crRNA and tracrRNA guide Cas9-mediated DNA interference in Streptococcus thermophilus

Genetic location of the tracrRNA-encoding sequence in the S. thermophilus DGCC7710 CRISPR3-Cas system

Sequence analysis of the DNA region upstream of the cas9 gene in the St-CRISPR3-Cas system revealed that, similarly to the S. pyogenes CRISPR-Cas system,²⁶ a putative tracrRNA (St-tracrRNA) is encoded upstream of the cas operon (Fig. 1A). In S. pyogenes, deep sequencing data²⁶ revealed 171 nt and 89 nt tracrRNAs (Sp-tracrRNAs) that would result from transcription at two distinct promoters to a shared transcriptional terminator (Fig. 1A). Northern blot analysis of the S. thermophilus LMD-9 strain, which contains a CRISPR3-Cas system nearly identical to that of S. thermophilus DGCC7710 CRISPR3-Cas system,³⁹ revealed a ~100 nt tracrRNA transcript.²⁶

To identify a putative transcriptional start position for the DGCC7710 St-tracrRNA transcript, we compared DNA fragments located between the translation initiation codon of Cas9 and the anti-repeat region of the tracrRNA (Fig. 1A). The distance between the cas9 start codon and tracrRNA is ~130 bp in DGCC7710 vs. ~200 bp in LMD-9 and ~300 bp in S. pyogenes. Within the shared frame, the corresponding DGCC7710 fragment is nearly identical to that of LMD-9. Therefore, we assumed that the transcription of St-tracrRNA (DGCC7710) could start at the same position as in LMD-9. The tracrRNA sequences (Fig. 1B) at the 3′-end were very similar (87% identical nucleotides) and contained Rho-independent transcription terminators. Based on this, we assumed that the estimated size of St-tracrRNA (DGCC7710) is ~100 nt, which would be consistent with the size of St-tracrRNA (LMD-9).²⁶

tracrRNA is necessary for in vivo DNA interference by St-CRISPR3-Cas

The pCRISPR3 plasmid, which carries a complete St-CRISPR3-cas locus, including a CRISPR3 array comprised of 12 repeat-spacer units and a tracrRNA-encoding fragment located upstream of the cas9 gene (Fig. 2A), provides interference against transformation of a donor pSP1 plasmid, which contains a proto-spacer identical to the SP1 spacer in the CRISPR3 array and the accompanying 5-NGGNG-3′ PAM sequence.^37,39 To establish whether tracrRNA is required for St-CRISPR3-Cas-mediated plasmid interference in E. coli, we generated pCRISPR3 plasmid variants with a compromised tracrRNA-encoding sequence (Fig. 2A). In the pCRISPR3-Δt plasmid, the entire tracrRNA coding sequence is deleted, while in the pCRISPR3-ΔtR variant the tracrRNA-encoding sequence is truncated at the 3′-end, to leave only the region from the transcription start site up to the end of the anti-repeat sequence (Fig. 2A). Next, we analyzed transformation efficiency of two recipient E. coli strains carrying tracrRNA-deficient pCRISPR3-Δt or pCRISPR3-ΔtR variants by the pSP1 donor plasmid. The pUC18 plasmid, which lacks a proto-spacer but contains multiple PAMs was used as a control in the plasmid transformation assay. We found that, in contrast to the pCRISPR3-carrying recipient strain, which was resistant to pSP1 transformation, the tracrRNA-deficient strains became permissive for transformation by pSP1 plasmid (Fig. 2B). Thus, the plasmid immunity provided by the heterologous St-CRISPR3-Cas system is compromised when either part of or the full tracrRNA-encoding sequence is eliminated, indicating that tracrRNA is necessary for Cas9-mediated interference.

To confirm that the St-CRISPR3-Cas-mediated plasmid interference is lost due to the tracrRNA gene deletion/truncation, we introduced a full-length tracrRNA in trans into the tracrRNA-deficient permissive recipient strain and evaluated pSP1 plasmid transformation efficiency. More specifically, the fragment encoding tracrRNA was cloned into the pCDF-DUET1 plasmid under T7 promoter control, the construct expressed in the pSP1 permissive E. coli BL21 (DE3) strain (Table S1) carrying the pCRISPR3-Δt plasmid (Fig. 2C) and transformation efficiency was evaluated by counting colonies on Ap, Str, Cm and IPTG-supplemented agar plates. Because the E. coli BL21 (DE3) strain is transformed less efficiently than RR1, 25 ng instead of 1 ng of pUC18 or pSP1 plasmid was used in the transformation assay. Under these conditions, we found that transformation by the pUC18 plasmid yielded ~1,000 colonies, while no colonies were obtained in the case of pSP1 plasmid (Fig. 2C). The same results were obtained in a recipient host, which carried the pCRISPR3 plasmid containing tracrRNA in cis (Fig. 2C). These results show that trans-complementation of tracrRNA converts a pSP1-permissive E. coli strain into a transformation-resistant (non-permissive) strain. Taken together, plasmid transformation assays demonstrate that tracrRNA is necessary for DNA interference provided by the St-CRISPR3-Cas effector complex, in vivo.

tracrRNA co-purifies with Cas9 protein

St-Cas9 is the sole Cas protein required for St-CRISPR3-Cas-mediated immunity.^37,39 St-Cas9 protein co-purifies with a 42 nt crRNA (Fig. 2D).³⁹ To probe whether tracrRNA also co-purifies with Cas9 and crRNA, we performed northern blot analysis using an anti-tracrRNA 36 nt oligodeoxynucleotide probe. Nucleic acids extracted from the Strep-Tactin-purified Cas9 preparation³⁹ hybridized with the anti-tracrRNA probe and were sensitive to RNase, but not to DNase treatment (Fig. 2D). The estimated size of the tracrRNA, which co-purified with Cas9 is ~65 nt. Minor amounts of longer tracrRNA intermediates were present (Fig. 2D). The northern blot revealed that Cas9 preparation expressed and purified in the absence of CRISPR3 repeat-spacer array sequence and tracrRNA, contains no crRNA or tracrRNA (Fig. S1).

Role of RNase III in pre-crRNA maturation

In S. pyogenes, crRNA maturation and effector complex assembly requires tracrRNA and a housekeeping RNase III.²⁶ During S. pyogenes crRNA maturation, the tracrRNA base pairs with pre-crRNA to form a double-stranded RNA that is recognized and cleaved by the host RNase III in the presence of Cas9.²⁶ Deletion of the rnc gene encoding RNase III in S. pyogenes abrogates DNA interference provided by the CRISPR-Cas system.

The E. coli RNase III is similar to St-RNaseIII (39.6% identical and 52.3% similar aa). To probe whether the E. coli RNase III is able to replace St-RNase III in the crRNA processing and contribute to the St-CRISPR3-Cas mediated immunity in the heterologous E. coli host, we performed pSP1 plasmid transformation assay in the rnc+ (E. coli RR1) and rnc- (E. coli HT115) strains⁴¹ (Table S1) carrying either a wild-type (with 12 distinct spacers in-between repeats) or a minimal (R-SP1-R) CRISPR locus (Figs. 2B and 3A, respectively). In-line with our previous findings,^37,39 the rnc+ host, carrying either a wild-type or a minimal CRISPR region, provided interference against pSP1 transformation with nearly equal efficiency. On the other hand, in the rnc- host, the transformation efficiency radically differed depending on whether the recipient strain carried a wild-type or a minimal (R-SP1-R) CRISPR array. The rnc- strain carrying a wild-type CRISPR array became permissive for pSP1 plasmid transformation while the rnc- strain carrying a minimal CRISPR array was non-permissive to pSP1 transformation (Fig. 3B). This indicates that RNase III is not necessary for the maturation of pre-crRNA containing a single CRISPR spacer. This reduces the cloning needs and system requirements for a single-target Cas9-mediated programmable cleavage.

In vitro reconstitution of the St-CRISPR3 Cas9-crRNA-tracrRNA effector complex

Analysis of the protein and nucleic acid content of the St-CRISPR3-Cas effector complex isolated from the heterologous E. coli strain revealed the presence of Cas9, a ~42 nt crRNA³⁹ and a ~65 nt tracrRNA (see above). Next, we aimed to reconstitute an effector complex in vitro, by combining these three individual components. To assemble a ternary Cas9-crRNA-tracrRNA complex (St-Cas9t), Cas9 was pre-incubated with equimolar amounts of a synthetic 42 nt crRNA and 78 nt tracrRNA corresponding to a mature form, obtained by in vitro transcription. Extra nucleotides (78 nt vs 65 nt) were introduced for in vitro transcription by a T7 RNA polymerase. The DNA cleavage activity of the reconstituted complex was monitored in vitro using the pSP1 plasmid, as described previously.³⁹ In the presence of a reconstituted ternary St-Cas9t complex, the pSP1 plasmid, which contains a proto-spacerSP1 sequence flanked by the 5′-NGGNG-3′PAM is converted into a linear form (Fig. 4A), indicating that both DNA strands are cleaved. If one of the complex components (Cas9, crRNA or tracrRNA) is missing, no cleavage of pSP1 is observed, indicating that all three components are necessary to form a functional St-Cas9t complex. On the other hand, the pUC18 plasmid which lacks the proto-spacer is not cleaved by the St-Cas9t. Hence, a catalytically competent Cas9-crRNA-tracrRNA ternary complex can be reconstituted in vitro by mixing individual components, similarly to the sp-Cas9 complex of S. pyogenes.³⁶

According to the plasmid interference assay, RNase III is not required for DNA interference in the E. coli host carrying a plasmid with a minimal R-SP1-R unit (Fig. 3). Therefore, we next investigated whether RNase III is important for St-Cas9t complex assembly in vitro. To assemble the complex, we first pre-incubated the Cas9 protein with an equimolar mixture of a 105 nt tracrRNA (101 nt putative tracrRNA transcript + 4 nt for in vitro transcription by the T7 RNA polymerase) and a 150 nt pre-crRNA (including 20 nt of the leader and a R-SP1-R-SP2 fragment, which contains the targeting spacer SP1 and a truncated SP2) transcripts, and then added RNase III. The pSP1 plasmid cleavage assay revealed that the complex assembled in vitro using pre-crRNA cleaved the pSP1 plasmid (Fig. 4B), generating a linear DNA form albeit at slower rate than a complex assembled using a 42 nt synthetic crRNA. Elimination of the RNase III from the reaction mix produced an active complex (Fig. 4B), consistent with in vivo data indicating that RNase III is not crucial for the assembly of the catalytically active complex when a short pre-crRNA (containing only one spacer) transcript is used.

Experiments with sp-Cas9 revealed that tracrRNA can be 3′-truncated in in vitro applications.³⁶ To define the minimal length of the 3′-end of tracrRNA required for crRNA-guided DNA cleavage by the St-Cas9t, we designed and generated a set of tracrRNA molecules truncated by stretches of 5 nt from the 3′ terminus. The overall size of tracrRNA variants varied between 33 and 78 nt (Fig. 4C). The truncated tracrRNA variants were used for in vitro reconstitution of the ternary St-Cas9t complex, followed by analysis of the pSP1 plasmid cleavage. We found that 38 nt and 33 nt tracrRNA variants do not support DNA cleavage by St-Cas9 (Fig. 4D), whereas 43, 48 and 53 nt variants showed decreased cleavage activity. On the other hand, tracrRNAs longer than 58 nt support efficient St-Cas9 cleavage of the pSP1 plasmid. This again reduces cloning requirements for a synthetic system, and simplifies the elements necessary for Cas9-mediated programmable cleavage.

Cas9 role in formation of the pre-crRNA:tracrRNA duplex

In the S. thermophilus CRISPR3-Cas and S. pyogenes effector complexes, the target site recognition is achieved by the crRNA while the Cas9 protein provides two active sites for the cleavage of opposite DNA strands in the proto-spacer.^36,39 Indirect evidence obtained in vivo in S. pyogenes suggests that Cas9 is also an essential protein for crRNA maturation and may facilitate formation and stabilization of the pre-crRNA:tracrRNA duplex.²⁶ However, experimental evidence and molecular details remain to be established.

To probe experimentally whether the S. thermophilus CRISPR3 system Cas9 protein promotes pre-crRNA and tracrRNA annealing, we produced tracrRNA3 (105 nt) and pre-crRNA3 (94 nt containing a single 36 nt repeat sequence) from CRISPR3 by in vitro transcription and monitored tracrRNA3:pre-crRNA3 duplex formation in the presence of Cas9 (Fig. 5). In a control set of experiments, tracrRNA1 (105 nt) and pre-crRNA1 (94 nt) from the homologous S. thermophilus DGCC7710 CRISPR1 system were used (Table S4). To monitor duplex formation, radioactively labeled tracrRNAs were incubated for 10 min with pre-crRNAs in the presence of a non-specific 2 kb RNA transcript and different amounts of the Cas9 protein of the CRISPR3 system. The reaction was quenched by adding Proteinase K and an excess of unlabeled tracrRNA and samples were analyzed using non-denaturing PAGE (Fig. 5). Under these experimental conditions, in the absence of Cas9, no tracrRNA:pre-crRNA duplex is formed. Increasing amounts of Cas9 promote tracrRNA3:pre-crRNA3 formation but have no effect with tracrRNA1 and pre-crRNA1 (Fig. 5). Taken together, these results indicate that Cas9 specifically promotes pre-crRNA3 and tracrRNA3 annealing and facilitates formation of the pre-crRNA3:tracrRNA3 duplex.

DNA manipulations

A DNA fragment encoding cas9 was PCR amplified from S. thermophilus DGCC7710 strain genomic DNA using primers GG-384 and GG-385 (Table S2) and pre-cleaved with Esp3I and XhoI. The resulting Esp3I-XhoI fragment was cloned into a pBAD24-CHis expression vector using NcoI and XhoI sites to generate a pBAD-Cas9 plasmid, which was used for the expression of the C-terminal (His)₆-tagged Cas9 protein variant. Full sequencing of cas9 in the pBAD-Cas9 plasmid revealed no difference with the original cas9 sequence.

To generate pCRISPR3-Δt and pCRISPR3-ΔtR plasmid variants lacking an entire tracrRNA-encoding gene or its 5′-terminal fragment, respectively, DNA fragments were PCR amplified from pCRISPR3 plasmid³⁷ using primer pairs T1/T2 and T1/T3, and cloned into pCRISPR3 plasmid over XbaI and SnaBI sites. To obtain the ptracrRNA plasmid, encoding tracrRNA under the control of a T7 RNA polymerase promoter, a DNA fragment was PCR amplified using the T4/T5 primer pair and cloned into the pCDF-DUET (Invitrogen) expression vector using NcoI and PstI sites.

Plasmid transformation

Plasmid transformation assays were performed as described previously.³⁷ The E. coli RR1 strain was transformed with 1 ng, BL21 (DE3) strain, with 25 ng and HT115 strain, with 100 ng of pUC18 or pSP1 plasmid (Table S1). The transformants were plated on LB agar supplemented with appropriate antibiotics and 0.1 mM IPTG (for BL21 (DE3) strain). All transformation experiments were repeated at least three times. Bars in the graphs represent mean values from three or more independent experiments ± 1 SD.

Expression and purification of St-Cas9

Cas9 was expressed in the E. coli DH10B strain. Cells were grown in LB broth medium supplemented with Ap (100 μg/ml) at 37°C to OD₆₀₀ of ~0.5. Cas9 expression was induced with 0.2% (w/v) arabinose for 5 h. Harvested cells were disrupted by sonication and cell debris removed by centrifugation. The supernatant was loaded onto the Ni²⁺-charged 5 ml HiTrap chelating HP column (GE Healthcare) and eluted with a linear gradient of increasing imidazole concentration. The fractions containing the Cas9-(His)₆ protein were pooled and subsequently loaded onto a heparin column, and eluted using a linear gradient of increasing NaCl concentration. The fractions containing Cas9 were pooled and dialysed against 10 mM Tris–HCl (pH 7.0), 300 mM KCl, 1 mM EDTA, 1 mM DTT, 50% (v/v) glycerol and stored at -20°C. The homogeneity of protein preparations was estimated by SDS-PAGE and western blotting against (His)₆-tag with anti-His antibodies (Novogen). Concentrations of Cas9-(His)₆ protein were determined by measuring absorbance at 280 nm using an extinction coefficient of 128,390 M⁻¹ cm⁻¹.⁵⁴

RNA production

RNAs were either purchased as synthetic oligoribonucleotides (Metabion) or synthesized by in vitro transcription using the TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific). For in vitro transcription, fragments containing a T7 promoter at the proximal end of the RNA coding sequence were PCR-generated using primers and templates listed in the Table S3. Resulting RNA fragments were purified using RNeasy MinElute Cleanup Kit (Qiagen). The tracrRNAs, dephosphorylated with FastAP phosphatase (Thermo Fisher Scientific) were radiolabeled with (γ-³³P)ATP (Hartmann Analytic) using T4 polynucleotide kinase (Thermo Fisher Scientific). Sequences of RNA fragments used in this study are provided in the Table S4.

In vitro assembly of the Cas9-RNA complexes

A ternary Cas9-crRNA-tracrRNA complex was assembled by mixing 100 nM of (His)₆-tagged Cas9 protein with a pre-annealed 42 nt crRNA and 78 nt tracrRNA duplex at 1:1 molar ratio and incubated in a buffer containing 10 mM TRIS-HCl (pH 7.5 at 37°C), 100 mM NaCl and 1 mM DTT at 37°C for 1 h.

To reconstitute crRNA maturation and complex assembly pathway in vitro, the (His)₆-tagged Cas9 protein (100 nM) was mixed with 150 nt pre-crRNA (100 nM) and 105 nt tracrRNA (200 nM) transcripts at 1:1:2 molar ratio and pre-incubated in a buffer containing 10 mM TRIS-HCl (pH 7.5 at 37°C), 100 mM NaCl and 1 mM DTT at 37°C for 30 min followed by addition of E. coli RNase III (0.1–2 µM; Ambion) and MgCl₂ (10 mM) and further incubated for additional 30 min at 37°C.

Reactions with plasmid substrates

Reactions on pSP1 and pUC18 plasmids were initiated by mixing plasmid DNA with a ternary Cas9-crRNA-tracrRNA complex (1:1 v/v ratio) and conducted at 37°C. Final reaction mixture contained 3 nM (500 ng) plasmid, 50 nM of the effector complex, 10 mM TRIS-HCl (pH 7.5 at 37°C), 100 mM NaCl, 1 mM DTT and 10 mM MgCl₂ in a 100 µl reaction volume. Aliquots were removed at timed intervals and quenched with phenol/chloroform. The aqueous phase was mixed with 3 × loading dye solution (0.01% bromphenol blue and 75 mM EDTA in 50% v/v glycerol) and reaction products analyzed by agarose gel electrophoresis and ethidium bromide staining.

RNA strand annealing activity assay

The annealing activity of Cas9 was assayed in 10 µl reaction volumes. Reactions were initiated by mixing 5 nM 94 nt pre-crRNA and varying concentrations of Cas9 with 5 nM 105 nt tracrRNA radiolabeled with (γ-³³P)ATP (Hartmann Analytic) and 1 µg nonspecific 2 kb RNA transcript in 10 mM TRIS-HCl (pH 7.5 at 37°C), 100 mM NaCl, 1 mM DTT, 0.05 mg/ml BSA reaction buffer and conducted for 10 min at 37°C. In order to inactivateCas9, disrupt protein-nucleic acid complexes and prevent spontaneous pre-crRNA annealing to labeled tracrRNA, reactions were terminated by mixing 5 µl reaction aliquots with 20 µl of solution containing 0.625 mg/ml Proteinase K (Thermo Fisher Scientific), 0.625% SDS, 12.5% glycerol and RNA trap (187.5 nM solution of unlabeled tracrRNA). Reaction mixtures were incubated for additional 5 min at 37°C and analyzed in non-denaturing 8% PAGE. Gels were dried and visualized by a FLA-5100 phosphor-imager (Fujifilm).

Northern blot analysis

Northern blot analysis was performed as described previously.³⁹ Cas9-bound RNA was isolated from Strep-Tactin-purified Cas9-RNA complex³⁹ using the miRNeasy Mini kit (Qiagen). The RNA was probed with a (γ-³²P)ATP (Hartmann Analytic) labeled 36 nt oligodeoxynucleotide (GG-322) (Table S2) complementary to tracrRNA or 42 nt oligodeoxynucleotide (GG-321) (Table S2) complementary to crRNA. The size of tracrRNA was estimated by comparison with ³³P-labeled Decade RNA marker (Ambion) and RNA transcripts of different lengths.