Two HEPN domains control CRISPR RNA processing and target cutting in Cas13d

In order to understand the structural basis of pre-crRNA processing and mature crRNA recognition by Cas13d, we solved the crystal structure of the uncultured Ruminococcus sp. Cas13d (UrCas13d, also known as RspCas13d)-crRNA complex at 2.15 Å resolution using the single-wavelength anomalous diffraction method (Supplementary Table 1). The binary complex was obtained by the selenomethionine (SeMet)-derived UrCas13d (R288A/R823A) mutant co-expressed with CRISPR RNA in vivo. The SeMet-labeled structure was subsequently used as a model to solve the crystal structure of the unlabeled UrCas13d (R288A/R823A) binary complex at 1.86 Å resolution using the molecular replacement method.

In the binary complex, except for NTD, the predominant secondary structures of UrCas13d are α-helices (Supplementary Fig.). NTD contains a seven-stranded, antiparallel β-sheet and short α-helix. HEPN-1 is composed of ten α-helices, and HEPN-2 domain is composed of eleven α-helices. Two HEPN domains interact with each other primarily through helix-α3 and helix-α28. Notably, the long helix-α28 of HEPN-2 domain extends to the opposite sides of the binary complex and roughly runs parallel to the crRNA structure. Helical-1 domain contains eight α-helices. Helical-2 domain contains eight α-helices, and wraps around helix-α28 of HEPN-2 domain. 1

A mature 50-nt crRNA was determined in the binary complex (Fig. 1b, e and Supplementary Fig. 2). It originated from an in vivo transcript encoded by a co-expressed CRISPR template, which was further processed by UrCas13d itself in E. coli cells. The crRNA consists of a 30-nt repeat region (C(-30)-C(-1)) and 20-nt spacer region (U(1)-A(20)) (Fig. 1e). The repeat region displays a stem-loop architecture. Its secondary structure is remarkably similar to the predicted one, with the exception of base pairing of A(-29) (Fig. 1b and Supplementary Fig. 3). The stem region shows a distorted A-form RNA duplex, which includes seven Watson-Crick base pairs and a wobble G-U base pair (Fig. 1b). The loop region contains four nucleotides (A(-20)-U(-17)). Interestingly, the nucleotides U(-8)-C(15) display four successive U-shaped turns (Fig. 1e).

In addition, a tetra-hydrated Mg²⁺ was found to coordinate with two backbone oxygen atoms of residues R137 and E140 from a loop region of NTD (Fig. 2d, e and Supplementary Movie 3). This (Mg(H₂O)₄)²⁺ ion forms hydrogen bonds with backbone atoms of residues T135, R137, S141, and S142, which helps to maintain the conformation of this loop region. Meanwhile, the side chains of H136 and S138 point outwards and form hydrogen bonds with the phosphate groups of U(-24) and G(-25), respectively. The H136A mutation slightly decreases target RNA cleavage (Fig. 2c). These imply that the (Mg(H₂O)₄)²⁺ ion is involved in the conformational stabilization of the crRNA repeat region.

To investigate whether metal ions are essential for pre-crRNA processing and target cleavage by UrCas13d, we carried out the cleavage assays with different types of metal ions. A pre-crRNA containing a repeat-spacer-repeat sequence was designed to serve as the processing target. The results clearly show that all tested divalent metal ions including Mg²⁺ are unnecessary for pre-crRNA processing, but essential for target RNA cleavage (Fig. 2f, g).

Although the nucleotides A(-5)-C(-1) of the repeat region are unpaired, they are recognized in both the sugar-phosphate backbone-dependent and base-specific manners (Fig. 4a, b). The nucleotide A(-5) stacks with U(-6) and the side chain of N906 hydrogen bonds with the 2′-hydroxyl group of A(-5) (Fig. 4a). Residues N900 and R903 interact with the phosphate group of A(-5). The purine ring of A(-4) is flipped outwards to form hydrogen bond with the side chain of D914, and stacks with the side chains of R744 and R915. The side chains of K181 and R903 interact with the phosphate group of A(-4), and the side chain of R915 interacts with the 2′-hydroxyl group of A(-4) (Fig. 4a). The side chain of K181 also interacts with the phosphate group of A(-3). Furthermore, the purine ring of A(-2) is recognized by the side chains of K56, K61, N79 and N527 (Fig. 4b). The pyrimidine ring of C(-1) is recognized by G78 and R431, and partially stacks with the side chain of F75. The 2′-hydroxyl group of C(-1) hydrogen bonds with the side chain of K52.

To understand the functional importance of residues involved in recognition of the repeat region, we performed mutational analysis. As shown in Supplementary Fig. 6, single mutations of all aforementioned residues did not influence pre-crRNA processing to a detectable level. In contrast, single mutations D178A, K181A, and K524A mentioned above, as well as single mutations K61A, R744A, and R903A, greatly decrease target RNA cleavage. In addition, single mutations K56A, R145A, and Q171A also decrease target RNA cleavage to a certain level (Fig. 4c).

The crRNA extends over the long axis of protein and interacts with every domain of UrCas13d (Fig. 1c, d). A large number of positively charged residues interact with the phosphodiester backbone of crRNA (Fig. 3). Notably, the nucleotides U(-24)-A(-13) of the stem-loop structure protrude from UrCas13d and are mostly exposed to the environmental solvent (Fig. 4d). Deletion of four Watson-Crick base pairs (U(-24)-A(-21)/U(-16)-A(-13)) within this exposed region (designated the Δstem mutant) has no obvious effect on pre-crRNA processing and target RNA cleavage (Fig. 4e, f). Moreover, the time-course study shows that, when compared with pre-crRNA or crRNA, their respective Δstem mutants have a similar efficiency in pre-crRNA processing and target RNA cleavage by UrCas13d and its mutants (Supplementary Fig. 7). The rest of the stem duplex is bound inside the channel formed by NTD, HEPN-1, and HEPN-2 domains. The sugar-phosphate backbone of this bound duplex interacts with the side chains of S53, S54, S138, R145, K799, N800, and K905 (Figs. 2e, 3 and 4b). NTD resides on the stem duplex and contributes most interactions with the crRNA repeat region (Fig. 3).

To investigate the functional importance of the nucleotides U(-8)-C(-1) from the repeat region, we carried out various substitutions and deletions in this region and further inspected their pre-crRNA processing and target RNA cleavage capabilities. The results show that single deletions of A(-4), C(-7), and U(-8) decrease pre-crRNA cleavage, and single mutations A(-2)G, A(-4)U, A(-4)G, and C(-7)G have the similar effect (Fig. 4g and Supplementary Fig. 8a). Notably, single mutation or deletion of A(-2), A(-4), C(-7), or U(-8) significantly compromises target RNA cleavage (Fig. 4h). Therefore, both structure and sequence of the repeat region are critical for target RNA cleavage.

To further explore the functional importance of residues involved in interactions with the spacer region, the mutational studies were performed. The results indicate that single mutations N515A, N620A, and K736A unambiguously decrease target RNA cleavage (Fig. 5e, f).

The previous study has found that two R-X₄-H HEPN ribonuclease motifs of Cas13d are responsible for target RNA cleavage and that the mutation of four catalytic residues in R-X₄-H motifs abolishes the cleavage activity²⁰. We further observed that a single R288A or H828A mutation of UrCas13d was sufficient to abrogate target RNA cleavage, the H293A mutant had a greatly decreased cleavage activity, whereas the R823A mutant still showed detectable activity to a certain degree (Fig. 5f). In the present model, the distance between the backbone Cα atoms of residues R288 and R823 of UrCas13d is ~15.3 Å (Fig. 6c and Supplementary Movie 5). The corresponding distances in LshCas13a and BzCas13b are ~17.8 Å and ~18.6 Å, respectively (Fig. 6d, e). This implies that, in comparison with LshCas13a and BzCas13b, UrCas13d requires a rather small conformational rearrangement to achieve its activation, which is consistent with the compact architecture of Cas13d.

PAM-adjacent seed region within the crRNA spacer is known to be essential for target recognition in Cas9, Cas12a, and Cas12b^32–36. Unlike these Cas effectors, Cas13a possesses a central seed region^17,26,37. In the case of the UrCas13d binary complex, both internal nucleotides U(4)-U(8) and 3′-end nucleotides A(14)-A(20) within the spacer region are exposed and accessible to the environmental solvent. This may allow UrCas13d to use the internal or 3′-end nucleotides of the spacer to interrogate target RNA. To identify the mismatch tolerance between the spacer and target RNA, we prepared 11 target RNA variants by tiling 2-nt mismatches across the 22-nt targeted region and performed the cleavage assays (Fig. 6f and Supplementary Fig. 8b). The results demonstrated that mismatches of the nucleotides C(5)-U(8) and U(13)-C(22) abolished target cleavage (Fig. 6g). This suggests that correct base pairings within two separate spacer regions (an internal and a 3′-end region) are essential for target RNA cleavage.

The full-length gene encoding UrCas13d was synthesized (Supplementary Table 2) and inserted between NdeI and XhoI sites of the pET-30b vector. CRISPR template was synthesized (Supplementary Table 3) and cloned between NdeI and XhoI sites of the pCDFDuet-1 vector.

To acquire the binary complex assembled in vivo, vectors encoding the UrCas13d R288A/R823A mutant and CRISPR template were co-transformed into E.coli Rosetta (DE3) cells (Novagen). The cells were then grown at 37 °C for 3 h until the cell concentration reached OD₆₀₀ of approximately 0.6, which was followed by induction with 0.2 mM IPTG at 18 °C for 12 h. The cells were harvested and lysed in buffer A (25 mM Tris-HCl (pH 7.5), 500 mM NaCl, 5 mM β-mercaptoethanol, 1 mM phenylmethanesulfonyl fluoride (PMSF)). The lysate was centrifuged and the supernatant was loaded onto a Ni-NTA Superflow (QIAGEN) column. The nickel column was washed with buffer A containing 10 mM imidazole, after which the binary complex was eluted with buffer A containing 50 mM imidazole. Afterwards, the eluted target complex was successively purified with a Q FF column and a Heparin HP column (GE Healthcare) using a salt gradient between buffer B1 (25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM MgCl₂, 2 mM dithiothreitol (DTT)) and buffer B2 (25 mM Tris-HCl (pH 7.5), 1 M NaCl, 2 mM MgCl₂, 2 mM DTT) in both steps. The target complex was then dialyzed against buffer C1 (25 mM Tris-HCl (pH 7.5), 200 mM NaCl, 2 mM MgCl₂, 2 mM DTT) and further purified using a Superdex 200 Increase (10/300 GL) column (GE Healthcare). Fractions containing the pure binary complex were collected, concentrated and aliquoted before storing at −80 °C for later crystallization uses. The SeMet-labeled binary complex was expressed in E. coli Rosetta (DE3) cells grown in M9 minimal medium supplemented with SeMet, Lys, Phe, Thr, Val, Leu, and Ile. The SeMet UrCas13d binary complex was purified following the procedure described above.

Full-length UrCas13d gene was expressed in E. coli Rosetta (DE3) cells (Novagen). After the cells were harvested and lysed in buffer A, the target protein was purified with a Ni-NTA Superflow (QIAGEN) column using buffer A and then with the SP FF column (GE Healthcare) using a salt gradient between buffer B3 (25 mM Tris-HCl (pH 7.5), 250 mM NaCl, 2 mM DTT) and buffer B4 (25 mM Tris-HCl (pH 7.5), 1 M NaCl, 2 mM DTT). Fractions of UrCas13d were dialyzed against buffer B3, concentrated, aliquoted and stored for later pre-crRNA or target RNA cleavage assays.

T7 RNA polymerase was used in this study for in vitro synthesis of RNA fragments. Briefly, the synthetic DNA fragments encoding the sequences of pre-crRNA, crRNA, and target RNA were individually cloned into a modified pUC-119 vector between StuI and HindIII sites⁴⁰. After the recombinant vector was amplified in E.coli DH5α cells (Novagen), it was linearized by HindIII, and then purified using phenol-chloroform extraction and ethanol precipitation. Transcription reactions were performed at 37 °C for 3 h, in a buffer composed of 100 mM HEPES-K (pH 7.9), 20 mM MgCl₂, 30 mM DTT, 3 mM NTPs, 2 mM spermidine, 30 ng/μl linearized template, and 100 μg/ml home-made T7 RNA polymerase. The reaction sample was resolved in denaturing polyacrylamide gel containing 8 M urea. Target RNA bands were excised from the gel and eluted using the Elutrap system (GE Healthcare). The eluted RNA was collected by ethanol precipitation, dissolved in DEPC-treated H₂O and preserved at −80 °C.

In order to generate desired mutants (Supplementary Tables,, and), PCR reactions using mutation-specific oligonucleotides and vectors encoding the wild-type UrCas13d, pre-crRNA, crRNA, or target RNA as templates were carried out. The results of mutagenesis were verified by sequencing. 3 4 5

The SeMet-labeled UrCas13d R288A/R823A binary complex assembled in vivo was crystallized by the hanging-drop vapor diffusion method at 16 °C. Crystals were obtained by mixing 1 μl of complex solution (A_{280 nm} = 11.8) and 1 μl of reservoir solution (0.5% (w/v) Tryptone, 0.1 M HEPES sodium (pH 7.0), 16% (w/v) Polyethylene glycol 3350). Crystals grew to their full size within 6 days, after which they were harvested, cryoprotected in reservoir solution supplemented with 20% (v/v) glycerol and flash-frozen.

In addition, the unlabeled UrCas13d R288A/R823A binary complex assembled in vivo was crystallized by the hanging-drop vapor diffusion method at 16 °C. Crystals were obtained by mixing 1 μl of complex solution (A_{280 nm} = 11.7) and 1 μl of reservoir solution (8% (v/v) Tacsimate (pH 7.0), 17% (w/v) Polyethylene glycol 3350). Crystals grew to their full size within 7 days, after which they were harvested, cryoprotected in reservoir solution supplemented with 20% (v/v) glycerol and flash-frozen.

X-ray diffraction data was collected at Shanghai Synchrotron Radiation Facility (SSRF) on beamline BL-17U1. Datasets were processed automatically by the program XIA2 integrated into data collection platform of the beamline⁴¹, and scaled by the program AIMLESS in CCP4^42,43. The crystal structure of the SeMet-labeled UrCas13d R288A/R823A binary complex was solved by single-wavelength anomalous dispersion method with the program AutoSol in PHENIX⁴⁴. An asymmetric unit of this SeMet-labeled crystal structure contains two binary complexes. Next, using the SeMet-labeled structure as a model, the crystal structure of the unlabeled UrCas13d R288A/R823A binary complex was solved by the molecular replacement method with the program Phaser-MR in PHENIX⁴⁴. An asymmetric unit of this crystal structure contains one binary complex. Model building was carried out in COOT⁴⁵. Multiple rounds of refinement were performed with REFMAC in CCP4^43,46. All figures of the structures were prepared using PyMOL (http://pymol.org↗). Data collection and refinement statistics are listed (Supplementary Table 1).

Pre-crRNA and the mutants were designed to contain two repeat regions separated by a spacer region (Supplementary Table 3). Reaction mixtures were prepared by incubating 9.6 μM purified UrCas13d with 4.8 μM pre-crRNA or the pre-crRNA mutants in the assay buffer (25 mM Tris-HCl (pH 7.5), 250 mM NaCl, 2 mM MgCl₂, 2 mM DTT). Total volume of the reaction mixture was adjusted to 20 μl. The cleavage was allowed to proceed at 37 °C for 30 min, after which the reactions were terminated in liquid nitrogen and subsequently quenched at 75 °C for 5 min by adding 2× loading buffer (2× TBE buffer, 12 M urea). Samples were then run on a 20% PAGE TBE-urea denaturing gel and the cleavage products were visualized with the toluidine blue staining. For pre-crRNA cleavage in the presence of different metal ions or ethylenediaminetetraacetic acid (EDTA), the assay buffer contained 2 mM divalent metal ions or 10 mM EDTA without 2 mM DTT. Each cleavage assay was independently carried out three times to verify the repeatability of the results.

Reaction mixtures were prepared by incubating 1.6 μM UrCas13d with 0.8 μM crRNA or the crRNA mutants in the assay buffer (25 mM Tris-HCl (pH 7.5), 250 mM NaCl, 2 mM MgCl₂, 2 mM DTT) at 37 °C for 15 min, and then a final concentration of 16 μM target RNA or the mismatched target RNAs was mixed with the prepared UrCas13d and crRNA sample at 20:2:1 molar ratio. Total volume of the reaction mixture was adjusted to 20 μl. The cleavage assays were then carried out at 37 °C for 45 min. For target RNA cleavage in the presence of different metal ions or EDTA, the assay buffer contained 2 mM divalent metal ions or 10 mM EDTA without 2 mM DTT. Each cleavage assay was independently carried out three times to verify the repeatability of the results. Reactions were terminated and samples were analyzed as described for the pre-crRNA cleavage assay.

Further information on research design is available in thelinked to this article. Nature Research Reporting Summary

Supplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Supplementary Movie 5 Reporting Summary

Source Data