mediated immunity. ble of base-pairing with the 5 tag sequence of the crRNA Type III RNase and DNase activities must be tightly prevented DNase activity but not target RNA destruction regulated so that undesirable cellular toxicity or host cell (9,11). In contrast, the identity of three nucleotides within death can be prevented before or during an immune re- the PFS immediately 3 of the target RNA protospacer sponse. The molecular details for how each of the RNases, (previously referred to as the rPAM [RNA protospacer- DNase and cOA signal generation activities of Type III sys- adjacent motif] (9)), was critical for DNase activity in vitro tems become specifically activated by target RNA binding and anti-plasmid immunity in vivo, but not target RNA is currently not fully understood. Several studies have re- cleavage (9). Whether or not Pfu crRNPs are capable of Nucleic Acids Research, 2020, Vol. 48, No. 8 4421 producing cOA second messenger compounds had not yet with 0.5 mM iso-propyl-β-D-thiogalactopyranoside (IPTG) been tested. Furthermore, deletion of the csx1 gene in Pfu at 24 C overnight. Cells were pelleted then resuspended in did not disrupt anti-plasmid immunity (9). These findings lysis buffer (40 mM Tris–HCl (pH 7.5), 500 mM NaCl, raised the key question as to whether cOA-mediated acti- 10 mM Imidazole) containing one protease inhibitor tablet vation of Csx1 RNase activity is important for conferring (Roche) and lysed via sonication. Cell lysates underwent a immunity by the Pfu Cmr crRNP effector complex. In this thermal precipitation by incubating in a 75 Cbeadbath study, we expand our understanding of the molecular mech- for 20 minutes. Insoluble material was removed by cen- anisms of action of the well-characterized Type III-B (Cmr) trifugation at 14,000 rpm for 20 minutes at 4 C and fil- system of Pyrococcus furiosus (Pfu) and further define how tered through a 0.8 #2;M syringe filter (Corning Incorpo- the Csx1 RNase and Cmr2 DNase activities are regulated rated). Proteins were purified using gravity affinity chro- by target RNA elements and second messenger signaling matography and either Ni-NTA resin (Cmr2, Cmr3, Cmr4, events. Cmr5, Cmr6, Csx1; Thermo Scientific) or Talon Cobalt resin (Cmr1–1; Clontech). Cell lysates were rotated with pre-rinsed and equilibrated resin for 1 h at 4 C. The pro- MATERIALS AND METHODS teins were then washed with the lysis buffer and wash buffer P. furiosus strains and growth conditions (40 mM Tris–HCl (pH 7.5), 500 mM NaCl, 20 mM Im- idazole). The proteins were then eluted using four differ- All P. furiosus strains utilized in this study are listed in Sup- ent elution buffers containing increasing amounts of imida- plemental Table S1. Strains were grown at 90 C under strict zole (50, 100, 200, 500 mM). Wildtype and HEPN mutant anaerobic conditions using defined medium as previously Csx1 proteins were dialyzed and underwent a second round detailed (60). Cultures were grown in 5 or 20 mL volumes of gravity affinity chromatography. Buffer exchange was and inoculated with either 1% inoculum or a single isolated performed using Slide-A-Lyzer Dialysis Cassettes (Thermo colony. Cultures were incubated for 16–24 h and plates were Scientific) in elution buffer lacking Imidazole. Protein con- incubated for 64 h. Uracil (20 #2;M) and/or 5-fluoroorotic centrations were assessed using Qubit protein concentration acid (5-FOA, 2.75 mM) were supplemented in the media assays (Invitrogen) and purity was assessed via SDS-PAGE for selection or counterselection for the pyrF marker gene. and Coomassie blue staining analysis. P. furiosus strains were produced using homologous re- combination of transformed SOE-PCR (splicing by over- lap extension-polymerase chain reaction) constructs as pre- Csx1 mutagenesis viously reported (9). Complementation strains were gen- The wildtype gene encoding Csx1 was subcloned into a erated by adding a modified wildtype cmr2 or csx1 gene modified pET24D vector. The Csx1 HEPN mutant (Csx1- back onto the P. furiosus genome that contained restriction HEPN ) contains a H436A mutation and was created as sites that enabled detection of the introduced genes from previously described (37). CARF domain mutants were cre- wildtype genes and did not affect the protein coding po- ated using the wildtype vector and either inverse PCR or tential. A BclI-HF (5 -TGATAA-3 → 5 -TGATCA-3 )re- Quikchange site-directed mutagenesis (Stratagene). Muta- striction site was introduced into cmr2 and a NruI-HF (5 - genesis primers are provided in Supplemental Table S3. TCGGGA-3 → 5 -TCGCGA-3 ) restriction site was intro- Amino acid residues 121–127 were deleted from Csx1 us- duced into csx1. Primers used to make the complementa- ing inverse PCR to create a mutant form of the protein pre- tion strains can be found in Supplemental Table S3. All dicted to not be able to organize a functional CARF motif strains underwent at least three rounds of strain purification (Csx1-CARF ). Quikchange mutagenesis was used to cre- using minus uracil selective media. Successful strain gener- ate site-specific mutations within the predicted cOA binding ation was confirmed via PCR amplification of the Pfu gene pocket of the CARF domain of Pfu Csx1. Two cOA bind- of interest and DNA sequencing (Eurofins Genomics). ing mutants were created: Csx1-INAA (I169A, N170A) and Csx1-INQQ (I169Q, N170Q). Successful mutagenesis for Recombinant protein expression and purification all mutant plasmids was confirmed via DNA sequencing The genes encoding P. furiosus Cmr1–6 and Csx1 pro- (Eurofins Genomics). teins were amplified via PCR and cloned into modified ver- Purification of Csx1-HEPN was performed as described sions of pET24-D (Cmr4, Cmr5, Csx1), pET101-D (Cmr1– above. Purification of wildtype, Csx1-CARF , Csx1-INAA 1) and pET200-D (Cmr2, Cmr3, Cmr6) as previously de- and Csx1-INQQ proteins was performed using a batch tailed (17,37). All constructs contain a 6x-histidine tag method. The batch method involved affinity purifying the on either the N-terminus (Cmr2–6, Csx1) or C-terminus proteins by incubating the soluble lysate with Ni-NTA resin (Cmr1–1) of the corresponding protein. Recombinant pro- for 1 h at 4 C then performing subsequent washes and elu- tein expressions were performed in E. coli BL21-RIPL cells tions in a 15 ml centrifuge tube and spinning at 4000 rpm for (DE3, Novagen). Expression cultures for wildtype and mu- 2 mins in between each step. Protein concentrations were as- tant proteins were grown in 1 l (Cmr1–1, Cmr4, Cmr5, sessed using Qubit assays and purity was assessed via SDS- Csx1), 2 l (Cmr2, Cmr3) or 4 l (Cmr6) cultures at 37 C. PAGE and Coomassie blue staining analysis. Luria broth (Cmr1–1, Cmr2–Cmr5, Csx1; Research Prod- ucts International (RPI)) or Terrific broth (Cmr6; RPI) Preparation of RNA and DNA substrates medium was used for cultures and supplemented with ei- ther 50 #2;g/ml kanamycin sulfate (Cmr2-Cmr6, Csx1) or Synthetic RNAs (7.01 crRNA and 7.01 Target RNA) were 100 #2;g/ml ampicillin (Cmr1–1) for plasmid selection. Cul- purchased from Integrated DNA Technologies and DNAs tures were grown to an OD of 0.7 at 37 C then induced from Eurofins Genomics. The RNA and DNA sequences 600 4422 Nucleic Acids Research, 2020, Vol. 48, No. 8 can be found in Supplemental Table S3. Synthetic target added to each reaction and incubated for 30 min at 37 C RNA was 5 -end labeled using #3;- P-ATP, gel purified, prior to gel electrophoresis. Cyclic oligoadenylate produc- eluted, extracted, and precipitated as previously described tion assays were performed by assembling the Cmr crRNPs (37). 7.01 crRNA was gel purified, eluted, extracted, and as described above using RNA assay buffer and 50 nM of precipitated prior to using in assays. 5 -end labeled, double- 7.01 crRNA. After assembly, 0.5 mM of ATP (NEB), 5 nM stranded DNA substrate was prepared by annealing com- of #4;- P-ATP (3000 Ci/mmol; Perkin Elmer), and 100 nM plementary DNA oligonucleotides as previously described of target RNA was added to the reaction and incubated (37) and gel purification. for 1 h at 70 C. Unless otherwise indicated, cOA produc- Target RNAs with the indicated PFS sequences were cre- tion assays were completed with target RNA containing a ated by in vitro transcription using T7 RNA polymerase 5 -GGG-3 PFS sequence. All Reactions were stopped by and the MEGAshortscript T7 kit (Invitrogen) as described adding Gel Loading Buffer II (Life Technologies) and vi- (9). DNA templates with a T7 phage promoter sequence sualized by using 7M urea denaturing 15% polyacrylamide were generated by amplifying a target plasmid listed in Sup- gels followed by autoradiography. Cyclic oligoadenylate re- plemental Table S2 with IVT primers listed in Supplemen- actions were also ran on 8M urea denaturing 20% poly- tal Table S3 and gel purified using the Zymoclean Gel Re- acrylamide sequencing gels. Decade Markers (Life Tech- covery Kit (Zymo Research). Following synthesis, target nologies) and partial alkaline hydrolysis ladders (Ambion) RNAs were subsequently gel purified from denaturing gels, of poly A RNA were generated as previously described eluted, extracted, and precipitated prior to adding to the as- (34,37). says. Target RNA concentrations were determined using the Qubit RNA BR Assay Kit (Invitrogen) and quality was as- Csx1 In vitro activation assays sessed using 7M urea denaturing 15% polyacrylamide gels and ethidium bromide staining (Supplemental Figure S2). Activation with the native Cmr produced cOA. Csx1 RNase activity assays were performed as previously reported (37). Plasmid interference assay Ribonuclease activity of Csx1 was assessed by incubating 500 nM of Csx1 with 0.5–1.5 nM of radiolabeled target Plasmid transformation interference assays were performed RNA in assay buffer (20 mM Tris–HCl (pH 7.5), 200 mM as previously described (9). Liquid cultures of P. furio- NaCl) for 1 h at 70 C. In order to assess activation of Csx1, sus strains were allowed to reach mid-to-late log phase 10 or 20 nM of Csx1 was incubated with radiolabeled target of growth. 100 #2;l of liquid culture was transformed with RNA and assay buffer in the presence of unlabeled cOA for 1 ng of either Target plasmid (pJE65; containing a tran- 1h70 C. Unlabeled native cOA was generated as described scribed protospacer matching the 7.01 crRNA) or No Tar- above by omitting #4;- P-ATP. The unlabeled cOA was then get (pJE47) control plasmid. The transformations were in- extracted similarly to published methods (61). Five reac- cubated for 15–45 min at room temperature prior to plat- tion volumes of phenol/chloroform/isoamyl alcohol (PCI, ing. Each transformation mixture was split between two 125:24:1 at pH 4.5; Ambion) was added to the reaction and plates and spread onto solid defined media lacking uracil. vortexed for 30 seconds. The mixture was then centrifuged The plates were incubated at 90 C in an anaerobic chamber. at 20 000 rpm at 4 C then the aqueous layer incubated with Plates were observed for colony growth and counted after vfi e reaction volumes of chloroform (Fisher Scientific) vor- 64 h of incubation. Results shown represent two replicates. texed and centrifuged. The aqueous layer was extracted, aliquoted, and stored at −80 C in single use aliquots. All Cmr crRNP in vitro activity assays reactions were stopped by adding Gel Loading Buffer II and visualized by using 7M urea denaturing 15% polyacry- RNA and DNA nuclease activity assays were performed similarly to methods previously described (9,17). Purified lamide gels followed by autoradiography. recombinant Cmr proteins were first incubated with 7.01 crRNA to form crRNPs. For RNase assays, Cmr crRNPs Activation with synthetic cA and cA . Csx1 ribonuclease 4 6 were assembled by preincubating 500 nM of each Cmr pro- activation assays with cA and cA species was performed 4 6 tein (50 nM of Cmr2) with RNA assay buffer (20 mM Tris– as described above except synthetic cOAs from BIOLOG HCl (pH 7.5), 250 mM NaCl, 1.5 mM MgCl ), and 12.5 Life Science Institute were used. Several concentrations of nM of 7.01 crRNA for 25 min at 70 C. After the prein- cA and cA were tested as indicated in the figure legend. 4 6 cubation, 0.5–1.5 nM of radiolabeled synthetic 7.01 target RNA was added to the reaction and incubated for one hour Activation with Csx1-treated cOA. Cmr crRNP-generated at 70 C. One unit of Proteinase K (NEB) was then added cOA was incubated with or without 600 nM of wildtype ◦ ◦ to each reaction and incubated for 30 min at 37 C prior Csx1 for 30 min at 70 C. Reaction products were then ex- to gel electrophoresis. For DNase assays, the Cmr crRNP tracted to deproteinize the samples (as described above) and was assembled in DNA assay buffer (20 mM Tris–HCl (pH incubated with 20 nM of Csx1, 5 -end labeled target RNA, 7.5), 250 mM NaCl, 1.5 mM MgCl , 200 #2;M NiCl )and and assay buffer for 1 h at 70 C. 2 2 50 nM of 7.01 crRNA for 25 min at 70 C. After the prein- cubation, 100 nM of 7.01 target RNA and 1 nM of radi- Mass spectrometry olabeled dsDNA was added to the reaction and incubated for one hour at 70 C. Unless otherwise indicated, DNase Unlabeled cOA production assays were performed as de- assays were completed with target RNA containing a 5 - scribed above except the reactions were incubated for 2 h. GGG-3 PFS sequence. One unit of Proteinase K was then Liquid chromatography high resolution mass spectrometry Nucleic Acids Research, 2020, Vol. 48, No. 8 4423 (LC-HRMS) analysis was performed on a Thermo Scien- ity of Csx1 (via HEPN motif and activated by Cmr2-Palm tific Velos Pro instrument equipped with HESI source and GGDD motif) are each sufficient for conferring highly ef- Dionex UltiMate 3000 chromatography system as previ- fective anti-plasmid immunity in Pfu. ously described (41). RESULTS P. furiosus Cmr crRNPs produce cyclic oligoadenylate second Effective anti-plasmid immunity is achieved by either Csx1 messengers RNase or Cas10 DNase activity Next, we addressed whether the Pfu Cmr system functioned through generating cOA signaling molecules as has been Distinct Type III-A or III-B systems have shown consider- observed for other bacterial and archaeal Type III systems able variability in the need for Cas10 (Csm1/Cmr2) DNase (31–33). In vitro reconstituted Cmr crRNPs were assayed and/or Csm6/Csx1 RNase activities for anti-plasmid and for their ability to generate cOA compounds as well as to anti-viral immunity (9,32,34,35,38,47,62). In our earlier in support previously observed RNase and DNase activities vivo work with the Pfu Cmr (III-B) system, we found that (Figure 2)(9,17). Wildtype as well as four different func- individual mutations of either the HD (H13A/D14A) or tional mutants of the Cmr crRNP complex (Cmr2-HD , Palm domain GGDD (D673A/D674A) motifs of Cmr2 Cmr2-Palm , Cmr4-D26N and Cmr2-HD /Palm ), were (Cmr2-HD or Cmr2-Palm mutants) or a single dele- m m m m m tion of the csx1 gene (Csx1 ), did not interfere with anti- assembled in vitro (Figure 2A) and tested. As expected, tar- plasmid immunity but double mutations in the Cmr2 HD get RNAse activity was observed for all Cmr crRNP com- and GGDD Palm motifs (Cmr2-HD /Palm )prevented plexes except those harboring a mutation in the Cmr4 sub- m m immunity (9) (see Figure 1A for overview of the crRNA unit (Cmr4-D26N) (Figure 2B). Moreover, DNase activity and Cas protein components). This early work was per- was only observed for wildtype, Cmr2-Palm , and Cmr4- formed prior to knowledge that the Palm domain of some D26N complexes but not Cmr2 mutants in which the HD Cas10 superfamily proteins can catalyze conversion of ATP motif was mutated (Figure 2C). To test for cOA produc- to cOA signaling molecules that activate the ribonuclease tion, the same vfi e complexes were incubated with #4;- P- activity (HEPN domain) of Csm6/Csx1 HEPN ribonucle- ATP and the products of the reactions were separated by ases (31–33,41,42,51,52). This new information motivated denaturing polyacrylamide gel electrophoresis. Conversion us to perform a more systematic in vivo mutational analy- of the #4;- P-ATP to slower migrating products indicative of cOA compounds was only observed for complexes with an ses in which specific combinations of double mutants were intact Cmr2 Palm GGDD motif. Additionally, the presence tested to more fully address if the Cmr2 DNase and Csx1 of the target RNA was required for cOA production (Fig- RNase activities were important for anti-plasmid immunity ure 2D). These findings reveal that the Pfu Type III-B Cmr (Figure 1). system possesses the highly conserved activity of cOA gen- Anti-plasmid immunity was assayed in vivo by transform- eration which is catalyzed by the Cmr2 Palm GGDD motif ing a Pfu strain containing wildtype or mutant versions of and is dependent upon interactions between crRNPs and the Cmr system with a target plasmid that harbors a tran- complementary target RNA. scribed protospacer matching an endogenous crRNA (7.01; the first crRNA from CRISPR locus 7) or empty plasmid control (Figure 1B). As we observed previously, the anti- plasmid immunity observed with wildtype Cmr was unaf- Pfu Csx1 ribonuclease activity is activated by cyclic oligoad- fected when Cmr effector complexes contained mutations enylate produced by the Cmr complex in either the Cmr2 DNase HD active site (Cmr2-HD ) or Palm GGDD motif (Cmr2-Palm ), or if the csx1 gene Next we sought to determine if the cyclic oligoadenylate was deleted from the genome (Csx1 )(Figure 1C and (9)). compounds produced by Pfu Cmr crRNPs, are capable of As expected, immunity was absent for a strain that lacked stimulating the ribonuclease activity of Pfu Csx1 in vitro the entire Cmr complex and Csx1 (null strain) or when (Figure 2E). Previously, we found that recombinant Pfu the Cmr2-HD /Palm double mutant was re-tested. Im- Csx1 is capable of cleaving RNA substrates via the HEPN m m munity was disrupted when the Pfu csx1 gene was deleted motif but only when high concentrations (e.g. 500 nM) of or contained a mutation within the RNase catalytic mo- the Csx1 enzyme is used (37). Therefore, we used levels of tif of csx1 (Csx1-HEPN ; H436A) in conjunction with a the Csx1 protein (20 nM) that showed no or low RNase ca- second mutation within the DNase catalytic site of Cmr2 pacity to test if cOA produced by Cmr crRNPs could stimu- (Cmr2-HD mutation) (Figure 1C and (9)). These same late the RNase activity of Csx1. ATP (unlabeled) was incu- csx1 mutations did not prevent immunity when combined bated with wildtype and mutant Cmr crRNPs and the prod- with a Cmr2-Palm mutation (GGDD; predicted to block ucts of the reactions were added to reactions containing cOA generation). Cmr-mediated immunity was rescued by Csx1 and radiolabeled substrate RNA (Figure 2E). Cleav- restoring either wildtype cmr2 or csx1 genes in the Cmr2- age of the radiolabeled substrate RNA by Csx1 depended HD /Csx1 double mutant strains (Cmr2c and Csx1c are upon Cmr crRNP complexes having an intact Cmr2-Palm Cmr2-HD /Csx1 strains complemented with wt cmr2 or GGDD motif (Figure 2D and E). The results indicate that wt csx1, respectively. Figure 1C). Taken together, the results Cmr2 Palm domain-mediated cOA production by the Cmr show that both the DNase activity of the Cmr effector cr- crRNP complex is a potent activator of the RNA degrada- RNPs (via HD domain of Cmr2) as well as the RNase activ- tion capacity of Pfu Csx1. 4424 Nucleic Acids Research, 2020, Vol. 48, No. 8 Cmr crRNP Cmr2 Cmr6 Cmr1-1 Cmr3 Cmr4 Cmr5 ssRNA dsDNA BC : : M Cmr crRNP M Cmr crRNP * 20 cOA Production D ssRNA : Rxn Product Cmr crRNP : Csx1 M ++ + + + + ++ + + : MM2 + Target RNA 20 20 - cOA ATP Figure 2. The ribonuclease activity of Pfu Csx1 is activated by cOA species produced by the Pfu effector complex in a Cmr2 Palm domain-dependent manner. (A) Purification of Pfu complex proteins. HD and Palm mutations are located within Cmr2. The D26N mutation is located within Cmr4. (B) RNase activity of each Pfu Cmr complex. Radiolabeled 7.01 Target RNA was incubated with Cmr complexes and reaction products were visualized by urea-PAGE. Radiolabeled RNA size standards (M) in nucleotides were used. The black arrow indicates the full-length substrate and the black asterisks indicate cleavage products. (C) DNase activity of each Pfu Cmr complex. Each complex was incubated with dsDNA (label located on DNA Target 2) and reactions were visualized as described in part B. (D) cOA production activity of the Pfu Cmr crRNP. The vfi e Cmr complexes were tested for cOA production in presence (+) and absence (–) of 7.01 Target RNA. Reaction products were run on a denaturing gel. A radiolabeled alkaline hydrolysis ladder (M2) and size standard (M) were added to the urea-PAGE analysis. (E) Activation of wildtype Csx1 by Pfu cOAs. Csx1 was tested for ribonuclease activity under activating (20 nM Csx1) conditions. Reactions were visualized as described in part B. - - - - Cmr2-HDm Cmr2-Palmm Cmr2-HDm HDm/Palmm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm Cmr2-HDm Cmr2-Palmm HDm/Palmm WT WT Cmr4-D26N Cmr4-D26N WT WT Cmr4-D26N WT Cmr4-D26N WT Cmr4-D26N Nucleic Acids Research, 2020, Vol. 48, No. 8 4425 Pfu Csx1 is activated by cA species produced by Cmr com- served that, when paired with a Cmr2-HD mutation that plexes disrupts DNA cleavage, the Csx1-CARF and Csx1-INQQ mutants both led to a loss of anti-plasmid immunity (Figure The identity of the compounds generated after addition 4F). As expected, wild type Csx1 but not the Csx1-HEPN of ATP to wildtype Cmr complexes (Figure 2D) was de- supported anti-plasmid immunity. Collectively, the in vitro termined by mass spectrometry using established meth- and in vivo results support a key role for the CARF domain ods (41). We found that the Pfu Cmr complex primarily of Pfu Csx1 in triggering cOA-stimulated RNase activity produces cyclic-triadenylate (cA ) and cyclic-tetraadenylate and reveal important CARF domain residues responsible (cA ) species (Figure 3A). The data reveal that cA is the 4 4 for mediating cOA-triggered, Pfu Csx1 RNase activation. most abundant species and is approximately twice as abun- dant as cA ; there were traces of cA and cA which made 3 5 6 Pfu Csx1 cleaves and inactivates cOA using its adenosine- up less than 5% of the total cOA species. To address which specific HEPN RNase active site form(s) of the generated cOA leads to the activation of Csx1 ribonuclease activity, varying concentrations of commer- Once cOAs are produced by Type III crRNPs, it is not well cially available synthetic cA and cA compounds were in- understood how cOA levels are controlled to prevent un- 4 6 cubated with Pfu Csx1 in the presence of 5 -end labeled necessary destruction of vital cellular RNAs by trans-acting substrate RNA (cA and cA compounds were not com- Csx1 (or related Csm6) that could lead to host cell toxicity 3 5 mercially available and so could not be tested). cA but not or death during the immune response. Recently, a new class cA stimulated the ribonuclease activity of Csx1 (Figure 3). of CARF-domain containing proteins called ring nucleases, These results indicate that cA is both the dominant cA were discovered in Sulfolobus solfataricus and found to ex- species produced by Pfu Cmr effector complexes and a po- hibit cOA nuclease activity that halts cOA-triggered Csx1 tent activator of Pfu Csx1 RNA cleavage activity. activity (45). S. solfataricus Csx1 itself was not able to de- grade cOA molecules (45). In contrast, Csm6 of Thermo- The CARF domain of Pfu Csx1 is needed for cOA ribonucle- coccus onnurineus exhibited an intrinsic ability to utilize its ase activation CARF domain to both bind and cleave cOA which gen- erates inactive, linear di-adenylate products with 2 ,3 cyclic The X-ray structure of Pfu Csx1 has been solved (63) and re- phosphate termini (A >p) (44). Previous characterization veals two major domains that are conserved amongst Type of Pfu Csx1 revealed an adenosine specificity for endori- III-affiliated Csx1 or Csm6 RNases: the HEPN (RNase ac- bonuclease activity conferred by the HEPN ribonuclease tive site) and CARF domains (shown in other systems to motif (37). That knowledge combined with the lack of a selectively bind cOA for allosteric activation of the HEPN CARF domain-containing ring nuclease homologs in P. fu- RNase activity) (Figure 4A). Previously, we found that mu- riosus, led us determine if Pfu Csx1 was capable of control- tation to the HEPN motif (Csx1-HEPN ; H436A) dis- ling its own ribonuclease activity by recognizing and de- rupted the ribonuclease activity of Pfu Csx1 (37). To investi- grading cOA (Figure 5). Indeed, we found that wildtype gate if the CARF domain is required for the cOA-stimulated Csx1 protein efficiently converted native cOA substrates RNase activity of Pfu Csx1, we examined the effects of into products with relative mobilities consistent with inac- Csx1 mutants predicted to disrupt CARF function and tive linear di-adenylate (A > p) products (Figure 5Aand cOA binding (Figure 4A and B). Csx1-CARF contains (43,45)). The conversion of cOA into the presumed A >p a deletion of residues 121–127 predicted to be critical for products was abolished by mutations in the HEPN do- CARF domain assembly. The Csx1-INAA and Csx1-INQQ main but not mutations in the CARF domain (Figure 5A). mutants are predicted to prevent specific binding of cOA Moreover, we found that the Csx1-mediated cOA cleav- and contain a double mutation of residues Isoleucine169 age products failed to activate Csx1 ribonuclease activity and Asparagine170 to alanine or glutamine (Figure 4A (Figure 5B). The results indicate the adenosine-specific Pfu and Supplemental Figure S3). We previously showed that Csx1 endoribonuclease responsible for target RNA destruc- high (500 nM) concentrations of wildtype Csx1 cleave RNA tion (37), also utilizes its HEPN active site to recognize, without cOA activation (37). We further analyzed Pfu Csx1 cleave and inactivate cOA molecules providing an autoreg- activity and determined that lowering the concentration to ulation negative feedback control mechanism that limits 10–20 nM resulted in a loss of detectable RNase activity Csx1 RNase activity. by Csx1 in the absence of cOA (Figure 4). All three Csx1 CARF mutants disrupted the ability of low levels of Csx1 Activation of Cmr2-mediated DNA nuclease activity and (10 nM) to support RNA cleavage activity in response to cOA production is dependent upon distinct target PFS ele- cOA (Figure 4C and D) as is observed for the wildtype Csx1 ments enzyme (Figure 4C). None of the CARF mutants impaired the ability of Csx1 to cleave RNAs when high amounts of Prior investigation into regulatory mechanisms controlling enzyme (500 nM) were assayed showing that the mutations the function of Pfu Cmr complexes revealed the important per se did not negatively impact HEPN functionality (Fig- role of three nucleotides within the PFS immediately ad- ure 4C and D). In contrast, the Csx1-HEPN mutant failed jacent to the crRNA/target RNA protospacer interaction to efficiently cleave RNA at either low (10 nM) or high (500 (see Figure 6A) (9). Successful anti-plasmid immunity in nM) concentrations and with or without cOA addition (Fig- vivo and Cmr2-mediated DNA cleavage (via the HD do- ure 4E). main) in vitro required a PFS with a NGN, NNG, or NAA The impact of CARF domain mutations was further sequence (9) and Table 1. In contrast, target RNA cleav- tested in vivo on anti-plasmid immunity (Figure 4F). We ob- age by Pfu crRNPs (via Cmr4 backbone subunit) occurs 4426 Nucleic Acids Research, 2020, Vol. 48, No. 8 Figure 3. cA is the relevant activator for Pfu Csx1. (A) UV chromatogram (258 nm) and MS chromatogram (extracted ion chromatogram for m/z 494.6, 659.1, 823.6, 988.2) for cA samples from Pfu Cmr wildtype complexes. (B) Pfu Csx1 activation assays with synthetic cOA. Wildtype Csx1 (20 nM) was incubated with increasing concentrations of synthetic cA and cA . For each activator, 1.85, 18.5 and 185 pM concentrations were used. A radiolabeled 4 6 alkaline RNA size marker (M) was added to the urea-PAGE analysis. The black arrow indicates the full-length RNA substrate and the black asterisks indicate cleavage products. Dotted lines are indicative of noncontiguous data being omitted from the gel. Table 1. Summary of PFS requirements for Pfu Cmr activities independent of the PFS (9,11). Given the newly observed cOA generation activity for Pfu Cmr complexes revealed in CRISPR defense DNase cOA this study (Figure 2D), we examined whether the first three Target RNA PFS (5 -3 ) (in vivo) (in vitro) (in vitro) positions of the PFS were important for cOA production. NNN GGG + + + Specifically, we addressed whether a large panel of target AAA + + − RNAs that differed only in having distinct PFS elements CCC −− − could activate DNA cleavage (Figure 6B) or cOA produc- UUU −− − NGG GGG + + + tion (Figure 6C) in vitro in parallel reactions. As expected, AGG + + + only target RNAs containing a NGN, NNG, or NAA pro- CGG + + + tospacer flanking sequence activated DNA cleavage (Figure UGG + + + 6B). In contrast, a much smaller subset of the same target GGN GGG + + + GGA + + + RNAs activated cOA production (Figure 6C). The target GGC + + (+) RNAs eliciting strong cOA production activity contain a GGU + + − PFS consensus of NGR sequence. Weak cOA generation ac- UNG UGG + + + tivity was observed for target RNAs containing a GGC or UAG + + (+) UAG sequence within the first three positions of the PFS. A UCG + + − UUG + + − summary of the results for all PFS elements tested on either NUG GUG + + − DNase or cOA generation activities are provided in Table 1. AUG + + − The results reveal a difference in specificity for PFS elements CUG + + − needed for activating cOA generation vs. DNase activity for UUG + + − NGU GGU + + − the Pfu Cmr crRNP. AGU + + − CGU + + − UGU + + − DISCUSSION UNU UGU + + − UAU −− − The Pyrococcus furious Type III-B (Cmr) effector crRNP UCU −− − was the first example of a CRISPR system that identifies UUU −− − and pairs with RNA transcripts rather than DNA strands YUY CUU −− − of invaders (17). A decade of subsequent in vivo and in vitro UUC −− − research has revealed the detailed structure and organiza- Recap of PFS dependent CRISPR defense, DNase activity, and cOA pro- tion of the Pfu Cmr effector crRNPs (20,21,30,64,65)and duction activity results observed for target RNAs used in this study and determined that the system employs highly versatile strate- previously reported (9). + indicates the activity was observed, – denotes gies to combat invading mobile genetic elements (Figure the activity was not observed, and (+) indicates the activity was weakly 7). Previous work showed that immunity provided by the observed. Nucleic Acids Research, 2020, Vol. 48, No. 8 4427 Figure 4. Activation of Pfu Csx1 is CARF domain dependent. (A) Ribbon structure of Pfu Csx1 (PDB: 4EOG) indicating the CARF and HEPN domains. Residue H436 was mutated in the HEPN domain (green). Residues 121–127 were deleted from the CARF domain (blue), and residues I169 and N170 were mutated within the CARF domain (red). (B) SDS-PAGE analysis of Pfu Csx1 protein purifications. WT refers to protein purified using a column column method and WT refers to protein purified using a batch method. ( C–E) In vitro activation of Csx1 mutants. Wildtype Csx1 along with four batch Csx1 mutants (CARF , INAA, INQQ, HEPN ) was tested for ribonuclease activity under high (500 nM Csx1) and low (10 nM Csx1) concentration m m conditions. A radiolabeled alkaline RNA size marker (M) was added to the urea-PAGE analysis. The black arrow indicates the full length RNA substrate and the black asterisks indicate cleavage products. (F) In vivo plasmid silencing assay results for CARF domain mutants. Strains were transformed with the target plasmid (blue) or no target plasmid (gray). Mean colonies obtained per #2;g of plasmid DNA are plotted for two replicates. Error bars indicate standard deviation of replicates. 4428 Nucleic Acids Research, 2020, Vol. 48, No. 8 mune response by activating the trans-acting Csx1 ribonu- clease via their CARF domains (Figures 2-4). Furthermore, the Csx1 HEPN RNase active site, responsible for target RNA destruction via cleaving after adenosine residues (37), also degrades its cognate cyclic-tetra-AMP (cA )activator to switch off the signaling pathway and to limit the activ- ity of Csx1 through autoregulation (Figures 5 and 7). Our in vivo mutational analyses show that either the DNase ac- tivity (via HD domain of Cmr2) or RNase activity (via HEPN motif of Csx1 and triggered by cOA generation by the GGDD motif of Cmr2 Palm domain using ATP pre- cursors) leads to robust anti-plasmid immunity. Our finding that the DNase and cOA generation capacities are activated by distinct 3 protospacer flanking sequences (PFS) of the target RNA highlight that the two target-RNA stimulated enzymatic processes are differentially regulated by distinct allosteric control mechanisms. Taken together with results obtained from other investigated bacterial or archaeal Type III-B (Cmr) and III-A (Csm) systems, our results contribute to making a compelling case for Type III systems as the most complex and highly regulated CRISPR systems dis- covered thus far. Role of trans-acting Csx1 endoribonuclease in conferring anti-plasmid immunity Our previous investigation into the requirement of Pfu Csx1 in anti-plasmid immunity through single csx1 gene muta- tional analysis, failed to uncover its role (9). The more com- prehensive gene mutational analysis performed here reveals that the Csx1 RNase activity or the Cmr2 DNase activity each can independently provide robust anti-plasmid immu- nity (Figure 1). We show that immunity remains unchanged by disrupting either the DNase activity of the Pfu crRNP (via mutation of the HD motif of Cmr2), the RNase ac- tivity of Csx1 (via mutation of the HEPN motif or csx1 gene deletion), or the ability to generate cOA second mes- sengers (via the Palm domain of Cmr2) needed to acti- vate Csx1 RNase activity (Figure 1). In contrast, double mutations that block both DNase and Csx1 RNase activ- Figure 5. Degradation of cOAs by Pfu Csx1 inactivates the cOA activators. (A) Sequencing gel reactions of Csx1 proteins incubated with Pfu cOAs. ity (either directly or through interfering with cOA signal- Five different Pfu Csx1 proteins were incubated with radiolabeled cOAs ing), are required to prevent immunity (Figures 1 and 4). for two timepoints-5 and 20 mins. Urea-page analysis on sequencing gels The results reveal that Cmr2 DNase and Csx1 RNase ac- was performed. Radiolabeled RNA size marker (M) and alkaline hydroly- tivities serve a redundant function in conferring highly ef- sis ladder (M) and were included in the analysis. (B) Effect of Csx1 treat- fective anti-plasmid immunity by the Type III-B Pfu Cmr ment on cOA activation activity. Unlabeled Pfu cOAs were treated with 600 nM of wildtype Csx1 or no protein for 30 min. The plus extraction set system. In comparison, other studied Type III-A or III- of reactions were tested for activation of wildtype Csx1 by incubating 20 B systems have shown a dominant role for Csm6/Csx1 nM of wildtype Csx1 with decreasing concentrations of A >P or cOA –– 2 4 activity (32,34,41,47,62) or Cmr2/Csm1 DNase activity −1 −2 −4 roughly 1.25, 0.6, 0.6 × 10 ,0.6 × 10 , and 0.6 × 10 uM. #2; indicates (14,35,38,50,62) to execute effective immunity against plas- a putative designation for this reaction product. Native cOA without the mids (14,34,38,47,62) or phages (32,35). Type III-C, D, E 30 min incubation or extraction was utilized as a control (N). and F systems are predicted to contain Cmr2 or Csm1 (i.e. Cas10) subunits that lack functional DNase or cOA gen- Pfu Cmr crRNP effector complexes, requires that the in- erating capacities and further work on these systems is re- vasive DNA undergo transcription to produce the target quired to understand if and how they execute immunity RNA required for triggering destruction of both the target against mobile genetic elements (3,4). RNA transcript and invading genome by intrinsic Cmr cr- The exact mechanism of action of Pfu Csx1 for anti- RNP RNase (Cmr4) and DNase (Cmr2 HD domain) activi- plasmid immunity remains to be determined. For example, ties, respectively (9,11,17,27). Here we demonstrate that Pfu it is unclear if Pfu Csx1 functions through selectively de- Cmr effector crRNP complexes also produce cyclic oligoad- stroying the invading target RNA (to provide specific im- enylate second messenger compounds that amplify the im- munity) and/or by destroying host RNA transcripts (lead- Nucleic Acids Research, 2020, Vol. 48, No. 8 4429 Cmr6 Cmr5 Cmr5 Cmr5 5’ Cmr4 Cmr4 Cmr4 Cmr4 Cmr1 Cmr3 3’ crRNA Cmr2 3’ 5’ Target RNA PFS (+1,2,3) crRNA 5’-AUUGAAAGUUGUAGUAUGCGGUCCUUGCGGCUGAGAGCACUUCAG-3’ Target RNA 3’-..UAGGGAGCCAANNNAACAUCAUACGCCAGGAACGCCGACUCUCGUGAAGUCUCCUAGG..-5’ +3+2+1 PFS dsDNA PFS (5’-3’) M + ++ ++ + + + + ++ + ++ ++ + + + + ++ ++ : Cmr crRNP cOA PFS (5’-3’) + + + + + + + ++ + + + + + + + + ++ + + + + + : MM2 Cmr crRNP cOA ATP Figure 6. Pfu Cmr DNase and cOA production activities are PFS dependent. (A) Composition of the Pfu Cmr crRNP. The Cmr effector complex is composed of Cmr1–6 and a mature crRNA (black & orange) containing a 5 tag (black). The Pfu Cmr complex is activated by binding of complementary target RNA (purple) to the crRNA. The target RNA contains a PFS (red) within it that is located adjacent (3 ) to the target sequence. (B)TargetRNA dependent activation of DNase activity by the Pfu complex. Cmr crRNP complexes were incubated with radiolabeled double-stranded DNA in the presence of 23 different target RNAs. Each target RNA contains a different three nucleotide sequence in the PFS region of the target RNA. Urea-page analysis on sequencing gels was performed. Radiolabeled RNA size markers (M) were included in the analysis. The black arrow indicates the full-length RNA substrate and the black asterisks indicate cleavage products. (C) Target RNA dependent activation of cOA production activity by the Pfu complex using the same target RNAs mentioned in part B. Products were analyzed as described in part B. An alkaline hydrolysis ladder (M2) was included in the urea-PAGE analysis. ing to cellular dormancy or death) (Figure 7). We are un- the Type III-B system of S. epidermidis was found to ei- able to distinguish these two different scenarios given that ther selectively cleave target RNA (including at regions out- inducible gene expression systems have yet to be established side of the RNA protospacer segment) while sparing cel- for Pfu that would enable testing which RNAs are being de- lular RNAs (35), or to destroy both target RNA and cel- stroyed as a function of activation of the Cmr crRNP activi- lular RNAs (38). It is conceivable that low levels of target ties over time. Given that Pfu Csx1 was shown to cleave var- RNA and second messenger signaling may limit Csx1 and ious RNAs without specificity in vitro (except that it cleaves Csm6 RNase activity to target RNA destruction, while el- after adenosine residues (37)) and that Csx1 or related Csm6 evated target RNA concentrations and high cOA signaling are not normally stably associated with Type III effector cr- may lead to more indiscriminate RNA degradation. Selec- RNPs (15,17,21–23,34), to be selective for the target RNA tive targeting of invader RNAs would promote immunity would involve conditional recruitment of Csx1 to the cr- and maintain viable host cells. In contrast, indiscriminate RNP when the complex is engaged in interaction with the RNA destruction would prevent the growth or kill infected target RNA protospacer. The related Csm6 RNase from host cells, which would effectively prevent the spread of the - - - - GGG AAA CCC UUU UGG UUG UGU CGG AGG GGU GGC GGA UCG UAG CUG CUU UCU CGU AUG UUC UAU AGU GUG GGG AAA CCC UUU UGG UUG UGU CUU UCU UUC CGG AGG GGU GGC GGA UCG UAU UAG CUG CGU AUG AGU GUG 4430 Nucleic Acids Research, 2020, Vol. 48, No. 8 Dormacy or Viral/Plasmid Immunity Cell Death Type III crRNP 5’ ATP 5’ Cmr4 crRNA CARF Palm 3’ Csx1 5’ cOA PFS HEPN Cmr2 Cellular Viral HD RNA RNA Viral DNA 5’ 3’ 3’ 3’ 5’ 3’ RNAP Figure 7. Model for regulation of the RNAse, DNase, and cOA synthetase activities of the Pfu Type III-B effector crRNP in providing anti-viral immunity. Inactive Type III-B crRNPs become activated upon interaction with expressed viral RNA. crRNA-guided base-pairing to the viral RNA and Cas protein recognition of the PFS results in conformational changes that trigger three important activities: viral RNA cleavage by Cmr4, viral DNA cleavage by the HD domain of Cmr2 (a member of the Cas10 superfamily), and ATP dependent production of cyclic oligoadenylates (cOA) by the Palm domain of Cmr2. The cOA second messengers allosterically regulate the trans-acting Csx1 enzyme by binding to the CARF domain to trigger HEPN RNase activity. Clearance of the viral DNAs and RNAs provides immunity and restores the crRNPs back to the inactive state. Under conditions of controlled immunity, Csx1 degrades and inactivates cOA signaling molecules using the HEPN domain, to provide an auto-inhibitory off-switch of Csx1 RNase activation. Inefficient degradation of cOA molecules by the Csx1 or failure to efficiently destroy the viral RNAs or viral genome may cause Csx1 to destroy both viral and cellular RNA molecules, leading to cellular dormancy or death. invader to other host cells in the local environment (Figure onnurineus (Ton), was found to be capable of autoregulat- 7). ing RNase activity through binding and cleaving cA (44). However, unlike Pfu Csx1, Ton cA cleavage of cA into in- 4 4 active linearized adenosine dinucleotide products (A >P) Second messenger signaling pathway leading to Csx1 RNase depended upon the CARF rather than HEPN domain de- activation and deactivation spite evidence from high resolution structural studies show- ing that cOA binds to both CARF and HEPN domains of We show for the first time that the Pfu Cmr complex gen- Ton Csm6 (44). Interestingly, the HEPN domains of Ton erates primarily cA signaling molecules (Figure 3A) that Csm6 and Pfu Csx1 each cleave RNA with a strict adeno- significantly stimulate the RNase activity of Csx1 (Figures sine specificity ( 37,44). Furthermore, Thermus thermophilus 2-4). Either cA or cA have been found to be the rele- 4 6 Csm6 exhibits an intrinsic ability to degrade its cognate cA vant activators for various characterized bacterial or ar- using its CARF domain but not its HEPN domain (43). chaeal Csx1 and Csm6 enzymes (31–33,41,42,44,51). Inter- Studies examining RNA cleavage patterns of various estingly, the HEPN RNase active site of Pfu Csx1 (which Csx1 and Csm6 HEPN endoribonucleases revealed that is adenosine-specific) cleaves and inactivates cOA signaling many are either specific for cleaving after adenosines ( 37,44) molecules (Figure 5) in addition to executing the destruc- or purines (adenosines and guanosines) (31,34). These cat- tion of adenosine-containing, target RNA transcripts (Fig- alytic properties indicate that other Csx1 or Csm6 enzymes ures 2–5 and (37)). Previous studies have shown that the may utilize their HEPN RNase active site to degrade cOA CARF domain is the cOA sensor domain (31,32,44,46,51) molecules. Thus, Pfu Csx1 HEPN domain-mediated de- and accordingly, mutation of the CARF domain of Pfu struction of its cognate cOA activator reveals a novel and Csx1 blocks activation of Csx1 RNase activity. However, distinct auto-catalytic control mechanism that limits Csx1 Pfu Csx1 CARF domain inactivation did not abolish cleav- RNase activity that may be common to other Type III sys- age and inactivation of cOA second messengers (Figure 5). tems. Our finding that the HEPN RNase motif, rather than CARF domain of Pfu Csx1, can cleave and inactivate cOA signaling molecules expands the number of mechanisms ca- Influence of target RNA interactions in controlling DNase pable of down-regulating Type III-induced cOA signaling and RNase activities of Pfu effector crRNPs pathways. For example, Sulfolobus islandicus Csx1 appears to lack the ability to itself degrade cOA molecules and in- All characterized Type III CRISPR–Cas systems are specif- stead utilizes extrinsic CARF domain-containing factors ically activated through identifying and interacting with (called ring nucleases) to switch off Csx1 RNase activa- target RNA. Base-pairing interactions between crRNAs tion (45). Similar to our finding with Pfu Csx1, Csm6 within crRNP effector complexes and matching target from another hyperthermophlic archaeon, Thermococcus RNA protospacer elements, has been shown to trigger Nucleic Acids Research, 2020, Vol. 48, No. 8 4431 major conformational changes within Type III complexes vating the DNase or cOA synthesis. These results have pro- (25,49,50). In turn, less understood, minor target RNA- vided insight that the two important activities for immunity induced conformational changes lead to activation of sev- are differentially controlled by distinct molecular mecha- eral enzymatic activities of the complexes including RNase, nisms. Additional evidence supporting the notion that dis- DNase and cOA generation needed for downstream ac- tinct conformational changes triggered by PFS/protein in- tivation of Csx1 and Csm6 HEPN endoribonucleases teractions separately influence either DNase or cOA pro- (25,49,50,52). duction were observed with S. thermophilus Type III-A sys- Some but not all of these enzymatic activities are further tem where Csm1 mutations at residues thought to transmit regulated by a ∼eight base-pair PFS region of the target allosteric effects of the PFS were found to impair DNase RNA that is located just 3 of the target RNA protospacer activity only, cOA synthesis only, or both activities (25). Of region. For all Type III systems evaluated thus far, includ- note, the Pfu spacer acquisition machinery responsible for ing Pfu (9), if the target RNA PFS is complementary to the addition of new spacers into CRISPR arrays, recognized the 5 tag of the crRNA (as is the case if anti-sense transcrip- invading DNAs having a 5 -NGG-3 consensus PAM (pro- tion of the CRISPR locus occurs (27,66)) then base-paring tospacer adjacent motif) (70,71). In turn, our results pre- between residues of the crRNA tag and the PFS of the tar- dict that the resultant 5 -NGG-3 RNA PFS of expressed in- get RNA apparently prevents the ability of the crRNPs to vader genomes would trigger both DNA cleavage and cOA allosterically activate DNase and cOA production (9,10,12– activities of Pfu Cmr effector crRNPs (Figure 6 and Table 14,54,67,68). In contrast, cleavage of the target RNA (at 1). A functional coupling between PAM recognition during regular 6 nucleotide intervals throughout target RNA pro- CRISPR spacer acquisition and Type III interference acti- tospacer region) by Cmr4 (III-B) or Csm4 (III-A) crRNP vation may boost the specificity of certain Type III systems backbone RNases is unaffected by crRNA 5 tag/target to compensate for an observed high degree of tolerance to RNA PFS interactions (9,11)(Figure 7). mismatches in the rest of the target RNA (69). There are apparent differences in PFS requirements for activation of DNase and cOA generation activities for dis- tinct Type III systems. For example, a simple lack of base- Role for the RNAse activity of Cmr4 subunit? pairing between 5 crRNA tag and target RNA PFS, ap- pears to be sufficient for activating these enzymatic activ- We were unable to determine the possible in vivo role of the ities for the Staphylococcus epidermidis Type III-A system RNase activity of Cmr4 on anti-plasmid immunity. Cmr4 (69). In contrast, the identity of PFS sequences, especially is the backbone subunit of the III-B crRNP (see Figure 1) within the +1, +2 and/or +3 positions have been found to be that cleaves the bound target RNA at regular six nucleotide important for conferring immunity for Type III systems of intervals (27). Despite repeated attempts, we failed to gen- Pyrococcus furiosus (9), Streptococcus thermophilus (25,48) erate a strain in which the RNase activity of Cmr4 was inac- and Thermotoga maritima (54). For most studied Type III tivated (via D26N mutation that blocks target RNA cleav- systems, the PFS requirements have not been thoroughly in- age in vitro (Figure 2)). Given that we were able to read- vestigated. ily create the catalytically inactive Cmr4 mutant in a strain Based on recent Type III crRNP structural studies, the in which both the DNase and Csx1 RNase activities were PFS sequences in the +1, +2, +3 positions are predicted to also inactivated (i.e. the Cmr2-HD /Csx1 strain that can- interact with Csm1/Cmr2 (Cas10) and Csm4/Cmr3 (5 tag not execute anti-plasmid immunity (Figure 1)), we suspect interacting) subunits (25,48,49). We propose that in systems that preventing Cmr4 RNase activity results in host cell where the identity of protospacer flanking sequences mat- lethality brought about host cell DNAase activity and/or ter for activating function such as Pfu,PFS RNA/protein cOA signaling and Csx1 RNase activation as the result of contacts play a key role in initiating the conformational generation of constitutively active Type III-B crRNPs. This changes that ultimately trigger activation of the Cas10 HD would occur if there was recognition of endogenous P. fu- domains to cleave invader DNA and the Cas10 Palm do- riousus RNA(s) exhibiting complementarity to any of the main GGGDD motif to convert ATP into cOA second mes- 200 known crRNAs that are assembled into P. furious type sengers. Interestingly, our systematic analyses of the impact III-B crRNPs (55,56). Consistent with this possibility, it has of varying the PFS elements in the +1, +2 and +3 positions been shown that cleavage of the crRNA-bound target RNA of the target RNA, revealed that a given PFS can activate by Cmr4 (and related Csm3 protein of III-A crRNPs) stim- DNase activity, cOA production, both activities or neither ulates rapid release of the target RNA into solution to tem- activity (independent of the ability of these PFS nucleotides porally control (switch off) the DNase and cOA produc- to base-pair with the 5 crRNA tag element) (Figure 6 and tion activities of the complex (9,10,13). Failure to cleave Table 1). Furthermore, we found a direct correlation be- the target RNA that is required for triggering the activi- tween PFS elements that were previously found to support ties of the complex, is expected to drive the effector crRNPs anti-plasmid immunity in vivo (9) and those found to also into a long-lasting nuclease-active state. The scenario is fur- be required for in DNase activity in vitro (Figure 6 and Ta- ther supported by findings with the S. epidermidis III-A ble 1). In contrast, only a subset of PFS elements required complex that exhibits a hyperactive immunity phenotype in in vivo, resulted in cOA production in vitro (Figure 6 and vivo when target RNA cleavage was disrupted through in- Table 1). ducible expression of the comparable catalytically inactive All Pfu PFS elements that triggered activity, shared a Csm3 mutant (14). Taken together, we view it likely that the purine-rich character within position +2 and +3 of the tar- RNase activity of P. furiousus Cmr4 normally contributes get RNA PFS appears to responsible for differentially acti- to both the destruction and turnover of the target RNA and 4432 Nucleic Acids Research, 2020, Vol. 48, No. 8 plays an important role in immunity and regulation of type and structural analyses of Pfu Type III-B crRNPs in the III-B crRNP activity. presence and absence of substrate target RNAs and DNAs and with and without Csx1, would provide important in- sight into specific molecular mechanisms that control the Contribution of DNase vs. RNase activities to type III immu- timing and specificity of Type III crRNP-mediated immu- nity nity. The ability of type III CRISPR–Cas systems to act both at the level of DNase and RNase degradation for invader im- SUPPLEMENTARY DATA munity, sets them apart from all know other CRISPR types. Moreover, it is clear that there is great variation in the rel- Supplementary Data are available at NAR Online. ative importance of these two distinct nuclease activities in conferring immunity between different type III systems as ACKNOWLEDGEMENTS well as within a system for targeting specific invaders. Here, we find that anti-plasmid immunity for the P. furiosus III-B We thank members of the Terns and White laboratories for system is robust when only employing DNase activity (me- helpful discussions and guidance. We also express gratitude diated by the HD domain of Cmr2) or RNAse activity (me- to Janet Westpheling for providing critical guidance on gen- diated by the cOA-regulated, HEPN domain of Csx1) (Fig- erating the cmr2 and csx1 complementation strains. ure 1). This is in contrast to in vivo results obtained for sev- eral other type III systems including Sulfolobus islandicus (47), S. epidermidis (62,72), S. thermophilus (72), Lactococ- FUNDING cus lactis (72) shown to rely primarily on Csx1 RNase ac- National Institutes of Health (NIH) [R35GM118160 to tivity given that plasmid immunity was abolished by single M.P.T., R01GM097330 to S.B. and 1F31GM125365 to deletion or mutational inactivation of the csx1 or csm6 gene. K.F.]; Biotechnology and Biological Sciences Research Yet other systems appear to solely utilize the DNase activity Council [REF: BB/S000313/1 to M.F.W.]. Funding for as exemplified by the recently characterized Lactobacillus open access charge: NIH grant. delbrueckii found to confer anti-plasmid immunity despite Conflict of interest statement. None declared. lacking a Csm6 gene and not being capable of producing cOA (73). Anti-plasmid immunity for the S. aureus III-B system, like P. furious, appears to depend on both DNase REFERENCES and RNase targeting but a partial interference phenotype 1. Jackson,R.N., van Erp,P.B., Sternberg,S.H. and Wiedenheft,B. (2017) is observed with csx1 gene knockout, indicating relatively Conformational regulation of CRISPR-associated nucleases. Curr. weak DNase activity (74). Finally, when anti-viral immu- Opin. Microbiol., 37, 110–119. nity was examined, the DNAse activity alone was found be 2. Hille,F., Richter,H., Wong,S.P., Bratovic,M., Ressel,S. and protective except under conditions of excess viral genome Charpentier,E. (2018) The biology of CRISPR–Cas: backward and forward. Cell, 172, 1239–1259. copy numbers where Csm6 RNase activity becomes neces- 3. Makarova,K.S., Wolf,Y.I., Iranzo,J., Shmakov,S.A., Alkhnbashi,O.S., sary to provide robust immunity (75). The natural high de- Brouns,S.J.J., Charpentier,E., Cheng,D., Haft,D.H., Horvath,P. et al. gree of functional variation amongst type III systems is also (2020) Evolutionary classification of CRISPR–Cas systems: a burst highlighted by the prediction that type III-C through III- of class 2 and derived variants. Nat. Rev. Microbiol., 18, 67–83. 4. Makarova,K.S., Wolf,Y.I., Alkhnbashi,O.S., Costa,F., Shah,S.A., F systems are predicted to encode Cas10 (Cmr2 or Csm1) Saunders,S.J., Barrangou,R., Brouns,S.J., Charpentier,E., Haft,D.H. proteins that lack the ability to cut DNA or generate cOA et al. (2015) An updated evolutionary classification of CRISPR–Cas signalling molecules (3). systems. Nat. Rev. Microbiol., 13, 722–736. In addition to these functional differences, the regulatory 5. Brouns,S.J., Jore,M.M., Lundgren,M., Westra,E.R., Slijkhuis,R.J., mechanism that govern spatiotemporal control of the activ- Snijders,A.P., Dickman,M.J., Makarova,K.S., Koonin,E.V. and van der Oost,J. (2008) Small CRISPR RNAs guide antiviral defense in ities of type III systems have also shown to be quite differ- prokaryotes. Science, 321, 960–964. ent. For example, we show here clear evidence that the P. 6. Garneau,J.E., Dupuis,M.E., Villion,M., Romero,D.A., furiousus III-B system utilizes bipartite recognition of tar- Barrangou,R., Boyaval,P., Fremaux,C., Horvath,P., Magadan,A.H. get RNA (through crRNA base-pairing and likely protein- and Moineau,S. (2010) The CRISPR/Cas bacterial immune system cleaves bacteriophage and plasmid DNA. Nature, 468, 67–71. mediated PFS interaction) in stimulating the DNase and 7. Zetsche,B., Gootenberg,J.S., Abudayyeh,O.O., Slaymaker,I.M., RNase activities of the crRNP (Figure 6). In contrast, cr- Makarova,K.S., Essletzbichler,P., Volz,S.E., Joung,J., van der Oost,J., RNA pairing alone with the target RNA appears to trig- Regev,A. et al. (2015) Cpf1 is a single RNA-guided endonuclease of a ger activity for the S. epidermidis III-A system (69). Collec- class 2 CRISPR–Cas system. Cell, 163, 759–771. tively, the high degree of functional variability argues that 8. Abudayyeh,O.O., Gootenberg,J.S., Konermann,S., Joung,J., Slaymaker,I.M., Cox,D.B., Shmakov,S., Makarova,K.S., diverse type III systems must be extensively characterized to Semenova,E., Minakhin,L. et al. (2016) C2c2 is a single-component understand the full spectrum of modes of action and regu- programmable RNA-guided RNA-targeting CRISPR effector. latory mechanisms operating in type III immunity. Science, 353, aaf5573. In summary, our findings significantly contribute to our 9. Elmore,J.R., Sheppard,N.F., Ramia,N., Deighan,T., Li,H., Terns,R.M. and Terns,M.P. (2016) Bipartite recognition of target understanding of the multiple mechanisms that ensure that RNAs activates DNA cleavage by the Type III-B CRISPR–Cas Type III CRISPR systems become appropriately activated system. Genes Dev., 30, 447–459. and then rapidly deactivated to achieve maximum benefit 10. Estrella,M.A., Kuo,F.T. and Bailey,S. (2016) RNA-activated DNA to host cells experiencing infection from invading mobile cleavage by the Type III-B CRISPR–Cas effector complex. Genes genetic elements. In future studies, a detailed biochemical Dev., 30, 460–470. Nucleic Acids Research, 2020, Vol. 48, No. 8 4433 11. Hale,C.R., Cocozaki,A., Li,H., Terns,R.M. and Terns,M.P. (2014) 31. Kazlauskiene,M., Kostiuk,G., Venclovas,C., Tamulaitis,G. and Target RNA capture and cleavage by the Cmr type III-B Siksnys,V. (2017) A cyclic oligonucleotide signaling pathway in type CRISPR–Cas effector complex. Genes Dev., 28, 2432–2443. III CRISPR–Cas systems. Science, 357, 605–609. 12. Han,W., Li,Y., Deng,L., Feng,M., Peng,W., Hallstrom,S., Zhang,J., 32. Niewoehner,O., Garcia-Doval,C., Rostol,J.T., Berk,C., Schwede,F., Peng,N., Liang,Y.X., White,M.F. et al. (2017) A type III-B Bigler,L., Hall,J., Marraffini,L.A. and Jinek,M. (2017) Type III CRISPR–Cas effector complex mediating massive target DNA CRISPR–Cas systems produce cyclic oligoadenylate second destruction. Nucleic Acids Res., 45, 1983–1993. messengers. Nature, 548, 543–548. 13. Kazlauskiene,M., Tamulaitis,G., Kostiuk,G., Venclovas,C. and 33. Rouillon,C., Athukoralage,J.S., Graham,S., Gruschow,S. and Siksnys,V. (2016) Spatiotemporal control of Type III-A CRISPR–Cas White,M.F. (2018) Control of cyclic oligoadenylate synthesis in a type immunity: coupling DNA degradation with the target RNA III CRISPR system. Elife, 7, e36734. recognition. Mol. Cell, 62, 295–306. 34. Foster,K., Kalter,J., Woodside,W., Terns,R.M. and Terns,M.P. (2019) 14. Samai,P., Pyenson,N., Jiang,W., Goldberg,G.W., Hatoum-Aslan,A. The ribonuclease activity of Csm6 is required for anti-plasmid and Marraffini,L.A. (2015) Co-transcriptional DNA and RNA immunity by Type III-A CRISPR–Cas systems. RNA Biol., 16, cleavage during Type III CRISPR–Cas immunity. Cell, 161, 449–460. 1164–1174. 35. Jiang,W., Samai,P. and Marraffini,L.A. (2016) Degradation of phage 15. Staals,R.H.J., Agari,Y., Maki-Yonekura,S., Zhu,Y., Taylor,D.W., van transcripts by CRISPR-Associated RNases enables Type III Duijn,E., Barendregt,A., Vlot,M., Koehorst,J.J., Sakamoto,K. et al. CRISPR–Cas immunity. Cell, 164, 710–721. (2013) Structure and activity of the RNA-targeting Type III-B 36. Niewoehner,O. and Jinek,M. (2016) Structural basis for the CRISPR–Cas complex of Thermus thermophilus. Mol. Cell, 52, endoribonuclease activity of the type III-A CRISPR-associated 135–145. protein Csm6. RNA, 22, 318–329. 16. Tamulaitis,G., Kazlauskiene,M., Manakova,E., Venclovas,C., 37. Sheppard,N.F., Glover,C.V. 3rd, Terns,R.M. and Terns,M.P. (2016) Nwokeoji,A.O., Dickman,M.J., Horvath,P. and Siksnys,V. (2014) The CRISPR-associated Csx1 protein of Pyrococcus furiosus is an Programmable RNA shredding by the type III-A CRISPR–Cas adenosine-specific endoribonuclease. RNA, 22, 216–224. system of Streptococcus thermophilus. Mol. Cell, 56, 506–517. 38. Rostol,J.T. and Marraffini,L.A. (2019) Non-specific degradation of 17. Hale,C.R., Zhao,P., Olson,S., Duff,M.O., Graveley,B.R., Wells,L., transcripts promotes plasmid clearance during type III-A Terns,R.M. and Terns,M.P. (2009) RNA-guided RNA cleavage by a CRISPR–Cas immunity. Nat. Microbiol., 4, 656–662. CRISPR RNA-Cas protein complex. Cell, 139, 945–956. 39. Anantharaman,V., Makarova,K.S., Burroughs,A.M., Koonin,E.V. 18. Tamulaitis,G., Venclovas,C. and Siksnys,V. (2017) Type III and Aravind,L. (2013) Comprehensive analysis of the HEPN CRISPR–Cas immunity: major differences brushed aside. Trends superfamily: identification of novel roles in intra-genomic conflicts, Microbiol., 25, 49–61. defense, pathogenesis and RNA processing. Biol. Direct, 8,15. 19. Goldberg,G.W., Jiang,W., Bikard,D. and Marraffini,L.A. (2014) 40. Makarova,K.S., Anantharaman,V., Grishin,N.V., Koonin,E.V. and Conditional tolerance of temperate phages via Aravind,L. (2014) CARF and WYL domains: ligand-binding transcription-dependent CRISPR–Cas targeting. Nature, 514, regulators of prokaryotic defense systems. Front. Genet., 5, 102. 633–637. 41. Gruschow,S., Athukoralage,J.S., Graham,S., Hoogeboom,T. and 20. Spilman,M., Cocozaki,A., Hale,C., Shao,Y., Ramia,N., Terns,R., White,M.F. (2019) Cyclic oligoadenylate signalling mediates Terns,M., Li,H. and Stagg,S. (2013) Structure of an RNA silencing Mycobacterium tuberculosis CRISPR defence. Nucleic Acids Res., complex of the CRISPR–Cas immune system. Mol. Cell, 52, 146–152. 47, 9259–9270. 21. Osawa,T., Inanaga,H., Sato,C. and Numata,T. (2015) Crystal 42. Nasef,M., Mufyfl ,M.C., Beckman,A.B., Rowe,S.J., Walker,F.C., structure of the CRISPR–Cas RNA silencing Cmr complex bound to Hatoum-Aslan,A. and Dunkle,J.A. (2019) Regulation of cyclic a target analog. Mol. Cell, 58, 418–430. oligoadenylate synthesis by the Staphylococcus epidermidis 22. Rouillon,C., Zhou,M., Zhang,J., Politis,A., Beilsten-Edmands,V., Cas10-Csm complex. RNA, 25, 948–962. Cannone,G., Graham,S., Robinson,C.V., Spagnolo,L. and 43. Athukoralage,J.S., Graham,S., Gruschow,S., Rouillon,C. and White,M.F. (2013) Structure of the CRISPR interference complex White,M.F. (2019) A Type III CRISPR ancillary ribonuclease CSM reveals key similarities with cascade. Mol. Cell, 52, 124–134. degrades its cyclic oligoadenylate activator. J. Mol. Biol., 431, 23. Staals,R.H., Zhu,Y., Taylor,D.W., Kornfeld,J.E., Sharma,K., 2894–2899. Barendregt,A., Koehorst,J.J., Vlot,M., Neupane,N., Varossieau,K. 44. Jia,N., Jones,R., Yang,G., Ouerfelli,O. and Patel,D.J. (2019) et al. (2014) RNA targeting by the type III-A CRISPR–Cas Csm CRISPR–Cas III-A Csm6 CARF domain is a ring nuclease complex of Thermus thermophilus. Mol. Cell, 56, 518–530. triggering stepwise cA4 cleavage with ApA>p formation terminating 24. Taylor,D.W., Zhu,Y., Staals,R.H., Kornfeld,J.E., Shinkai,A., van der RNase activity. Mol. Cell, 75, 944–956. Oost,J., Nogales,E. and Doudna,J.A. (2015) Structural biology. 45. Athukoralage,J.S., Rouillon,C., Graham,S., Gruschow,S. and Structures of the CRISPR-Cmr complex reveal mode of RNA target White,M.F. (2018) Ring nucleases deactivate type III CRISPR positioning. Science, 348, 581–585. ribonucleases by degrading cyclic oligoadenylate. Nature, 562, 25. You,L., Ma,J., Wang,J., Artamonova,D., Wang,M., Liu,L., Xiang,H., 277–280. Severinov,K., Zhang,X. and Wang,Y. (2019) Structure studies of the 46. Molina,R., Stella,S., Feng,M., Sofos,N., Jauniskis,V., CRISPR-Csm complex reveal mechanism of co-transcriptional Pozdnyakova,I., Lopez-Mendez,B., She,Q. and Montoya,G. (2019) interference. Cell, 176, 239–253 e216. Structure of Csx1-cOA4 complex reveals the basis of RNA decay in 26. Hale,C., Kleppe,K., Terns,R.M. and Terns,M.P. (2008) Prokaryotic Type III-B CRISPR–Cas. Nat. Commun., 10, 4302. silencing (psi)RNAs in Pyrococcus furiosus. RNA, 14, 2572–2579. 47. Deng,L., Garrett,R.A., Shah,S.A., Peng,X. and She,Q. (2013) A 27. Hale,C.R., Majumdar,S., Elmore,J., Pfister,N., Compton,M., novel interference mechanism by a type IIIB CRISPR-Cmr module Olson,S., Resch,A.M., Glover,C.V. 3rd, Graveley,B.R., Terns,R.M. in Sulfolobus. Mol. Microbiol., 87, 1088–1099. et al. (2012) Essential features and rational design of CRISPR RNAs 48. Guo,M., Zhang,K., Zhu,Y., Pintilie,G.D., Guan,X., Li,S., that function with the Cas RAMP module complex to cleave RNAs. Schmid,M.F., Ma,Z., Chiu,W. and Huang,Z. (2019) Coupling of Mol. Cell, 45, 292–302. ssRNA cleavage with DNase activity in type III-A CRISPR-Csm 28. Carte,J., Wang,R., Li,H., Terns,R.M. and Terns,M.P. (2008) Cas6 is revealed by cryo-EM and biochemistry. Cell Res., 29, 305–312. an endoribonuclease that generates guide RNAs for invader defense 49. Jia,N., Mo,C.Y., Wang,C., Eng,E.T., Marraffini,L.A. and Patel,D.J. in prokaryotes. Genes Dev., 22, 3489–3496. (2019) Type III-A CRISPR–Cas Csm complexes: assembly, periodic 29. Ramia,N.F., Spilman,M., Tang,L., Shao,Y., Elmore,J., Hale,C., RNA cleavage, DNase activity regulation, and autoimmunity. Mol. Cocozaki,A., Bhattacharya,N., Terns,R.M., Terns,M.P. et al. (2014) Cell, 73, 264–277. Essential structural and functional roles of the Cmr4 subunit in RNA 50. Wang,L., Mo,C.Y., Wasserman,M.R., Rostol,J.T., Marraffini,L.A. cleavage by the Cmr CRISPR–Cas complex. Cell Rep., 9, 1610–1617. and Liu,S. (2019) Dynamics of Cas10 govern discrimination between 30. Benda,C., Ebert,J., Scheltema,R.A., Schiller,H.B., Baumgartner,M., self and Non-self in Type III CRISPR–Cas immunity. Mol. Cell, 73, Bonneau,F., Mann,M. and Conti,E. (2014) Structural model of a 278–290. CRISPR RNA-silencing complex reveals the RNA-target cleavage 51. Han,W., Stella,S., Zhang,Y., Guo,T., Sulek,K., Peng-Lundgren,L., activity in Cmr4. Mol. Cell, 56, 43–54. Montoya,G. and She,Q. (2018) A Type III-B Cmr effector complex 4434 Nucleic Acids Research, 2020, Vol. 48, No. 8 catalyzes the synthesis of cyclic oligoadenylate second messengers by 64. Cocozaki,A.I., Ramia,N.F., Shao,Y., Hale,C.R., Terns,R.M., cooperative substrate binding. Nucleic Acids Res., 46, 10319–10330. Terns,M.P. and Li,H. (2012) Structure of the Cmr2 subunit of the 52. Jia,N., Jones,R., Sukenick,G. and Patel,D.J. (2019) Second messenger CRISPR–Cas RNA silencing complex. Structure, 20, 545–553. cA4 formation within the composite Csm1 palm pocket of Type III-A 65. Shao,Y., Cocozaki,A.I., Ramia,N.F., Terns,R.M., Terns,M.P. and CRISPR–Cas Csm complex and its release path. Mol. Cell, 75, Li,H. (2013) Structure of the Cmr2-Cmr3 subcomplex of the Cmr 933–943. RNA silencing complex. Structure, 21, 376–384. 53. Marraffini,L.A. and Sontheimer,E.J. (2010) Self versus non-self 66. Lillestol,R.K., Shah,S.A., Brugger,K., Redder,P., Phan,H., discrimination during CRISPR RNA-directed immunity. Nature, Christiansen,J. and Garrett,R.A. (2009) CRISPR families of the 463, 568–571. crenarchaeal genus Sulfolobus: bidirectional transcription and 54. Johnson,K., Learn,B.A., Estrella,M.A. and Bailey,S. (2019) Target dynamic properties. Mol. Microbiol., 72, 259–272. sequence requirements of a type III-B CRISPR–Cas immune system. 67. Liu,T.Y., Iavarone,A.T. and Doudna,J.A. (2017) RNA and DNA J. Biol. Chem., 294, 10290–10299. Targeting by a Reconstituted Thermus thermophilus Type III-A 55. Majumdar,S., Zhao,P., Pfister,N.T., Compton,M., Olson,S., CRISPR–Cas System. PLoS One, 12, e0170552. Glover,C.V. 3rd, Wells,L., Graveley,B.R., Terns,R.M. and Terns,M.P. 68. Zhang,J., Graham,S., Tello,A., Liu,H. and White,M.F. (2016) (2015) Three CRISPR–Cas immune effector complexes coexist in Multiple nucleic acid cleavage modes in divergent type III CRISPR Pyrococcus furiosus. RNA, 21, 1147–1158. systems. Nucleic Acids Res., 44, 1789–1799. 56. Terns,R.M. and Terns,M.P. (2013) The RNA- and DNA-targeting 69. Pyenson,N.C., Gayvert,K., Varble,A., Elemento,O. and CRISPR–Cas immune systems of Pyrococcus furiosus. Biochem. Soc. Marraffini,L.A. (2017) Broad targeting specificity during bacterial Trans., 41, 1416–1421. Type III CRISPR–Cas immunity constrains viral escape. Cell Host 57. Majumdar,S. and Terns,M.P. (2019) CRISPR RNA-guided DNA Microbe, 22, 343–353. cleavage by reconstituted Type I-A immune effector complexes. 70. Shiimori,M., Garrett,S.C., Chambers,D.P., Glover,C.V.C. 3rd, Extremophiles, 23, 19–33. Graveley,B.R. and Terns,M.P. (2017) Role of free DNA ends and 58. Majumdar,S., Ligon,M., Skinner,W.C., Terns,R.M. and Terns,M.P. protospacer adjacent motifs for CRISPR DNA uptake in Pyrococcus (2017) Target DNA recognition and cleavage by a reconstituted Type furiosus. Nucleic. Acids. Res., 45, 11281–11294. I-G CRISPR–Cas immune effector complex. Extremophiles, 21, 71. Shiimori,M., Garrett,S.C., Graveley,B.R. and Terns,M.P. (2018) Cas4 95–107. nucleases define the PAM, length, and orientation of DNA fragments 59. Elmore,J., Deighan,T., Westpheling,J., Terns,R.M. and Terns,M.P. integrated at CRISPR Loci. Mol. Cell, 70, 814–824. (2015) DNA targeting by the type I-G and type I-A CRISPR–Cas 72. Ichikawa,H.T., Cooper,J.C., Lo,L., Potter,J., Terns,R.M. and systems of Pyrococcus furiosus. Nucleic Acids Res., 43, 10353–10363. Terns,M.P. (2017) Programmable type III-A CRISPR–Cas DNA 60. Lipscomb,G.L., Stirrett,K., Schut,G.J., Yang,F., Jenney,F.E. Jr., targeting modules. PLoS One, 12, e0176221. Scott,R.A., Adams,M.W. and Westpheling,J. (2011) Natural 73. Lin,J., Feng,M., Zhang,H. and She,Q. (2019) Characterization of a competence in the hyperthermophilic archaeon Pyrococcus furiosus novel type III effector provides new insights into the allosteric facilitates genetic manipulation: construction of markerless deletions activation and suppression of the Cas10 DNase. bioRxiv doi: of genes encoding the two cytoplasmic hydrogenases. Appl. Environ. https://doi.org/10.1101/2019.12.17.879585, 18 December 2019, Microbiol., 77, 2232–2238. preprint: not peer reviewed. 61. Rouillon,C., Athukoralage,J.S., Graham,S., Gruschow,S. and 74. Cao,L., Gao,C.H., Zhu,J., Zhao,L., Wu,Q., Li,M. and Sun,B. (2016) White,M.F. (2019) Investigation of the cyclic oligoadenylate signaling Identification and functional study of type III-A CRISPR–Cas pathway of type III CRISPR systems. Methods Enzymol., 616, systems in clinical isolates of Staphylococcus aureus. Int. J. Med. 191–218. Microbiol.: IJMM, 306, 686–696. 62. Hatoum-Aslan,A., Maniv,I., Samai,P. and Marraffini,L.A. (2014) 75. Pyenson,N.C. and Marraffini,L.A. (2017) Type III CRISPR–Cas Genetic characterization of antiplasmid immunity through a type systems: when DNA cleavage just isn’t enough. Curr. Opin. III-A CRISPR–Cas system. J. Bacteriol., 196, 310–317. Microbiol., 37, 150–154. 63. Kim,Y.K., Kim,Y.G. and Oh,B.H. (2013) Crystal structure and nucleic acid-binding activity of the CRISPR-associated protein Csx1 of Pyrococcus furiosus. Proteins, 81, 261–270.

Journal

Nucleic Acids ResearchPubmed Central

Published: Mar 21, 2020

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off