Prime editing has been widely used to generate targeted, site-specific, DNA sequence modifications including nucleotide substitutions, insertions, and deletions in a wide range of organisms (1). Prime editors (PEs) are composed of Cas9 nickase (nCas9) containing a H840A mutation fused to Moloney-murine leukemia virus reverse transcriptase (M-MLV RT). To serve as a template for directly editing the target site with reverse transcription, the engineered prime editing guide RNA (pegRNA) contains a primer binding site that marks the initiation site for reverse transcription and a reverse transcription template (RTT) sequence encoding the desired edit (2).

The first generation of PE (PE1) was designed by fusing M-MLV RT to either the N- or C-terminal end of nCas9 (H840A) indispensable for prime editing (2). Since its development, the PE system has been further improved mainly by modifying its protein components. The PE2 system was developed by introducing point mutations (D200N/L603W/T330P/T306K/W313F) in M-MLV RT that could improve its thermostability and cDNA synthesis efficiency (2). An engineered plant prime editor (ePPE) was also created by removing the RNase H domain of M-MLV RT (∆RNase H) and introducing a viral nucleocapsid (NC) protein (3). Furthermore, based on the finding that mismatch repair (MMR) interferes with prime editing, a dominant negative version of the human mutL homolog 1 protein (hMLH1dn), a component of the MutSα-MutLα MMR complex, was introduced into PE2 or PE3 to decrease undesired byproducts in the edited strand (PE4 or PE5) (4). The fusion of PEs to the RNA-binding N-terminal domain of La, a small RNA-binding exonuclease protection factor, could also improve prime editing (PE7) (5).

Additionally, the use of engineered pegRNA (epegRNAs) can greatly enhance prime editing efficiency (6). Stability and reverse transcription of pegRNAs can be enhanced by adding RNA secondary structures to pegRNA (6) and adjusting length of the primer binding site (7). A dual-pegRNA strategy, in which two separate pegRNAs introduce the same edits in each DNA strand, can also increase prime editing efficiency (7).

PEs have been improved and exploited mainly in monocot plants, including rice, wheat, and maize (8-13). However, their applications in dicot plants remain limited because of their low editing efficiencies in such plants. Here, based on previous strategies used to improve prime editing efficiency, we generated multiple combinations of PE components and pegRNAs and found that the v4e2 variant showed the highest prime editing efficiency in Arabidopsis thaliana, a model dicot, raising the possibility of prime editing in dicot plants.

A protoplast transient expression assay is a versatile, rapid, and high-throughput method for assessing genome editing efficiency (14, 15). To efficiently evaluate editing efficiencies of multiple PE variants, we optimized our protocol for transient expression assays by increasing protoplast yield and viability (16). To isolate protoplasts, two-week-old seedlings (Fig. 1A) were incubated with an enzyme solution at room temperature for 6 h (Fig. 1B) (16). A sucrose density gradient purification method was then employed to enrich viable protoplasts (Fig. 1C-E). Freshly prepared protoplasts were used strictly for polyethylene glycol (PEG)/Ca^2＋-mediated transfection of plasmid DNAs. Protoplasts transfected with an AtUBQ10::NLS-mCherry plasmid showed a transfection efficiency of approximately 46% at 18 h after transfection (Fig. 1F, G). An optimized transient expression assay was subsequently used for evaluating editing efficiencies of PE variants.

Previous studies have shown that prime editing efficiency can be improved by modifying PE components and/or engineering pegRNAs (2-6). Exploiting these strategies, we generated eight PE variants in this study using varying combinations of Arabidopsis codon-optimized nCas9 (R221K/N394K/H840A) (4), tomato codon-optimized 34-amino acid flexible linker as well as M-MLV RT (8), ∆RNase H (M-MLV lacking the RNase H domain) (3), dominant negative human MLH1 protein (hMLH1dn) (4), viral NC with nucleic acid chaperone activity (3), engineered pegRNA (epegRNA, generated by incorporating structured RNA motifs into the 3’ end of pegRNA) (6), and dual-pegRNAs (7) (Fig. 2A). We used 2×35S and AtU6 promoters to express PE variants and pegRNAs, respectively. Two different genomic regions (AtGL1 and AtPDS3) were chosen as target loci for editing to create a stop codon at the leucine 87 (L87) position of the AtGL1 protein and to disrupt the splicing acceptor site of the first intron of AtPDS3 (Fig. 2B).

To identify PE variant exhibiting the highest prime editing efficiency in the model system, we transiently expressed each PE variant construct into Arabidopsis protoplasts for three days. Genomic DNAs were then isolated from transfected protoplasts for targeted deep sequencing. Although prime editing efficiency varied in the context of sequence, v4e2 showed the highest editing frequency at target sites among all PE variants examined (Fig. 2C). Notably, the relative position of nCas9 and M-MLV RT influenced editing efficiency (2, 17). A v4e2 variant, in which M-MLV RT was fused to the C-terminus of nCas9, showed 2.4-fold and 11-fold higher efficiencies of desired editing at AtGL1 and AtPDS3, respectively, than v4Re2, in which M-MLV RT was fused to the N-terminus of nCas9. Furthermore, the dual-epegRNA strategy substantially enhanced prime editing efficiency. Transfection with v4e2 resulted in 10.9-fold and 5.5-fold higher desired editing efficiencies at AtGL1 and AtPDS3, respectively, than that with v4e1. Results of analyzing product purity of v4e2 showed that desired editing was more predominant than imprecise editing (Fig. 2D).

To determine whether our finding could be applied to plant systems, we transformed each AtGL1-targeting PE variant into Arabidopsis plants using a floral dip method (18) to generate stable transgenic plants. Genomic DNAs were then extracted from rosette leaves of T₁ plants for high-throughput sequencing analysis. Consistent with results of protoplast transfections assays, v4e2 showed the highest prime editing efficiency (up to 9.11%) in terms of average editing frequency in transgenic Arabidopsis plants (Fig. 3). Product purity of v4e2 in transgenic plants also appeared to be similar to that in Arabidopsis protoplasts. Although we did not try to obtain T₂-edited plants, it is highly likely that genome editing is inheritable in transgenic plants, albeit with a low editing efficiency, as observed in previous studies (19-21).

Overall, v4e2 was found to be the best PE among those tested in Arabidopsis. The v4e2 system could be used for prime editing in other dicot plants, although its editing efficiency is still low. Considering that differences in DNA repair mechanisms between mammals and plants may lead to their variable prime editing efficiencies (20, 22), we speculate that dicot plants likely have molecular processes different from those of monocot plants. Such differences can interfere with prime editing. It is important to elucidate dicot-specific genetic factors that hinder the action of PEs and to engineer PE systems that can bypass genetic hindrance in the future for optimizing prime editing in dicot plants.

Plasmid construction

All DNA fragments for plasmid construction were amplified with Primestar GXL DNA polymerase (Takara Bio, R050A) and cloned using the Gibson assembly method. The epegRNA scaffolds were designed based on a previous study (6), and dual-epegRNAs were designed using the PlantPegDesigner webtool (6, 7). All pegRNAs were commercially synthesized by Bionics, and all plasmids were confirmed by Sanger sequencing (Bionics). PE-expressing plasmids were constructed using pPPED (backbone plasmid) (Addgene #162468). A variant of nCas9 (R221K/N394K/H840A) was generated by PCR-based site-directed mutagenesis using a primer containing the desired mutation. Arabidopsis codon-optimized MLH1dn was commercially synthesized (Bionics) and cloned into pPPED by adding 2A-cleavage sequences and NLS, resulting in 2×35S::NLS-nCas9 (R221K/N394K/H840A)-linker-M-MLV RT-NLS-P2A-MLH1dn-NLS (v2e1, expressing a single epegRNA). To investigate the effect of the relative order of nCas9 and M-MLV RT on editing efficiency, the 2×35S::NLS-M-MLV RT-linker-nCas9 (R221K/N394K/H840A)-NLS-P2A-MLH1dn-NLS plasmid (v2Re1) was constructed. To construct the v3 variants, including 2×35S::NLS-nCas9 (R221K/N394K/H840A)-linker-viral NC-NLS-linker-M-MLV RT (∆RNase H)-NLS (v3e2, expressing dual-epegRNA) and 2×35S::NLS- M-MLV RT (∆RNase H)-linker-viral NC-NLS-linker-nCas9 (R221K/N394K/H840A)-NLS (v3Re2), the viral NC domain commercially synthesized by Bionics and M-MLV RT deleted for ribonuclease H (∆RNase H) were cloned into the pPPED backbone using Gibson assembly. The v4 variants containing the single epegRNA-expressing cassette including 2×35S::NLS-nCas9 (R221K/N394K/H840A)-linker-viral NC-NLS-linker-M-MLV RT (∆RNase H)-NLS-P2A-MLH1dn-NLS (v4e1) and 2×35S::NLS- M-MLV RT (∆RNase H)-linker-viral NC-NLS-linker-nCas9 (R221K/N394K/H840A)-NLS-P2A-MLH1dn-NLS (v4Re1) were generated by combinational reconstruction. Subsequently, the v4 variants, namely, v4e2 and v4Re2, were constructed by replacing the single epegRNA-expressing cassette with the dual-epegRNA-expressing cassette in v4e1 and v4Re1, respectively.

Plant growth conditions and plant transformation

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was used to perform all experiments. Surface-sterilized seeds were sown on half-strength Murashige and Skoog (MS) medium containing vitamins supplemented with 0.05% (w/v) 2-morpholinoethanesulfonic acid monohydrate (MES·H₂O), 1% (w/v) sucrose, and 0.8% (w/v) plant agar. After sowing, the seeds were stratified for 3 days at 4°C in the dark. Seedlings were grown under long-day (16 h light/8 h dark) conditions using white fluorescent lamps (120 μE m⁻² s⁻¹) at 23°C.

Arabidopsis plants were transformed with Agrobacterium tumefaciens strain GV3101 harboring individual AtGL1-targeting PE variants using the floral dip method (18) to generate transgenic (T₁) lines.

Plasmid DNA preparation

A single colony of DH5α carrying a certain plasmid was picked and grown overnight in a test-tube containing 2 ml of LB medium supplemented with 100 mg/L kanamycin on an orbital shaker (250 rpm) at 37°C. Then, 0.5-ml of the cell suspension was used to inoculate 40 ml of fresh TB medium supplemented with 100 mg/L kanamycin and grown in 250-ml baffled flasks at 37°C for 16 h with shaking (250 rpm). The overnight culture was transferred to a 50-ml tube. Bacterial cells were collected by centrifugation at 6,500 × g for 10 min and resuspended in 4 ml of TEG buffer (25 mM Tris-HCl, 10 mM EDTA, and 50 mM glucose [pH 8.0]). Then, 8 ml of cell lysis solution (0.2 N NaOH and 1.0% (w/v) SDS) was added to the bacterial cell suspension, mixed gently by inverting 5-7 times, and incubated at room temperature for 10 min. The cell lysate was neutralized with 6 ml of potassium acetate (KOAc, 3 M) solution containing 11.5% (v/v) acetic acid and incubated at room temperature for 10 min. After centrifugation at 14,000 × g for 10 min, the supernatant was passed through two layers of Miracloth (Millipore, 475855). The filtrate was collected in a new 50-ml tube, mixed with 10 ml of isopropyl alcohol, and incubated at room temperature for 10 min. Plasmid DNA was collected by centrifugation at 14,000 × g for 10 min. The DNA pellet was washed with 70% (v/v) ethanol, resuspended in 0.6 ml of TE buffer (25 mM Tris-HCl and 10 mM EDTA [pH 8.0]), and transferred into a 2-ml tube. To remove bacterial RNA, the suspension was mixed with 300 μl of 4.2 M CaCl₂ solution and incubated at room temperature for 5 min. The tube was centrifuged at 12,000 × g for 10 min. The supernatant was transferred into a new 2-ml tube and mixed with 70 μl of 3 M NaOAc and 0.5 ml of isopropyl alcohol. After incubation at room temperature for 10 min, the tube was centrifugated at 12,000 × g for 10 min. The pellet was washed twice with 70% (v/v) ethanol and resuspended in 0.5 ml of TE buffer. To selectively precipitate the plasmid DNA, an equal volume of PEG-NaCl solution (20% [w/v] PEG 8000 and 0.5 M NaCl) was added to the plasmid DNA and mixed gently by inverting 5-7 times. Immediately after mixing, the tube was centrifugated at 12,000 × g for 10 min. The pellet containing purified plasmid DNA was washed twice with 70% (v/v) ethanol and resuspended in TE buffer. Plasmid DNA concentrations were determined using BioDrop (Biochrom, BD1776) and adjusted to a concentration of 2 μg/μl.

Protoplast preparation

Protoplast isolation and purification were performed as described previously (16), with slight modifications. Briefly, 2-week-old Arabidopsis seedlings were soaked in 20 ml of 0.5 M mannitol solution (pH 5.8) at 23°C for 1 h. The plasmolyzed seedlings were immersed in 20 ml of enzyme solution containing 2% (v/v) Viscozyme L, 1% (v/v) Celluclast 1.5 L, and 1% (v/v) Pectinex ultra SP-L and incubated on a rotating shaker (50-60 rpm) for 6 h. Intact protoplasts were collected using sucrose density gradient centrifugation.

Protoplast transfection assay

Protoplast transfection was performed as described previously (14), with the following modification. Purified protoplasts were resuspended in MMG solution (0.4 M mannitol, 15 mM MgCl₂ and 4 mM MES [pH 5.7]) at a concentration of 4 × 10⁶ protoplasts/ml. Twenty micrograms of plasmid DNA was mixed with 300 μl of protoplasts. Then, an equal volume (300 μl) of PEG solution (40% [w/v] PEG 4000, 0.1 M CaCl₂, and 0.2 M mannitol) was added to the protoplasts, mixed, and incubated for 5 min at room temperature. After incubation, 1 ml of W5 solution (154 mM NaCl, 125 mM CaCl₂, 5 mM KCl, 2 mM MES, and 5 mM glucose [pH 5.7]) was added to the protoplasts and gently mixed. Protoplasts were collected by centrifugation at 80 × g for 5 min, washed with 1 ml of W5 solution, centrifuged again at 50 × g for 5 min, and resuspended in 100 μl of PIM medium (Gamborg B5 medium containing vitamins, 20 g/L sucrose, 60 g/L myo-inositol, 2 mg/L 6-BAP, and 0.5 mg/L α-NAA [pH 5.8]). The transfected protoplasts were incubated at 23°C in the dark.

To evaluate transfection efficiency, protoplasts were transfected with pCMU-NUCr (Addgene #61168) (23), incubated for 18 h, and then stained with fluorescein diacetate (FDA) (24). The viability and transfection efficiency of protoplasts were assessed using a CQ1 confocal quantitative image cytometer (Yokogawa).

To assess prime editing efficiency, the transfected protoplasts were incubated for 3 days. Subsequently, the protoplasts were collected and subjected to genomic DNA extraction using the cetyltrimethylammonium bromide (CTAB) method for high-throughput sequencing.

High-throughput sequencing

To perform targeted amplicon sequencing, genomic regions containing the target sites were PCR-amplified using primers, including adapters that are compatible with Illumina index barcodes. Amplified DNAs were purified using Expin PCR SV mini (GeneAll, 103-102), and purified libraries were then sequenced using the Miniseq sequencing system (Illumina). After sequencing, paired-end data were analyzed using a PE-analyzer (http://www.rgenome.net/pe-analyzer/) (25). Sequencing data were mostly analyzed using the computing server at the Genomic Medicine Institute Research Service Center.

This work was supported by the Basic Science Research (NRF-2022R1A2B5B02001266 to P.J.S.; NRF-2021M3A9H3015389 to S.B.) and Basic Research Laboratory (NRF2022R1A4A3024451) programs of the National Research Foundation of Korea, and the New Breeding Technologies Development Program (RS-2024-00322275 to P.J.S.) of the Rural Development Administration.

The authors have no conflicting interests.