The TTNN PAM sequence is recognized by Lb2Cas12a

We first explored the mechanism of crRNA-dependent dsDNA cleavage by Lb2Cas12a. In the presence of Mg²⁺, wild-type Lb2Cas12a (MW, 141 kDa) exhibits strong catalytic activity against the dsDNA substrate; 20 nt crRNA was used as the guide strand (S1A–S1C Fig). Sequencing analysis revealed the production of staggered ends downstream of PAM at nucleotides 14–17 in the nontargeted (noncomplementary) strand and at nucleotide 23 in the targeted (complementary) strand (S1B Fig). The breakage sites of the nontarget strand are consistent and observed in its orthologs [31], which may be due to the repetitive endonucleolytic cleavage of the nontarget strand. The targeted strand forms a heteroduplex with crRNA and is stabilized by the REC lobe. We determined that Lb2Cas12a recognizes and cleaves target dsDNA at the 5′-TTNN-3′ PAM (N = A, T, G, or C); this recognition sequence was deemed most effective when compared with other cleavage sequences (S1D Fig). Of note, others have demonstrated that Acidaminococcus sp. (As)Cas12a and Lachnospiraceae bacterium ND2006 (Lb)Cas12a orthologs do not recognize the 5′-TTTT-3′ PAM [32]. Together, these observations indicate that Lb2Cas12a minimally requires dinucleotide PAM.

Overall structures of Lb2Cas12a

The crystal structure of full-length Lb2Cas12a (residues 1–1,206 aa) was determined at 3.1-Å resolution (S1 Table) (PDB ID: 8H9D). Strikingly, both apo and RNA-bound Lb2Cas12a are present in the asymmetric unit of the crystal. Each Lb2Cas12a monomer comprises an α-helical recognition lobe (34–492 aa) followed by a nuclease (NUC) lobe (1–33 aa, 493–1,206 aa) (Fig 1A–1C). These 2 lobes are naturally connected by flexible loops (Fig 1B and 1C). The 20 nucleotides of the crRNA in the complex are well defined in the electron density map (Fig 1D and S2A and S2B Fig). Notably, there are differences in the conformations between the apo and RNA-bound forms of Lb2Cas12a. The 2D class averages for negative staining of apo Lb2Cas12a indicated multiple conformations, whereas Lb2Cas12a-crRNA persisted with a uniform, compact conformation (S3A Fig), which is similar to the study of orthologs [25] of Lb2Cas12a. In addition, after incubation with trypsin, apo Lb2Cas12a degraded faster than Lb2Cas12a-crRNA, which is indicative of a different conformation between the apo and RNA-bound forms (S3B Fig). These results suggest that apo Lb2Cas12a exists dynamically between elongated and compact forms and that crystal packing may favor the capture of the elongated form in the asymmetric unit.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Crystal structures of Lb2Cas12a.

(A) Bar diagram showing the domain organization of Lb2Cas12a. Different domains are distinguished by color. (B) Overall structure of the apo Lb2Cas12a, BH domain is disordered in apo and not modeled. (C) Overall structure of the Lb2Cas12a-crRNA complex. (D) Schematic diagram of the Lb2Cas12a-crRNA, black line: Canonical base pairs; gray line: noncanonical base pairs. The structure-related figures in this paper are prepared by CueMol2 program.

https://doi.org/10.1371/journal.pbio.3002023.g001

The elongated structure of apo Lb2Cas12a

In apo Lb2Cas12a, the N-terminal REC1 and REC2 domains (REC lobe) and the C-terminal wedge (WED), RuvC, and Nuc domains (NUC lobe) (Fig 1A) are arranged in an elongated conformation, with the REC1 domain and Nuc domain at each end (Fig 1B). Two flexible loops (Asn33–Asn45 and Lys478–Thr498) connect the REC and NUC lobes. The WED domain connects the REC2 and RuvC domains and forms 2 positively charged cavities within these domains (S4A Fig). Using a DALI structural homolog search [33], we were unable to identify any similar structure for the full-length apo Lb2Cas12a in the PDB database (S2 Table). However, structural homologs were identified for both the independent NUC and REC lobes (S3 and S4 Tables). These results indicate that apo Lb2Cas12a exhibits a unique, elongated conformation, with distinct spatial arrangement of the domains.

The Lb2Cas12a+crRNA binary complex adopts a compact structure

The overall 3D structure of the Lb2Cas12a+crRNA binary complex adopts a triangular, compact conformation (Fig 1C). The REC and NUC lobes lie adjacent to each other in 3D space, creating a positively charged cavity in the center (Fig 1C and S4B Fig). Trp871 of the RuvC II domain, lying close to the Bridge Helix (BH) region (Leu477–Lys493), is anchored in the hydrophobic pocket of the REC2 domain (S5A Fig). The first helix (Trp871–Tyr899) of the RuvC II domain is analogous to the suspension helix of orthologous protein structures that forms a loop upon binding to DNA [29,34]. Lys869, located between the BH and RuvC II domains, forms electrostatic interactions with the REC2 domain. These interactions stabilize the compact conformation, as reported elsewhere for orthologous proteins [35,36] (S5 Table). crRNA is anchored in the positively charged cavity and forms the pseudoknot of the 5′ handle through intramolecular base-pairing; this interaction is coordinated by hydrated magnesium ions and highly conserved WED domain residues (Fig 2A and 2B and S2B Fig). Indeed, alanine mutations at these conserved sites (Arg17, Leu696, Ile766, and Ile746) nearly completely abolish dsDNA cleavage activity (Fig 2C and S2C Fig). These observations indicate that conserved WED domains coordinate the interactions of the REC and NUC lobes with crRNA to maintain a compact conformation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Recognition of the crRNA by Lb2Cas12a.

(A) Interactions between the 5′ pseudoknot of crRNA and Lb2Cas12a residues. (B) Interactions between the hydrated magnesium ions and crRNA, Mg²⁺ and H₂O are indicated by gray and green, respectively. (C) dsDNA cleavage activity analysis of Lb2Cas12a mutants in RNA recognition region. C: Control; only target dsDNA, S: Substrate, P: Cleaved product.

https://doi.org/10.1371/journal.pbio.3002023.g002

Comparison of apo Lb2Cas12a and Lb2Cas12a+crRNA binary complex structures

Superposition of the apo Lb2Cas12a and Lb2Cas12a+crRNA binary complex structures demonstrates a 103° rotation of the REC lobe with respect to the α-helix of the WED-I domains of the crRNA complex in its compact conformation (Fig 3A). Residues from the region Met493–Leu523 of the WED domain participate in major conformational changes to facilitate the structural rearrangements associated with this rotation (Fig 3B, S1 Video). The structural comparison further demonstrates a conformational rearrangement of the linker region between the REC2 and RuvC domains, although there were no structural changes observed within the domains (S6 Table). The flexible loop (Leu477–Lys493) connecting the REC and RuvC domains of apo Lb2Cas12a transforms into an α-helix in the compact conformation, folding back at the C-terminus of the REC2 domain. Meanwhile, in this compact conformation, a β-strand (Lys500–Asn504) in the apo Lb2Cas12a becomes a short loop (Lys500–Asn504) (Fig 3B), a change that is likely supported by the interaction of residue Lys500 with the crRNA backbone. Significantly, the loop-turned-α-helix (Leu477–Lys493) harbors a hydrophobic patch that interacts with Trp871, linking the REC and NUC lobes through the BH and RuvC II suspension helix (S5A Fig). Without crRNA, residues Tyr484, Leu491, and Thr492 in the REC2 domain remain separated and do not interact with Trp871. PISA (Protein Interfaces, Surfaces and Assemblies) analysis indicates that the Lb2Cas12a+RNA complex gains a buried area of 2,436 Ų compared with apo Lb2Cas12a. Collectively, crRNA binding and rearrangement of the hinge loop precedes the formation of the PAM-binding channel.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Structural rearrangement of Lb2Cas12a is triggered by binding crRNA.

(A) Superposition of NUC lobe of apo Lb2Cas12a and Lb2Cas12a-crRNA, structure of apo and RNA bound Lb2Cas12a represented as ribbon diagram, Asp200 and Gly524 in apo and RNA bound form shown as sticks. (B) N-terminus of RuvC domain in apo and RNA bound form, helices represented by cylinder, key residues shown as sticks.

https://doi.org/10.1371/journal.pbio.3002023.g003

crRNA binding

Next, we analyzed the extensively positively charged RNA-binding pocket (S4A Fig) of Lb2Cas12a, anticipating that the positive charges in this region (Lys686–His699, Gly883–Asn900) play a role in maintaining its elongated conformation (S5B Fig). Upon interacting with crRNA, the positively charged pocket within the NUC lobe wraps around the crRNA. Subsequently, the bound crRNA backbone alters the surface charges of the pocket. These interactions of the REC lobe with crRNA result in a compact triangular conformation (Fig 1C and S4B Fig) that is stabilized by a hydrophobic cluster among Trp871, Tyr484, Leu491, and Thr492 and several hydrogen-bonding contacts through residues Asp39, Tyr149, Asn504, Glu875, and Thr492 (S5A Fig).

Next, to understand the role of electrostatic surface potential changes caused by the binding of crRNA, we assessed the role of pH. Through gel filtration and dynamic light scattering (DLS) analyses, we noted that the overall size of apo Lb2Cas12a depends on the pH of the solution (S6A and S7A–S7C Figs); for example, at pH 7.4, the Lb2Cas12a+crRNA complex is smaller than apo Lb2Cas12a (S6B and S7B and S7D Figs). To further assess the involvement of surface charge in these size discrepancies, we mutated 4 basic residues (Arg864Glu-Lys866Glu-Arg868Glu-Lys869Glu) of the BH domain at the ridge of the positively charged cavities. This quadruple mutant was unable to bind crRNA; this observation is similar to that with the PI domain-deletion mutant (ΔPhe558~Thr660) (S8 Fig). However, through DLS (S7F and S7G Fig) and gel filtration (S6C Fig), we identified that the size of the quadruple mutant remained unchanged by the presence or absence of crRNA. Collectively, these observations indicate that the electrostatic environment and RNA binding contribute to the observed conformational rearrangement.

RNA–DNA complex formation drives rearrangements of the dynamic REC lobe

Next, we determined the cryo-EM structure of the Lb2Cas12a+crRNA+DNA ternary complex at 3.95 Å resolution (Fig 4A, S7 Table) (PDB ID 8I54). This structure is reminiscent of the conformation of Cas12a+crRNA+DNA ternary complexes reported for other Cas12a orthologs (Fig 4, S2 Table). Of note, the ternary complex contains an RNA–DNA heteroduplex of only 14 bp that mimics the cleaved heteroduplex that would exist in situ (Fig 4B); this 14-bp heteroduplex is shorter than the 20-bp heteroduplex used previously to study Cas12a orthologs in target cleavage conformation assays [26,28,34,37].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Cryo-EM structures of Lb2Cas12a-crRNA-DNA complex.

(A) Overall structure of the Lb2Cas12a-crRNA-DNA, at 3.95 Å resolution. These figures are prepared by Chimera program; different domains are distinguished by color, surface transparency: 70%. (B) crRNA-DNA duplex in Cryo-EM density map. (C) The guide strand length of crRNA required for Lb2Cas12a to achieve cleavage activity. None: only target dsDNA, S: target dsDNA substrate, P: cleavage product.

https://doi.org/10.1371/journal.pbio.3002023.g004

In our Cas12a+crRNA+DNA ternary complex structure, the REC domain is rotated compared with the binary complex (S9A Fig), forcing the REC linker (261Leu–273Ser) toward the 14-bp heteroduplex. The RNA–DNA duplex passes through a central channel formed by the REC and NUC lobes, which results in the expansion of the central channel (S4C Fig). Unlike FnCas12a [28,29], the REC and NUC lobes in Lb2Cas12a are not connected through a Nuc domain stem (Arg1056–Asn1066), which permits flexibility between these lobes in Lb2Cas12a. The 14-bp heteroduplex is not sufficiently long to stabilize the DNA-REC2 contact through the finger helix (234Lys~244Gln) [24,28]. Consistently, the 14-nt guide strand of crRNA is the minimum length for efficient cleavage of target dsDNA by Lb2Cas12a (S4C Fig). Finally, in the ternary complex, the PI domain is rotated by 38° and lies closer to the PAM nucleotides of the dsDNA (S10A Fig). Based on these structures, we hypothesize that residues Lys575 and Lys518 of Lb2Cas12a play a role in PAM recognition (S10B Fig). Consistently, mutation of Lys575 or Lys518 weakens or abolishes dsDNA cleavage activity (S10C Fig). Rearrangement of the hinge region is also observed in the transition from the binary to the ternary complex.

Further comparison confirms that the REC lobe is highly flexible (S9B Fig), which is consistent with its orthologs [28]. This finding suggests that there is movement of the domains during heteroduplex formation. These conformational changes are consistent with the DLS and gel filtration results, in which the ternary complex exhibits a slightly larger size than the binary complex at a constant pH of 7.4 (S6B and S7D and S7E Figs). Thus, our comparison of the pre-cleavage binary complex with the post-cleavage ternary complex reveals that bilobed movement drives the structural rearrangement of the ternary complex.

The mechanism of DNA cleavage

The Lb2Cas12a catalytic site for DNA cleavage is within the RuvC domain, which is formed by 3 conserved residues: Asp815, Glu906, and Asp1161 (S13A and S14 Figs). We demonstrated that alanine substitution of these 3 residues abolished dsDNA cleavage activity (S13B Fig). In addition, alanine substitution of the conserved catalytic residue Arg1124 in the Nuc domain significantly weakened dsDNA cleavage activity (S13B Fig), suggesting that the Nuc domain assists with double strand cleavage by the RuvC domain. After targeted cleavage of dsDNA and formation of the RNA–DNA duplex, Lb2Cas12a then initiates trans-cleavage activity (S13C Fig). Consistently, the “Lid” structure (S13A Fig), which mediates trans-cleavage activity in previously reported ortholog studies [28], was also observed in Lb2Cas12a, indicating that the Cas12a family of proteins adopts a similar mechanism to activate trans-cleavage activity. Notably, the presence of Mn²⁺ triggers indiscriminate ssDNA cleavage activity of apo Lb2Cas12a (S13C Fig), which is consistent with previous studies [38]. Overall, we demonstrate that the RuvC catalytic residues Asp815, Glu906, and Asp1161, along with Arg1124 of the Nuc domain (S13C Fig), are responsible for the trans-cleavage activity of Lb2Cas12a, which is consistent with a previous study in which the Nuc domain assists the RuvC domain in achieving trans-cleavage [35].

Lb2Cas12a expression and purification

Lb2Cas12a gene was purchased from Addgene and reconstituted in the pET28b vector. Lb2Cas12a or mutants with pET 28b-N-6×His tag was heterologously expressed in BL21(DE3) pLysS E. coli bacteria. The engineered bacteria in LB medium were placed in a shaker incubator at 37°C, 220 rpm until it reached an OD600 = 0.6, then added IPTG to a final concentration of 0.4 mM and cultured for 16 h at 16°C, 220 rpm. The harvested bacteria were resuspended in lysis buffer (1 mM DTT, 25 mM KH₂PO₄ (pH7.0), 10% glycerol, 500 mM NaCl, 25 mM Tris (pH 7.4), 1 mM EDTA). The high-pressure homogenizer (Avestin) with a cooling system was used for the lysis. The lysed bacteria were centrifuged at 40,000 rpm, 4°C for 1 h by Beckman Ultracentrifuge Type 45Ti. The supernatant was incubated with RNase A at 16°C for 15 min to remove endogenous RNA. Subsequently, the harvested supernatant was purified using Ni²⁺ affinity chromatography column (GE Healthcare) combined with ÄKTA purification system (GE Healthcare). Further purification was carried out by size-exclusion chromatography using the HiLoad Superdex 200 26/60 prep column (GE Healthcare). The purified protein was dialyzed against 20 mM Tris-HCl, 100 mM NaCl (pH 7.4) solution until use.

Crystallization and structure determination

For crystallization, the purified Lb2Cas12a (10 mg/ml) and crRNA were mixed in a molar ratio of 1:2 and incubated at 4°C for 20 min. The Lb2Cas12a-crRNA complex and screening solution were mixed in 1:1 ratio and crystallized at 20°C by hanging drop vapor diffusion method. Lb2Cas12a-crRNA crystals were grown in 10 mM MgCl₂, 0.1 M sodium cacodylate (pH 6.8), 17% PEG1000, and 1 mM DTT. The complex crystals were briefly soaked in cryoprotectant solution containing 25% w/v D- (+)-Glucose monohydrate and flash cooled at 100 K. Diffraction data were collected at the NSRRC, Taiwan TPS05A beamline at 0.99 Å wavelength. HKL2000 program [40] was used for data processing. The Matthews coefficient was 3.2 ų/Da [41] with 61.55% solvent content and 2 molecules in the asymmetric unit. The complex structure of Lb2Cas12a-crRNA was determined by molecular replacement using Phenix-Phaser program [42] and the individual REC and NUC coordinates of Lachnospiraceae bacterium ND2006 LbCas12a (PDB: 5ID6) were used as the search models. Several rounds of model building were done using COOT program [43] followed by refinement using Phenix-Refine [44]. The final model had good stereochemistry, with 99.5% residues falling within the allowed regions of the Ramachandran plot. Although the average B factor is around 90Ų, the model has good electron density map well covering the model (S15 Fig). Molecular graphic images were prepared using CueMol2 and pymol program.

Cryo-EM sample preparation and data collection

Purified Lb2Cas12a, crRNA with 20 nt guide sequences, and DNA were mixed in a ratio of 1:1.5:2 and remove excess nucleic acid by gel filtration chromatography. The fresh sample was purified at 0.5 mg/ml in buffer containing 20 mM HEPES-Na (pH 7), 150 mM NaCl, 10 mM MgCl₂, 5 mM DTT. The cryo-EM data were collected at CBIS CryoEM Facility, National University of Singapore. Four microliters of sample were applied on glow-discharged UltrAufoil R1.2/1.3 (Quantifoil) and blotted for 1 s in 22°C with 100% humidity, a wait time of 15 s, a drain time of 0 s, and a force of −5 using FEI Vitrobot Marc IV. The grid was plunge-frozen in liquid ethane cooled by liquid nitrogen. The frozen-hydrated grid was loaded into Titan Krios cryo-electron microscope equipped with Gatan K3 direct-electron counting camera and operated at 300 keV, and 35-frame movies were collected at 81,000× magnification in counting mode with a physical pixel size of 1.105 Å/pixel. The images were recorded at defocus range of 0.5 to 2.5 μm. The exposure time was 3.49 s. The dose was 45 e/Å per movie stack. The 2,560 stacks of 35-frame movies were collected, using SerialEM program (FEI; Thermo Fisher Scientific).

Image processing

The micrographs were pre-processed by Relion-3.1.1, and 2D- and 3D-classifications were done in cryoSPARC-3.2.0 [45]. The movie frames were aligned by MotionCor2, using Relion’s own implementation. Contrast transfer function was estimated by CTFFIND-4.1. The particles were first LoG-picked and then template-picked on the same set of micrographs. The duplicate picks were removed from the combined particle sets. The particles were extracted with box size 64² pix² by 4-fold binning (4.420 Å/pix) in Relion-3.1.1 and imported to cryoSPARC-3.2.0. Suboptimal particle images were removed by multiple rounds of 2D classification and class selection. The selected particles were exported to Relion-3.1.1 using pyem command csparc2star.py–copy-micrograph-coordinates and re-extracted with box size 128² pix² by 2-fold binning (2.210 Å/pix). The particle duplicates were removed by inter-particle 30 Å cutoff. The particles were again imported to cryoSPARC-3.2.0 and subjected to initial 3D modeling by Ab-initio Reconstruction with the number of models 3 and class similarity 0, followed by Heterogeneous Refinement. The particle set belonging to the least represented class was discarded. The particle sets belonging to remaining two 3D-classes were retained and subjected to another round of Heterogeneous Refinement using the previous three 3D-classes as the volume input. After 3 times Heterogeneous Refinement, the best 3D-class reaches 5.0 Å resolution (S11 and S12 Figs). The particles were exported again using csparc2star.py to Relion-3.1.1 and re-extracted with un-binned box size 256² pix² (1.105 Å/pix). The volume output belonging to the best 3D-class of Heterogeneous Refinement was rescaled to angpix 1.105 and re-sized to box size 256³ pix³ using relion_image_handler and imported to Relion-3.1.1 as 3D-reference. The re-extracted particles and the imported volume were subjected to 3D auto-refinement without masking. Following mask creation and post-processing, the images were Ctf-refined 3 times in order: (1) beam tilt; (2) anisotropic magnification; and (3) defocus per particle and astigmatism per micrograph. The images were Bayesian polished with the trained optimal parameters on the original output of MotionCorr2 in Relion-3.1.1. The polished particles were used for the second 3D-auto refinement without masking. The volume output from the first 3D-auto refinement was used as reference map with initial low-pass filter 40 Å. The volume from the second 3D auto-refinement was used to create a mask with low-pass filter 20 Å, initial binarization threshold 0.004 and extension of binary map threshold by 7 pixels, adding a soft edge of 8 pixels. The new mask and the output from the second 3D auto-refinement were subjected to post-processing with automatic sharpening, resulting in 4.28 Å map. The post-processing output was used for the second round of Ctf-refinement and Bayesian polishing as described above. The polished particles, the mask, and output volume from the second 3D auto-refinement before sharpening were subjected to the third 3D auto-refinement, which resulted in map resolution 4.49 Å before sharpening and 4.09 Å after sharpening. The second polished particles were imported to cryoSPARC-3.2.0 and performed Ab initio Reconstruction and Non-uniform Refinement without further particle sorting. The map quality was improved to 3.95 Å.

Lb2Cas12a/crRNA/DNA model building

The initial model was prepared by using SWISS-MODEL online suite based on the structure 5XUS as a template. Phenix dock-in-map [46] was used for initial model building followed by manual model building in COOT [47] using the cryo-EM map. After initial model building, the model was refined against the EM-derived maps using the phenix.real-space-refinement tool from the PHENIX software package [46], employing rigid body, local grid, NCS, and gradient minimization. The model was used to sharpen the map in CCPEM-1.6.0, and further rebuilt by Flex-EM and ISOLDE-1.0.1. This model was then subjected to additional rounds of manual model-building and refinement which resulted in a final model-to-map cross-correlation coefficient of 0.72 for Lb2Cas12a/crRNA/DNA model. Stereo-chemical properties of the model were evaluated by Molprobity [48]. Molecular graphic images were prepared using CueMol2 (http://www.cuemol.org/en/) and Chimera programs [49].

RNA substrate in vitro transcription

The primers were designed to amplify DNA fragments with T7 promotor. In vitro transcription was based on standard procedures of RiboMAX Large Scale RNA Production System T7 (Promega). The reaction mixture was further purified by 7% polyacrylamide 8 M urea denaturation TBE gel and classical ethanol precipitation [50].

Dynamic light scattering (DLS)

The purified Lb2Cas12a or mutant was incubated with RNA, DNA, or RNA/DNA-duplex. Subsequently, the apo or complex was dialyzed against the pH 6.2, pH 7.4, and pH 8.0 buffers (20 mM Tris-HCl, 100 mM NaCl). Further purification was performed by size exclusion chromatography using a HiLoad 16/600 Superdex 200 prep grade preparative column (GE Healthcare). Subsequently, 5 μl sample (1 mg/ml) was used for DLS (DynaPro NanoStar) experiments that was run at 4°C.

Electrophoretic mobility shift assay

Lb2Cas12a wild type or mutant (0–400 nM) was incubated with crRNA (100 nM) in the buffer that contained 20 mM HEPES 10 mM MgCl₂, 100 mM NaCl, 1 mM DTT, 4% glycerol, pH 7.4 at 4°C for 15 min. Subsequently, the incubation mixture with native PAGE loading dye was used to perform 7% native polyacrylamide gel electrophoresis. The gel was stained with SYBR Safe at room temperature for 5 min and imaged via Imaging via ChemiDoc Imaging System (Bio-Rad Laboratories).

dsDNA cleavage assay

To generate Biotin-tagged substrates, DNA duplex was amplified by PCR reaction using forward or reverse primers labeled with 5′ Biosg (synthesized from IDT). The 20 μl cleavage reaction mixture containing 20 mM Tris-HCl, 50 mM NaCl, 1 mM DTT, 10 mM MgCl₂ (pH 7.4), 100 nM crRNA, 50 nM dsDNA, and 100 nM Lb2Cas12a WT or mutant was incubated at 37°C for 20 min. The reaction mixture was subjected to denaturation TBE-urea 7% PAGE and imaged using SYBR Safe stain and ChemiDoc MP Imaging System. The cleaved products were excised from the gel to perform sequencing analysis. The biotin-tagged cleavage products were detected by Chemiluminescent Nucleic Acid Detection Module (Thermo, 89880) and visualized by ChemiDoc Imaging System. The reaction was performed in independent triplicates. Oligonucleotides used in this assay are shown in S8 Table.

ssDNA cleavage assay

Single-stranded DNA activity was detected using M13mp18 ssDNA (New England Biolabs) as the substrate. The RNA-DNA duplex-triggered ssDNA cleavage reaction was performed according to the previous protocols [17]. The Mn²⁺-triggered ssDNA cleavage reaction was performed as follows: the 20 μl reaction mixture containing 30 nM M13mp18 ssDNA, 20 nM Lb2Cas12a, 20 mM Tris-HCl, 50 mM NaCl, 1 mM DTT (pH 7.4), 10 mM MgCl₂ or CaCl₂ or MnCl₂ or CuCl₂ was incubated at 37°C for 15 min. The reaction mixture was separated by 1.2% agarose gel. The gel was stained with SYBR Gold and the image was captured with ChemiDoc Imaging System.

2D Classification of negative staining apo Lb2Cas12a particles and Lb2Cas12a-crRNA

The concentration of apo Lb2Cas12a and Lb2Cas12a-crRNA were diluted to 0.02 mg/ml to perform negative staining, 5 μl sample was loaded onto grid (carbon film 300 mesh copper, EMS) for 60 s and excess sample was removed with filter paper. Then, 5 μl stain solution (Uranyless EM Stain, EMS) was applied on the grid for 30 s and removed with filter paper. The grid was dried at room temperature for 10 min. The images were captured by Tecnai 12 (FEI 120kV, LaB6, magnification 52Kx).

Trypsinization

Approximately 5 μg of purified Lb2Cas12a was incubated with crRNA at a molar ratio of 1:1 for 15 min on ice. Subsequently, 5 μg apo Lb2Cas12a and Lb2Cas12a-crRNA complex were incubated with 0.1 μg trypsin at room temperature in the 10 μl reaction buffer containing 30 mM HEPES, 150 mM NaCl, 1 mM DTT (pH 7.4), 10 mM MgCl₂, 0.02 μg trypsin, 5 μg apo Lb2Cas12a, or Lb2Cas12a-crRNA complex. Five replicates were set up for each apo Lb2Cas12a and Lb2Cas12a-crRNA complex reaction mixture. The reaction was terminated at 1 min, 3 min, 5 min, 10 min, and 15 min by adding SDS-PAGE loading buffer and heating at 100°C for 2 min. The reaction product was resolved by SDS-PAGE gel and imaged with ChemiDoc Imaging System.