Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Mar 1;69(5):893-905.e7.
doi: 10.1016/j.molcel.2018.01.032.

CRISPR RNA-Dependent Binding and Cleavage of Endogenous RNAs by the Campylobacter Jejuni Cas9

Free PMC article

CRISPR RNA-Dependent Binding and Cleavage of Endogenous RNAs by the Campylobacter Jejuni Cas9

Gaurav Dugar et al. Mol Cell. .
Free PMC article


Cas9 nucleases naturally utilize CRISPR RNAs (crRNAs) to silence foreign double-stranded DNA. While recent work has shown that some Cas9 nucleases can also target RNA, RNA recognition has required nuclease modifications or accessory factors. Here, we show that the Campylobacter jejuni Cas9 (CjCas9) can bind and cleave complementary endogenous mRNAs in a crRNA-dependent manner. Approximately 100 transcripts co-immunoprecipitated with CjCas9 and generally can be subdivided through their base-pairing potential to the four crRNAs. A subset of these RNAs was cleaved around or within the predicted binding site. Mutational analyses revealed that RNA binding was crRNA and tracrRNA dependent and that target RNA cleavage required the CjCas9 HNH domain. We further observed that RNA cleavage was PAM independent, improved with greater complementarity between the crRNA and the RNA target, and was programmable in vitro. These findings suggest that C. jejuni Cas9 is a promiscuous nuclease that can coordinately target both DNA and RNA.

Keywords: CRISPR; Campylobacter jejuni; Cas9; RIP-seq; RNA binding proteins; RNA cleavage; crRNA; genome editing; non-coding RNA; post-transcriptional regulation.

Conflict of interest statement


C.L.B. is a co-founder and scientific advisory board member of Locus Biosciences and previously submitted provisional patent applications on CRISPR technologies.


Figure 1
Figure 1. Cas9 co-immunoprecipitates with cellular RNAs in C. jejuni
(A) Western blot analysis of coIP samples of C. jejuni WT and cas9-3xFLAG strains using anti-FLAG antibody. The amount of protein samples loaded according to OD600 of bacteria is indicated. GroEL served as loading control. (B) RNA-seq reads from the control and Cas9-3xFLAG coIP libraries mapped to the CRISPR locus of C. jejuni strain NCTC11168. Each of the five repeats carries a -10 promoter element that drives transcription of the downstream crRNA(s). Black arrows indicate transcriptional start sites from Dugar et al., 2013. C. jejuni crRNAs are ~37 nt in length, with ~24 nt of spacer-derived guide sequence and ~13 nt of repeat sequence, which remains base paired with tracrRNA upon processing (Dugar et al., 2013). The sequence in orange corresponds to the processed part of the tracrRNA and the full interaction of the crRNA repeat sequence with tracrRNA is shown below along with the RNase III-dependent cleavage site. Enrichment values of the four crRNAs and the tracrRNA are reported representing the relative reads of each RNA in the Cas9-3xFLAG coIP versus the control coIP (see also Table S2). (C) Enrichment and genomic location of RNA peaks. CjCas9 dependent peaks were identified based on >5-fold enrichment of cDNA read counts in the Cas9-3xFLAG coIP compared to the control coIP followed by manual curation (see Methods). RNA peaks are colored based on common RNA classes. The two most enriched peaks (truA, +775157; Cj0571 antisense, -533042) can be attributed to few reads in the control coIP (see Table S2).
Figure 2
Figure 2. CjCas9 binds endogenous RNAs through base-pairing with crRNAs
(A) Consensus motifs for two subsets of enriched peaks predicted by MEME. The complementarity of the respective motifs with either crRNA3 (left) or crRNA4 (right) is depicted below. Base pairing between the crRNA repeat part and tracrRNA is also shown. (B) Examples of enriched peaks in Cj1321 mRNA and nifU mRNA in the Cas9-3xFLAG coIP compared to the control coIP that are predicted to base pair with crRNA3 (left, Cj1321 mRNA) or crRNA4 (right, nifU mRNA). The cj1321 start codon is highlighted in bold. RNA-seq reads of the Cas9-3xFLAG coIP and the control coIP for the C. jejuni NCTC11168 crRNA3-deletion mutant are also shown. (C) Predicted binding affinities of the enriched RNA peaks to each of the four crRNA guides predicted by NUPACK. Each dot represents the predicted binding affinity between the RNA peak and the crRNA, where the color indicates the crRNA (crRNA1, gray; crRNA2, yellow-green; crRNA3, turquoise; crRNA4, dark blue). RNA peak-crRNA pairs with no predicted base-pairing interactions appear at the bottom. RNA peaks are ordered based on the crRNA guide with the highest predicted binding affinity. RNA peaks that are not predicted to bind to any crRNA are shown on the far right. Highlighted in peach are RNA peaks that were not enriched more than 5-fold in the coIP experiment with the crRNA3-deletion strain. Values above each grouped set of RNA peaks indicate the percentage of RNA peaks that are highlighted in peach. Asterisks designate two RNA peaks (in the fumC and kpsM mRNAs) that shifted between the WT and ΔcrRNA3 coIP, resulting in loss of all predicted crRNA binding sites.
Figure 3
Figure 3. Some co-purified RNAs undergo CjCas9-dependent cleavage
(A) 5′ end visualization (red) of mapped cDNA reads from whole transcriptome libraries from C. jejuni NCTC11168 WT as well as Δcas9, ΔCRISPR-tracrRNA, ΔtracrRNA, and ΔcrRNA3 mutant strains. The mapped, full-length cDNA reads from the control and Cas9-3xFLAG coIP libraries are shown below. The black dotted lines mark the potential sites of Cas9-dependent mRNA cleavage. The light and dark blue boxes below indicate the predicted binding sites of crRNA3 and crRNA4, respectively. (B) Analysis of the 35 RNA peak cleavage positions relative to the predicted crRNA binding sites. The vertical axis indicates the number of RNA peaks. The horizontal axis indicates the distance (in nucleotides) between the 5′ end of the crRNA and the cleavage site in the RNA peak. A distance of 0 corresponds to a cleavage site in the RNA peak across from the 5′ end of the crRNA. Distances were determined by counting the number of bases between the cleavage site and the closest nt base paired with the crRNA in the RNA peak, and then taking this distance relative to the 5′ end of the crRNA (see Methods for details). The crRNA was selected that bound closest to the cleavage site. In all but one case, this crRNA exhibited the highest predicted binding affinity for that RNA peak. See Table S2 for binding locations and energetics. For three peaks (right), potential cleavage sites were detected, but no corresponding crRNA binding sites were predicted.
Figure 4
Figure 4. crRNA3-dependent cleavage of the Cj1321 mRNA requires the HNH domain of CjCas9
(A) RNA-seq reads from C. jejuni NCTC11168 differential RNA-seq (dRNA-seq) (Dugar et al., 2013), as well as coIP libraries mapped to the cj1321 locus. The dRNA-seq data comprise two cDNA libraries, generated from RNA with (+TEX) or without (-TEX) prior treatment with terminator exonuclease. The two processed and stably accumulating fragments R1 and R2 of Cj1321 mRNA are indicated by black bars along with the crRNA3-Cas9 dependent cleavage site. (B) Northern blot analysis of the Cj1321 mRNA, crRNA3, tracrRNA, and 5S rRNA in the indicated WT and CRISPR mutant strains. RNAs were probed with radiolabeled oligonucleotide probes and marker sizes are indicated on the left. (C) Northern blot and Western blot analysis of Cas9, GroEL and RNAs mentioned in (B) in strains complementing the Δcas9 deletion with different nuclease mutant versions as indicated on top. Cas9 complementations are expressed under control of the PmetK promoter in the unrelated rdxA locus. “+” indicates which Cas9 nuclease domain is still intact.
Figure 5
Figure 5. DNA targeting of the Cj1321 locus by CjCas9 requires a PAM and extensive crRNA complementarity
(A) Plasmid clearance assays in E. coli. Assays were conducted by encoding variants of the cj1321 locus in an E. coli plasmid and then measuring the transformation efficiency in cells expressing CjCas9 and the crRNA3 sgRNA or a non-targeting sgRNA. The variants included the natural locus (cj1321) as well as mutant variants, where the natural locus was mutated to be completely complementary to the crRNA3 guide (CC), to contain the consensus PAM (+PAM), or both. Values represent the geometric mean and S.E.M. of experiments from three separate colonies. (B) In vitro cj1321 DNA cleavage assay with CjCas9 (final concentration of 300 nM). The assay was conducted for 5 min with 63 bp-long cj1321 wildtype dsDNA construct or a mutant variant (final concentration of 4 nM) that includes a PAM sequence (cj1321 + PAM). Both DNA strands are 5′-end radiolabelled. One of three sgRNAs (final concentration of 300 nM) with either no (NC), full (CC), or partial complementarity (cr3, sequence of WT crRNA3) were provided. Base pairing interactions with cj1321 are indicated below each sgRNA sequence. The upper cleavage fragment corresponds to the non-target strand and the lower fragment to the target strand. The results are representative of three independent experiments.
Figure 6
Figure 6. Enhancing and reprogramming RNA cleavage by CjCas9 in vitro
(A) In vitro cleavage of 5′-end radiolabelled in vitro-transcribed Cj1321 mRNA (final concentration of 4 nM) with CjCas9 (final concentration of 1 μM) and guide RNAs (final concentration of 1 μM) from Figure 5B. The Cj1321 mRNA GUG start codon is marked in green, scissors indicate the cleavage site, and base pairing interactions with the Cj1321 mRNA are indicated above each sgRNA. RNase T1 indicates a ladder for G residues, OH for all nucleotides. The results are representative of three independent experiments. (B) In vitro cleavage of an in vitro-transcribed flaA mRNA fragment. Three sgRNAs targeting the fragment are shown in color. The sgRNA CC with full complementarity to Cj1321 mRNA from (A) does not bind to flaA mRNA and serves as a control. The results are representative of two independent experiments. (C) Model of double-stranded DNA and single-stranded RNA cleavage by CjCas9. (Left) Schematic drawing of dsDNA (green) cleavage (scissors) by the HNH and RuvC domains of CjCas9, that depends on the presence of a PAM (yellow) and full complementarity of the crRNA (blue) to the target strand of the dsDNA (green). (Right) PAM-independent ssRNA (Cj1321, red) cleavage by the Cas9 HNH domain with partial complementarity to the crRNA3 guide (blue).

Similar articles

See all similar articles

Cited by 22 articles

See all "Cited by" articles

Publication types

MeSH terms