CRISPR-Associated Primase-Polymerases are implicated in prokaryotic CRISPR-Cas adaptation

Abstract

CRISPR-Cas pathways provide prokaryotes with acquired "immunity" against foreign genetic elements, including phages and plasmids. Although many of the proteins associated with CRISPR-Cas mechanisms are characterized, some requisite enzymes remain elusive. Genetic studies have implicated host DNA polymerases in some CRISPR-Cas systems but CRISPR-specific replicases have not yet been discovered. We have identified and characterised a family of CRISPR-Associated Primase-Polymerases (CAPPs) in a range of prokaryotes that are operonically associated with Cas1 and Cas2. CAPPs belong to the Primase-Polymerase (Prim-Pol) superfamily of replicases that operate in various DNA repair and replication pathways that maintain genome stability. Here, we characterise the DNA synthesis activities of bacterial CAPP homologues from Type IIIA and IIIB CRISPR-Cas systems and establish that they possess a range of replicase activities including DNA priming, polymerisation and strand-displacement. We demonstrate that CAPPs operonically-associated partners, Cas1 and Cas2, form a complex that possesses spacer integration activity. We show that CAPPs physically associate with the Cas proteins to form bespoke CRISPR-Cas complexes. Finally, we propose how CAPPs activities, in conjunction with their partners, may function to undertake key roles in CRISPR-Cas adaptation.

Fig. 1. Bioinformatic analysis of CRISPR-associated Prim-Pols.

a A phylogenetic tree (bootstrap n = 100) from multiple sequence alignments of all identified CAPP proteins (Supplementary Data 1 and 2). Branch colours and outer ring colours indicate phyla from the National Center for Biotechnology (NCBI) taxonomy database. The inner ring is composed of coloured protein domain annotations of CAPP proteins from the NCBI Conserved Domains Database (CDD; abbreviated terms are used to simplify colour coding, see Supplementary Fig. 1 for full description). Bootstrap values of 100% are indicated with bold branches. b Coloured heatmap of keywords from gene names, upstream and downstream of the CAPP genes, indicating their occurrence (%) at certain positions relative to CAPP. c Classification of CAPP proteins based on their sequence homology, forming two major classes CAPP_A and CAPP_B. CAPP_A is further divided to three types, based on the protein domain architectures, CAPP-TPR, CAPP-RT and CAPP-Helicase. For each class, an example of NCBI CDD derived domain architecture is shown (left), together with its corresponding operonic region (right) showing the neighbouring genes present in each operon. CRISPR repeat array – white textured arrow. TPR – tetratricopeptide repeat – magenta domain, RT – reverse transcriptase – yellow domain, VirE_N – virulent protein E N-terminal like – green domain, helicase - blue domain.

Fig. 2. MpCAPP and DbCAPP are DNA-dependent DNA polymerases.

a MpCAPP possesses DNA polymerase extension activity (left panel). 1, 5, 25 and 125 nM MpCAPP wild-type (WT) or 125 nM D177A + D179A (AxA) mutant protein was added into 30 nM DNA substrate (DNA template + labelled DNA primer) and 100 µM dNTPs in MpPolBuffer. DbCAPP DNA polymerase activity (right panel). 1, 5, 25 and 125 nM DbCAPP wild-type (WT) or 125 nM D86A + D88A mutant (AxA) was added into 30 nM substrate and 100 µM dNTPs in DbPolBuffer. The reactions were incubated at 37 °C for 30 min and resolved by denaturing PAGE. b Time course of MpCAPP and DbCAPP polymerase extension activities. In all, 50 nM protein was incubated with 500 nM substrate, 100 µM dNTPs at 37 °C and the generated pyrophosphate was detected using the Malachite Green-based coupled polymerase assay. Circle – MpCAPP WT, open circle – MpCAPP AxA, square – DbCAPP WT, open square – DbCAPP AxA, grey line: standard deviation of n = 3 technical replicates. Data are presented as mean values. c CAPP polymerase extension activities with dNTPs. Both MpCAPP and DbCAPP have the strongest polymerase activities with DNA templates and dNTPs. In all, 50 nM protein was added into 50 nM substrate and 100 µM dNTPs in MpPolBuffer or DbPolBuffer. The reaction was incubated at 37 °C for 30 min. C – control reaction without protein, Green star – labelled-DNA primer without extension, Red star – labelled-RNA primer without extension, nts – nucleotide length of DNA markers. d CAPP polymerase activities with NTPs. MpCAPP and DbCAPP show weaker polymerase activity on DNA template with NTPs. In total, 50 nM protein was added into 50 nM substrate and 100 µM NTPs in buffers as described for panel c. The reaction was incubated at 37 °C for 30 min. e Time course of MpCAPP and DbCAPP polymerase specificities. 50 nM protein was incubated with 500 nM substrate, 100 µM dNTPs or NTPs at 37 °C and the generated pyrophosphate was detected by using Malachite Green-based coupled polymerase assay. Red – MpCAPP, green – DbCAPP, circle – dNTPs, triangle – NTPs, grey line – standard deviation of n = 3 technical replicates. Data are presented as mean values.

Fig. 3. CAPPs require ribonucleotides for primer synthesis.

a MpCAPP needs both NTPs and dNTPs for primer synthesis. In all, 4 µM MpCAPP wild-type protein was added into 10 ng/µl circular M13 ssDNA substrate in presence of 2.5 µM dNTPs (FAM-labelled dUTP) + 2.5–100 µM non-labelled NTPs and MpPrimBuffer. The reaction was incubated at 50 °C for 30 min and the products were resolved by denaturing PAGE. b MpCAPP FAM-labelled UTP incorporation is very poor. In total, 4 µM MpCAPP was added into 10 ng/µl circular M13 ssDNA substrate in presence of 2.5 µM NTPs (FAM-labelled UTP) + 2.5–100 µM non-labelled dNTPs and MpPrimBuffer. The reaction was incubated at 50 °C for 30 min. c MpCAPP priming is purine ribonucleotide dependent. In all, 1 µM MpCAPP was added into 1 µM mixed-sequence ssDNA substrate (oKZ53) in presence of 2.5 µM dNTPs (FAM-labelled dCTP) and 1 mM unlabelled NTPs as indicated in the figure. The reaction was incubated at 50 °C for 10 min. C – control reaction without protein, no – no NTPs, all – all NTPs, black arrow – signal of Cy5-labelled template. d DbCAPP primase activity is stimulated by addition of purine ribonucleotides. In total, 1 µM DpCAPP protein was added into 1 µM ssDNA (oKZ53) in presence of 2.5 µM dNTPs (FAM-labelled dCTP) and 1 mM unlabelled NTPs in DbPrimBuffer. The reaction was incubated at 50 °C for 10 min. Annotations identical as for panel c. nts – nucleotide length of DNA markers. Results shown are representative of three independent repeats (3a–d).

Fig. 4. CAPPs initiate primer synthesis with a 5′ ribonucleotide and prefer purines.

a γ-phosphate labelled GTP incorporation during MpCAPP de novo primer synthesis is sequence-dependent. In total, 1 µM of MpCAPP was added into the reaction with MpPrimBuffer containing 1 µM ssDNA substrates as indicated, 100 µM dNTP mix and 10 µM γ-phosphate Atto488-labelled GTP. The reactions were incubated at 50° for 30 min. The products were resolved on 20% urea-PAGE gel. C – control reaction without protein, nts – nucleotide length of DNA markers. b γ-phosphate labelled GTP incorporation during DbCAPP de novo primer synthesis is sequence-dependent. In total, 1 µM of DbCAPP was added into the reaction with DbPrimBuffer containing 1 µM DNA substrates as indicated, 100 µM dNTP mix and 10 µM γ-phosphate Atto488-labelled GTP. The reactions were incubated at 50° for 30 min. The products were resolved on 20% urea-PAGE gel. Annotations identical as for panel a. Results shown are representative of three independent repeats (4a, b).

Fig. 5. CAPPs possess synthesis-dependent strand displacement activity.

a MpCAPP possesses strong strand displacement activity. Polymerase assay was performed with 50 nM protein and 50 nM dsDNA substrates, containing a gap of the indicated lengths (0 – nick, 1, 2, 3, 5 or OH – over-hang). C – no enzyme, WT – wild-type, AxA – active site mutant, Green star – labelled-DNA primer without extension. b DbCAPP possesses strong strand displacement activity. Polymerase assay was performed on dsDNA substrates as described in panel a. Annotations identical as for panel a. c MpCAPP is able to dismantle a half-site integration intermediate – replication fork using its strand-displacement activity. Assay was performed with the indicated concentrations of proteins with or without dNTPs in presence of replication fork (RF) substrate. The products were resolved on native PAGE gel. WT (wild type) and AxA (active site mutant). d Left panel – Schematic representation of post-synaptic DNA substrate used in the MpCAPP-displacement assay (right). Right panels – MpCAPP is able to fully displace Cas1-Cas2 post-synaptic DNA substrate. Displacement assay was performed on 30 nM post-synaptic substrate at 50 °C for 30 min. C – no enzyme, WT – wild-type, AxA – active site mutant, Green signal – Atto550, Red signal – Cy5, Red star – Cy5-labelled leader, Green star – Atto550-labelled Spacer B, Red and green arrows – full extension products after CAPP strand displacement synthesis (Cy5- and Atto550-labelled, respectively), nts – nucleotide length of DNA markers. Results shown are representative of three independent repeats (a–d).

Fig. 6. MpCAPP domains involved in self-association.

a Schematic representaion of MpCAPP with its domains highlighted – three tetratricopeptide repeat (TPR) repeats in magenta, archaeo-eukaryotic primase (AEP) catalytic core (aa. ~110–340) in red and primase C-terminal (PriCT), containing conserved cysteine residues, in orange found in C-terminal domain (CTD). b MpCAPP shows self-association in the yeast two-hybrid assay. Full-length (FL) MpCAPP and its fragments (ΔCTD – aa. 1-360, TPR – aa. 1-100, AEP – aa. 100-360, and CTD – aa. 360-546) were either fused with GAL4 DNA-binding domain (BD) or its activation domain (AD) and their interaction with the indicated counterparts or empty vector (V) was established on selective plates lacking leucine, tryptophan and histidine or adenine. Addition of 10 mM 3-amino-triazole (3-AT) was also used to increase stringency of the histidine reporter. Results shown are representative of three independent repeats. c Schematic summarising the interactions observed: +++++ stands for very strong interaction, +++ for strong interaction, ++ for medium interaction, + for weak interaction and – for no interaction. T stands for toxic growth. d Graphical summary of the observed interactions.

Fig. 7. MpCas1 site-specific integration and disintegration activities.

a Schematic representation of the MpCRISPR array, before and after prespacer integration, used in the Cas1-integration assay (panel b). b 26 nM CRISPR array (PCR-synthetised) was incubated with wild-type Cas1 (WT) in presence of Cas2, increasing concentration of IHF and 200 nM prespacer (Cy5-labelled) in integration buffer (10 mM Bis-Tris Propane; pH 7, 10 mM MgCl₂, 100 mM NaCl, 0.5 mM TCEP and 0.1 mg/ml BSA) for 90 min at 50 °C. After Proteinase K digestion, the products were resolved on denaturing urea-PAGE. Green signal – Atto550 (CRISPR array), Red signal – Cy5 (prespacer) Red dot – prespacer, Green dot – CRISPR array without any integration, Red and green arrows– products after prespacer integration. c Schematic representation of branched substrates used in Cas1 transesterification assay (panel d). RF – replication fork. d MpCas1 prefers transesterification of the 5′-flap over RF and other tested branched structures. Increasing concentration of Cas1 was incubated with 100 nM branched substrates in buffer containing 10 mM Bis-Tris Propane; pH 7, 10 mM MgCl₂, 10 mM NaCl and 0.3 mg/ml BSA for 30 min at 50 °C. After Proteinase K digestion the products were resolved by denaturing urea-PAGE. Green signal – Atto550; Red signal – Cy5; Blue signal – FAM, nts – nucleotide length of DNA markers. Results shown are representative of three independent repeats (7b, d).

Fig. 8. CAPPs form a complex with Cas1 and Cas2.

a MpCAPP interacts with MpAgo and MpCas1 in yeast two-hybrid assays (Y2H) and MpCas2 self-associates. GAL4 DNA-binding domain (BD), activation domain (AD), empty vector (V). b Interactions between MpCAPP full-length (FL) and itself or fragments (FL – aa. 1-546, TPR – aa. 1–100, AEP – aa. 101–359, CTD – aa. 360–546, ΔCTD – aa. 1–359), MpCas1, MpCas2 and MpAgo were evaluated by Y2H. c MpCAPP forms a complex with MpCas1, MpCas2 and MpAgo proteins and directly interacts with MpAgo and MpCas1 in pull-down assays. Bait (MpCAPP-MBP or MBP) was pre-bound to the amylose beads and washed before adding prey (MpAgo, MpCas1 and/or MpCas2). I – prey input (the same input was used for the test and control experiment), FT – flow-through, B – bound protein. Test and control samples were run on the same gel (Supplementary Fig. 19). d Graphical summary of MpCAPP-Cas1-Cas2-Ago interactions. e DbCAPP interacts with DbCas1 and DbCas2, DbCas1 interacts with DbCas2 and DbRecD in Y2H. DbCas1 and DbCas2 self-associate. f Graphical summary of DbCAPP-Cas1-Cas2-RecD interactions. TPR – tetratricopeptide repeat – magenta domain, AEP – archaeo-eukaryotic primase – red domain, PriCT – primase C terminal – orange domain, MBP – maltose binding protein. Results shown are representative of three independent repeats in 8a–c, e.

Fig. 9. Prospective roles of CAPP during CRISPR-cas adaptation.

a–c Roles CAPPs may play during prespacer synthesis. a First role of CAPP during prespacer synthesis: 1. CAPP starts de novo primer synthesis and extension on the short ssDNA (e.g. nuclease degradation, CAPP synthesis or Cas1-transesterification ssDNA products). 2. Cas1-Cas2 complex binds the CAPP-synthesised dsDNA (prespacer) ready for integration. b Second CAPP role in prespacer synthesis: 1. Two CAPP molecules start synthesis on both DNA strands. 2. Strand displacement synthesis by another CAPP displaces the newly synthesised strands. 3. These complementary ssDNA strands could be annealed by Cas1-Cas2 complex (and the process could continue as described in panel a2). c Third role of CAPP in prespacer synthesis: 1. CAPP recognises nicked DNA. 2. CAPPs displacement synthesis activity creates 5′-flap structures. 3. Cas1 recognises 5′-flap and creates short ssDNA substrate via its transesterification activity and the process continues as described in panel a1. d Role of CAPP in the resolution of the post-synaptic complex. 1. Prespacer bound by Cas1-Cas2 complex and docks with the CRISPR array in readiness for new spacer integration. 2. Cas1 cleaves the first repeat strand (next to the leader) by transesterification and ligates the 3′-prime strands of the prespacer. 3. CAPP binds to the nicks on the Cas1-Cas2 spacer integration intermediate. 4. CAPP starts strand displacement synthesis of the repeat strands, resolving the post-synaptic complex. 5. The new spacer is ligated to complete spacer integration into the CRISPR array.