CRISPR RNA maturation by trans-encoded small RNA and host factor RNase III.

CRISPR/Cas systems constitute a widespread class of immunity systems that protect bacteria and archaea against phages and plasmids, and commonly use repeat/spacer-derived short crRNAs to silence foreign nucleic acids in a sequence-specific manner. Although the maturation of crRNAs represents a key event in CRISPR activation, the responsible endoribonucleases (CasE, Cas6, Csy4) are missing in many CRISPR/Cas subtypes. Here, differential RNA sequencing of the human pathogen Streptococcus pyogenes uncovered tracrRNA, a trans-encoded small RNA with 24-nucleotide complementarity to the repeat regions of crRNA precursor transcripts. We show that tracrRNA directs the maturation of crRNAs by the activities of the widely conserved endogenous RNase III and the CRISPR-associated Csn1 protein; all these components are essential to protect S. pyogenes against prophage-derived DNA. Our study reveals a novel pathway of small guide RNA maturation and the first example of a host factor (RNase III) required for bacterial RNA-mediated immunity against invaders.

Organisms of all kingdoms of life have evolved RNA-guided immunity mechanisms to protect themselves against genome invaders1-6. In bacteria and archaea, CRISPR/Cas (clustered, regularly interspaced short palindromic repeats/CRISPR-associated proteins) constitutes an adaptive RNA-mediated defence system which targets invading phages or plasmids in three steps: (1) adaptation via integration of viral or plasmid DNA-derived spacers into the CRISPR locus, (2) expression of short guide crRNAs (CRISPR RNAs) consisting of unique single repeat-spacer units and (3) interference with the invading cognate foreign genomes by mechanisms that are yet to be fully understood7-27. A key event in CRISPR activation is the maturation of the active crRNAs from the CRISPR precursor transcript (pre-crRNA)28,29. Three Cas proteins, Cse3 (CasE), Cas6 and Csy4, have been identified as endoribonucleases that cleave within the repeat sequences of pre-crRNA to generate the mature cRNAs28-31. However, their homologues are missing in many CRISPR/Cas subtypes, suggesting the existence of alternate crRNA maturation pathways involving other Cas proteins and/or fundamentally different RNA processing events. Here, our study of the human pathogen Streptococcus pyogenes uncovered a novel pathway of CRISPR activation wherein a trans-encoded small RNA, the host endoribonuclease III and the CRISPR-associated Csn1 protein are responsible for the production of the active crRNAs.

CRISPR/Cas systems in S. pyogenes

Our analysis of S. pyogenes genome sequences revealed the presence of two CRISPR/Cas loci of two different subtypes, CRISPR01 (system II, Nmeni/CASS4 subtype) and CRISPR02 (system I-C, Dvulg/CASS1 subtype)32,33, each having a distinct set of repeats and cas genes (Fig. 1a, Supplementary Fig. 1 and Supplementary Table 1). Almost all of the associated CRISPR spacers show homology to chromosomal prophage sequences34-38 (Supplementary Tables 2-5), indicating that the CRISPR/Cas systems of S. pyogenes target lysogenic phages.

Figure 1

A newly identified tracrRNA is required for crRNA maturation in S. pyogenes

a, dRNA-seq reveals expression of tracrRNA and crRNAs. Sequence reads of cDNA libraries derived from untreated and TEX-treated total RNA are shown. Vertical axis, relative amounts of sequenced cDNAs. The absence of ~75 nt tracrRNA form and 39-42 nt crRNA fragments in the TEX-treated cDNA library indicates that they are generated by processing. Genomic organization of tracrRNA and CRISPR01/Cas (csn1-cas1-cas2-csn2) loci. Transcriptional start sites and a terminator are indicated. (Left) tracrRNA (red) is encoded on the minus strand and detected as 171, 89 and ~75 nt tracrRNA species. Black rectangle, 36 nt sequence stretch complementary to CRISPR01 repeat. (Right) pre-crRNA is encoded on the plus strand. Rectangles, CRISPR01 repeats; diamonds, CRISPR01 spacers; 511, 66 and 39-42 nt, pre-crRNA and processed crRNAs. b, Base-pairing of tracrRNA with a CRISPR01 repeat is represented. Cleavages observed by dRNA-seq and leading to the formation of short overhangs at the 3′ ends of the processed RNAs are indicated by black arrows. Open arrow, cleavage in the spacer sequence. c, tracrRNA and pre-crRNA are co-processed in vivo. Northern blot analysis of S. pyogenes total RNA: strains and probes are indicated (Supplementary Figs 2 and 4). Left panel: Processing of tracrRNA into the ~75 nt form is abolished in Δpre-crRNA and re-established upon complementation with pre-crRNA. Right panel: Processing of pre-crRNA into mature crRNA forms (39-42 nt) is abrogated in ΔtracrRNA. Trans complementation of ΔtracrRNA with 171 or 89 nt tracrRNA restores the processing.

To examine the in vivo expression of CRISPR01 and CRISPR02, we analysed S. pyogenes strain SF370 (M1 serotype) by differential RNA sequencing (dRNA-seq)39. The most abundantly recovered small RNA species were CRISPR01 crRNAs originating from a ~511 nt pre-crRNA (Fig. 1a; Supplementary Fig. 2a, b and Supplementary Table 6), confirming that the CRISPR01 locus is active. In contrast, the CRISPR02 locus seems not to be expressed (Supplementary Fig. 3).

We detected six crRNAs from CRISPR01 which were 39 to 42 nt in length and likely processed species, as judged by their depletion in the dRNA-seq library for primary transcripts. The individual crRNAs appeared to result from double cleavage, one within the repeat and the other within the spacer, and carry a 20 nt spacer-derived 5′-guide sequence and a 19-22 nt repeat-derived 3′-sequence (Supplementary Fig. 2a, b). The latter sequence is distinct from the crRNA-tag (8 nt of the upstream repeat sequence) located in 5′ of mature crRNAs produced by the Cse3 (CasE) and Cas6-encoding CRISPR/Cas subtypes of Escherichia coli, Pyrococcus furiosus and Staphylococcus epidermidis28,31, providing evidence for the diversity of crRNA-tags among CRISPR/Cas systems and perhaps also for the underlying crRNA maturation and immunity mechanisms

tracrRNA directs pre-crRNA processing

Strikingly, dRNA-seq also detected abundant RNA species transcribed 210 nt upstream, on the opposite strand of the CRISPR01-associated genes and the leader-repeat-spacer array (Fig. 1a, b; Supplementary Fig. 4a, b and Supplementary Table 6); we refer to these abundant transcripts as tracrRNA (-activating CRISPR RNA). Northern blot probing detected four tracrRNA forms with approximate lengths of 171, 89, 75 and 65 nt, all of which were present throughout growth, notwithstanding a slightly decreased abundance of the longer transcripts in late stationary phase (Fig. 1c and Supplementary Fig. 4b-d). According to our dRNA-seq data, the 171 and 89 nt forms corresponded to primary transcripts whereas the shorter ~75 nt species resulted from processing of those longer tracrRNAs (Fig. 1a and Supplementary Fig. 4b).

Remarkably, both the 171 and 89 nt tracrRNAs contain a 25 nt stretch with almost perfect (one mismatch) complementarity to all repeats of CRISPR01 (Fig. 1b and Supplementary Fig. 5), predicting their potential base-pairing with pre-crRNA. Moreover, the tracrRNA and pre-crRNA processing sites detected by dRNA-seq fell in the putative RNA duplex region, indicative of co-processing of the two RNAs upon pairing. In support of this prediction, tracrRNA processing into the ~75 nt form was absent in a Δpre-crRNA mutant. Conversely, we did not detect mature crRNAs in a ΔtracrRNA strain, suggesting that tracrRNA is essential for the processing of pre-crRNA (Fig. 1c, Supplementary Figs. 2c and 4c). Trans complementation with the long tracrRNA species restored pre-crRNA processing in ΔtracrRNA bacteria, and revealed that the 89 nt form of tracrRNA suffices for co-processing (Fig. 1c and Supplementary Fig. 2c). Together, these findings reveal a novel function of a bacterial non-coding RNA such that a trans-encoded small RNA (tracrRNA) directs the maturation of another non-coding RNA (pre-crRNA) to yield the active species (crRNAs).

crRNA maturation requires RNase III and Csn1

According to our dRNA-seq data, the co-processed tracrRNA and pre-crRNA carry short 3′ overhangs reminiscent of cleavage by the endoribonuclease RNase III22,40-44 or the related eukaryotic Dicer and Drosha enzymes1,4-6,45. Since none of the Cas proteins of CRISPR01 contains an RNase III-like motif, we hypothesized that the endogenous RNase III—a general RNA processing factor40,42,46 encoded by the conserved rnc gene of the host—was recruited to cleave tracrRNA and pre-crRNA upon base-pairing. In support of our prediction, tracrRNA and pre-crRNA co-processing was abrogated in an Δrnc mutant of S. pyogenes (Fig. 2), yet restored by trans-complementation of RNase III expression (Supplementary Fig. 6).

Figure 2

Co-processing of tracrRNA and pre-crRNA requires both endogenous RNase III and Csn1 in vivo

Northern blot analysis of tracrRNA (a) and pre-crRNA (b) expression: strains and probes are indicated (Supplementary Figs 6 and 8). Processing of tracrRNA (a) into a ~75 nt form and pre-crRNA (b) into 39-42 nt mature crRNA forms is abolished in Δrnc, Δcas/csn and Δcsn1 (refer to Supplementary Figs 6 and 8).

To directly demonstrate that the paired RNAs are substrates of this nuclease, tracrRNA and pre-crRNA were synthesized, annealed in vitro and incubated with E. coli RNase III. Whereas neither of the two RNAs alone was cut by the nuclease, their annealing promoted the expected singular RNase III cleavage in either RNA (Fig. 3a-c). Consistent with their shared complementarity to CRISPR01 repeats, both the 171 nt and 89 nt tracrRNAs promoted RNase III cleavage of pre-crRNA within the repeat to produce intermediate crRNA species, and both were converted to the ~75 nt tracrRNA species in the process (Fig. 3b, c). Mutations in the complementarity regions of tracrRNA or pre-crRNA hindered co-processing with the respective wild-type RNA partner, yet RNase III cleavage was fully restored when the compensatory tracrRNA and crRNA mutants were combined (Supplementary Fig. 7), corroborating that RNA duplex formation underlies the observed processing. Others have noticed that the repeats of CRISPR/Cas subtype II (Nmeni/CASS4) lack the potential to form stem-loop structures47; our findings suggest that tracrRNA overcomes this deficiency by providing an inter-molecular RNA structure for pre-crRNA processing. Taken together, RNase III serves as a host factor in tracrRNA-mediated crRNA maturation, and constitutes the first example of a non-Cas protein that is recruited to CRISPR activity.

Figure 3

tracrRNA directs pre-crRNA cleavage by RNase III in vitro

a, Schematic representation of tracrRNA89 corresponding to 89-nt long tracrRNA, and crRNA213 and crRNA148 corresponding to a 213-nt long leader-repeat-spacer1-repeat-spacer2 fragment and a 148-nt long spacer1-repeat-spacer2-repeat-spacer3 fragment, respectively. b, Identification of tracrRNA89 binding sites on crRNA148*. 5′ end-labeled crRNA148* (~10 nM) was subjected to lead(II), RNase III and RNase T1 cleavage in the absence (lanes 1, 4, 7) and presence of cold tracrRNA89 (final concentration in lanes 2, 5, and 7: ~50 nM; lanes 3, 6, and 9: ~500 nM). Lane C: untreated crRNA148*; Lane T1: RNase T1 digest of crRNA148* under denaturating conditions; Lane OH: alkaline ladder; cleaved G residues are labeled. Vertical bars: crRNA148 region protected by tracrRNA89. Arrows denote specific RNase III cleavages in the two repeat regions of crRNA148 in the presence of tracrRNA89. c, Identification of crRNA148 and crRNA213 binding sites on tracrRNA89. 5′ end-labeled tracrRNA89* (~10 nM) was subjected to RNase T1, lead(II) and RNase III cleavage in the absence (lanes 1, 6, 11) and presence of cold crRNA148 or crRNA213 (final concentration in lanes 2, 4, 7, 9, 12 and 14: ~50 nM; lanes 3, 5, 8, 10, 13 and 15: ~500 nM). Lanes C, T1 and OH, positions of cleaved Gs and vertical bars: as above but referring to tracrRNA89* in the presence of cold crRNA148 or crRNA213.

Next, we entertained the possibility that—in addition to RNase III—Cas proteins facilitate the co-processing of the duplex RNA in vivo. Intriguingly, deletion of the csn1-cas1-cas2-csn2 operon impaired the processing of both tracrRNA and pre-crRNA (Fig. 2 and Supplementary Fig. 8a). In-frame deletions of any of the operon’s four genes then revealed Csn1 as the only Cas protein required for the production of mature crRNAs and concomitant tracrRNA cleavage. This was further supported by restored tracrRNA and pre-crRNA processing upon ectopic expression of Csn1 in Δcas-csn or Δcsn1 mutants (Fig. 2 and Supplementary Fig. 8b-d).

Csn1 (or COG3513) is a large, likely multi-domain protein32,33 of unknown function except that it is essential for CRISPR-mediated immunity in Streptococcus thermophilus8. Here, we propose a model wherein Csn1 acts as a molecular anchor facilitating the base-pairing of tracrRNA with pre-crRNA for subsequent recognition and cleavage of pre-crRNA repeats by the host RNase III (Fig. 4). Because Csn1 has predicted motifs of RuvC-like (RNase H fold) and McrA/HNH nucleases32,33, it might also mediate the second cleavage to occur at a fixed distance within the spacers. Furthermore, Csn1 might help protect tracrRNA and pre-crRNA against other host ribonucleases, as suggested by the strongly reduced accumulation of tracrRNA in the absence of csn1 (Fig. 2 and Supplementary Fig. 8a, b). Collectively, our results show that in the absence of Cse3 (CasE), Cas6 or Csy4 proteins, CRISPR01 crRNA maturation is achieved by the concerted action of three novel factors, a trans-encoded small RNA, a host-encoded ribonuclease and a Cas protein previously not implicated in pre-crRNA cleavage.

Figure 4

Model for tracrRNA-mediated crRNA maturation involving RNase III and Csn1

Black, repeat; green, spacer. tracrRNA can bind with almost perfect complementarity to each repeat sequence of the pre-crRNA. The resulting RNA duplex is recognized and site-specifically diced by RNase III in the presence of Csn1, releasing the individual repeat-spacer-repeat units (1st processing event). The generated units undergo further processing within the spacer sequences resulting in mature crRNA species consisting of unique spacer-repeat sequences (2nd processing event) by a yet-to-be elucidated mechanism. Csn1 may also be involved in the silencing of invading sequences.

CRISPR immunity against prophage sequences

To further investigate the role of tracrRNA in CRISPR01-mediated immunity, we developed a plasmid-based read-out system that mimics infection with protospacer-containing lysogenic phages (a protospacer is a DNA target sequence that matches a CRISPR spacer). We assayed transformation rates of a plasmid carrying a protospacer of the speM exotoxin gene, expected to be a target because of 100% identity to the second spacer of CRISPR01 (Spyo1h_002; Supplementary Table 2). Consistent with this protospacer being recognized by CRISPR01, wild-type S. pyogenes was protected from plasmids containing the speM gene, with or without its endogenous promoter region. Protection was specific since the wild-type strain was readily transformed with variants of the parental backbone plasmid as control (Fig. 5 and Supplementary Fig. 9). Importantly, in contrast to the wild-type strain, the Δpre-crRNA, ΔtracrRNA, Δrnc and Δcsn1 mutants invariably tolerated the speM plasmid (Fig. 5 and Supplementary Fig. 9). Together, these results demonstrate that tracrRNA, RNase III and Csn1 are essential in CRISPR01-mediated immunity of S. pyogenes against lysogenic phages, and further suggest that the tracrRNA/CRISPR01/Cas system, in concert with the host RNase III, limits horizontal virulence gene transfer among pathogenic streptococcal species34-38.

Figure 5

Both tracrRNA and pre-crRNA confer immunity against acquisition of a protospacer gene derived from a lysogenic phage

Transformation efficiencies of S. pyogenes with speM protospacer containing plasmid (pEC287) and reference “backbone” plasmid (pEC85) (Supplementary Fig. 9). Strains: S. pyogenes WT (SF370), Δpre-crRNA, ΔtracrRNA, Δrnc and Δcsn1. Graph bars, mean values of colony forming units (CFU) per μg of plasmid DNA; error bars, standard deviation (SD); n≥3. pEC287 is tolerated by the Δpre-crRNA, ΔtracrRNA, Δrnc and Δcsn1 mutants but not by the WT strain. As control, transformants in all strains were obtained with the backbone plasmid (Supplementary Fig. 9).

tracrRNA homologues in CRISPR/Cas systems

How widely spread is the tracrRNA-mediated CRISPR activation? Sequence analysis revealed anti-CRISPR repeat sequences, thus candidate tracrRNA homologues, in the vicinity of system II (Nmeni/CASS4) CRISPR/Cas loci of other bacterial genomes (Supplementary Table 7). We probed selected loci of Listeria innocua, Neisseria meningitidis, Streptococcus mutans and S. thermophilus (Fig. 6 and Supplementary Figs 12-16) and consistently observed both expression and processing of the homologous tracrRNAs and respective pre-crRNAs (Supplementary Figs 12-16). In addition, our analysis indicates potential co-evolution of the tracrRNA anti-repeat and CRISPR repeat sequences (Supplementary Table 7 and Supplementary Fig. 11). Thus, RNA base-pairing might generally determine crRNA maturation in type II CRISPR/Cas systems, and based on RNA probing results, these systems seem to be constitutively activated to target and affect the maintenance of invader genomes.

Figure 6

tracrRNA-mediated crRNA maturation is conserved among different bacterial species

tracrRNA-mediated crRNA maturation is inherent to the type II (Nmeni/CASS4) CRISPR/Cas systems. Type II (Nmeni/CASS4) loci from S. pyogenes SF370, S. mutans UA159, L. innocua Clip11262, N. meningitidis Z2491 and S. thermophilus LMD-9 (Nmeni/CASS4a); red, tracrRNA; rectangles, repeats; diamonds, spacers.

No putative tracrRNA homologue was found in the vicinity of other CRISPR/Cas subtypes, and the two additional degenerated repeats identified near the type III-A (Mtube/CASS6) CRISPR/Cas locus in S. epidermidis RP62a25 lacked a corresponding tracrRNA homologue (Supplementary Fig. 17). Thus, the requirement of a trans-encoded small RNA for pre-crRNA processing into active crRNAs is a general RNA maturation mechanism shared by the type II (Nmeni /CASS4) CRISPR/Cas systems that lack the cse3 (casE), cas6 or csy4 gene but possess csn1. Whether all of the type II CRISPR/Cas loci require RNase III as a host factor remains to be seen.

In summary, trans RNA-mediated activation of crRNA maturation to confer sequence-specific immunity against parasite genomes represents a novel RNA maturation pathway, and highlights the unfolding diversity and complexity of molecular mechanisms of CRISPR/Cas systems9-14,26,28,29. Importantly, CRISPR loci have been generally regarded as autonomous genetic units, encoding all the proteins and RNAs required for their activity. Our identification of RNase III as the first host factor in CRISPR activity raises the exciting possibility that a recruitment of non-Cas proteins from the host chromosome might contribute to the rapid evolutionary diversification of CRISPR/Cas systems.

We suggest that Csn1 together with RNase III forms a microprocessor complex responsible for tracrRNA-mediated pre-crRNA processing (Fig. 4). As such, the requirement of RNase III in the process seems reminiscent of the key roles of related nucleases (Dicer, Drosha) in eukaryotic RNA-protein complexes that mediate the production of small interfering RNAs and maturation of microRNAs. However, the eukaryotic pathways employing RNase III-like enzymes for pre-RNA processing do not rely on trans-encoded RNA factors. More studies are needed to determine whether an RNase III-mediated activation of a small effector RNA by co-processing with a trans-acting non-coding RNA is also used in other biological systems.

METHODS SUMMARY

Details of bacteria (culture conditions, transformation), DNA (biocomputational analysis, plasmid construction, in-frame gene deletion mutants) and RNA manipulation (cDNA library construction [vertis Biotechnologie AG], RNA expression analysis, RNA structure probing and hydrolysis) are provided as Supplementary Information34,39,48. In brief, half of DNase I-treated SF370 total RNA was enriched for primary transcripts by treatment with the Terminator™ 5′-phosphate-dependent exonuclease (TEX) (Epicentre), which degrades RNAs with a 5′P (processed RNAs) but not primary transcripts with a 5′PPP RNA39. cDNA librairies were constructed from both untreated and TEX-treated RNA39. Following 454 pyrosequencing, cDNAs were mapped to the SF370 genome and visualized using the Affymetrix Integrated Genome Browser39. Strains, plasmids and primers are listed in Supplementary Tables 8, 9 and 10, respectively.