Efficient and accurate prime editing system in plants

Prime editing (PE), a new Clustered regularly interspaced short palindromic repeats (CRISPR)-based tool, enables a broad spectrum of genetic changes. PE combines a fusion protein of a Cas9 nickase (nCas9) and M-MLV reverse transcriptase with a PE guide (peg)RNA composed of a single guide (sg)RNA and a 3′ extension containing the reverse transcription (RT) template and primer-binding site (PBS). When initially developed, PE, PE2, which uses a single pegRNA and engineered PE harboring five mutations, and PE3, which uses the engineered PE coupled with single pegRNA and additional nicking sgRNA targeting the nonedited strand, were developed in a stepwise manner, establishing PE3 as a highly reproducible method in mammalian cells (Anzalone et al., 2019). However, efficiency varied considerably among target sites and cell types (Anzalone et al., 2019). Efficient PE systems have now been applied widely to programmable and precise genome editing in a variety of organisms, including several plant species (Chen & Liu, 2023). Efforts to further optimize effective PE in mammalian (Anzalone et al., 2022; Choi et al., 2022) and plant (J. Li et al., 2022; Jiang et al., 2022) cells have encompassed several approaches, including transient inhibition of p53 (M. Li et al., 2022) or mismatch repair-related protein (Chen et al., 2021), separation of nCas9 and RT enzyme (B. Liu et al., 2022), addition of T5 exonucleases to the RT enzyme (Liang et al., 2023), engineering the RT enzyme (Zong et al., 2022), treatment with Histone Deacetylase (HDAC) inhibitor (N. Liu et al., 2022), and improving the pegRNA by using circular RNA (B. Liu et al., 2022). Among these previous reports, major breakthroughs have been achieved using (1) engineered pegRNA (epegRNA) with an RNA pseudoknot sequence at the 3′ end to enhance stability and prevent degradation (Nelson et al., 2022) and (2) paired pegRNAs (Lin et al., 2021) to edit different DNA strands simultaneously by RT of two templates. The increased PE frequency highlights the concomitant introduction of undesired edits (roughly classified into two types) occurring in addition to desired mutations at target sites (Anzalone et al., 2019; Nelson et al., 2022; Zong et al., 2022). One type of undesired mutation occurring via PE is the insertion/deletion (indel) mutations and/or tandem duplications induced by endogenous DNA repair pathways. Fusion of an nCas9 variant (H840A + N854A) with M-MLV RT enzyme led to the decreased introduction of unwanted indels and increased frequency of correct edits compared with nCas9 (H840A) in mammalian cells (Lee et al., 2023). A second type of undesired mutation is scaffold-derived by-products generated by incorporation into the genome of reverse transcription products of the scaffold sequence of the pegRNA. Chen et al. (2021) recoded the 3′ end of the pegRNA scaffold to avoid sequence homology with the genomic target, and reported a reduced incidence of scaffold-derived byproducts; however, this approach, although promising, has not yet been fully investigated. Recently, we determined cryogenic electron microscopy structures of the nSpCas9-based PE-pegRNA-target DNA complex, revealing that M-MLV-RT enzyme can extend reverse transcription beyond the end of the RT template into the scaffold region (Shuto et al., 2024). Furthermore, an in vitro PE assay using PE2 revealed that the RT products were consistently three nucleotides longer than the length of the RT template sequence (Shuto et al., 2024). Therefore, we developed a high-fidelity pegRNA (pegRNA-HF) by modifying the scaffold sequence to replace the last 3-nt of the pegRNA scaffold (5′-UGC-3′) with 3-nt of the target sequence adjacent to the RT template, changing a matched pair in the stem-loop-forming region (Shuto et al., 2024). However, undesired scaffold-derived mutations in mammalian cells are much less frequent than the indels induced by error-prone repair (Chen et al., 2021; Shuto et al., 2024); therefore, it was not clear whether pegRNA-HF could suppress scaffold-derived mutations in mammalian cells (Shuto et al., 2024). Here, we present an efficient PE system in rice, developed using a combination of engineered pegRNA (epegRNA) to enhance stability and prevent degradation of the 3′ extension, and paired pegRNAs to target different DNA strands, synthesizing edited DNA sequences simultaneously by RT of each template. Using this approach, we observed that undesired scaffold-derived mutations occurred more frequently in regenerated rice plants than indels resulting from DNA double-strand break (DSB) repair. To improve the fidelity of the precise edits, we next tested a PE system using paired epegRNA-HFs (PE with paired epegRNA-HFs). Our results show that PE with paired epegRNA-HFs in rice avoids the introduction of undesired scaffold-derived byproducts into the target locus, resulting in the establishment of an efficient and accurate PE system in plant cells. The pegRNA cloning plasmid, pOsU6_BbsIx2_poly, was constructed by removing the SpCas9 scaffold sequence from the gRNA cloning vector, pU6gRNA-oligo (Mikami et al., 2015), which has two BbsI sites between the OsU6 promoter and the gRNA scaffold-poly T sequence, using site-directed mutagenesis with the primer set listed in Supporting Information Table S1. The epegRNA cloning plasmid, pOsU6_BbsIx2_tQ1_polyT, which carries an epegRNA acceptor cassette (OsU6 promoter:BbsIx2:tevopreQ1:polyT) was constructed as follows: fragment (BbsIx2:evopreQ1) was synthesized by overlap extension PCR (Nelson & Fitch, 2011) with the paired oligo DNA carrying a 20 bp overlap (Table S1) and inserted into the BbsI sites of pOsU6_BbsIx2_poly using an In-Fusion Snap cloning kit (TaKaRa, Shiga, Japan). The PE expression vector, pZH/PEwt (Os Opt), was constructed from the CRISPR/Cas9 expression vector described in a previous study as pPZP_ZmUbi:ABE8e_nSpCas9-NG (Endo et al., 2025). The ABE8e_nSpCas9-NG coding sequence was digested with SmaI/KpnI from the pPZP_ZmUbi:ABE8e_nSpCas9-NG vector and was replaced with a rice codon-optimized SpCas9 coding sequence (Mikami et al., 2015) amplified by PCR using the primer sets listed in Table S1 using the In-fusion HD cloning system (TaKaRa). The SpCas9 expression vector was digested with MluI/PvuII, and the N863A mutation was introduced by inserting a SpCas9-N863A fragment digested from a nuclease-dead SpCas9 (SpdCas9) expression vector using a conventional ligation method. Rice codon-optimized M-MLV RT coding sequence (SV40 NLS_engineered M-MLV RT (D200N/L603W/T330P/T306K/W313F) (Anzalone et al., 2019)_SV40 NLS) was synthesized by GeneArt Gene Synthesis and introduced into the KpnI site located at the C-terminal region of nSpCas9-N863A using the In-fusion HD cloning system (TaKaRa). The pegRNA and nicking gRNA for PE3, paired pegRNAs, and epegRNAs were designed with pegFinder (http://pegfinder.sidichenlab.org/) (Chow et al., 2021) and PlantPegDesinger (http://www.plantgenomeediting.net/) (Lin et al., 2021), respectively. The 8-nt linkers between PBS and tevopreQ1 for epegRNAs were designed using pegLIT (https://peglit.liugroup.us/) (Nelson et al., 2022). In this study, we designed the pegRNA and nicking gRNA for PE3 and paired pegRNAs, epegRNAs, and epegRNA-HFs to introduce S627I (aGt to aAt) into the rice acetolactate synthase (OsALS, LOC_Os02g30630) gene. To introduce R115H (cGc to cAc) into the rice Orange (OsOr, LOC_Os02g43500) gene, we designed paired pegRNAs, epegRNAs, and epegRNA-HFs. We designed the paired epegRNAs and epegRNA-HFs to introduce F140H (TTt to CAt) and D20stop (GaC to TaG) into the rice 4-hydroxyphenylpyruvate dioxygenase inhibitor sensitive 1-like (OsHSL1, LOC_Os06g07932) and the rice granule-bound starch synthase (OsGBSS or OsWaxy, LOC_Os06g04200) gene, respectively. To construct the nicking sgRNA expression plasmid for the PE3 system targeting the OsALS gene, 20-nt annealed oligonucleotide pairs for the target sequences shown in Table S1 were cloned into the BsaI site of the sgRNA expression vector described in a previous study as pU3gRNA-oligo (Mikami et al., 2015). To construct the pegRNA expression cassettes, the complementary oligonucleotides, pegRNA_5′F/5′R and pegRNA_3′F/3′R (Table S1), which consist of the 20-nt pegRNA protospacer::5′ half of scaffold sequence and the 3′ half of scaffold sequence::RT template::PBS, respectively, were annealed and ligated together into the BbsI site of pOsU6_BbsIx2_poly. To construct the epegRNA expression cassette targeting OsALS and OsOr, the complementary oligonucleotides, pegRNA_5′F/5′R and pegRNA_3′F_Lt/3′R_Lt (Table S1), which consist of the 20-nt pegRNA protospacer::5′ half of scaffold sequence and the 3′ half of scaffold sequence::RT template::PBS::8-nt linker, respectively, were annealed and ligated together into the BbsI site of pOsU6_BbsIx2_tQ1_polyT. To construct the epegRNA targeting OsHSL1 and OsWaxy or the epegRNA-HF targeting the OsALS, OsOr, OsHSL1, and OsWaxy expression cassettes, the oligonucleotide pairs listed in Table S1, including the 20-nt complementary sequence each 5′ and 3′ end, respectively, were synthesized and subjected to overlap extension PCR. The resulting fragments (the 20-nt pegRNA protospacer::original or modified scaffold sequence::RT template::PBS::8-nt linker) were introduced into the BbsI site of pOsU6_BbsIx2_tQ1_polyT using the In-fusion HD cloning system (TaKaRa). For the PE3 system targeting the OsALS gene, the nicking sgRNA and pegRNA expression cassettes were amplified by PCR with the primers shown in Table S1, and introduced sequentially into the PmeI and AscI/PacI sites, respectively, of pZH/PEwt (Os Opt) vector using the In-Fusion Snap cloning kit (TaKaRa). For the PE with paired pegRNAs, epegRNA, or epegRNA-HF system, the paired pegRNA, epegRNA, or epegRNA-HF expression cassettes were amplified by PCR and inserted concomitantly into the AscI/PacI site of pZH/PEwt (Os Opt) vector using the In-Fusion Snap cloning kit (TaKaRa). These binary vectors were transformed into Agrobacterium tumefaciens EHA105 (Hood et al., 1993) by electroporation (Bio-Rad). Rice (Oryza sativa L. cv Nipponbare) was used in this study. Four-week-old rice scutellum-derived calli grown on N6D medium were used for Agrobacterium-mediated transformation following the procedure of Toki et al. (2006). After 3 d cocultivation, transformed calli were washed and transferred to N6D medium containing 50 mg l−1 hygromycin B (FUJIFILM Wako Pure Chemical, Osaka, Japan) and 25 mg l−1 meropenem (Wako Pure Chemicals) for 2 wk. Antibiotic-resistant clonal callus lines were transferred to fresh N6D medium containing the same antibiotics used in first-round selection and 25 mg l−1 meropenem and were cultured for a further 1 wk. Transgenic callus lines carrying the desired mutations in the target gene were transferred to regeneration medium with 20 μg l−1 1-naphthalene acetic acid and 2 mg l−1 kinetin containing 25 mg l−1 meropenem and cultured for 4 wk for plant regeneration. Genomic DNA was extracted from clonal propagated transgenic calli according to the protocol for rapid DNA extraction reported by Kasajima et al. (2004). Around 96 (for PE3 and paired pegRNAs) or 48 (paired epegRNAs or epegRNA-HFs) independent transgenic callus lines were cloned and subjected to genomic DNA extraction for Cleaved Amplified Polymorphic Sequence (CAPS) or sequencing analysis to detect mutations in the target genes in each transformation experiment. The PE frequency in the transgenic rice calli was calculated as the ratio of the number of calli carrying the detectable targeted mutation via PE to the number of transgenic calli analyzed. More than three biological replicates were performed and calculated values are shown as dots on bar graphs indicating mean percentage scores. Transgenic callus lines carrying the desired mutation in the target gene were transferred to regeneration medium. Genomic DNA was extracted from regenerated plants and was subjected to direct sequencing analysis to detect the existence of desired and undesired mutations in the target gene. To detect the S627I mutation (AGT to ATT), which generates a recognition site for the restriction enzyme MfeI (CAATTG, I627 shown in italic), in the OsALS gene via PE, PCR amplifications were performed with KOD one PCR master mix (Toyobo, Osaka, Japan) using the primer sets shown in Table S1. PCR products were digested by the restriction enzyme MfeI and analyzed using MCE-202 MultiNA with a DNA-500 kit (Shimadzu, Kyoto, Japan). To detect the mutations introduced into the OsOr, OsHSL1, and OsWaxy gene via PE, and to confirm the existence of the desired mutation and undesired scaffold-derived byproducts in the regenerated plants, PCR amplifications were performed with KOD one PCR master mix (Toyobo) using the primer sets listed in Table S1. The PCR products were purified with AMPure XP (Beckman Coulter, Brea, CA, USA) and subjected to sequencing analysis using a Sanger sequencing kit (Thermo Fisher Scientific, Waltham, MA, USA). Seeds of wild-type (WT) rice (Nipponbare) and T1 progeny plants carrying a homozygous S627I mutation in the OsALS gene were sown on ½-strength Murashige & Skoog medium with or without 0.25 μM bispyribac (Kumiai Chemical Industry Co., Tokyo, Japan). After a 12-d treatment period, photographs of seedlings were taken. WT (Nipponbare) or T1 with a homozygous R115H mutation in the OsOr gene were subjected to carotenoid analysis; seeds were germinated on N6D medium for 4 wk then transferred to ReIII medium. After a 1-wk culture on ReIII medium, calli were collected and frozen in liquid nitrogen. Seeds and calli were ground in liquid nitrogen and suspended in 100 μl of methanol; 50 μl of 50 mM Tris–HCl (pH 7.5), 150 μl of 1 M NaCl, and 200 μl of chloroform were added to the solution, mixed for 5 min, and centrifuged. The lower phase was collected and dried by centrifugal evaporation. The pellet was suspended with ethyl acetate and analyzed using High-Performance Liquid Chromatography with a Photo Diode Array (HPLC-PDA). HPLC analysis was performed using a Waters Alliance 2695-2996 system with a photo diode array detector as described previously (Takemura et al., 2015). A TSKgel ODS-80s column (4.6 × 150 nm, 5 μm; Tosoh, Tokyo, Japan) was used. Samples were eluted at a flow of at with two A 5 and B A was for 5 min, by a from A to B for 5 min, and then B for were by and with are shown with each at When two was used and are with For was used and are with different To the efficiency of PE3 with that of PE systems coupled with paired pegRNAs or epegRNAs, we introduced a S627I mutation into the rice acetolactate synthase gene et al., a recognition site for the restriction enzyme MfeI at the target site nicking sgRNA and a single pegRNA were designed for PE3 by and OsU6 respectively, in a PE3 vector harboring expression cassettes for and a gene analysis of PCR products from transgenic calli revealed that of these PE3 vectors induced the desired mutation in transgenic calli at frequency of transgenic calli transformed with a PE vector harboring paired pegRNAs targeting a different DNA by OsU6 the desired mutation in the OsALS gene to whether the epegRNA Nelson et al., 2022) PE efficiency in rice, we introduced PE vectors harboring paired epegRNAs expression cassettes into rice calli Around of transgenic calli PE + paired epegRNAs were shown to cells carrying of the S627I mutation in the OsALS gene Sanger sequencing of plants regenerated from independent callus lines with a of cells revealed that the S627I mutation in the OsALS gene was in plants PE + paired pegRNAs, and undesired mutations, as indels and were in these plants (Table However, although homozygous S627I mutations were in plants PE + paired epegRNAs, additional by-products with desired mutations (Table mutations occurring via PE can classified as scaffold-derived byproducts by RT of the scaffold sequence or indels at the nicking site of the gRNA In this study, mutations at the nicking site were observed (Table scaffold-derived byproducts were in regenerated plants PE with paired epegRNAs, although in a of plants (Table and not as the of incorporation of the last at the 3′ end of the sgRNA scaffold To undesired scaffold-derived we designed paired epegRNA-HFs. The last 3-nt of the epegRNA-HF scaffold was replaced with 3-nt of the target sequence adjacent to the RT template to for or for changing a matched pair to for or for in the stem-loop-forming region analysis with PCR the OsALS revealed that the S627I mutation frequency of PE coupled with epegRNA-HFs was with that of epegRNAs Furthermore, we analyzed regenerated plants from each of the independent callus lines PE + paired epegRNAs or respectively, for the existence of the desired mutation and undesired scaffold-derived by-products at the OsALS by direct sequencing of PCR The scaffold-derived by-products were concomitantly with the desired mutation in at one plant in of lines PE with regenerated plants PE and paired epegRNA-HFs the desired mutation in the homozygous or with undesired mutations at the target in of the lines tested (Table We that T1 from plants carrying a homozygous S627I mutation in the OsALS gene a compared with wild-type plants We next designed paired pegRNAs, epegRNAs and epegRNA-HFs an R115H mutation into the rice Orange gene the protein that carotenoid mutation of which has been with in et al., 2015). sequencing of PCR products the target site revealed on the sequencing to in and of transgenic callus lines pegRNAs and epegRNAs in mutation frequency were between transgenic calli with epegRNAs and epegRNA-HFs Using pegRNAs, regenerated plants with a R115H mutation a mutation and undesired byproducts (Table We a or homozygous R115H mutation in the OsOr gene from to in the plants regenerated from or independent callus lines epegRNAs or respectively. However, we observed undesired by-products in a of regenerated plants epegRNAs or epegRNA-HFs (Table were undesired byproducts in plants epegRNA-HFs than in plants epegRNAs (Table T1 were from regenerated plants harboring a R115H mutation in the OsOr gene. of the OsOr gene in T1 that the R115H mutation introduced via PE was from plants to T1 from the of seeds with the homozygous R115H mutation in the OsOr gene more in as compared with WT with this was in calli not in WT calli or seeds of WT and by HPLC We further the and of PE with paired epegRNAs and epegRNA-HFs at two in epegRNAs and epegRNA-HFs were designed to F140H and to a in the OsHSL1 and OsWaxy of the F140H mutation into the OsHSL1 gene was to to the dioxygenase et al., 2024). OsWaxy has been used as a gene for genome editing in several plant species of these genes has been reported to to et al., 2015). were in PE in transgenic calli between epegRNAs and epegRNA-HFs at the OsHSL1 and OsWaxy genes Sequence analysis of regenerated plants from transgenic calli PE with epegRNAs revealed in addition to desired mutations, scaffold-derived undesired mutations were in several lines lines in OsHSL1 and lines in (Table the the of undesired mutations in the regenerated plants the PE with epegRNA-HFs was in these target genes lines in OsHSL1 and lines in (Table study has a more efficient PE system by PE with paired epegRNAs in rice has been reported that the efficiency of PE with single epegRNA has been by the of a promoter to a and a promoter in rice (J. Li et al., 2022). with PE with single epegRNA, the PE with paired epegRNAs system developed in this study has the that it can applied to as precise fragment and or of DNA fragments in mammalian and plant cells (Anzalone et al., 2022; Choi et al., 2022; et al., et al., 2025). efficiency the introduction of undesired mutations was in regenerated plants (Table In this study, mutations at the nicking site were observed at the target sites analyzed in regenerated to of the target sites by epegRNAs of mutations at the target site using paired epegRNAs targeting at different DNA strands adjacent to the mutation undesired scaffold-derived byproducts were in a of regenerated plants PE with paired epegRNAs (Table For of undesired scaffold-derived the last at the 3′ end of the sgRNA scaffold is used as an RT template, resulting in the incorporation of an additional or or into target genes et al., 2022; et al., 2023). et al. used RT designed to the scaffold-derived byproducts at the of the to a However, revealed that scaffold-derived by-products not the last at the 3′ end of the sgRNA scaffold further in epegRNAs are to the of scaffold-derived Chen et al. (2021) designed a recoded pegRNA, changing the 3′ end of the pegRNA scaffold sequence to avoid sequence homology with the genomic The 3′ end of the pegRNA scaffold sequences used in this study sequence homology with the sequence the target However, undesired scaffold-derived mutations were in these Table indicating that this not undesired scaffold-derived In this study, we could prevent the introduction of undesired scaffold-derived by-products by using epegRNA-HF in rice (Table introduction of the R115H mutation into the OsOr gene, scaffold-derived by-products were observed in a of regenerated plants PE with epegRNA-HFs (Table indicating that PE could introduce scaffold-derived by-products longer than three nucleotides via RT of scaffold were in mammalian cells (Shuto et al., 2024). Our that the epegRNA-HF not to prevent the incorporation of scaffold-derived undesired mutations in To prevent the introduction of undesired scaffold-derived byproducts further of for the stability of the of the of the RT template, and the region of the tested in We observed a PE frequency from to in transgenic rice calli PE with paired epegRNA-HFs at several target sites in rice, to that of PE with paired epegRNAs has been reported that the PE frequency on the target locus, the of desired mutations, and the PBS or RT in plant and mammalian cells (Lin et al., et al., 2023). the the PE frequency varied in regenerated plants independent lines (Table In we propagated independent transgenic callus lines the selection However, the introduction of a mutation via PE and repair of mutations via the mismatch repair is to by in plant cells the of callus resulting in a of the desired mutation in the independent callus lines the was the of regenerated plants carrying the mutation a of mutations in each varied PE efficiency on the target or each transgenic a for in PE efficiency by using a combination of PE and a promoter (J. Li et al., 2022; Li et al., 2022) to engineering the RT enzyme (Zong et al., 2022), and mismatch repair-related protein (Chen et al., 2021) to these in the we that the efficient and accurate PE described to not plant is in by a from the of for and and are in by a a and for a We are to and for and for Japan) for on the and for and have a to this designed this study and the and collected the and the for the performed the analysis of carotenoid edited the the results and on the to the in the are present in the and/or the Supporting Information S1 of OsALS gene via PE3, PE with paired pegRNAs or of OsOr gene via PE with paired pegRNAs or analysis of rice calli and sequencing from regenerated plants harboring the desired mutation and scaffold-derived mutation at the target Table S1 of primers used in this study. Table of of the OsALS gene in regenerated plants PE with paired pegRNAs Table of of the target gene in regenerated plants PE with epegRNAs or epegRNA-HFs. Table of of the OsOr gene in regenerated plants PE with paired pegRNAs is not for the or of Supporting Information by the than to the The is not for the or of by the than to the for the The with to in and in