Strains, plasmids, primers and nucleic acid substrates can be found in Supplementary Table 1. The mutant constructs used in this paper were prepared by point-directed mutagenesis using the Q5 Site-Directed Mutagenesis Kit (NEB). Primers were designed and annealing temperatures were determined using the NEBase Changer tool. All primers were ordered through IDT. Plasmid ligation was achieved with KLD Enzyme Mix (NEB) and completed plasmids were sequenced to verify correct alterations (Plasmidsaurus).
Expression and purification
N-terminal hexa-his-tagged Cas12a2 and various mutant plasmids were transformed into chemically competent Escherichia coli NiCo21 cells (NEB). A single colony from transformation was selected for starter culture in 20 ml LB medium grown at 37 °C overnight (16–18 h). Each starter culture was used to inoculate 1.0 l of TB medium, and then cultures were grown up to an optical density at 600 nm of 0.6 at 37 °C. Cultures were cooled in ice for 15 min before being grown at 18 °C for a further 16–18 h followed by collection by centrifugation. Cell pellets were stored at −80 °C or used immediately for protein purification.
Cell pellets were resuspended in lysis buffer (25 mM Tris pH 7.2, 500 mM NaCl, 2 mM MgCl2, 10 mM imidazole, 10% glycerol) treated with protease inhibitors (2 μg ml−1 aprotinin, 10 μM leupeptin, 0.2 mM AEBSF, 1.0 μg ml−1 pepstatin) and 1 mg ml−1 lysozyme and incubated for 30 min on ice with shaking. Cells were lysed by sonication and clarified by centrifugation. Clarified lysate was batch-bound to nickel resin for 30 min at 4 °C and then allowed to flow through. Lysate was then passed again over nickel resin twice. Nickel resin was washed with nickel wash buffer (25 mM Tris pH 7.2, 2 M NaCl, 2 mM MgCl2, 10 mM imidazole, 10% glycerol) and eluted with nickel elution buffer (25 mM Tris pH 7.2, 500 mM NaCl, 2 mM MgCl2, 250 mM imidazole, 10% glycerol). Nickel elutions were desalted into low-salt buffer (25 mM Tris pH 7.2, 50 mM NaCl, 2 mM MgCl2, 10% glycerol) using either HiPrep 26/10 desalting column (Cytiva) or Pd10 Sephadex G25M Columns (Cytiva) depending on purification size. Protein samples were then loaded over an ion-exchange column (HiTrap SP HP or HiTrap Q HP column (Cytiva) depending on the isoelectric point of the construct) using low-salt buffer and eluted with a gradient of high-salt buffer (25 mM Tris pH 7.2, 1.0 M NaCl, 2 mM MgCl2, 10% glycerol).
Peak elutions were then pooled and concentrated to about 1.0 ml. During this process, protein to be used for biochemistry was desalted by refilling the concentrator with low-salt buffer twice to exchange out the high-salt buffer. Concentrated protein for cryo-EM was loaded over a HiLoad 26/600 Superdex 200 pg column (GE Healthcare) using SEC buffer (25 mM HEPES pH 7.2, 150 mM NaCl, 2 mM MgCl2, 5% glycerol). Peak fractions were then pooled and concentrated again. Concentrated protein was either flash frozen in liquid nitrogen or used in complex formation for cryo-EM.
Complex formation for cryo-EM
Before crRNA was added to Cas12a2, RNA was incubated at 65 °C for 3 min followed by cooling 1 °C min−1 to room temperature. Binary complex was formed for cryo-EM by combining protein and synthetic crRNA in a 1:1.2 molar ratio in SEC buffer (25 mM HEPES pH 7.2, 150 mM NaCl, 2 mM MgCl2, 5% glycerol) and incubating at 24 °C for 10 min. Unbound crRNA was then separated from the binary complex over a Superdex 200 10/300 increase GL sizing column (Cytiva) into Cryo-EM Buffer (12.5 mM HEPES pH 7.2, 150 mM NaCl, 2 mM MgCl2). Eluted protein was concentrated to 30 μM in a 100-kDa MWCO spin concentrator (Corning) and flash frozen in liquid nitrogen.
Far-UV circular dichroism spectroscopy
Protein samples of Cas12a2 mutants were prepared at a concentration in the range 0.3–0.5 mg ml−1 determined by nanodrop in circular dichroism (CD) buffer (20 mM Tris pH 7.2, 100 mM NaCl). Far-UV CD readings used a Jasco-J1500 spectropolarimeter. The CD spectra were obtained from 260–190 nm using a scanning speed of 50 nm min−1 (with a 2 s response time and accumulation of three scans). Melting curves of Cas12a2(ΔPI) samples (in sealed quartz cuvettes with 0.1 cm path length) were obtained by monitoring the CD signal at 222 nm every 1 °C over a 10–90 °C temperature range, using a temperature ramp of 15 °C h−1. The CD signal was converted to molar ellipticity by Jasco Spectra Manager software.
Plasmid curing assay
Plasmid curing assays were conducted as described previously1. In short, immune system plasmids were prepared with Cas12a2 and a 3× CRISPR repeat and transformed into BL21 AI cells by heat-shock transformation. Cells expressing the immune system were then made electrocompetent41 and immediately transformed by electroporation with 50 ng of either target or non-target plasmid. Transformations were recovered for 18 h in 450 μl LB medium containing 1 mM IPTG, 0.2% l-arabinose and antibiotics for the immune system plasmid. Recovered transformations were then serially diluted in LB medium between 101 and 106 and spotted in 10-μl drops on LB agar plates containing 1 mM IPTG, 0.2% l-arabinose and antibiotics for both immune system and target or non-target plasmids. Colonies were counted in the highest countable spot in the dilution series and the relative transformation efficiency was calculated between the target and non-target plasmid.
Binary complex of either WT Cas12a2 or Cas12a2(ΔPI) with crRNA was combined with various targets (FL, ∆PFS, ∆5 and ∆10) to final reaction conditions of 600 nM Cas12a2, 720 nM crRNA and 300 nM FAM-labelled target in 1× NEB 3.1 Buffer (50 mM Tris pH 7.9, 100 mM NaCl, 10 mM MgCl2, 100 μg ml−1 BSA). Binary complex was first formed by combining WT Cas12a2 or Cas12a2(ΔPI) with crRNA in a 1:1.2 molar ratio and incubating for 30 min at room temperature with NEB 3.1 buffer as a 2× master mix. Binary complex and target RNAs were then combined to their final reaction concentration and incubated at room temperature for 1.0 h before quenching with 1:1 (v/v) phenol–chloroform pH 4.5 mixed by flicking.
Reactions were quenched with phenol–chloroform pH 4.5 and mixed by flicking followed by spinning down for 30 s. Reaction products were run on a 12% fully denaturing formaldehyde (FDF) polyacrylamide gel electrophoresis (PAGE) gel as described by previously42 with minor modifications. Loading dye was replaced with 30% glycerol, and gels were run at 50 V for 15 min before increasing the voltage to 150 V.
Target protection assay
Binary complex of Cas12a2 and crRNA was combined with FAM-labelled FL target RNA to a final reaction condition of 600 nM Cas12a2, 720 nM crRNA, 300 nM target RNA in 1× NEB 3.1 buffer (50 mM Tris pH 7.9, 100 mM NaCl, 10 mM MgCl2, 100 μg ml−1 BSA). Binary complex was first formed as described in the activation assay. The 2× master mix was then combined with the target RNA and incubated at 37 °C. Samples were taken at time points 5, 15, 30, 60 and 120 min, and quenched in pH 4.5 phenol–chloroform. Samples were then run on a 12% FDF–PAGE gel as described in the activation assay. Completed gels were imaged for FAM fluorescence and then stained with SYBR Gold and imaged again to show unlabelled RNA species.
Effect of ratios on cleavage
FAM-labelled FL target RNA (300 nM) was combined with a range of binary complex concentrations (600, 300, 150, 75 and 37.5 nM) to achieve complex/target ratios of 2:1, 1:1, 1:2, 1:4 and 1:8. Binary complex was first formed by combining Cas12a2 (1,200 nM) with crRNA in a 1:1.2 molar ratio and incubating for 30 min at 37 °C with NEB 3.1 buffer as a 2× master mix. Binary complex was then serially diluted in 2× NEB 3.1 buffer (100 mM Tris pH 7.9, 200 mM NaCl, 20 mM MgCl2, 200 μg ml−1 BSA) to form a 2× master mix for each complex/target ratio. Each 2× master mix was combined with target RNA and incubated at 37 °C for 1.0 h followed by quenching with phenol–chloroform pH 4.5. Quenched samples were visualized as described in the activation assay.
Reactions containing 600 nM Cas12a2 (WT or mutants) and 720 nM crRNA were combined with 300 nM FL target RNA and 300 nM FAM-labelled non-target RNA, ssDNA or dsDNA. Binary complex was first formed as described previously as a 4× master mix. The master mix was combined with target and non-target substrates and incubated at 37 °C for 1.0 h and quenched with phenol–chloroform pH 4.5. Samples were then separated on a 12% 7 M urea PAGE gel and imaged for FAM fluorescence.
To test the effect of pre-incubation of target with Cas12a2, 100 nM Cas12a2–crRNA complex was incubated for 2 h at room temperature with 200 nM target RNA in NEB 3.1 buffer (50 mM Tris-HCl pH 7.9, 100 mM NaCl, 10 mM MgCl2, 100 μg ml−1 BSA). After incubation, Cas12a2/crRNA/target mixture was combined with 1 μM RNAse Alert or DNAse Alert (IDT), and the reactions were allowed to proceed for 60 min at room temperature. These reactions were compared to the same reaction condition with target RNA added simultaneously (no incubation) with RNAse Alert or DNAse Alert. Fluorescent signal was tracked with a Synergy H4 plate reader (BioTek) and data were plotted in GraphPad Prism.
Plasmid cleavage assay
Plasmid cleavage reactions were prepared by combining 14 nM Cas12a2 (or mutants) with 14 nM crRNA and 25 nM target RNA in NEB 3.1 buffer (50 mM Tris-HCl pH 7.9, 100 mM NaCl, 10 mM MgCl2, 100 μg ml−1 BSA). Protein was preheated at 37 °C for 15 min before the addition of 7 nM supercoiled pUC19 plasmid. Samples were taken at time points of 1, 2, 5, 10, 20, 30, 45 and 60 min and quenched in pH 8.0 phenol–chloroform. Quenched reactions were mixed by flicking followed by centrifugation. Samples were loaded on 1% agarose gels and visualized with ethidium bromide.
Cryo-EM sample preparation and data acquisition and processing
Flash-frozen Cas12a2 binary complex was rapidly thawed. A 4 µl volume of the binary complex was applied to C-flat holey carbon grids (2/2, 400 mesh) that had been plasma-cleaned for 30 s in a Solarus 950 plasma cleaner (Gatan) with a 4:1 ratio of O2/H2. Grids were blotted with Vitrobot Mark IV (Thermo Fisher) for 2 s, blot force 4 at 4 °C and 100% humidity, and plunge-frozen in liquid ethane. Data were collected on an FEI Glacios cryo-TEM equipped with a Falcon 4 detector. Data were collected in SerialEM v3.8, with a pixel size of 0.94 Å, a defocus range of −1.5 to −2.5 µm and a total exposure time of 15 s resulting in a total accumulated dose of 40 electrons Å−2 that was split into 60 electron event representation fractions. Motion correction, contrast transfer function (CTF) estimation and particle picking was carried out on-the-fly using cryoSPARC Live v4.0.0-privatebeta.2 (ref. 43). A total of 1,577 videos were collected, of which 1,159 were accepted on the basis of meeting the criteria of a CTF fit of 5 Å or better. All subsequent data processing was carried out in cryoSPARC v3.2 (ref. 44).
From 987,122 particles picked, 214,647 were selected from a single round of two-dimensional (2D) classification. These particles were subjected to ab initio reconstruction (three classes) followed by heterogeneous refinement resulting in a final subset of 97,470 particles that yielded a 3.46-Å-resolution structure from non-uniform refinement. Re-extraction of this subset of particles in a 320-pixel box size, splitting of particles into 4 exposure groups and carrying out per-group CTF refinement and per-particle defocus optimization as implemented in non-uniform refinement45 resulted in a 3.2-Å-resolution reconstruction that was used for modelling.
For the ternary complex, a rapidly thawed Cas12a2 binary complex fraction was supplemented with a fourfold excess of heat-annealed (90 °C for 5 min, and rapidly cooled to 4 °C) PFS-containing RNA target and incubated at room temperature (about 25 °C) for 30 min before vitrification, which was carried out in an identical manner to that for the binary complex as described above. Data were collected using an FEI Titan Krios cryo-electron microscope equipped with a K3 Summit direct electron detector (Gatan). Images were recorded with SerialEM46 with a pixel size of 0.81 Å. A total accumulated dose of 70 electrons Å−2 during a 6-s exposure was fractionated into 80 frames. A total of 6,940 micrographs were collected, of which 6,614 with CTF fits of 5 Å or better were retained. On-the-fly processing was carried out as described above.
A total of 3,515,037 particles were picked, of which 2,212,319 were selected after 2D classification. Multiple rounds of ab initio reconstruction and heterogeneous refinement resulted in a subset of 192,639 particles that were reconstructed to 2.92-Å-resolution using non-uniform refinement as described above. This map was then used for modelling.
For the quaternary complex, ternary complex was prepared as described above, and incubated with heat-annealed phosphothioate dsDNA duplex for 30 min at room temperature. A 2.5 µl volume of complex was applied to C-flat grids (1.2/1.3, 300 mesh) and blotted for 6 s, blot force 0 at 4 °C and 100% humidity before vitrification. Data were collected on an FEI Glacios cryo-TEM equipped with a Falcon 4 detector, as described for the binary complex. A total of 1,755 videos were collected, of which 1,539 had CTF fits of 5 Å or better and were retained for subsequent processing. On-the-fly motion correction, CTF estimation and particle picking were carried out as described above.
A total of 1,692,368 particles were picked, of which 425,770 were retained after a single round of 2D classification. A single round of ab initio reconstruction (3 classes) followed by heterogeneous refinement yielded a subset of 260,958 particles that were reconstructed to 2.97 Å resolution using non-uniform refinement. Additional rounds of ab initio reconstruction and heterogeneous refinement were used to further classify particles, resulting in a final subset of 104,857 particles. Extraction of said particles with a 384-pixel box, splitting particles into 9 exposure groups, and reconstruction using non-uniform refinement with per-group CTF refinement and per-particle defocus optimization resulted in a 2.74-Å-resolution reconstruction that was then used for modelling.
Model building and figure preparation
A Cas12a binary complex (Protein Data Bank (PDB) 5NG6)20 was rigid body fitted into the Cas12a2 binary complex map. Although most of the model did not correspond to the map, the RuvC and WED domains generally were consistent. However, pairwise blast of the two proteins revealed multiple gaps or inserts for the relative insertions. However, a single 20-residue hairpin of the Cas12a RuvC domain fitted the Cas12a2 map well and had no gaps or insertions in the pairwise blast. This fragment was isolated and rigid body fitted into the Cas12a2 map, and the sequence was mutated to the corresponding region of Cas12a2 using Coot v1.0 (ref. 47). This was then used as a fiducial to build the rest of the complex de novo using in Coot. Attempts at using AlphaFold2 (AF2)48 to generate fragments to fit in the map were unsuccessful as adjacent residues within the WED and RuvC domains were separated by protein sequence and the REC1 and REC2 domain boundaries were not obvious from the sequence alone. However, AF2 was used to validate modelling of small structural domains after-the-fact, for which smaller, compact regions of the model were folded using AF2, and fitted into the map, indicating correct modelling. This was particularly useful when the de novo model contained gaps due to local flexibility.
The 5′ crRNA handle was built using the Cas12a 5′ crRNA as a template (PDB 5NG6). In the binary complex, the seven-nucleotide 3′ seed region was modelled de novo as polyU as it was not possible to unambiguously determine nucleotide identity.
Once fully modelled, Isolde v1.4 (ref. 49) was used to improve the fit of the model to the map, and real-space refinement as implemented within Phenix v1.19 (ref. 50) was carried out to optimize model geometry.
For the Cas12a2 ternary complex, the RuvC, WED, and part of the insert domains were in the same conformation as in the binary complex. These were rigid body fitted into the ternary complex map. The REC1, REC2 and the C-terminal half of the insertion domain were separately fitted as rigid bodies into the ternary complex map. The PI domain structure was predicted using AF2, and then manually connected to the rest of the model. The crRNA–target RNA duplex was modelled as ideal A-form RNA within Coot, and manually connected to the 5′ crRNA handle. The target RNA 3′ PFS was modelled de novo. Coot was used to fit in gaps within the model, and Isolde was then used to improve the quality of model before real-space refinement as described above.
For the quaternary complex, the ternary complex structure was rigid body fitted into the map, and then flexibly fitted using Isolde. The ZR domain structure was predicted using AF2, and manually connected to the rest of the model. The dsDNA duplex was modelled de novo, with one strand modelled as polyT and the other as polyA as it was not possible to unambiguously determine nucleotide identity. Mg2+ and Zn2+ ions and an activating H2O were modelled manually using the sharpened map. Isolde and real-space refinement were carried out as described above.
All structural figures and videos were generated using ChimeraX v1.0 (refs. 51,52), apart from the modevectors, which were generated in PyMol v2.5.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.