RNA-activated DNA cleavage by the Type III-B CRISPR–Cas effector complex

The T. maritima Cmr complex cleaves ssRNA

In vitro, the Cmr complexes from P. furiosus, S. solfataricus, and T. thermophilus cleave ssRNA that is complementary to the crRNA (Hale et al. 2009; Zhang et al. 2012; Staals et al. 2013). To determine the substrate specificity of the T. maritima Cmr complex, we incubated the complex with crRNA8.3 and a selection of nucleic acid targets that were radiolabeled at their 5′ ends (5′ labeled). Reactions were incubated for 10 min at 80°C, the optimal growth temperature of T. maritima (Huber et al. 1986), and the products were analyzed by denaturing PAGE followed by autoradiography. Noncomplementary ssRNA, complementary ssDNA, and dsRNA were not cleaved (Fig. 2A). Only complementary ssRNA (the ssRNA8.3 target) was cleaved, with the cleavage site located 14 nt from the 3′ end of the paired crRNA (Fig. 2). Cleavage appeared to be sequence-independent, as a ssRNA target complementary to a crRNA with a different sequence (crRNA8.4 and ssRNA8.4) also resulted in the same 14-nt cleavage product (Fig. 2A). Cleavage also required that the 5′ handle was intact and had the correct repeat-derived sequence (Fig. 2B), in agreement with results from the P. furiosus Cmr complex (Hale et al. 2012), and, like other Type III complexes, was dependent on the presence of magnesium or manganese ions (Supplemental Fig. S3A; Hale et al. 2009; Zhang et al. 2012; Staals et al. 2013, 2014; Tamulaitis et al. 2014; Samai et al. 2015).

Open in a separate window

Figure 2.

The T. maritima Cmr complex cleaves ssRNA. (A) The indicated 5′ labeled substrates were incubated with (+) or without (−) the Cmr complex, and the products were resolved on a denaturing polyacrylamide gel. (B) Requirement of the crRNA and determinant of its repeat-derived 5′ handle. (C,D) 5′ and 3′ labeled ssRNA8.3 targets were incubated with the Cmr complex for the indicated times and resolved on denaturing polyacrylamide gels. In all panels, “T1” denotes the T1 ladder, and “OH” denotes the hydroxide ladder. (E) Schematic depicting the cleavage sites on the ssRNA8.3 target in relation to crRNA8.3. The sizes of the cleavage products are indicated. Note that the 3′ end labeling reaction added one additional nucleotide to the 3′ end of the RNA target.

Other Cmr complexes can cleave their RNA targets at up to five sites that are separated by 6-nt intervals (Staals et al. 2013; Hale et al. 2014). To determine the frequency and spacing of target cleavage by the T. maritima Cmr complex, we monitored cleavage of a 5′ labeled ssRNA8.3 target as a function of time. The results show that the target was cleaved at four sites (sites 1–4), each separated by 6 nt, producing 32-, 26-, 20-, and 14-nt products (Fig. 2C). Cleavage at site 1 is minimal compared with the other three sites, which was also observed with the P. furiosus and T. thermophilus Cmr complexes (Staals et al. 2013; Hale et al. 2014). The appearance and disappearance of the cleavage products over time suggests that the target is cleaved sequentially, starting predominantly at site 2 and then proceeding to site 3 and then to site 4 (Fig. 2E). After ∼5 min, the target is almost completely cleaved to a single prominent 14-nt product, which is much faster than observed with P. furiosus and T. thermophilus Cmr complexes (Hale et al. 2009; Staals et al. 2013). To confirm that cleavage is sequential, we repeated this experiment but used a 3′ labeled target. In this experiment, we observed a single prominent 12-nt product (Fig. 2D), corresponding to cleavage at site 2 (note that the 3′ end labeling reaction added one additional nucleotide to the 3′ end of the RNA target), suggesting that cleavage primarily begins at site 2.

To gain further insights into the determinants for RNA cleavage, we tested activity on RNA targets that would result in base-pairing-disrupting mismatches when paired with crRNA8.3. RNA targets that result in mismatches at position 20 and/or position 21, which flank the scissile phosphate at site 3, do not inhibit cleavage of the target at any site. In fact, cleavage of these mismatched targets appears to be enhanced relative to the fully complementary target, as, with these targets, we observed an increased proportion of the site 4 product relative to the fully complementary target (Supplemental Fig. S4A, experiments 1–4). Thus, perfect complementarity is not required at site 3, suggesting that RNA cleavage by the Cmr complex, like the related Csm complexes (Staals et al. 2014; Tamulaitis et al. 2014), is not strictly dependent on complete complementarity between the crRNA and RNA target.

It has been shown that the 2′-hydroxyl group adjacent to the scissile phosphate is crucial for the RNA cleavage activity of a chimeric Cmr complex reconstituted from Archaeoglobus fulgidus Cmr1–3 and P. furiosus Cmr4–6 (Osawa et al. 2015). We confirmed the importance of this hydroxyl group to cleavage by the T. maritima Cmr complex. An ssRNA8.3 target containing a 2′-deoxyribose at position 20 was not cleaved at site 3 (Supplemental Fig. S4A, experiment 5), but, in a control experiment, a target with a 2′-deoxyribose at position 21 was cleaved at site 3 (Supplemental Fig. S4A, experiment 6). Furthermore, when any of the four cleavage sites were blocked with a 2′-deoxyribose, we observed no cleavage at that site (Supplemental Fig. S4B, experiments 7–11). A modified ssRNA8.3 target (ssRNA8.3*) that contained 2′-deoxyribose base substitutions at each of the four cut sites (positions 14, 20, 26, and 32) was not cleaved by the Cmr complex at any site (Supplemental Fig. S4B, experiment 12). To confirm that the 2′-deoxyribose modifications in ssRNA8.3* do not interfere with binding to the Cmr complex, we performed competition experiments. Here, we monitored cleavage of the 5′ labeled ssRNA8.3 target using either unlabeled ssRNA8.3* as a specific competitor or ssRNA8.4 as an unspecific competitor. Only the unlabeled crRNA8.3* inhibited cleavage of the target (Supplemental Fig S4C, lane 3), thereby providing support for the Cmr complex binding to crRNA8.3*.

DNA cleavage by the Cmr complex

Genetic experiments implicate the Cmr complex in targeting transcriptionally active plasmid DNA in cells (Deng et al. 2013). However, Cmr complexes have only been observed to cleave RNA and not DNA in vitro (Fig. 2A; Hale et al. 2009; Zhang et al. 2012; Staals et al. 2013) and in vivo (Hale et al. 2012; Zebec et al. 2014). We reasoned that these differing observations could be reconciled if the DNA nuclease activity of the Cmr complex required the presence of a ssRNA transcript. To test this, we incubated the T. maritima Cmr complex with crRNA8.3, an ssRNA8.3 target (mimicking a transcript), and one of two different ssDNA substrates (substrates A and B), which were 5′ labeled (Fig. 3A). We initially screened ssDNA as a potential substrate because Csm1 (which, like Cmr2, is a Cas10 family protein) has ssDNA nuclease activity (Jung et al. 2015) and is the only subunit of a Type III effector complex with known DNA nuclease activity. The Cmr complex cut the DNA substrate weakly at multiple sites only when it and the ssRNA8.3 target were added to the reaction at the same time (Fig. 3B, lanes 4,11). No cleavage was observed if the Cmr complex was preincubated with the ssRNA8.3 target for 10 min before addition of the ssDNA (Fig. 3B, lanes 3,10). This suggested that cleavage of the ssRNA target may inhibit cleavage of the ssDNA substrate. No DNA cleavage was observed in the absence of the Cmr complex, in the absence of the complementary ssRNA8.3 target (Fig. 3B), or with an ssRNA target noncomplementary to crRNA8.3 (Fig. 3B, lanes 7,14). Cleavage was dependent on the presence of manganese ions but not magnesium ions (Supplemental Fig. S3B) and was specific to ssDNA, as no cleavage of either ssRNA or dsDNA was observed under the same reactions conditions (Fig. 3B, lanes 15–18).

Open in a separate window

Figure 3.

DNA cleavage by the Cmr complex. (A) Schematic of DNA substrates A and B. Arrowheads indicate the mapping of the cleavage sites onto the sequences. (B) 5′ labeled ssDNA substrates (substrates A and B), an ssRNA (whose sequence is equivalent to ssDNA substrate A), or a dsDNA (substrate A annealed to a complementary oligo) were incubated with the Cmr complex (or a variant Cmr complex containing Cmr4 D26A) in the presence of ssRNA8.3 targets and then analyzed by denaturing PAGE. An asterisk indicates the use of the noncleavable ssRNA8.3* target, and “M” denotes the marker lane.

To further investigate the effects of RNA cleavage on the DNA cleavage activity of the Cmr complex, we monitored DNA cleavage under two regimes where the ssRNA target is not cut. First, we monitored DNA cleavage using the noncleavable ssRNA8.3* as the RNA target. Second, we monitored DNA cleavage using a Cmr complex that was reconstituted with Cmr4 harboring a D26A mutation (Benda et al. 2014; Ramia et al. 2014a; Zhu and Ye 2015). This variant of the T. maritima Cmr complex specifically binds to (Supplemental Fig. S5, lanes 3,8) but is unable to cleave (Supplemental Fig. S6A, lane 3) a complementary RNA target. Under both of these regimes, cleavage of the ssDNA was greatly enhanced (Fig. 3B, lanes 5,6, 12,13), suggesting that cleavage of the RNA target inhibits cleavage of ssDNA by the Cmr complex.

RNA target binding licenses ssDNA cleavage

Our results suggest that RNA target binding licenses ssDNA cleavage by the Cmr complex. To gain additional insight into this licensing, we investigated RNA cleavage more closely. First, we wished to determine whether the RNA products dissociate from the Cmr complex following cleavage or whether they remain bound. Thus, the Cmr complex pre-equilibrated with an equimolar amount of crRNA8.3 was incubated with either equimolar or twofold or fourfold excess of the ssRNA8.3 target. Cleavage was monitored over time by denaturing PAGE followed by autoradiography. If RNA products did not dissociate from the complex following cleavage, then, in reactions with target in excess, the amount of target cleaved should equal the amount of Cmr complex in the reaction. However, at all ratios of Cmr complex to target tested, ∼100% of the ssRNA8.3 target was cleaved (Fig. 4A), indicating that the RNA products dissociate from the complex following cleavage. In a second set of experiments, the Cmr complex, pre-equilibrated with 50 nM crRNA8.3 (1:1 stoichiometry), was incubated with 100 nM target ssRNA8.3 and 1 nM ssDNA substrate A (5′ radiolabeled). Cleavage of the DNA was monitored over time until the accumulation of product had plateaued (Fig. 4B). Note that, under these conditions (100 nM ssRNA8.3 target), the DNA cleavage activity of the Cmr complex was more robust than in previous experiments, where the concentration of the ssRNA8.3 target was much lower (1 nM). Once the accumulation of product had plateaued, fresh complementary (ssRNA8.3) or, in a control experiment, noncomplementary (ssRNA8.4) RNA targets were added to the reactions, and we continued to monitor DNA cleavage (Fig. 4B). Addition of fresh complementary ssRNA8.3 reinitiated cleavage of the ssDNA, whereas the noncomplementary ssRNA8.4 had no effect, suggesting that the RNA products had dissociated from the complex. If the RNA products are not released after cleavage, then addition of fresh target should have no effect. Together, these data indicate that the presence of an intact ssRNA target complementary to the crRNA activates a ssDNA nuclease activity in the Cmr complex. Cleavage of the ssRNA target and dissociation of the resulting fragments then prevent DNA cleavage.

Open in a separate window

Figure 4.

RNA turnover by the Cmr complex. (A) 5′ labeled ssRNA8.3 target was incubated with the Cmr complex in the presence of unlabeled crRNA8.3 at the indicated concentrations, and cleavage was monitored over time. “T1” denotes the T1 ladder, and “C” denotes a control lane containing only the labeled ssRNA8.3 target. (B) 5′ labeled ssDNA substrate A was incubated with the unlabeled ssRNA8.3, the Cmr complex, and crRNA8.3 for 30 min before either fresh ssRNA8.3 or ssRNA8.4 was added to the reaction (pulse), which was then monitored for a further 30 min. Samples were analyzed by denaturing PAGE. Below the gel panel is a quantification of the cleavage activity over time. The point at which the pulse was added is indicated.

In Type III-A systems, base pairing between the 5′ handle of the crRNA and the 3′ flanking sequence of the RNA target has been shown to have no effect on RNA target cleavage, but it inhibits cleavage of the DNA (Marraffini and Sontheimer 2010; Samai et al. 2015). We tested the effects of complementarity to the 5′ handle on RNA and DNA cleavage by the T. maritima Cmr complex and found that the RNA was cleaved, albeit to a lesser extent (Supplemental Fig. S6A, lane 5), and that cleavage of DNA was almost undetectable (Supplemental Fig. S6B, lane 3). Thus, activation of DNA cleavage by the Cmr complex also requires a lack of complementarity between the RNA target and the crRNA 5′ handle.

Sequence specificity for the ssDNA target

The Cmr complex was able to cleave two ssDNA substrates with distinct sequences, suggesting that the ssDNA nuclease activity of the Cmr complex is not sequence-specific. However, cleavage of each substrate produced a distinct band pattern (Fig. 3B), indicating some sequence preference at the sites of cleavage. To investigate this further, we mapped the length of the cleavage products for the two different ssDNA substrates (Fig. 3A). This mapping revealed that the Cmr complex cleaved the DNA after every thymidine and not after any other nucleotide. Thymidine specificity was further verified using a series of DNA substrates (derived from substrate A) where cleavage was blocked at the sites where thymidines were replaced with adenosine (Fig. 5A). A ssDNA substrate completely lacking thymidine was not cleaved by the Cmr complex (Fig. 5A).

Open in a separate window

Figure 5.

Determinants for DNA cleavage by the T. maritima Cmr complex. (A) 5′ labeled ssDNA substrates with the indicated thymidine-to-adenosine substitutions were incubated with (+) or without (−) the Cmr complex and analyzed by denaturing PAGE. (B) 5′ labeled ssDNA substrates with the indicated substitutions at position 18 were incubated with (+) or without (−) the Cmr complex and analyzed by denaturing PAGE. (C) dsDNA substrates containing 2-, 4-, 6-, 8-, and 10-nt “bubbles” were incubated with the Cmr complex and analyzed by denaturing PAGE. Below the gel is a schematic of the “bubble” substrate.

In addition to the observed thymidine specificity, cleavage between a thymidine and a cytidine appeared to be less efficient than cleavage between a thymidine and any other nucleotide (see the 17-nt product of substrate A and the 22-nt produce of substrate B in Fig. 3B). Substrate A contains a single thymidine–cytidine dinucleotide sequence, located at positions 17–18 (Fig. 3A). We mutated this cytidine (C18) to each of the other 3 nt and monitored cleavage of the modified DNA substrate. Cleavage at position 17 of these modified substrates was more robust than in substrate A (Fig. 5B), thus confirming that cleavage between a thymidine and a cytidine is less efficient than cleavage between thymidine and any other nucleotide.

Cleavage of ssDNA in the context of dsDNA

To determine whether the T. maritima Cmr complex can cleave ssDNA in the context of dsDNA, we generated a series of dsDNA substrates that contained mismatched “bubble” regions. The “bubble” regions ranged in size from 2 to 10 nt and consisted of a polythymidine sequence on one strand (the T strand) and a polycytidine sequence on the opposite strand (the C strand) (Fig. 5C). The double-stranded regions of these substrates were designed such that they did not melt at the reaction temperature, 80°C. The “bubble” substrates (labeled at the 5′ end of their T strands) were incubated with the T. maritima Cmr complex crRNA8.3 and an ssRNA8.3 target and then analyzed by denaturing PAGE and autoradiography. Each of the substrates was cleaved multiple times within the bubble region (Fig. 5C), suggesting that the Cmr complex can cleave ssDNA regions as short as 2 nt in the context of dsDNA.

Pull-down assay

Purified Cmr proteins that were tagged (bait) and untagged (prey) were incubated in 100 μL of binding buffer (100 mM KCl, 20 mM Tris-HCl at pH 8.0, 10 mM imidazole, 1 mM TCEP) for 30 min at 37°C. Twenty microliters of nickel sulfate-charged IMAC resin (BioRad) equilibrated in binding buffer was added to the Cmr protein sample and incubated for 1 h at 4°C on a rotating plate. The resin was washed three times with 0.5 mL of binding buffer and further incubated with elution buffer (binding buffer containing 250 mM imidazole) for 30 min at 4°C. Proteins eluted from the beads were analyzed on 4%–20% miniprotean TGX precast gels (BioRad) stained with Coomassie.

Preparation of labeled oligonucleotides

RNA and DNA oligonucleotides were purchased from Sigma Aldrich (Supplemental Tables 1, 2). They were 5′ radiolabeled with T4 polynucleotide kinase (New England BioLabs) and γ-³²PATP (Perkin Elmer) in 1× T4 polynucleotide kinase buffer for 30 min at 37°C. The substrates were resolved on a denaturing polyacrylamide gel, visualized by autoradiography, excised from the gel, and placed in a 0.5-mL solution of 0.3 M sodium acetate overnight at 4°C followed by ethanol precipitation and resuspension in RNA storage solution (Ambion) for RNA or 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA for DNA. To generate dsDNA or dsRNA targets, labeled oligonucleotide with twice the molar excess of the complementary unlabeled oligonucleotide was incubated for 10 min at 95°C followed by slow cooling to room temperature. Complete annealing was confirmed by nondenaturing PAGE.

Radiolabeling of the 3′ end of ssRNA was performed as described previously (Huang and Szostak 1996). Briefly, ssRNA was annealed to a short ssDNA oligonucleotide complementary to the 3′ end of ssRNA, generating a TG overhang. The duplex was then incubated with a 3′–5′ exonuclease-deficient Klenow fragment (New England Biolabs) and α-³²PdATP (Perkin Elmer) in 1× buffer 2 (New England Biolabs) for 2 h at 37°C. Following the reaction, the ssDNA was removed by denaturing PAGE.

RNA cleavage assays

Cmr2–6 was formed by pull-down (see above). Unless otherwise indicated, 250 nM Cmr2–6, 250 nM Cmr1, and 100 nM crRNA were incubated for 30 min at 80°C in reaction buffer (100 mM KCl, 20 mM HEPES at pH 7, 1 mM MnCl₂, 0.5 mM ATP, 1 mM TCEP). The reaction (80°C) was initiated by addition of 1 nM (unless otherwise indicated) radiolabeled target. The reaction was stopped at 10 min or the indicated time point with the addition of quencher dye (90% formamide, 2.5% glycerol, 0.01% SDS, 0.01% bromophenol blue, 0.01% xylene cyanol, 1 mM EDTA) and heated for 10 min at 95°C. Radiolabeled RNA ladders were generated with RNase T1 (New England Biolabs) or alkaline hydrolysis buffer (50 mM sodium carbonate at pH 9.2, 1 mM EDTA). The samples are then run on 20% polyacrylamide denaturing gels and visualized by phosphorimaging.

DNA cleavage assay

The Cmr complex was assembled with a crRNA as described above. This complex was then incubated with both a target RNA (1 nM, unless otherwise indicated) and 5′ labeled ssDNA (1 nM, unless otherwise indicated) substrate in reaction buffer for 10 min at 80°C. The DNA was isolated by phenol–chloroform extraction and analyzed by denaturing PAGE and autoradiography.

We thank Jennifer M. Kavran for critical reading of the manuscript. This work was supported by National Institutes of Health grant GM097330 to S.B., and a Ruth L. Kirschstein National Institutes of Health F31 fellowship (GM105364) to M.A.E.