Skip to main content
NCBI home page
Howard Hughes Medical Institute Author Manuscripts logoLink to Howard Hughes Medical Institute Author Manuscripts
. Author manuscript; available in PMC: 2018 Mar 29.
Published in final edited form as: Cell. 2017 Aug 24;170(6):1224–1233.e15. doi: 10.1016/j.cell.2017.07.037

A Broad-Spectrum Inhibitor of CRISPR-Cas9

Lucas B Harrington 1,12, Kevin W Doxzen 2,12, Enbo Ma 1, Jun-Jie Liu 1,3, Gavin J Knott 1, Alireza Edraki 4, Bianca Garcia 5, Nadia Amrani 4, Janice S Chen 1, Joshua C Cofsky 1, Philip J Kranzusch 6,7, Erik J Sontheimer 4, Alan R Davidson 5,8, Karen L Maxwell 5, Jennifer A Doudna 1,2,3,9,10,11,13,*
PMCID: PMC5875921  NIHMSID: NIHMS952151  PMID: 28844692

SUMMARY

CRISPR-Cas9 proteins function within bacterial immune systems to target and destroy invasive DNA and have been harnessed as a robust technology for genome editing. Small bacteriophage-encoded anti-CRISPR proteins (Acrs) can inactivate Cas9, providing an efficient off-switch for Cas9-based applications. Here we show that two Acrs, AcrIIC1 and AcrIIC3, inhibit Cas9 by distinct strategies. AcrIIC1 is a broad-spectrum Cas9 inhibitor that prevents DNA cutting by multiple divergent Cas9 orthologs through direct binding to the conserved HNH catalytic domain of Cas9. A crystal structure of an AcrIIC1-Cas9 HNH domain complex shows how AcrIIC1 traps Cas9 in a DNA-bound but catalytically inactive state. By contrast, AcrIIC3 blocks activity of a single Cas9 ortholog and induces Cas9 dimerization while preventing binding to the target DNA. These two orthogonal mechanisms allow for separate control of Cas9 target binding and cleavage and suggest applications to allow DNA binding while preventing DNA cutting by Cas9.

INTRODUCTION

CRISPR systems provide bacteria and archaea with adaptive immunity against foreign DNA and RNA (Barrangou et al., 2007; Brouns et al., 2008; Marraffini and Sontheimer, 2008). To initiate immunity, CRISPR-associated (Cas) proteins integrate fragments of invading DNA into the host genome at the CRISPR locus, where they serve as transcription templates for the synthesis of RNA that directs Cas nucleases to cleave infectious nucleic acids (Garneau et al., 2010; Hale et al., 2009). Class 2 CRISPR systems are streamlined versions that require only a single protein to target foreign DNA or RNA (Makarova et al., 2015; Shmakov et al., 2017). CRISPR-Cas9, the most abundant and diverse of Class 2 CRISPR proteins, exists in three subtypes of which type IIA and IIC are more common compared to the relatively rare type IIB (Chylinski et al., 2014; Shmakov et al., 2017). The programmable nature of Cas9 has made it a powerful tool for gene editing and genomic modulation in a wide range of organisms.

In response to these robust prokaryotic immune systems, phages have evolved proteins that bind to and inactivate Cas proteins as they search for foreign nucleic acids (Bondy-Denomy et al., 2013, 2015; Pawluk et al., 2014). Although only a small number of these anti-CRISPRs (Acrs) have been discovered to date, phylogenetic analysis suggests that Acrs are widespread and likely play a significant role in the evolution of Cas proteins (Houte et al., 2016; Pawluk et al., 2016a). In addition to their native functions, Acrs that inhibit Cas9 nucleases allow for control of Cas9 in genome editing applications (Pawluk et al., 2016b; Rauch et al., 2017). Specifically, three unique Acrs that target the type IIC Cas9 from Neisseria meningitidis (NmeCas9) have been identified (AcrIIC1, 2 and 3) along with four that target select type IIA Cas9 orthologs (AcrIIA1, 2, 3 and 4). While some of these Acrs have been shown to inhibit NmeCas9 and SpyCas9 in mammalian cells (Pawluk et al., 2016b; Rauch et al., 2017), their ability to inactivate other Cas9 orthologs used for genome editing remains unknown. Understanding this specificity as well as the mechanisms by which they disable Cas9 will be critical for their successful deployment as modulators of Cas9 in human and other cell types. Apart from applications, this mechanistic information is also fundamental to understanding how these Acrs have evolved to target distinct Cas9 orthologs and what evolutionary pressures they impose on CRISPR systems.

Here we investigated the inactivation of Cas9 by AcrIIC1 and AcrIIC3, uncovering unique mechanisms for both. We focused on these two Acrs because of their potent inhibition of NmeCas9 in human cells (Pawluk et al., 2016b). Our data show that AcrIIC1 blocks DNA cleavage by multiple Cas9 orthologs without impacting DNA binding, effectively transforming catalytically active Cas9 into catalytically inactive dCas9. This mechanism is accomplished by AcrIIC1 binding directly to the HNH nuclease domain of Cas9, obscuring the active site and restricting conformational changes required for cleavage. AcrIIC3, by contrast, inhibits only a single Cas9 ortholog by blocking DNA binding. AcrIIC3 also causes Cas9 to dimerize, possibly contributing to its ability to interfere with target recognition and suggesting a mechanism distinct from that observed for AcrIIA4 (Dong et al., 2017; Shin et al., 2017). Together, AcrIIC1 and AcrIIC3 enable either broad-spectrum or selective inhibition of Cas9 orthologs respectively. The different mechanisms of these two Acrs allow separate control of binding to and cleavage of DNA by Cas9. Moreover, these mechanisms reveal vulnerabilities of Cas9 that are susceptible to inhibition, shedding light on the evolutionary arms race between bacteriophage and bacteria.

RESULTS

AcrIIC1 inhibits diverse Cas9 orthologs

Phylogenetic analysis revealed that AcrIIC1 is part of an unusually diverse family of Acr proteins (Figure S1A). Mirroring this diversity, the bacterial genomes containing AcrIIC1 include Cas9 orthologs that span a large portion of the type IIC Cas9 tree (Figure 1A). Based on its phylogenetic distribution, we hypothesized that AcrIIC1 would be more promiscuous than other Acr proteins with respect to the Cas9 orthologs it can inhibit.

Figure 1. AcrIIC1 inhibits diverse Cas9 orthologs while AcrIIC2 and AcrIIC3 are highly specific.

Figure 1

Open in a new tab

(A) Unrooted phylogenetic tree of Cas9. Cas9 orthologs targeted by Acrs are indicated with circles at ends of branches (closed circles, Cas9 orthologs naturally targeted by an Acr; open circles, Cas9 orthologs which have been shown experimentally to be inhibited by an Acr but without naturally occurring AcrIIC1 orthologs). For branches containing multiple Acrs of a given type only one circle is shown for simplicity (phylogeny adapted form Burstein et al., 2017).

(B) DNA cleavage assays conducted by various Cas9 orthologs in the presence of AcrIIC1, AcrIIC2 and AcrIIC3. (–Cas9, no Cas9 added; +Cas9, Cas9 and sgRNA added; Cje, Campylobacter jejuni; Nme, Neisseria meningitidis; Geo, Geobacillus stearothermophilus; Spy, Streptococcus pyogenes).

(C) (Left) Cartoon depicting experiment to test inhibition of Cas9 orthologs by AcrIIC1 in HEK293 cells. (Right) T7E1 assay analyzing indels produced by CjeCas9 and NmeCas9 shows that CjeCas9 genome editing is inhibited by AcrIIC1Nme but not AcrIIC3Nme. See also Figure S1.

To test this idea, we conducted cleavage assays using various type IIC Cas9 orthologs previously shown to function in human cells (Esvelt et al., 2013; Harrington et al., 2017; Hou et al., 2013; Kim et al., 2017). We found that in addition to NmeCas9, the AcrIIC1 from Neisseria meningitidis (AcrIIC1Nme) exhibits robust inhibition of the Cas9 proteins from Geobacillus stearothermophilus (GeoCas9) and Campylobacter jejuni (CjeCas9) (Figure 1B and S1B). CjeCas9 and GeoCas9 are 36% and 42% identical to NmeCas9 respectively and represent diverse branches of the type IIC Cas9 phylogeny (Figure 1A). By contrast, AcrIIC2 and AcrIIC3 were both highly specific for NmeCas9, having no noticeable impact on CjeCas9- or GeoCas9-catalyzed DNA cleavage (Figure 1B and S1B).

To determine whether inhibition by AcrIIC1 can disable CjeCas9 in genome-editing applications, we transfected HEK293T cells with plasmids expressing NmeCas9, CjeCas9 or SpyCas9 and their respective single-guide RNAs (sgRNAs) in the presence or absence of a gene encoding AcrIIC1 (Figure 1C and S1C). Similar to the biochemical cleavage assays, we observed that in cells, CjeCas9 is inhibited by AcrIIC1 but not by AcrIIC3. Expressing AcrIIC1Nme or the AcrIIC1 from Brackiella oedipodis (AcrIIC1Boe) resulted in efficient inhibition of CjeCas9, indicating that this promiscuity is not unique to the AcrIIC1Nme ortholog (Figure S1C). In similar cell-based assays, we found that AcrIIC1 is also a potent inhibitor of GeoCas9 ribonucleoprotein complexes (RNPs) in mammalian cells (Figure S1D), revealing that AcrIIC1 can also function when delivered as an expressed protein. The robust inhibition of both CjeCas9 and GeoCas9, in addition to NmeCas9, suggested that AcrIIC1 exploits a conserved feature of the Cas9 protein.

AcrIIC1 traps the DNA-bound Cas9 complex

Acrs can potentially inhibit Cas9 proteins at multiple distinct steps including guide RNA binding, target DNA binding or target cleavage. To determine the step at which AcrIIC1 inhibition occurs, we biochemically tested each of these possible mechanisms. First, we measured the binding affinity of NmeCas9 for its sgRNA in the presence and absence of AcrIIC1 (Figure S2A) and found that RNP assembly was unaffected by AcrIIC1. Next, we conducted equilibrium binding measurements of NmeCas9–sgRNA to its target DNA. Surprisingly, we found that NmeCas9 DNA binding was unimpeded by the presence of AcrIIC1, indicating that AcrIIC1 selectively blocks DNA cleavage (Figure S2B-C). Titrating AcrIIC1 and AcrIIC3 in a cleavage assay revealed that both are capable of inhibiting NmeCas9 even at low concentrations (Figure S2D). We conducted end-labeled cleavage assays to determine if cutting of both the target and non-target DNA strands is inhibited to the same degree (Figure 2A and 2B). Here we found that AcrIIC1 strongly inhibits cleavage of both DNA strands but with a subtle difference in kinetics. Although slow cleavage of the non-target DNA strand catalyzed by the RuvC active site is observed, target-strand cleavage catalyzed by the HNH domain is undetectable. These results suggested that AcrIIC1 traps Cas9 in its DNA-bound state, while inhibiting DNA cleavage. We tested this hypothesis by conducting gel shift assays using catalytically active GeoCas9 with and without AcrIIC1. In the absence of AcrIIC1, GeoCas9 cleaved its DNA target substrate at concentrations above ~30 nM (Figure 2C). However, when AcrIIC1 was included in the reaction, Cas9 did not cleave the target DNA even though DNA binding was unaffected. This remarkable mechanism is distinct from the recently studied AcrIIA2 and AcrIIA4 anti-CRISPR proteins, which function as inhibitors of DNA binding by SpyCas9 (Dong et al., 2017; Shin et al., 2017). The unique ability of AcrIIC1 to trap Cas9 on its DNA target in a catalytically inactivate state effectively transforms the wild-type Cas9 into its catalytically inactive variant dCas9 (Jinek et al., 2012).

Figure 2. AcrIIC1 traps the DNA-bound Cas9 complex.

Figure 2

Open in a new tab

(A) Cartoon of Cas9-mediated double-stranded DNA cleavage. Guide RNA (black) is duplexed to the DNA target strand (red), which is splayed from the DNA non-target strand (blue) adjacent to the PAM sequence (yellow). The HNH and RuvC nuclease domains (black triangles) cleave the target strand and non-target strand, respectively.

(B) Radiolabeled cleavage assays conducted using GeoCas9 to measure AcrIIC1 inhibition of cleavage on the target and non-target strands. Cas9–sgRNA RNP was complexed with or without AcrIIC1 and added to radiolabeled target DNA duplex with each strand labeled separately. The lanes for a given condition correspond to increasing time (0-30min) from left to right. Black triangles indicate cleavage products.

(C) Analysis of GeoCas9 binding and cleavage in the presence or absence of AcrIIC1 analyzed on a non-denaturing gel with the non-target strand labeled. GeoCas9 RNP concentration was varied in the absence or presence of excess AcrIIC1. The top band corresponds to GeoCas9 bound to the target DNA, the middle band is free DNA, and the lower band is cleaved DNA (concentration series correspond to 0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512nM of GeoCas9 RNP). See also Figure S2.

AcrIIC1 binds to the HNH domain of Cas9

The ability of AcrIIC1 to inhibit multiple Cas9 orthologs without preventing DNA binding suggested that it targets a conserved region of Cas9 involved in DNA cleavage. To determine which region of Cas9 interacts with AcrIIC1, we generated Cas9 truncations and tested their abilities to bind to AcrIIC1 using size exclusion chromatography (Figure 3A, 3B and S3A). Although many NmeCas9 truncations were insoluble, we took advantage of the thermostable GeoCas9 (Harrington et al., 2017) to generate soluble truncations. AcrIIC1 was able to associate with GeoCas9 without either the REC or PAM-interacting domains (Figure 3A), leaving the two nuclease domains as potential interacting partners. Truncating further to remove the RuvC domain allowed us to identify the HNH domain as being sufficient for AcrIIC1 binding to Cas9 (Figure 3A, 3B and S3A). Cas9 binding to RNA and DNA has been shown to be independent of the HNH domain(Sternberg et al., 2015; Yamada et al., 2017). In line with this, complexing of GeoCas9 with AcrIIC1 revealed that AcrIIC1 is able to interact with Cas9 irrespective of the presence of sgRNA or sgRNA and target DNA (Figure S3B). To determine if AcrIIC1 interacts specifically with the HNH domain, we exchanged the HNH domain of a Cas9 ortholog that does not interact with AcrIIC1 (Actinomyces naeslundii, AnaCas9) with an ortholog that does (GeoCas9). Here we found that the GeoCas9 chimera containing the AnaCas9 HNH domain no longer bound to AcrIIC1, while the AnaCas9 with the GeoCas9 HNH domain substitution was able to interact with AcrIIC1 (Figure 3A, 3B and S3C). These results indicated that the HNH domain is the primary site of interaction for AcrIIC1.

Figure 3. AcrIIC1 binds to Cas9 HNH domain.

Figure 3

Open in a new tab

(A) Domain schematics of GeoCas9 truncations and Cas9 chimeras, designed to identify the Cas9 binding interface of AcrIIC1. Constructs 1-10 were incubated with AcrIIC1, fractionated over an S200 size-exclusion column and analyzed by SDS-PAGE. Constructs that bound to AcrIIC1 are indicated with a (+) and constructs that showed no interaction are indicated with a (−). The chimeric Cas9 proteins (7-10) were generated by switching the HNH domains of a Cas9 that is not inhibited by AcrIIC1 (AnaCas9) and a Cas9 that is inhibited (GeoCas9).

(B) Fractions from the S200 runs in Figure S3A were separated on a 4-20% SDS-PAGE gel. Numbers above the gel correspond to the construct or chimera numbers from Figure 2A.

(C) Elution from an S75 size exclusion column of NmeCas9 HNH domain (purple), AcrIIC1 (orange) or the two incubated together with 2-fold excess AcrIIC1(red). See also Figure S3.

To examine whether this interaction also occurs with NmeCas9, we purified the HNH domain of NmeCas9 and tested its ability to bind to AcrIIC1 using size exclusion chromatography (Figure 3C). While AcrIIC1 and the HNH domain eluted at similar volumes in isolation, applying the two proteins to the size exclusion column together resulted in a large shift in elution volume, indicative of protein association. Importantly, the eluted HNHNme–AcrIIC1 complex remains in the included volume of the column, indicating that the large shift is not due to aggregation. Further analysis of AcrIIC1 binding to NmeCas9 by isothermal titration calorimetry (ITC) demonstrated an equilibrium binding affinity of AcrIIC1 6.3 ± 3.4 nM with a stoichiometry of one AcrIIC1 for each Cas9 (Figure S3D). The HNH nuclease domain is highly conserved across all Cas9 proteins (Figure S3E) and controls cleavage of both strands of the target DNA (Dagdas et al., 2017; Sternberg et al., 2015). Although the Cas9 HNH nuclease domain is directly responsible for cleavage of the target strand of the DNA (Gasiunas et al., 2012; Jinek et al., 2012), conformational activation of the HNH domain is a prerequisite for activating cleavage of the non-target strand by the RuvC nuclease domain (Sternberg et al., 2015). The ability of AcrIIC1 to bind to the most conserved domain of Cas9 explains its ability to robustly inhibit related Cas9 orthologs (Figure 1B) and its wide phylogenetic distribution (Figure 1A).

Structure of AcrIIC1 bound to the Cas9 HNH domain

To better understand how AcrIIC1 has evolved to bind to multiple Cas9 proteins, we determined a 1.5 Å resolution crystal structure of AcrIIC1Nme bound to the HNH domain of NmeCas9. The overall structure shows that AcrIIC1 binds directly to the HNH active site (Figure 4A), restricting it from accessing the target DNA. AcrIIC1 binds to the active site interface of the HNH domain through several ionic and hydrogen-bonding interactions. Critically, the HNH domain active site residues H588 and D587 hydrogen bond to AcrIIC1 residue S78 and to the backbone amine of C79, respectively (Figure 4B), possibly excluding the divalent cation necessary for target-strand DNA cleavage (Jinek et al., 2012). Mapping amino acid conservation onto the structure revealed that residues within the binding interface of both the HNH domain and AcrIIC1 are highly conserved (Figure 4D, S4A, S4B). In contrast to this observed conservation, antagonistic binding interfaces often evolve rapidly, leading to lower conservation (Franzosa and Xia, 2011), suggesting that AcrIIC1 is targeting a highly conserved surface in order to limit the chance for the host to escape inhibition.

Figure 4. Structure of AcrIIC1 bound to the NmeCas9 HNH domain.

Figure 4

Open in a new tab

(A) (Top) Cartoon depiction of NmeCas9 (grey) bound to a guide RNA (black.) The black outline of the HNH domain (purple) indicates the binding interface to AcrIIC1. (Bottom) Crystal structure of NmeCas9 HNH domain bound to AcrIIC1 (PDB:5VGB). Catalytic residues are depicted as sticks.

(B) Occlusion of HNH active site residues (purple) through hydrogen bonding with AcrIIC1 (orange). HNH catalytic residues H588 and D587 form hydrogen bonds (black dotted line) with S78 and the backbone amine of C79 of AcrIIC1, respectively. 2mFo-DFc electron density map is shown for interacting residues and contoured at 1.8 σ.

(C) Plaquing of E. coli phage Mu targeted by GeoCas9 in the presence of wild-type AcrIIC1 or the S78A AcrIIC1 mutant. Mutation of S78A results in nearly complete inactivation of AcrIIC1’s inhibitory effect on GeoCas9.

(D) Binding interfaces of NmeCas9 HNH domain and AcrIIC1 show residue conservation. Conservation was calculated using multiple sequence alignments of AcrIIC1 orthologs and Cas9 HNH domains. Conserved residues are colored red (1, 100% sequence identity) and non-conserved residues are colored white (0).

(E) Model of AcrIIC1 inhibiting cleavage of both target and non-target strands. NmeCas9 HNH domain (purple) was modeled into a “docked” position using dsDNA-bound SpyCas9 structure (PDB: 5F9R) as a reference for a homology model of NmeCas9. Placement of AcrIIC1 (orange) between the HNH domain and the target strand (red) prevents target cleavage and activation of the RuvC domain for non-target strand (dark blue) cleavage. See also Figure S4, S5 and Table S1.

Comparative structural homology searches of AcrIIC1 against protein structure databases using DALI and Vast revealed that AcrIIC1 adopts a novel protein fold (Holm and Laakso, 2016). The β1β2β3α1α2β4β5 fold of AcrIIC1 comprises a five-stranded beta bundle interspaced by two alpha-helices. The beta bundle is the conserved core feature found in all AcrIIC1 orthologs while the internal loops connecting beta-strands and alpha-helices vary in length and composition across species (Figure S4B). All of the HNH-interacting residues occur within these variable loop regions, revealing how AcrIIC1 can evolve to target divergent HNH domains without compromising structural integrity.

For AcrIIC1 to effectively prevent Cas9 from cleaving the invading viral DNA, it must remain bound to the HNH domain for extended periods of time. This stable interaction is in part accomplished by multiple charged residues around the periphery of the active site that form an additional five hydrogen bonds with AcrIIC1 (Figure S4C). Interestingly, some interactions target conserved residues present in diverse Cas9 orthologs and other interactions appear to have evolved to target specific species. For example, S78 and E81 of AcrIIC1Nme interact with the highly conserved catalytic residues H588 and N616 of the HNHNme domain, respectively. By contrast, AcrIIC1Nme residue D14 and the backbone carbonyl of P39 interact with K551 and K549 of NmeCas9, which are mutated to a serine and glycine in AnaCas9. To assess the importance of individual amino acids for the biological function of AcrIIC1, we established an in vivo anti-CRISPR activity assay in E. coli. Plasmid-mediated expression of GeoCas9 and an sgRNA designed to target E. coli phage Mu (Morgan et al., 2002) led to a reduction in the plaquing efficiency of this phage by approximately 106-fold (Figure 4C). Co-expression of wild-type AcrIIC1 restored the full plaquing activity of phage Mu, implying that that GeoCas9 was completely inhibited by the anti-CRISPR. By contrast, the S78A mutant displayed very little anti-CRISPR activity in this assay as phage Mu plaquing in the presence of this mutant was barely above background (Figure 4C and S4D). Substitutions of other residues positioned in the HNH:AcrIIC1 interaction interface, such as M76 and E81, caused more modest reductions in anti-CRISPR activity (Figure S4D) while substitution of other interface residues caused no reduction in biological activity. Importantly, all mutant proteins were expressed at the same level as wild-type (Figure S4E). AcrIIC1 interacting residues on the active site interface of the NmeCas9 HNH domain closely align with those on the same interface of S. aureus Cas9 (SauCas9), but diverge from equivalent residues in SpyCas9 and AnaCas9 mainly near the N-terminus of the HNH domain (Figure S5A). Together with this structure, the high degree of structural similarity between the HNH domains of these species will enable rational engineering of AcrIIC1 to target specific Cas9 orthologs of interest.

Investigation of Cas9 target recognition and cleavage has uncovered several checkpoints along the interference pathway that ensure cleavage of the correct DNA sequence (Sternberg et al., 2015). The best understood of these checkpoints is mediated by the HNH domain, which undergoes a large rotation and translation to cleave the target DNA only when sufficient complementarity to the guide RNA is sensed (Dagdas et al., 2017; Sternberg et al., 2015). The structure presented here suggests that AcrIIC1 exploits one checkpoint in this process to ensure inhibited cleavage of both the target and non-target DNA strands. When modeled into a Cas9–sgRNA complex bound to dsDNA (Jiang et al., 2016), AcrIIC1 sterically blocks the HNH domain from rotating into position above the scissile phosphate (Figure 4E and Figure S5B). We propose that the inability to correctly dock the HNH domain in the presence of AcrIIC1 inhibits RuvC cleavage of the non-target strand, explaining how a small protein that allows dsDNA engagement can still inhibit the two separate nucleases of Cas9.

AcrIIC3 blocks DNA binding and induces NmeCas9 dimerization

In contrast to AcrIIC1, AcrIIC3 has few natural orthologs and is found only in Neisseria. In HEK293T cells, expression of AcrIIC3 leads to the inability of dNmeCas9 to localize to a genomic target (Pawluk et al., 2016b), suggesting that AcrIIC3 prevents NmeCas9 from binding to DNA. We tested this biochemically using fluorescence polarization, which detected a ~10 fold decrease in equilibrium DNA binding affinity of NmeCas9 in the presence of AcrIIC3 (71 ± 13.4 nM without AcrIIC3 versus 859 ± 149 nM with AcrIIC3) (Figure 5A). This reduced but not abolished binding affinity of NmeCas9 for DNA in the presence of AcrIIC3 may indicate that NmeCas9 can still interact with the PAM region of the DNA, but cannot achieve complete R-loop formation (Mekler et al., 2017).

Figure 5. AcrIIC3 blocks DNA binding and dimerizes Cas9.

Figure 5

Open in a new tab

(A) Equilibrium binding measurements of NmeCas9 to dsDNA using fluorescence polarization in the presence (blue) or absence (black) of AcrIIC3. Measurements were made in triplicate and the mean +/− S. D. is shown.

(B) Elution from a Superdex 200 10/300 size exclusion column for NmeCas9 (black), NmeCas9+AcrIIC1 (orange), and NmeCas9+AcrIIC3 (blue) showing a large shift in elution volume for NmeCas9-AcrIIC3, indicative of oligomerization.

(C) SAXs data for fractions collected from samples in (C). (Left) pair-distance distribution function for NmeCas9 alone (black), with AcrIIC1 (orange) or with AcrIIC3 (blue), indicating increased particle size upon AcrIIC3 binding. (Rg, radius of gyration; Vc, volume of correlation; Dmax, maximum dimension.)

(D) 2D class averages of NmeCas9-sgRNA monomers (left) and NmeCas9-sgRNA bound to AcrIIC3 (right). Scale bar is 10nm. See also Figure S2 and S6.

After incubating NmeCas9 and AcrIIC3 together we noted a large shift in elution volume by size exclusion chromatography (SEC) compared to either AcrIIC1-bound NmeCas9 or NmeCas9 alone (Figure 5B). This large shift suggested either a substantial conformational change or oligomerization of Cas9. Analysis of the fractions by small-angle X-ray scattering (SAXS) revealed that the AcrIIC3-bound NmeCas9 increased in size relative to NmeCas9 alone or NmeCas9–AcrIIC1, as indicated by an elongated pair distance distribution, increased radius of gyration (Rg) and volume of correlation (Vc) (Figure 5C, Figure S6A). Using the power-law relationship for protein SAXS (Rambo and Tainer, 2013), we estimated the mass of NmeCas9 alone and AcrIIC1-bound NmeCas9 to be ~110kDa, a slight underestimate of the theoretical masses of 124kDa and 137kDa, respectively. In contrast, the estimated mass of AcrIIC3-bound NmeCas9 was ~210kDa. Together, our solution studies suggest that AcrIIC3 induces dimerization of NmeCas9, possibly contributing to its ability to block DNA binding. Analysis of AcrIIC3 alone by SEC and native mass spectrometry suggested that this Acr is monomeric in solution (Figure S6B and S6C), although dimerization of two AcrIIC3 monomers upon binding to NmeCas9 is possible.

To visualize the dimerization of NmeCas9, we examined NmeCas9 and AcrIIC3-bound NmeCas9 using electron microscopy. Although the protein interaction surfaces could not be identified due to limited resolution, an overall shape of the AcrIIC3-bound NmeCas9 complex can be observed in the 2D class averages (Figure 5D). To obtain better-resolution 2D class averages, the particles were cross-linked to reduce the flexibility of the NmeCas9–sgRNA complex and reduce dissociation of the dimer. These data reveal an overall symmetrical complex with dimensions consistent with two Cas9 proteins (Figure 5D). Cas9 inhibition by AcrIIC3-induced dimerization is consistent with the independent evolution of Acrs that act by diverse mechanisms.

DISCUSSION

We investigated the functions of two anti-CRISPR proteins, AcrIIC1 and AcrIIC3, and found that they block Cas9 activity by distinct mechanisms (Figure 6). AcrIIC1, an 85 amino acid protein, inactivates a wide range of type IIC Cas9 orthologs by binding to and conformationally restraining the conserved HNH domain. The direct interaction with essential catalytic residues of the HNH domain limits the opportunity for Cas9 to mutate and escape inhibition by AcrIIC1, explaining the phylogenetic propagation of this inhibitor to target multiple Cas9 orthologs. Intriguingly, AcrIIC1 traps Cas9 in an inactive but DNA-bound state, effectively converting wild-type Cas9 into a catalytically inactive dCas9. In contrast, the 116-amino acid AcrIIC3 binds specifically to the NmeCas9 enzyme to trigger dimerization and prevent DNA binding. Both of these mechanisms are different from that of the anti-CRISPR protein AcrIIA4, which acts as a DNA mimetic that prevents DNA binding by occupying the PAM-recognition site within a small subset of related type IIA Cas9 orthologs (Dong et al., 2017; Shin et al., 2017).

Figure 6. Model of AcrIIC1 and AcrIIC3 inhibition of Cas9.

Figure 6

Open in a new tab

Cas9 assembles with its guide RNA to form the search complex. Phage encoded AcrIIC1 (orange) binds to Cas9, still allowing target dsDNA binding but occluding the HNH (purple) active site, and stopping cleavage of the target strand. AcrIIC1 also conformationally restricts HNH docking, stopping cleavage on the non-target strand. For AcrIIC3 (blue), Cas9’s target DNA binding is inhibited and Cas9 is caused to dimerize.

The CRISPR inhibition mechanisms determined in this study concur with two general strategies observed previously for blocking interference proteins in both type I and type II CRISPR systems. The first and most common mechanism is to target the CRISPR surveillance complex by disrupting DNA binding (AcrIIC3, AcrIIA4, AcrF1, AcrF2; (Bondy-Denomy et al., 2015; Chowdhury et al., 2017; Dong et al., 2017; Pawluk et al., 2016b; Rauch et al., 2017; Shin et al., 2017). The second is to target nucleases or nuclease domains, thereby allowing DNA binding but not cleavage (AcrIIC1, AcrF3; (Bondy-Denomy et al., 2015; Wang et al., 2016). Of the currently studied Acrs, the strategy of inhibiting crRNA assembly with Cas proteins has yet to be found. The absence of this mechanism is possibly because this strategy would be unable to interfere with CRISPR nucleases that were assembled prior to infection. Nonetheless, many CRISPR systems are tightly regulated and are often activated in response to cell density and other factors (Høyland-Kroghsbo et al., 2016; Patterson et al., 2016, 2017). Inhibition of crRNA binding could be an effective Acr method to inhibit those systems where RNP assembly coincides with phage infection, and such inhibitors may yet be discovered. Moreover, only Acrs that target CRISPR interference proteins have been found, despite the fact that the methods currently used to identify Acrs are capable of finding inhibitors of spacer acquisition and crRNA processing. Nonetheless, it is likely that in cases where interference proteins are linked to other steps in CRISPR adaptive immunity, such as acquisition for type IIA Cas9 or RNA processing for Cas12a, such Acrs exist (Fonfara et al., 2016; Heler et al., 2015; Wei et al., 2015).

The mechanism of HNH domain binding by AcrIIC1 is particularly interesting for several reasons. First, the high structural similarity of the Cas9 HNH domain across Cas9 orthologs implies that AcrIIC1-type inhibitors may be more widespread than current analysis has identified. There also may be other Acrs that have converged on this mechanism. Second, type IIC Cas9 orthologs are capable of tracrRNA- and PAM-independent DNA cleavage of single stranded DNA catalyzed by the HNH-domain (Ma et al., 2015; Zhang et al., 2015). In addition to inhibition of double-stranded DNA cleavage, AcrIIC1 would also be able to inhibit this single-stranded cleavage activity, whereas inhibitors of PAM binding such as AcrIIA4 may not be able to. Third, the fact that AcrIIC1 traps Cas9-guide RNA complexes in a catalytically inactive but DNA-bound state is consistent with additional roles for this inhibitor that include gene regulation rather than genome protection. Notably, efforts to engineer regulatory forms of Cas9 have utilized dCas9, a catalytically inactive mutant of the enzyme that retains RNA-programmed DNA binding activity. It would be exciting to determine whether bacteria natively employ AcrIIC1 to repurpose Cas9 as a gene regulator in cells. Whether or not this occurs in nature, it is an enticing possibility that AcrIIC1 can be employed in gene editing applications to obviate the need to generate separate dCas9 enzymes for gene regulatory purposes (Gilbert et al., 2014; Mali et al., 2013; Qi et al., 2013). Finally, we note that the HNH fold is not unique to Cas9 but is in fact common to many bacterial restriction enzymes (Vasu and Nagaraja, 2013). This raises the possibility that in addition to targeting CRISPR-based adaptive immunity, AcrIIC1 also inhibits restriction enzymes. In line with this hypothesis, we observed that no Cas9 ortholog is present in Pseudoaltermonas lipolytica despite the presence of an AcrIIC1. A blastp search of the P. lipolytica genome for Cas9 revealed an HNH restriction enzyme with homology to the Cas9 HNH domain (20% sequence identity of HNH domains). It may be that AcrIIC1 has evolved to inhibit both adaptive (CRISPR) and innate (restriction) immune systems by targeting this conserved protein domain.

Given the rapid evolution and resourcefulness of phage, it is likely that Acrs are much more widespread than is currently known. As the toolbox of proteins used to edit genomes continues to expand to include other Class 2 CRISPR systems, discovery of new Acrs can serve as potent tools to control these new systems. Continued analysis of the abundance of Acrs as well as their mechanisms will provide unique opportunities to regulate and disable CRISPR systems, and in the process illuminate the influence of Acrs on CRISPR diversity.

Supplementary Material

S1. Figure S1. AcrIIC1 inhibits diverse Cas9 orthologs, while AcrIIC2 and AcrIIC3 are highly specific, related to Figure 1.

(A) Unrooted phylogenetic tree of AcrIIC1.

(B) Kinetic measurement of DNA cleavage mediated by GeoCas9 in the presence or absence of type IIC Acrs.

(C) Genome editing mediated by NmeCas9, SpyCas9 and CjeCas9 in the presence of various Acrs. Human (HEK293T) cells were transfected with plasmids expressing NmeCas9, SpyCas9, or CjeCas9, along with cognate, previously validated sgRNAs targeting genomic sites. T7E1 digestion was used to detect editing. A Type I anti-CRISPR (AcrE2) was used as a negative control for inhibition. As reported previously, NmeCas9 genome editing (upper panel) was inhibited by AcrIIC1Boe, AcrIIC1Nme, and AcrIIC3Nme; full inhibition by AcrIIC2Nme in human cells generally requires higher amounts of cotransfected expression plasmid. Inhibition of SpyCas9 genome editing (middle panel) was observed only with AcrIIA4Lmo (Rauch et al., 2017). In contrast, CjeCas9 editing activity (bottom panel) was inhibited by AcrIIC1Boe and AcrIIC1Nme, but not by any of the other anti-CRISPRs.

(D) GeoCas9 RNP mediated editing of HEK293T cells in the presence and absence of AcrIIC1. Indels were analyzed by T7E1 digestions.

S2. Figure S2. Impact of AcrIIC1 and AcrIIC3 on Cas9 RNA and DNA binding, related to Figures 2 and 5.

(A) Filter-binding assays measuring the affinity of NmeCas9 to its guide in the presence and absence of AcrIIC1. Radiolabeled sgRNA was incubated with NmeCas9 in the presence (orange) or absence (black) of AcrIIC1. NmeCas9–sgRNA RNP formation was measured using a filter binding assay and fraction sgRNA bound was calculated and plotted against NmeCas9 concentration. When incubated with Cas9 prior to sgRNA binding, AcrIIC1 does not inhibit RNP formation.

(B) Equilibrium binding measurements of NmeCas9–sgRNA binding to dsDNA in the presence and absence of AcrIIC1, measured by fluorescence polarization.

(C) Equilibrium binding measurements of SpyCas9 in the presence and absence of AcrIIC3, related to Figure 5D.

S3. Figure S3. AcrIIC1 Binds to the Cas9 HNH domain, related to Figure 3.

(A) Superdex 200 10/300 traces of truncations used to identify the binding interface of AcrIIC1. Each trace included the indicated GeoCas9 truncation and excess AcrIIC1. Black asterisks indicate the fractions analyzed in Figure 3B. For construct #4 (REC lobe), multiple peaks resulted from bound contaminating nucleic acid species from the purification and both peaks were pooled and analyzed together.

(B) Superdex 200 10/300 traces for GeoCas9 complexed with the components indicated in the top right. Fractions indicated with asterisk were analyzed on SDS-PAGE (upper gel) gel and denaturing urea PAGE gel (lower gel) and components were added in the order listed.

(1, Apo GeoCas9; 2, GeoCas9 +AcrIIC1; 3, GeoCas9+sgRNA+AcrIIC1; 4, GeoCas9+sgRNA+DNA+AcrIIC1).

(C) (top) HNH domains between GeoCas9 (orange) and AnaCas9 (yellow) were swapped to create chimeric Cas9 proteins. AcrIIC1 was fused with GFP (to more easily visualize a change in elution volume once bound to Cas9) and run over a S200 size-exclusion column. Additionally, AcrIIC1–GFP detection at a wavelength of 495 nm offers another indication of change in AcrIIC1 elution volume. (bottom) SDS-PAGE gradient gel (4–20%) with protein samples from S200 elution peaks of chimeric Cas9 proteins incubated with AcrIIC1.

(D) Representative ITC trace for AcrIIC1 binding to the NmeHNH domain.

(E) Conservation of Cas9 mapped onto the approximate domain boundaries below using a non-redundant list of Cas9 orthologs from all Cas9 subtypes. HNH is highlighted in purple. Bar heights are proportional to protein identity, with yellow bars indicating highly conserved residues.

S4. Figure S4. Alignment of HNH domains and AcrIIC1 orthologs and structure of AcrIIC1–HNH, related to Figure 4.

(A) A multiple sequence alignment of selected HNH-domains. The multiple sequence alignment was generated using the extracted HNH domains of 6 Cas9 orthologs and the restriction enzyme from Pseudoaltermonas lipolytica (RE_HNH). Red boxes surround catalytic residues and black asterisks indicate catalytic residues involved in AcrIIC1 binding. Blue boxes surround other residues involved in AcrIIC1 interaction.

(B) A multiple sequence alignment of AcrIIC1 using 11 AcrIIC1 orthologs. Highest degree of conservation occurs within a beta barrel (marked by a grey arrow). Red boxes surround selected residues involved in the binding between the NmeCas9 HNH domain and AcrIIC1Nme. Red asterisks indicate AcrIIC1 residues that interact with the backbone of HNH. Black asterisks indicate AcrIIC1 residues that interact with catalytic residues of the HNH domain.

(C) Charged residues (depicted as sticks) surrounding the active site of the HNH domain (purple) form ionic and hydrogen binding interactions (depicted as dotted black lines) with AcrIIC1 (orange).

(D) Fold reduction in phage titer in response to GeoCas9 targeting of phage Mu in the presence of AcrIIC1 mutants. One representative plate is shown for each mutant tested. 10-fold serial dilutions of phage Mu lysate were spotted lawns of bacteria expressing the indicated AcrIIC1 mutant. The fold reductions shown in the bar graph were qualitatively evaluated from inspecting three replicates of each experiment.

(E) SDS-PAGE gel showing the expression levels of WT and mutant AcrIIC1s used for experiments in panel D. The approximate mass of AcrIIC1 is indicated on the right (ev, empty vector; E2, AcrIE2).

S5. Figure S5. Conservation of the Cas9 HNH domain in the context of full length Cas9, related to Figure 4.

(A) Crystal structure of NmeCas9 HNH domain bound to AcrIIC1 (PDB: 5VGB). The HNH is rotated 90˚C to show the active site interface with labeled residues (depicted as sticks) involved in AcrIIC1 binding. (Bottom) Crystal structures of the HNH domains from three Cas9 orthologs (Staphylococcus aureus, PDB: 5CZZ; Streptococcus pyogenes, PDB: 4CMP; Actinomyces naeslundii, PDB: 4OGE). RMSD values were generated using super alignment in PyMol. Comparison of crystallized Cas9 orthologs reveals strong similarity between AcrIIC1 interacting residues of NmeCas9 and SauCas9 HNH domains.

(B) (Left) Model of AcrIIC1 inhibiting cleavage of both target and non-target strands. NmeCas9 HNH domain (purple) was modeled into a “docked” position using the dsDNA-bound SpyCas9 structure (PDB: 5F9R) as a reference. Placement of AcrIIC1 (orange) between the HNH domain and the target strand (red) prevents target cleavage and activation of the RuvC domain for non-target strand (dark blue) cleavage. The black box shows a zoomed-in view of the NmeHNH-AcrIIC1 complex. (Right) Model of AcrIIC1 clashing with the RuvC domain when the HNH domain is in an “undocked” conformation. NmeCas9 HNH was placed in a undocked conformation using a GeoCas9 Phyre model as a reference. Placement of AcrIIC1 indicates steric clashing with the RuvC domain, indicating that binding of AcrIIC1 to the HNH domain must position the HNH domain between the docked and undocked position. The black box shows a zoomed-in view of the HNHNme-AcrIIC1Nme complex.

S6. Figure S6. AcrIIC3 is monomeric and dimerizes NmeCas9, related to Figure 5.

(A) Small-angle X-ray scattering (SAXS) curves of NmeCas9, NmeCas9+AcrIIC1 and AcrIIC3.

(B) Analysis of AcrIIC1 (yellow, top) and AcrIIC3 (blue, below) on a Superdex 200 10/300 size exclusion column.

(C) Native mass spectrometry of AcrIIC3. The estimated masses from deconvoluting the charge series are identified in the top right corner.

Table S1

Table S1. Data collection and refinement statistics, related to Figure 4.

Acknowledgments

We thank G. Meigs and the 8.3.1 beamline staff at the Advanced Light Source for assistance with data collection and members of the Doudna laboratory for comments and discussions. The 8.3.1 beamline is supported by UC Office of the President, Multicampus Research Programs and Initiatives grant MR-15-328599 and Program for Breakthrough Biomedical Research, which is partially funded by the Sandler Foundation. J.A.D. is an Investigator of the Howard Hughes Medical Institute. This research was supported in part by the Allen Distinguished Investigator Program, through The Paul G. Allen Frontiers Group, the National Science Foundation (MCB-1244557 to J.A.D.), the National Institutes of Health (GM115911 to E.J.S.) and the Canadian Institutes of Health Research (MOP-130482 to A.R.D and MOP-136845 to K.L.M.). L.B.H., K.W.D. and J.S.C. are supported by US National Science Foundation Graduate Research

Footnotes

SUPPLEMENTAL INFORMATION

Supplemental Information includes six figures and one table.

AUTHOR CONTRIBUTIONS

L.B.H., K.W.D., E.M., K.L.M, J.A.D. and A.R.D. designed experiments. L.B.H, A.E., N. A., and E.J.S. designed and performed cell-based assays. L.B.H, K.W.D., J.S.C. and J.C.C. purified proteins. L.B.H. designed and purified RNA and DNA substrates. L.B.H and E.M conducted in vitro cleavage and binding experiments. K.W.D and E.M. assembled complex and set trays for protein crystallization. K.W.D and L.B.H acquired diffraction data. P.J.K. and K.W.D. determined the experimental phases of crystallographic data and traced the initial models and K.W.D. completed model building. L.B.H. and G.J.K. conducted and analyzed SAXS experiments. J. L. conducted electron microscopy experiments. A.R.D, K.L.M and B.G. designed and executed phage plaquing experiments. L.B.H, K.W.D, and J.A.D. wrote the manuscript. All authors revised and agreed to the final manuscript.

References

  1. Adams PD, Afonine PV, Bunkóczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, et al. PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr Sect D Biol Crystallogr. 2010;66:213–221. doi: 10.1107/S0907444909052925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Barrangou R, Fremaux C, Deveau H, Richards M, Boyaval P, Moineau S, Romero DA, Horvath P. CRISPR provides acquired resistance against viruses in prokaryotes. Science (80-) 2007;315:1709–1712. doi: 10.1126/science.1138140. [DOI] [PubMed] [Google Scholar]
  3. Bondy-Denomy J, Pawluk A, Maxwell KL, Davidson AR. Bacteriophage genes that inactivate the CRISPR/Cas bacterial immune system. Nature. 2013;493:429–432. doi: 10.1038/nature11723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bondy-Denomy J, Garcia B, Strum S, Du M, Rollins MF, Hidalgo-Reyes Y, Wiedenheft B, Maxwell KL, Davidson AR. Multiple mechanisms for CRISPR–Cas inhibition by anti-CRISPR proteins. Nature. 2015;526:136–139. doi: 10.1038/nature15254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brouns SJJ, Jore MM, Lundgren M, Westra ER, Slijkhuis RJH, Snijders APL, Dickman MJ, Makarova KS, Koonin EV, van der Oost J. Small CRISPR RNAs Guide Antiviral Defense in Prokaryotes. Science (80-) 2008;321:960 LP–964. doi: 10.1126/science.1159689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burstein D, Harrington LB, Strutt SC, Probst AJ. New CRISPR-Cas systems from uncultivated microbes. Nat Publ Gr. 2017;542:237–241. doi: 10.1038/nature21059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chowdhury S, Carter J, Rollins MF, Golden SM, Jackson RN, Hoffmann C, Nosaka L, Bondy-Denomy J, Maxwell KL, Davidson AR, et al. Structure Reveals Mechanisms of Viral Suppressors that Intercept a CRISPR RNA-Guided Surveillance Complex. Cell. 2017;169:47–57.e11. doi: 10.1016/j.cell.2017.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chylinski K, Makarova KS, Charpentier E, Koonin EV. Classification and evolution of type II CRISPR-Cas systems. Nucleic Acids Res. 2014;42:6091–6105. doi: 10.1093/nar/gku241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dagdas YS, Chen JS, Sternberg SH, Doudna JA, Yildiz A. A conformational checkpoint between DNA binding and cleavage by CRISPR-Cas9. bioRxiv. 2017 doi: 10.1126/sciadv.aao0027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dong D, Guo M, Wang S, Zhu Y, Wang S, Xiong Z, Yang J, Xu Z, Huang Z. Structural basis of CRISPR–SpyCas9 inhibition by an anti-CRISPR protein. Nature. 2017:1–15. doi: 10.1038/nature22377. [DOI] [PubMed] [Google Scholar]
  11. Emsley P, Cowtan K. Coot: Model-building tools for molecular graphics. Acta Crystallogr Sect D Biol Crystallogr. 2004;60:2126–2132. doi: 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  12. Esvelt KM, Mali P, Braff JL, Moosburner M, Yaung SJ, Church GM. Orthogonal Cas9 proteins for RNA-guided gene regulation and editing. Nat Meth. 2013;10:1116–1121. doi: 10.1038/nmeth.2681. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fonfara I, Le Rhun A, Chylinski K, Makarova KS, Lécrivain AL, Bzdrenga J, Koonin EV, Charpentier E. Phylogeny of Cas9 determines functional exchangeability of dual-RNA and Cas9 among orthologous type II CRISPR-Cas systems. Nucleic Acids Res. 2014;42:2577–2590. doi: 10.1093/nar/gkt1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fonfara I, Richter H, Bratovič M, Le Rhun A, Charpentier E. The CRISPR-associated DNA-cleaving enzyme Cpf1 also processes precursor CRISPR RNA. Nature. 2016:1–19. doi: 10.1038/nature17945. [DOI] [PubMed] [Google Scholar]
  15. Franzosa EA, Xia Y. Structural principles within the human-virus protein-protein interaction network. Proc Natl Acad Sci. 2011;108:10538–10543. doi: 10.1073/pnas.1101440108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Garneau JE, Dupuis M-È, Villion M, Romero DA, Barrangou R, Boyaval P, Fremaux C, Horvath P, Magadán AH, Moineau S. The CRISPR/Cas bacterial immune system cleaves bacteriophage and plasmid DNA. Nature. 2010;468:67–71. doi: 10.1038/nature09523. [DOI] [PubMed] [Google Scholar]
  17. Gasiunas G, Barrangou R, Horvath P, Siksnys V. Cas9–crRNA ribonucleoprotein complex mediates specific DNA cleavage for adaptive immunity in bacteria. Proc Natl Acad Sci. 2012;109:E2579–E2586. doi: 10.1073/pnas.1208507109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gilbert LA, Horlbeck MA, Adamson B, Villalta JE, Chen Y, Whitehead EH, Guimaraes C, Panning B, Ploegh HL, Bassik MC, et al. Genome-Scale CRISPR-Mediated Control of Gene Repression and Activation. Cell. 2014;159:647–661. doi: 10.1016/j.cell.2014.09.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hale CR, Zhao P, Olson S, Duff MO, Graveley BR, Wells L, Terns RM, Terns MP. RNA-Guided RNA Cleavage by a CRISPR RNA-Cas Protein Complex. Cell. 2009;139:945–956. doi: 10.1016/j.cell.2009.07.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Harrington LB, Paez-espino D, Chen JS, Staahl BT, Kyrpides NC, Doudna J. A thermostable Cas9 with increased lifetime in human plasma. bioRxiv. 2017 doi: 10.1038/s41467-017-01408-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Heler R, Samai P, Modell JW, Weiner C, Goldberg GW, Bikard D, Marraffini LA. Cas9 specifies functional viral targets during CRISPR-Cas adaptation. Nature. 2015;519:1–16. doi: 10.1038/nature14245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Holm L, Laakso LM. Dali server update. Nucleic Acids Res. 2016;44:W351–W355. doi: 10.1093/nar/gkw357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hou Z, Zhang Y, Propson NE, Howden SE, Chu LF, Sontheimer EJ, Thomson JA. Efficient genome engineering in human pluripotent stem cells using Cas9 from Neisseria meningitidis. Proc Natl Acad Sci. 2013;110:15644–15649. doi: 10.1073/pnas.1313587110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. van Houte S, Ekroth AKE, Broniewski JM, Chabas H, Ashby B, Gandon S, Mike Boots SP, Buckling AJ, Westra ER. The diversity-generating benefits of a prokaryotic adaptive immune system. Nature. 2016;532:385–388. doi: 10.1038/nature17436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Høyland-Kroghsbo NM, Paczkowski J, Mukherjee S, Broniewski J, Westra E, Bondy-Denomy J, Bassler BL. Quorum sensing controls the Pseudomonas aeruginosa CRISPR-Cas adaptive immune system. Proc Natl Acad Sci U S A. 2016;114:201617415. doi: 10.1073/pnas.1617415113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jiang F, Taylor DW, Chen JS, Kornfeld JE, Zhou K, Thompson AJ, Nogales E, Doudna JA. Structures of a CRISPR-Cas9 R-loop complex primed for DNA cleavage. Science (80-) 2016;8282:1–8. doi: 10.1126/science.aad8282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A Programmable Dual-RNA–Guided DNA Endonuclease in Adaptive Bacterial Immunity. Sci. 2012;337:816–821. doi: 10.1126/science.1225829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kabsch W. Xds. Acta Crystallogr Sect D Biol Crystallogr. 2010;66:125–132. doi: 10.1107/S0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kim E, Koo T, Park SW, Kim D, Kim K, Cho H, Song DW, Lee KJ, Jung MH, Kim S, et al. In vivo genome editing with a small Cas9 orthologue derived from Campylobacter jejuni. Nat Commun. 2017;8:1–12. doi: 10.1038/ncomms14500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Konarev PV, Volkov VV, Sokolova AV, Koch MHJ, Svergun DI. PRIMUS : a Windows PC-based system for small-angle scattering data analysis. J Appl Crystallogr. 2003;36:1277–1282. [Google Scholar]
  31. Konarev PV, Petoukhov MV, Volkov VV, Svergun DI. ATSAS 2.1, a program package for small-angle scattering data analysis. J Appl Crystallogr. 2006;39:277–286. doi: 10.1107/S0021889812007662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lander GC, Stagg SM, Voss NR, Cheng A, Fellmann D, Yoshioka C, Irving C, Mulder A, Lau P, Potter CS, et al. Image Processing. Access. 2009;166:95–102. doi: 10.1016/j.jsb.2009.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lin S, Staahl B, Alla RK, Doudna JA. Enhanced homology-directed human genome engineering by controlled timing of CRISPR/Cas9 delivery. Elife. 2014;3:1–13. doi: 10.7554/eLife.04766. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Liu JJ, Bratkowski MA, Liu X, Niu CY, Ke A, Wang HW. Visualization of distinct substrate-recruitment pathways in the yeast exosome by EM. Nat Struct Mol Biol. 2014;21:95–102. doi: 10.1038/nsmb.2736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Ma E, Harrington LB, O’Connell MR, Zhou K, Doudna JA. Single-Stranded DNA Cleavage by Divergent CRISPR-Cas9 Enzymes. Mol Cell. 2015;60:398–407. doi: 10.1016/j.molcel.2015.10.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Makarova KS, Wolf YI, Alkhnbashi OS, Costa F, Shah SA, Saunders SJ, Barrangou R, Brouns SJ, Charpentier E, Haft DH, et al. An updated evolutionary classification of CRISPR-Cas systems. Nat Rev Microbiol. 2015 doi: 10.1038/nrmicro3569. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mali P, Aach J, Stranges PB, Esvelt KM, Moosburner M, Kosuri S, Yang L, Church GM. CAS9 transcriptional activators for target specificity screening and paired nickases for cooperative genome engineering. Nat Biotechnol. 2013;31:833–838. doi: 10.1038/nbt.2675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Marraffini LA, Sontheimer EJ. CRISPR interference limits horizontal gene transfer in staphylococci by targeting DNA. Science (80-) 2008;322:1843–1845. doi: 10.1126/science.1165771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mekler V, Minakhin L, Severinov K. Mechanism of duplex DNA destabilization by RNA-guided Cas9 nuclease during target interrogation. Proc Natl Acad Sci. 2017 doi: 10.1073/pnas.1619926114. 201619926. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Morgan GJ, Hatfull GF, Casjens S, Hendrix RW. Bacteriophage Mu Genome Sequence : Analysis and Comparison with Mu-like Prophages in Haemophilus. Neisseria and Deinococcus. 2002 doi: 10.1006/jmbi.2002.5437. [DOI] [PubMed] [Google Scholar]
  41. Nuñez JK, Kranzusch PJ, Noeske J, Wright AV, Davies CW, Doudna JA. Cas1-Cas2 complex formation mediates spacer acquisition during CRISPR-Cas adaptive immunity. Nat Struct Mol Biol. 2014;21:528–534. doi: 10.1038/nsmb.2820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Patterson AG, Jackson SA, Taylor C, Evans GB, Salmond GPC, Przybilski R, Staals RHJ, Fineran PC. Quorum Sensing Controls Adaptive Immunity through the Regulation of Multiple CRISPR-Cas Systems. Mol Cell. 2016;64:1–7. doi: 10.1016/j.molcel.2016.11.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Patterson AG, Yevstigneyeva MS, Peter C. Regulation of CRISPR–Cas adaptive immune systems. Curr Opin Microbiol. 2017;37:1–7. doi: 10.1016/j.mib.2017.02.004. [DOI] [PubMed] [Google Scholar]
  44. Pawluk A, Bondy-denomy J, Cheung VHW, Maxwell KL, Davidson R. A New Group of Phage Anti-CRISPR Genes Inhibits the Type I-E CRISPR-Cas System of Pseudomonas aeruginosa. 2014;5:1–7. doi: 10.1128/mBio.00896-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pawluk A, Staals RHJ, Taylor C, Watson BNJ, Saha S, Fineran PC, Maxwell KL, Davidson AR. Inactivation of CRISPR-Cas systems by anti-CRISPR proteins in diverse bacterial species. Nat Microbiol. 2016a;1:16085. doi: 10.1038/nmicrobiol.2016.85. [DOI] [PubMed] [Google Scholar]
  46. Pawluk A, Amrani N, Zhang Y, Sontheimer EJ, Maxwell KL, Davidson AR, Pawluk A, Amrani N, Zhang Y, Garcia B, et al. Naturally Occurring Off-Switches for CRISPR-Cas9 Article Naturally Occurring Off-Switches for CRISPR-Cas9. Cell. 2016b;167:1–10. doi: 10.1016/j.cell.2016.11.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, Lim WA. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell. 2013;152:1173–1183. doi: 10.1016/j.cell.2013.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Rambo RP, Tainer JA. Accurate assessment of mass, models and resolution by small-angle scattering. Nature. 2013;496:477–481. doi: 10.1038/nature12070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rauch BJ, Silvis MR, Hultquist JF, Waters CS, McGregor MJ, Krogan NJ, Bondy-Denomy J. Inhibition of CRISPR-Cas9 with Bacteriophage Proteins. Cell. 2017;168:150–158.e10. doi: 10.1016/j.cell.2016.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Semenyuk AV, Svergun DI. GNOM. A program package for small-angle scattering data processing. J Appl Crystallogr. 1991;24:537–540. doi: 10.1107/S0021889812007662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Shin J, Jiang F, Liu JJ, Bray NL, Rauch BJ, Baik SH, Nogales E, Bondy-Denomy J, Corn JE, Doudna JA. Disabling Cas9 by an anti-CRISPR DNA mimic. bioRxiv. 2017 doi: 10.1126/sciadv.1701620. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Shmakov S, Smargon A, Scott D, Cox D, Pyzocha N, Yan W, Abudayyeh OO, Gootenberg JS, Makarova KS, Wolf YI, et al. Diversity and evolution of class 2 CRISPR–Cas systems. Nat Rev Microbiol. 2017;15:169–182. doi: 10.1038/nrmicro.2016.184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Stamatakis A. RAxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics. 2014;30:1312–1313. doi: 10.1093/bioinformatics/btu033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sternberg SH, LaFrance B, Kaplan M, Doudna JA. Conformational control of DNA target cleavage by CRISPR-Cas9. Nature. 2015;527:1–14. doi: 10.1038/nature15544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Terwilliger TC. Reciprocal-space solvent flattening. Acta Crystallogr. Sect D Biol Crystallogr. 1999;55:1863–1871. doi: 10.1107/S0907444999010033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Vasu K, Nagaraja V. Diverse functions of restriction-modification systems in addition to cellular defense. Microbiol Mol Biol Rev. 2013;77:53–72. doi: 10.1128/MMBR.00044-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wang X, Yao D, Xu JG, Li AR, Xu J, Fu P, Zhou Y, Zhu Y. Structural basis of Cas3 inhibition by the bacteriophage protein AcrF3. Nat Struct Mol Biol Advance on. 2016:4–7. doi: 10.1038/nsmb.3269. [DOI] [PubMed] [Google Scholar]
  58. Wei Y, Terns RM, Terns MP, Terns MP, Terns MP. Cas9 function and host genome sampling in type II-A CRISPR–cas adaptation. Genes Dev. 2015;29:356–361. doi: 10.1101/gad.257550.114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Yamada M, Watanabe Y, Gootenberg JS, Hirano H, Ran FA, Nakane T, Ishitani R, Zhang F, Nishimasu H, Nureki O. Crystal Structure of the Minimal Cas9 from Campylobacter jejuni Reveals the Molecular Diversity in the CRISPR-Cas9 Systems. Mol Cell. 2017;65:1109–1121.e3. doi: 10.1016/j.molcel.2017.02.007. [DOI] [PubMed] [Google Scholar]
  60. Zhang Y, Rajan R, Seifert HS, Mondragón A, Sontheimer EJ. DNase H Activity of Neisseria meningitidis Cas9. Mol Cell. 2015;60:242–255. doi: 10.1016/j.molcel.2015.09.020. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1. Figure S1. AcrIIC1 inhibits diverse Cas9 orthologs, while AcrIIC2 and AcrIIC3 are highly specific, related to Figure 1.

(A) Unrooted phylogenetic tree of AcrIIC1.

(B) Kinetic measurement of DNA cleavage mediated by GeoCas9 in the presence or absence of type IIC Acrs.

(C) Genome editing mediated by NmeCas9, SpyCas9 and CjeCas9 in the presence of various Acrs. Human (HEK293T) cells were transfected with plasmids expressing NmeCas9, SpyCas9, or CjeCas9, along with cognate, previously validated sgRNAs targeting genomic sites. T7E1 digestion was used to detect editing. A Type I anti-CRISPR (AcrE2) was used as a negative control for inhibition. As reported previously, NmeCas9 genome editing (upper panel) was inhibited by AcrIIC1Boe, AcrIIC1Nme, and AcrIIC3Nme; full inhibition by AcrIIC2Nme in human cells generally requires higher amounts of cotransfected expression plasmid. Inhibition of SpyCas9 genome editing (middle panel) was observed only with AcrIIA4Lmo (Rauch et al., 2017). In contrast, CjeCas9 editing activity (bottom panel) was inhibited by AcrIIC1Boe and AcrIIC1Nme, but not by any of the other anti-CRISPRs.

(D) GeoCas9 RNP mediated editing of HEK293T cells in the presence and absence of AcrIIC1. Indels were analyzed by T7E1 digestions.

NIHMS952151-supplement-S1.pdf (2MB, pdf)
S2. Figure S2. Impact of AcrIIC1 and AcrIIC3 on Cas9 RNA and DNA binding, related to Figures 2 and 5.

(A) Filter-binding assays measuring the affinity of NmeCas9 to its guide in the presence and absence of AcrIIC1. Radiolabeled sgRNA was incubated with NmeCas9 in the presence (orange) or absence (black) of AcrIIC1. NmeCas9–sgRNA RNP formation was measured using a filter binding assay and fraction sgRNA bound was calculated and plotted against NmeCas9 concentration. When incubated with Cas9 prior to sgRNA binding, AcrIIC1 does not inhibit RNP formation.

(B) Equilibrium binding measurements of NmeCas9–sgRNA binding to dsDNA in the presence and absence of AcrIIC1, measured by fluorescence polarization.

(C) Equilibrium binding measurements of SpyCas9 in the presence and absence of AcrIIC3, related to Figure 5D.

NIHMS952151-supplement-S2.pdf (556.5KB, pdf)
S3. Figure S3. AcrIIC1 Binds to the Cas9 HNH domain, related to Figure 3.

(A) Superdex 200 10/300 traces of truncations used to identify the binding interface of AcrIIC1. Each trace included the indicated GeoCas9 truncation and excess AcrIIC1. Black asterisks indicate the fractions analyzed in Figure 3B. For construct #4 (REC lobe), multiple peaks resulted from bound contaminating nucleic acid species from the purification and both peaks were pooled and analyzed together.

(B) Superdex 200 10/300 traces for GeoCas9 complexed with the components indicated in the top right. Fractions indicated with asterisk were analyzed on SDS-PAGE (upper gel) gel and denaturing urea PAGE gel (lower gel) and components were added in the order listed.

(1, Apo GeoCas9; 2, GeoCas9 +AcrIIC1; 3, GeoCas9+sgRNA+AcrIIC1; 4, GeoCas9+sgRNA+DNA+AcrIIC1).

(C) (top) HNH domains between GeoCas9 (orange) and AnaCas9 (yellow) were swapped to create chimeric Cas9 proteins. AcrIIC1 was fused with GFP (to more easily visualize a change in elution volume once bound to Cas9) and run over a S200 size-exclusion column. Additionally, AcrIIC1–GFP detection at a wavelength of 495 nm offers another indication of change in AcrIIC1 elution volume. (bottom) SDS-PAGE gradient gel (4–20%) with protein samples from S200 elution peaks of chimeric Cas9 proteins incubated with AcrIIC1.

(D) Representative ITC trace for AcrIIC1 binding to the NmeHNH domain.

(E) Conservation of Cas9 mapped onto the approximate domain boundaries below using a non-redundant list of Cas9 orthologs from all Cas9 subtypes. HNH is highlighted in purple. Bar heights are proportional to protein identity, with yellow bars indicating highly conserved residues.

NIHMS952151-supplement-S3.pdf (9.9MB, pdf)
S4. Figure S4. Alignment of HNH domains and AcrIIC1 orthologs and structure of AcrIIC1–HNH, related to Figure 4.

(A) A multiple sequence alignment of selected HNH-domains. The multiple sequence alignment was generated using the extracted HNH domains of 6 Cas9 orthologs and the restriction enzyme from Pseudoaltermonas lipolytica (RE_HNH). Red boxes surround catalytic residues and black asterisks indicate catalytic residues involved in AcrIIC1 binding. Blue boxes surround other residues involved in AcrIIC1 interaction.

(B) A multiple sequence alignment of AcrIIC1 using 11 AcrIIC1 orthologs. Highest degree of conservation occurs within a beta barrel (marked by a grey arrow). Red boxes surround selected residues involved in the binding between the NmeCas9 HNH domain and AcrIIC1Nme. Red asterisks indicate AcrIIC1 residues that interact with the backbone of HNH. Black asterisks indicate AcrIIC1 residues that interact with catalytic residues of the HNH domain.

(C) Charged residues (depicted as sticks) surrounding the active site of the HNH domain (purple) form ionic and hydrogen binding interactions (depicted as dotted black lines) with AcrIIC1 (orange).

(D) Fold reduction in phage titer in response to GeoCas9 targeting of phage Mu in the presence of AcrIIC1 mutants. One representative plate is shown for each mutant tested. 10-fold serial dilutions of phage Mu lysate were spotted lawns of bacteria expressing the indicated AcrIIC1 mutant. The fold reductions shown in the bar graph were qualitatively evaluated from inspecting three replicates of each experiment.

(E) SDS-PAGE gel showing the expression levels of WT and mutant AcrIIC1s used for experiments in panel D. The approximate mass of AcrIIC1 is indicated on the right (ev, empty vector; E2, AcrIE2).

NIHMS952151-supplement-S4.pdf (4.7MB, pdf)
S5. Figure S5. Conservation of the Cas9 HNH domain in the context of full length Cas9, related to Figure 4.

(A) Crystal structure of NmeCas9 HNH domain bound to AcrIIC1 (PDB: 5VGB). The HNH is rotated 90˚C to show the active site interface with labeled residues (depicted as sticks) involved in AcrIIC1 binding. (Bottom) Crystal structures of the HNH domains from three Cas9 orthologs (Staphylococcus aureus, PDB: 5CZZ; Streptococcus pyogenes, PDB: 4CMP; Actinomyces naeslundii, PDB: 4OGE). RMSD values were generated using super alignment in PyMol. Comparison of crystallized Cas9 orthologs reveals strong similarity between AcrIIC1 interacting residues of NmeCas9 and SauCas9 HNH domains.

(B) (Left) Model of AcrIIC1 inhibiting cleavage of both target and non-target strands. NmeCas9 HNH domain (purple) was modeled into a “docked” position using the dsDNA-bound SpyCas9 structure (PDB: 5F9R) as a reference. Placement of AcrIIC1 (orange) between the HNH domain and the target strand (red) prevents target cleavage and activation of the RuvC domain for non-target strand (dark blue) cleavage. The black box shows a zoomed-in view of the NmeHNH-AcrIIC1 complex. (Right) Model of AcrIIC1 clashing with the RuvC domain when the HNH domain is in an “undocked” conformation. NmeCas9 HNH was placed in a undocked conformation using a GeoCas9 Phyre model as a reference. Placement of AcrIIC1 indicates steric clashing with the RuvC domain, indicating that binding of AcrIIC1 to the HNH domain must position the HNH domain between the docked and undocked position. The black box shows a zoomed-in view of the HNHNme-AcrIIC1Nme complex.

NIHMS952151-supplement-S5.pdf (12.2MB, pdf)
S6. Figure S6. AcrIIC3 is monomeric and dimerizes NmeCas9, related to Figure 5.

(A) Small-angle X-ray scattering (SAXS) curves of NmeCas9, NmeCas9+AcrIIC1 and AcrIIC3.

(B) Analysis of AcrIIC1 (yellow, top) and AcrIIC3 (blue, below) on a Superdex 200 10/300 size exclusion column.

(C) Native mass spectrometry of AcrIIC3. The estimated masses from deconvoluting the charge series are identified in the top right corner.

NIHMS952151-supplement-S6.pdf (903KB, pdf)
Table S1

Table S1. Data collection and refinement statistics, related to Figure 4.

NIHMS952151-supplement-Table_S1.docx (95.1KB, docx)

RESOURCES