A spacer sequence ( Supplementary Table 1 ) was designed to recognise a 20-nucleotide site downstream of a Cas9 Protospacer Adjacent Motif (PAM) site (NGG) in the 5′-untranslated region of the CENH3 (BnCENH3) gene ( Figure 1A ). The spacer sequence was designed using the genomic sequence of the BnCENH3 gene from the canola line DH12075; the target sequences across the two BnCENH3 homeologs (A09 and C08) are shown in Figure 1A . The spacer sequence was also checked to confirm that mutations at the predicted restriction site generated by the RNP complex could, in theory, be detected by a dual-probe ddPCR assay designed simultaneously. The spacer sequence was incorporated into a sgRNA ( Supplementary Table 1 ), which was synthesised by Integrated DNA Technologies (IDT).

Custom Alt-R™ sgRNA and recombinant Alt-R S.p. HiFi Cas9 nuclease protein were purchased from IDT.

Mesophyll protoplasts were isolated from the DH12075 canola line using the method described by Sahab et al. (2019) with some modifications ( Supplementary Methods 1 ). Critical factors determining the efficiency of each step are described in Sahab et al. (2019). CRISPR–Cas9 RNPs were prepared by combining 60 µg of Cas9 protein with 20 µg of sgRNA in 60 µL of transfection buffer or NE Buffer 3.1 (New England Biolabs, Inc., Ipswich, MA, USA). The transfection buffer contained 0.4 M mannitol, 15 mM MgCl₂, and 4 mM morpholinoethanesulphonic acid (MES), pH 5.7. Aliquots of RNP were added to one million protoplasts, followed by 150 μL of PEG solution [polyethylene glycol (40% (w/v) PEG 4000 (source: Sigma, Darmstadt, Germany; Catalogue #81240) in 0.4 M mannitol, 0.1 M Ca (NO3)₂; pH 6–7], and the samples were gently tapped to mix all the ingredients uniformly. After 15 min of incubation at room temperature, 10 mL wash buffer was added to the protoplast suspension. The protoplast suspension was gently mixed, and the protoplasts were pelleted by centrifugation using a swing-out rotor at 70 g for 10 min. The supernatant was removed, and the protoplast pellet was resuspended in 1 mL of fresh wash buffer (154 mM NaCl, 125 mM CaCl₂, and 5 mM glucose; pH 5.8) (Menczel et al., 1981). Transfected protoplasts were incubated in a growth chamber for 48 h at 22°C with a 16-h photoperiod and an average light intensity of 300 μmol·m⁻²·s⁻¹. In the case of transient studies, transfected protoplast pools were directly sampled for DNA extraction after 48 h of incubation. Transfected protoplasts were immediately embedded in a bead-type matrix in order to recover stably edited events ( Supplementary Methods 1 ). A schematic representation of these processes is shown in Figure 1B . In brief, transfected protoplasts were embedded in sodium alginate bead-type cultures until visible microcalli were formed (~2–4 weeks). Initially, microcalli were placed on callus proliferation media. Subsequently, the surviving calli were transferred to shoot regeneration media to continue their development. This process facilitated the recovery of T0 plants with roots, which were transferred to a potting mix and cultivated under controlled glasshouse conditions ( Figure 1C ). The final recovery step, involving recovery of T0 plants from calli, took approximately 4–6 weeks. Overall, a 6–10-week time period was required to complete the recovery of all putative edited T0 events. Of the 1,200 T0 plants recovered, 525 were screened to identify edited events.

Genomic DNA was extracted from protoplasts using the DNAdvance™ kit (Beckman Coulter, Inc., Brea, CA, USA), with minor modifications. Briefly, centrifugation at 10,000 × g for 4 min at 4°C was performed to concentrate protoplast samples, and the supernatant was removed. Each pellet was resuspended in 100 µL DNAdvance™ Lysis Master Mix, vortex-mixed, and incubated for 15 min at 55°C. Samples were cooled on ice for 2 min, mixed with 50 µL of pre-chilled 7.5 M ammonium acetate, and incubated for 1 hour on ice. Following centrifugation at 10,000 × g for 10 min at 4°C, 100 µL of the supernatant from each sample was aspirated and combined with 50 µL of DNAdvance™ Bind 1 solution, vortex-mixed, and combined with 85 µL of DNAdvance™ Bind 2 solution. After vortexing, the samples were incubated for 5 min at ambient temperature and placed on a magnet for 4 min to separate the beads. The supernatant was discarded, and the beads were rinsed three times with 200 µL of 70% v/v ethanol before brief air-drying on the magnet. Genomic DNA was eluted in 30 µL of DNAdvance™ Elution Buffer.

Samples of immature leaf tissue (~100 mg) were collected from T0 plants grown in the glasshouse and placed in racked collection microtubes (Qiagen, Valencia, CA, USA). Leaf tissue was freeze-dried for 48 h prior to DNA extraction. Genomic DNA was isolated using the DNeasy™ 96 Plant Kit (Qiagen), according to the manufacturer’s protocol. Purified DNA samples were quantified using a NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) and diluted to 5 ng/μL for genotyping using a ddPCR assay.

A dual-probe ddPCR assay, also called a drop-off or ablation assay, was used to measure ablation efficiency ( Figure 2A ). The assay was designed using Sequencher 5.0 (Gene Codes, Ann Arbor, MI, USA) and the IDT PrimerQuest™ tool. Primer and probe details are shown in Supplementary Table 2 . A FAM-labelled probe spanning the predicted CRISPR–Cas9 ablation site in the CENH3 gene and a HEX-labelled probe matching another part of CENH3 were located within an amplicon flanked by a forward primer and a reverse primer. ddPCR assays were performed in a 24-µL reaction volume in Twin Tec™ 96-well microplates (Eppendorf) comprising 1× ddPCR Master Mix (no UTP, Bio-Rad Laboratories, Hercules, CA, USA), 0.25 µM HEX- and FAM-labelled probes (IDT), 0.5 µM forward and reverse primers (IDT), and 10 ng of genomic DNA. Reactions were prepared for PCR using an automated droplet generator (Bio-Rad Laboratories), sealed with pierceable foil heat seals (Bio-Rad Laboratories), and cycled using C1000 Touch™ thermocyclers (Bio-Rad Laboratories) with an initial 95°C denaturation for 10 min, followed by 39 cycles of denaturation at 95°C for 30 s, annealing at 59°C for 30 s, and extension at 72°C for 2 min, followed by a final extension at 72°C for 5 min and denaturation at 95°C for 10 min. A ramp rate of 2°C/sec was used for all the steps. Data were collected using a QX200™ Droplet Reader, and droplet clusters were analysed using QuantaSoft™ Analysis Pro software (Bio-Rad Laboratories).

Representative plants from each of the observed ddPCR clustering patterns identified from the assessment of the T0 events were characterised by amplicon sequencing. A detailed description of the approach used for amplicon library preparation, sequencing, and data analysis is provided in Supplementary Methods 2 .

PEG-mediated transfection of mesophyll protoplasts with CRISPR–Cas9 RNPs was performed with either New England Biolabs Buffer 3.1 or transfection buffer. Protoplast transfections without RNPs were used as a negative control. Experiments involving RNPs were performed with two biological replicates derived from independent protoplast preparations and at least three technical replicates ( Table 1A ). Analysis of genomic DNA from mesophyll protoplasts transfected with RNPs showed evidence of CRISPR–Cas9-mediated CENH3 allele editing ( Figure 2B ). As expected, the PEG-only negative control assay showed a large cluster of droplets with high amplitudes of both HEX and FAM fluorescence (orange data points), corresponding to wild-type CENH3 alleles. Assays that involved CRISPR–Cas9 RNP delivery showed smaller clusters of droplets having a high amplitude of HEX fluorescence but reduced FAM fluorescence (green data points), corresponding to CENH3 alleles that were mutated at the target site by CRISPR–Cas9 RNP activity ( Figure 2B ). The frequency of mutated CENH3 alleles ranged from 65% to 76% in the genomic DNA derived from protoplasts treated with RNPs ( Table 1A ; Supplementary Table 3 ). A slightly lower mutation frequency was observed when the RNPs were delivered in the transfection buffer than in NEB Buffer 3.1. Despite this, transfection buffer was used for RNP deliveries aimed at regenerating edited canola plants from protoplasts because of its established use in transformation and regeneration processes (Sahab et al., 2019).

Approximately 1,200 protoplast-derived T0 plants were regenerated from mesophyll protoplasts in tissue culture following transfection with CRISPR–Cas9 RNPs, of which 525 were tested for gene ablations. The high recovery of T0 plants was due to the utilisation of an established protoplast transformation and plant regeneration method (Sahab et al., 2019) in which critical factors determining the success of each key step had been resolved. Genotyping of individual plants by ddPCR showed that they had droplet frequencies corresponding to non-wild-type versions of the canola CENH3 gene between 0% and 100% ( Table 1B ; Supplementary Table 4 ). The DH12075 canola line used in this study is known to have two copies of the CENH3 gene, located on chromosomes A09 and C08, both of which were targeted by CRISPR–Cas9 RNPs. Therefore, on a per-plant basis, non-wild-type droplet frequencies of 0%, 25%, 50%, 75%, and 100% were expected to correspond to the editing of 0, 1, 2, 3, and 4 CENH3 alleles, respectively ( Figure 2C ). Overall, 68.2% of the plants were edited at the CENH3 loci, with 4.8%, 5.7%, 8.2%, and 50% having one, two, three, and four mutated CENH3 alleles, respectively. Amplicon-based sequence characterisation of representative samples revealed corresponding allelic mutations consistent with the ddPCR frequencies ( Figure 2C ). Based on the sequencing data, only one type of allele was observed in the unedited (wild-type) samples where the CRISPR binding site was intact, and the amplicon size was as expected (171 bp long). In contrast, the T0 plants had between one and four mutated alleles with small insertions and deletions ( Figure 2C ). The number of mutated alleles per T0 plant concurred with the observed percentage of ablation droplets for that plant in the ddPCR analysis. The observation that half of the plants had four mutated alleles indicated that the biallelic editing of both endogenous CENH3 loci was highly efficient.

Screening of genomic DNA from protoplasts transfected with CRISPR–Cas9 RNPs by ddPCR represents a robust and rapid in vivo method for testing the efficacy of sgRNA designs. Whilst rapid in vitro tests for assessing the efficacy of gRNA designs using CRISPR–Cas9 RNPs are commercially available to validate cut site specificity, they do not accurately predict in vivo performance because they are not as quantitative as ddPCR, and short target molecules of DNA lack the DNA methylation patterns of genomic DNA in living cells, which is known to inhibit Cas9 activity, if present (Přibylová et al., 2022). These limitations can be overcome with the use of a ddPCR-coupled protoplast assay that allows RNP activity to be tested on targets within the genomic DNA inside plant cells. Overall, this rapid screening method, combining protoplast transfection and a dual-probe ddPCR-based ablation assay, has the potential to test large numbers of sgRNA designs against multiple targets at a relatively low cost, provided that the protoplast system is amenable to high-throughput microplate-based transfection and recovery of a sufficient quantity of genomic DNA for genotyping.

When protoplasts were transfected with CRISPR–Cas9 RNPs, mutation frequencies of between 65% and 76% in CENH3 alleles were measured. Plants regenerated from transfected protoplasts had mutation frequencies of approximately 68% in CENH3 alleles. Of the 525 T0 events tested, approximately 50% had biallelic mutations at both CENH3 loci, i.e., mutations in all four CENH3 alleles ( Tables 1A , B ). These results compare favourably with CRISPR–Cas9 RNP delivery experiments in a range of crop plants (reviewed by Metje-Sprink et al., 2019). Murovec et al. (2018) achieved mutation rates of between 2.5% and 24.5% in B. oleracea and B. rapa protoplasts following the delivery of CRISPR–Cas9 RNPs targeting the phytoene desaturase gene and homologues of the FRIGIDA gene. Sidorov et al. (2021) reported a biallelic editing frequency of up to 50% using RNPs containing LbCas12a and a sgRNA targeting an intergenic region in B. napus protoplasts and regenerated plants. Our results support those of Sidorov et al. (2021) in demonstrating that a CRISPR–Cas9 or CRISPR–Cas12a RNP system involving direct delivery of RNPs to protoplasts can be used for efficient DNA-free genome editing in canola. We also showed that DNA extracted from transfected protoplasts and regenerated plants can be screened for editing efficiency rapidly and in high throughput using the robust and established digital PCR method ( Figure 2C ).

A CRISPR–Cas9 RNP approach to generate transgene-free plants with targeted mutations is likely to become a valuable tool to augment conventional breeding and germplasm development, with a lower regulatory burden than DNA-based genome editing methods. There is also no need for in-house vector construction, characterisation of exogenous DNA in edited plants, and its removal by segregation. The major impediments in generating transgene-free edited plants using direct RNP delivery have been the efficient delivery of RNPs into cells and the selection and recovery of edited plants in the absence of a selectable/screenable marker. In this study, we demonstrated that the effective delivery of CRISPR–Cas9 RNPs into cells can be achieved using a PEG-based protoplast transfection system in canola. Efficient RNP delivery leading to a high rate of cellular uptake appears to result in a high genome editing rate, presumably due to the efficacy of the RNP complex. The high editing efficiency, demonstrated by 62% of the recovered plants carrying mutations in at least one of the targeted CENH3 alleles, shows that the use of a selectable marker is not needed for the recovery of a large number of edited plants. Our study suggests that edited canola plants with a desired allelic profile can be readily recovered, making this system a robust method for the creation and recovery of canola plants with targeted biallelic mutations in multiple homeologs of multicopy genes using a completely DNA-free system.

In conclusion, we have demonstrated an efficient DNA-free approach for efficiently generating gene-edited canola germplasm, as well as a high-throughput in vivo screening assay for assessing the efficacy of new CRISPR gRNA designs. Further work will focus on the comprehensive sequence-based analysis of edited events, the development of multiplexing strategies, and the testing of elite canola germplasm to facilitate the deployment of this technology into breeding systems.