Optimized prime editing efficiently generates glyphosate-resistant rice plants carrying homozygous TAP-IVS mutation in EPSPS

Prime editing is a novel and universal CRISPR-Cas-derived precise genome-editing technology and has great potentials for applications in basic plant research and crop molecular breeding (Anzalone et al., 2019Anzalone A.V. Randolph P.B. Davis J.R. Sousa A.A. Koblan L.W. Levy J.M. Chen P.J. Wilson C. Newby G.A. Raguram A. et al.Search-and-replace genome editing without double-strand breaks or donor DNA.Nature. 2019; 576: 149-157Crossref PubMed Scopus (1346) Google Scholar). Although low efficiency has restrained the original prime editors (PEs) from being used as a routine tool for precise genome editing in plants, an iterative update of the PEs is removing this obstacle (Lin et al., 2021Lin Q. Jin S. Zong Y. Yu H. Zhu Z. Liu G. Kou L. Wang Y. Qiu J.L. Li J. et al.High-efficiency prime editing with optimized, paired pegRNAs in plants.Nat. Biotechnol. 2021; 39: 923-927Crossref PubMed Scopus (73) Google Scholar; Xu et al., 2022Xu W. Yang Y. Yang B. Krueger C.J. Xiao Q. Zhao S. Zhang L. Kang G. Wang F. Yi H. et al.A design optimized prime editor with expanded scope and capability in plants.Nat. Plants. 2022; 8: 45-52Crossref PubMed Scopus (16) Google Scholar). Recently, the Liu group reported three optimization strategies for improving prime editing efficiency (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar; Nelson et al., 2022Nelson J.W. Randolph P.B. Shen S.P. Everette K.A. Chen P.J. Anzalone A.V. An M. Newby G.A. Chen J.C. Hsu A. et al.Engineered pegRNAs improve prime editing efficiency.Nat. Biotechnol. 2022; 40: 402-410Crossref PubMed Scopus (61) Google Scholar). The first strategy is based on engineered prime editing guide RNAs (epegRNAs), which were generated by incorporating structured RNA motifs to the 3′ terminus of pegRNAs. This strategy enhances pegRNA stability and prevents degradation of the 3′ extension (Nelson et al., 2022Nelson J.W. Randolph P.B. Shen S.P. Everette K.A. Chen P.J. Anzalone A.V. An M. Newby G.A. Chen J.C. Hsu A. et al.Engineered pegRNAs improve prime editing efficiency.Nat. Biotechnol. 2022; 40: 402-410Crossref PubMed Scopus (61) Google Scholar). The second strategy is based on the optimized PE2 protein (PEmax), which harbors a SpCas9 variant with increased nuclease activity, an additional nuclear localization signal (NLS) sequences, and a new linker between nCas9 and reverse transcriptase (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). The third strategy is based on inhibition of DNA mismatch repair (MMR) in cells (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). In this work, we tested the optimized PEs generated with these three strategies in rice, demonstrating that the optimized PEs greatly improved prime editing efficiency in rice. We named the two optimized PEs ePE3max and ePE5max: the former is comprised of the PEmax protein, an epegRNA with evopreQ1 appended to its 3′ end, and a nicking sgRNA; the latter is comprised of the ePE3max system and a dominant negative OsMLH1 variant for inhibiting MMR. Using the two optimized PEs, we efficiently generated homozygous and heterozygous T173I, A174V, and P177S (TAP-IVS) mutation in EPSPS in rice, which lays a solid foundation for rice non-transgenic glyphosate-resistance breeding.TAP-IVS mutation (T102I, A103V, and P106S) is a naturally occurring EPSPS triple substitution found in the highly glyphosate resistant Amaranthus hybridus population from Argentina (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar). Less or little fitness cost associated with the homozygous TAP-IVS mutation accounted for the proliferation of the mutants within the population, even in absence of glyphosate (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar). Since non-transgenic glyphosate-resistance rice varieties have great values in agriculture, we wondered whether we were able to use the optimized PEs to efficiently generate the TAP-IVS mutation in OsEPSPS for rice non-transgenic glyphosate-resistance breeding (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar; Li et al., 2020Li H. Li J. Chen J. Yan L. Xia L. Precise modifications of both exogenous and endogenous genes in rice by prime editing.Mol. Plant. 2020; 13: 671-674Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar). Thus, we tested the optimized PEs for installation of the TAP-IVS edits in EPSPS in rice protoplasts. Our results indicate that the PEmax architecture improved prime editing efficiency and could synergize with the epegRNA (pegRNA-evopreQ1) in rice protoplasts (Figure 1A ).Since the ePE3max, which is comprised of the PEmax protein, an epegRNA, and a nicking sgRNA, achieved the highest editing efficiency (Figure 1A), we then asked whether we were able to further improve editing efficiency of the ePE3max by co-expressing dominant negative OsMLH1 variants. We generated codon-optimized OsMLH1dn, OsMLH1 E44A, OsMLH1dn E44A, and OsMLH1 1-393-2xNLS, corresponding to human MLH1dn, MLH1 E34A, MLH1dn E34A, and MLH1 1-335-NLS, respectively (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). These variants could integrate into the MMR complex in replacement of OsMLH1 with normal function and interfere the function of MMR, leading to inhibition of MMR. These OsMLH1 variants and the ePE3max constitute the ePE5max system. The ePE5max system did not further improve editing efficiency of the ePE3max in rice protoplasts (Figure 1A).To generate rice lines harboring the TAP-IVS mutation, we generated two optimized PEs, ePE3max-TAP and ePE5max-TAP, for rice transformation (Figure 1B). We also generated two control PEs, PE3-TAP and PE5-TAP, which harbor the original PE protein and pegRNA, for comparison with the two optimized PEs in rice transgenic lines. We transformed rice with these four PEs. We analyzed the edits by the Sanger sequencing of PCR fragments amplified from the transgenic lines (Figure 1C) and next-generation sequencing of PCR amplicons. Next-generation sequencing results indicate that the ePE3max and ePE5max significantly improved editing efficiency in rice transgenic lines compared with PE3 and PE5, respectively (Figure 1D). The ePE5max had similar efficiency of heritable mutations (homozygous and heterozygous) to the ePE3max but had higher efficiency of homozygous mutations than the ePE3max. Using the ePE5max, we achieved 6.1% (10/163) homozygous TAP-IVS mutant lines (Figure 1D; Supplemental Table 1).We transplanted the homozygous TAP-IVS mutant lines and transgenic lines without edits (control) to soil and sprayed 10 mM glyphosate on the seedlings. As expected, the mutant plants showed reliable tolerance compared with the control (Figure 1E). As we previously reported (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar), in this study, once again we observed the new type of byproducts in rice protoplasts and transgenic lines (Figure 1A; Supplemental Figure 1). This type of undesired edits, which we previously described as unbiased heteroduplex DNA repair-derived byproducts (Re-byproducts), was produced upon installation of multiple-base substitution edits (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar). Consistent with the role of nicking sgRNAs in PE3 (Anzalone et al., 2019Anzalone A.V. Randolph P.B. Davis J.R. Sousa A.A. Koblan L.W. Levy J.M. Chen P.J. Wilson C. Newby G.A. Raguram A. et al.Search-and-replace genome editing without double-strand breaks or donor DNA.Nature. 2019; 576: 149-157Crossref PubMed Scopus (1346) Google Scholar), the Re-byproducts relative to the desired edits induced by the ePE3max were much less than those by the ePE2max (Figure 1A). Since prime editing efficiency in the generation of homozygous TAP-IVS mutants has been sufficiently high, the optimized PEs enable researchers to perform in planta-directed evolution of the TAP site of EPSPS, which facilitates overcoming the potential fitness cost, if present, associated with the TAP-IVS mutation (Xu et al., 2021Xu R. Liu X. Li J. Qin R. Wei P. Identification of herbicide resistance OsACC1 mutations via in planta prime-editing-library screening in rice.Nat. Plants. 2021; 7: 888-892Crossref PubMed Scopus (16) Google Scholar).To further evaluate effects of dominant negative OsMLH1 variants on prime editing efficiency, we tested them at more target sites in rice protoplasts. On the whole, we still did not observe significant improvement of prime editing efficiency in rice protoplasts when PEs were co-expressed with OsMLH1 variants (Figures 1F and 1G; Supplemental Figures 2 and 3). Nevertheless, installation of the desired edits at additional 20 target sites demonstrated that both the PEmax architecture and epegRNAs greatly improved prime editing efficiency and that these two strategies could synergize with each other (Figures 1H and 1I; Supplemental Figures 4, 5, and 6). Not surprisingly, we observed that improved prime editing efficiency also led to high-frequency byproducts derived from the pegRNA scaffold (Figure 1H; Supplemental Figures 4, 5, and 6). To overcome this problem, in our previous report, we put forward the rule of termination for design of the reverse transcriptase template (rtT): 5′ terminus of genomic DNA sequences corresponding to the rtT should be terminated behind one to three specific bases in genomic DNA, and the bases should be C, GC, or TGC, which correspond to the 3′ terminus of the sgRNA scaffold (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar). We demonstrate that by using the rule of termination to re-design pegRNAs, we were able to eliminate the byproducts (Supplemental Figures 7 and 8). Since most of the pegRNA-derived byproducts resulted from mistaken reverse transcription of the only one base (C) of the sgRNA scaffold adjoining the rtT, the rule for design of the rtT was able to change most pegRNA-derived byproducts, if present, into desired sequences matching genomic DNA completely.Interestingly, we found that PE3 outperformed PE2 at most sites (Figures 1H and 1I; Supplemental Figures 4, 5, and 6), which seems inconsistent with previous reports (Lin et al., 2020Lin Q. Zong Y. Xue C. Wang S. Jin S. Zhu Z. Wang Y. Anzalone A.V. Raguram A. Doman J.L. et al.Prime genome editing in rice and wheat.Nat. Biotechnol. 2020; 38: 582-585Crossref PubMed Scopus (287) Google Scholar). We reasoned that the 35S-CmYLCV-U6 (35C) composite promoter (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar) we used in this study may account for the inconsistency. To provide evidence, we compared editing efficiency of PE2s and PE3s with pegRNAs driven by the U3 or 35C promoter at four target sites in rice protoplasts (Supplemental Figure 9A and 9B). As expected, the results indicated that the 35C promoter exhibited higher editing efficiency in the two types of PEs and a larger ratio of PE3 to PE2 in efficiency than the U3 promoter. We further reasoned that it was because we selected appropriate nicking sgRNAs that we achieved higher editing efficiency of PE3s than that of PE2s. To provide evidence, we tested the same PE2 in combination with different nicking sgRNAs to constitute PE3 in rice protoplasts. As expected, editing efficiency of PE3 depended on appropriate nicking sgRNAs, which should have the high nicking activity and appropriate positions (Supplemental Figure 9C).To provide evidence that PE3 outperforms PE2 in transgenic lines and that OsMLH1dn is functional at some targets in transgenic lines, we generated four types of PEs at four target sites for rice transformation. Similar to the results in protoplasts, PE3 and PE5 outperformed PE2 and PE4, respectively, in transgenic lines (Supplemental Figure 10; Supplemental Table 2). Since PE4 outperformed PE2 at the four targets (Supplemental Figure 10; Supplemental Table 2), the results suggest that as in protoplasts, OsMLH1dn is more or less effective at least at some targets in transgenic lines. Consistent with the previous report (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar), co-expression of MLH1dn decreased the frequency of insertion or deletion byproducts of PE3 (Figure 1D; Supplemental Figure 10; Supplemental Table 2), providing further evidence that OsMLH1dn is more or less effective at least at some targets. Since the MMR system is critical for maintaining genome stability (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar), the inhibited MMR activity resulting from the constitutive overexpression of MLH1dn may cause random insertion, deletion, and substitution of bases at the whole-genome level. This potential defect remains to be investigated before applying the ePE5max system in practice.In summary, we demonstrate that the PEmax architecture and the epegRNA greatly improved prime editing efficiency in rice, and inhibition of cellular activity of MMR by co-expressing dominant negative OsMLH1 could synergize with the PEmax architecture and the epegRNA at some targets. We also demonstrate that the pegRNA scaffold-derived byproducts, which increased along with enhanced prime editing efficiency, could be eliminated by using the termination rule for design of pegRNAs. Our results suggest that the optimized PEs are now able to remove the main obstacle to broad applications of prime editing in rice. In addition, by using the optimized PE systems, we efficiently generated heritable TAP-IVS mutation in EPSPS in rice, which lays a solid foundation for rice non-transgenic glyphosate-resistance breeding. The optimized PE also facilitates overcoming the potential fitness cost associated with the TAP-IVS mutation by directed evolution of the TAP site of EPSPS.Author contributionsQ.-J.C., Y. Zhou, and X.-C.W. designed the research. Y.J., Y.C., D.Q., J.W., C.X., W.S., Z.C., and Y. Zhang performed the experiments. Q.-J.C., Y.J., Y.C., Y. Zhou, and X.-C.W. wrote the manuscript. Q.-J.C. supervised the project.FundingThis work was supported by grants from the National Natural Science Foundation of China (grant nos. U19A2022 and 31872678 ) and the National Crop Breeding Fund (grant no. 2016YFD0101804 ). Prime editing is a novel and universal CRISPR-Cas-derived precise genome-editing technology and has great potentials for applications in basic plant research and crop molecular breeding (Anzalone et al., 2019Anzalone A.V. Randolph P.B. Davis J.R. Sousa A.A. Koblan L.W. Levy J.M. Chen P.J. Wilson C. Newby G.A. Raguram A. et al.Search-and-replace genome editing without double-strand breaks or donor DNA.Nature. 2019; 576: 149-157Crossref PubMed Scopus (1346) Google Scholar). Although low efficiency has restrained the original prime editors (PEs) from being used as a routine tool for precise genome editing in plants, an iterative update of the PEs is removing this obstacle (Lin et al., 2021Lin Q. Jin S. Zong Y. Yu H. Zhu Z. Liu G. Kou L. Wang Y. Qiu J.L. Li J. et al.High-efficiency prime editing with optimized, paired pegRNAs in plants.Nat. Biotechnol. 2021; 39: 923-927Crossref PubMed Scopus (73) Google Scholar; Xu et al., 2022Xu W. Yang Y. Yang B. Krueger C.J. Xiao Q. Zhao S. Zhang L. Kang G. Wang F. Yi H. et al.A design optimized prime editor with expanded scope and capability in plants.Nat. Plants. 2022; 8: 45-52Crossref PubMed Scopus (16) Google Scholar). Recently, the Liu group reported three optimization strategies for improving prime editing efficiency (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar; Nelson et al., 2022Nelson J.W. Randolph P.B. Shen S.P. Everette K.A. Chen P.J. Anzalone A.V. An M. Newby G.A. Chen J.C. Hsu A. et al.Engineered pegRNAs improve prime editing efficiency.Nat. Biotechnol. 2022; 40: 402-410Crossref PubMed Scopus (61) Google Scholar). The first strategy is based on engineered prime editing guide RNAs (epegRNAs), which were generated by incorporating structured RNA motifs to the 3′ terminus of pegRNAs. This strategy enhances pegRNA stability and prevents degradation of the 3′ extension (Nelson et al., 2022Nelson J.W. Randolph P.B. Shen S.P. Everette K.A. Chen P.J. Anzalone A.V. An M. Newby G.A. Chen J.C. Hsu A. et al.Engineered pegRNAs improve prime editing efficiency.Nat. Biotechnol. 2022; 40: 402-410Crossref PubMed Scopus (61) Google Scholar). The second strategy is based on the optimized PE2 protein (PEmax), which harbors a SpCas9 variant with increased nuclease activity, an additional nuclear localization signal (NLS) sequences, and a new linker between nCas9 and reverse transcriptase (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). The third strategy is based on inhibition of DNA mismatch repair (MMR) in cells (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). In this work, we tested the optimized PEs generated with these three strategies in rice, demonstrating that the optimized PEs greatly improved prime editing efficiency in rice. We named the two optimized PEs ePE3max and ePE5max: the former is comprised of the PEmax protein, an epegRNA with evopreQ1 appended to its 3′ end, and a nicking sgRNA; the latter is comprised of the ePE3max system and a dominant negative OsMLH1 variant for inhibiting MMR. Using the two optimized PEs, we efficiently generated homozygous and heterozygous T173I, A174V, and P177S (TAP-IVS) mutation in EPSPS in rice, which lays a solid foundation for rice non-transgenic glyphosate-resistance breeding. TAP-IVS mutation (T102I, A103V, and P106S) is a naturally occurring EPSPS triple substitution found in the highly glyphosate resistant Amaranthus hybridus population from Argentina (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar). Less or little fitness cost associated with the homozygous TAP-IVS mutation accounted for the proliferation of the mutants within the population, even in absence of glyphosate (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar). Since non-transgenic glyphosate-resistance rice varieties have great values in agriculture, we wondered whether we were able to use the optimized PEs to efficiently generate the TAP-IVS mutation in OsEPSPS for rice non-transgenic glyphosate-resistance breeding (Perotti et al., 2019Perotti V.E. Larran A.S. Palmieri V.E. Martinatto A.K. Alvarez C.E. Tuesca D. Permingeat H.R. A novel triple amino acid substitution in the EPSPS found in a high-level glyphosate-resistant Amaranthus hybridus population from Argentina.Pest Manag. Sci. 2019; 75: 1242-1251Crossref PubMed Scopus (44) Google Scholar; Li et al., 2020Li H. Li J. Chen J. Yan L. Xia L. Precise modifications of both exogenous and endogenous genes in rice by prime editing.Mol. Plant. 2020; 13: 671-674Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar). Thus, we tested the optimized PEs for installation of the TAP-IVS edits in EPSPS in rice protoplasts. Our results indicate that the PEmax architecture improved prime editing efficiency and could synergize with the epegRNA (pegRNA-evopreQ1) in rice protoplasts (Figure 1A ). Since the ePE3max, which is comprised of the PEmax protein, an epegRNA, and a nicking sgRNA, achieved the highest editing efficiency (Figure 1A), we then asked whether we were able to further improve editing efficiency of the ePE3max by co-expressing dominant negative OsMLH1 variants. We generated codon-optimized OsMLH1dn, OsMLH1 E44A, OsMLH1dn E44A, and OsMLH1 1-393-2xNLS, corresponding to human MLH1dn, MLH1 E34A, MLH1dn E34A, and MLH1 1-335-NLS, respectively (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). These variants could integrate into the MMR complex in replacement of OsMLH1 with normal function and interfere the function of MMR, leading to inhibition of MMR. These OsMLH1 variants and the ePE3max constitute the ePE5max system. The ePE5max system did not further improve editing efficiency of the ePE3max in rice protoplasts (Figure 1A). To generate rice lines harboring the TAP-IVS mutation, we generated two optimized PEs, ePE3max-TAP and ePE5max-TAP, for rice transformation (Figure 1B). We also generated two control PEs, PE3-TAP and PE5-TAP, which harbor the original PE protein and pegRNA, for comparison with the two optimized PEs in rice transgenic lines. We transformed rice with these four PEs. We analyzed the edits by the Sanger sequencing of PCR fragments amplified from the transgenic lines (Figure 1C) and next-generation sequencing of PCR amplicons. Next-generation sequencing results indicate that the ePE3max and ePE5max significantly improved editing efficiency in rice transgenic lines compared with PE3 and PE5, respectively (Figure 1D). The ePE5max had similar efficiency of heritable mutations (homozygous and heterozygous) to the ePE3max but had higher efficiency of homozygous mutations than the ePE3max. Using the ePE5max, we achieved 6.1% (10/163) homozygous TAP-IVS mutant lines (Figure 1D; Supplemental Table 1). We transplanted the homozygous TAP-IVS mutant lines and transgenic lines without edits (control) to soil and sprayed 10 mM glyphosate on the seedlings. As expected, the mutant plants showed reliable tolerance compared with the control (Figure 1E). As we previously reported (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar), in this study, once again we observed the new type of byproducts in rice protoplasts and transgenic lines (Figure 1A; Supplemental Figure 1). This type of undesired edits, which we previously described as unbiased heteroduplex DNA repair-derived byproducts (Re-byproducts), was produced upon installation of multiple-base substitution edits (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar). Consistent with the role of nicking sgRNAs in PE3 (Anzalone et al., 2019Anzalone A.V. Randolph P.B. Davis J.R. Sousa A.A. Koblan L.W. Levy J.M. Chen P.J. Wilson C. Newby G.A. Raguram A. et al.Search-and-replace genome editing without double-strand breaks or donor DNA.Nature. 2019; 576: 149-157Crossref PubMed Scopus (1346) Google Scholar), the Re-byproducts relative to the desired edits induced by the ePE3max were much less than those by the ePE2max (Figure 1A). Since prime editing efficiency in the generation of homozygous TAP-IVS mutants has been sufficiently high, the optimized PEs enable researchers to perform in planta-directed evolution of the TAP site of EPSPS, which facilitates overcoming the potential fitness cost, if present, associated with the TAP-IVS mutation (Xu et al., 2021Xu R. Liu X. Li J. Qin R. Wei P. Identification of herbicide resistance OsACC1 mutations via in planta prime-editing-library screening in rice.Nat. Plants. 2021; 7: 888-892Crossref PubMed Scopus (16) Google Scholar). To further evaluate effects of dominant negative OsMLH1 variants on prime editing efficiency, we tested them at more target sites in rice protoplasts. On the whole, we still did not observe significant improvement of prime editing efficiency in rice protoplasts when PEs were co-expressed with OsMLH1 variants (Figures 1F and 1G; Supplemental Figures 2 and 3). Nevertheless, installation of the desired edits at additional 20 target sites demonstrated that both the PEmax architecture and epegRNAs greatly improved prime editing efficiency and that these two strategies could synergize with each other (Figures 1H and 1I; Supplemental Figures 4, 5, and 6). Not surprisingly, we observed that improved prime editing efficiency also led to high-frequency byproducts derived from the pegRNA scaffold (Figure 1H; Supplemental Figures 4, 5, and 6). To overcome this problem, in our previous report, we put forward the rule of termination for design of the reverse transcriptase template (rtT): 5′ terminus of genomic DNA sequences corresponding to the rtT should be terminated behind one to three specific bases in genomic DNA, and the bases should be C, GC, or TGC, which correspond to the 3′ terminus of the sgRNA scaffold (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar). We demonstrate that by using the rule of termination to re-design pegRNAs, we were able to eliminate the byproducts (Supplemental Figures 7 and 8). Since most of the pegRNA-derived byproducts resulted from mistaken reverse transcription of the only one base (C) of the sgRNA scaffold adjoining the rtT, the rule for design of the rtT was able to change most pegRNA-derived byproducts, if present, into desired sequences matching genomic DNA completely. Interestingly, we found that PE3 outperformed PE2 at most sites (Figures 1H and 1I; Supplemental Figures 4, 5, and 6), which seems inconsistent with previous reports (Lin et al., 2020Lin Q. Zong Y. Xue C. Wang S. Jin S. Zhu Z. Wang Y. Anzalone A.V. Raguram A. Doman J.L. et al.Prime genome editing in rice and wheat.Nat. Biotechnol. 2020; 38: 582-585Crossref PubMed Scopus (287) Google Scholar). We reasoned that the 35S-CmYLCV-U6 (35C) composite promoter (Jiang et al., 2020Jiang Y.Y. Chai Y.P. Lu M.H. Han X.L. Lin Q. Zhang Y. Zhang Q. Zhou Y. Wang X.C. Gao C. et al.Prime editing efficiently generates W542L and S621I double mutations in two ALS genes in maize.Genome Biol. 2020; 21: 257Crossref PubMed Scopus (71) Google Scholar) we used in this study may account for the inconsistency. To provide evidence, we compared editing efficiency of PE2s and PE3s with pegRNAs driven by the U3 or 35C promoter at four target sites in rice protoplasts (Supplemental Figure 9A and 9B). As expected, the results indicated that the 35C promoter exhibited higher editing efficiency in the two types of PEs and a larger ratio of PE3 to PE2 in efficiency than the U3 promoter. We further reasoned that it was because we selected appropriate nicking sgRNAs that we achieved higher editing efficiency of PE3s than that of PE2s. To provide evidence, we tested the same PE2 in combination with different nicking sgRNAs to constitute PE3 in rice protoplasts. As expected, editing efficiency of PE3 depended on appropriate nicking sgRNAs, which should have the high nicking activity and appropriate positions (Supplemental Figure 9C). To provide evidence that PE3 outperforms PE2 in transgenic lines and that OsMLH1dn is functional at some targets in transgenic lines, we generated four types of PEs at four target sites for rice transformation. Similar to the results in protoplasts, PE3 and PE5 outperformed PE2 and PE4, respectively, in transgenic lines (Supplemental Figure 10; Supplemental Table 2). Since PE4 outperformed PE2 at the four targets (Supplemental Figure 10; Supplemental Table 2), the results suggest that as in protoplasts, OsMLH1dn is more or less effective at least at some targets in transgenic lines. Consistent with the previous report (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar), co-expression of MLH1dn decreased the frequency of insertion or deletion byproducts of PE3 (Figure 1D; Supplemental Figure 10; Supplemental Table 2), providing further evidence that OsMLH1dn is more or less effective at least at some targets. Since the MMR system is critical for maintaining genome stability (Chen et al., 2021Chen P.J. Hussmann J.A. Yan J. Knipping F. Ravisankar P. Chen P.F. Chen C. Nelson J.W. Newby G.A. Sahin M. et al.Enhanced prime editing systems by manipulating cellular determinants of editing outcomes.Cell. 2021; 184: 5635-5652.e29Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar), the inhibited MMR activity resulting from the constitutive overexpression of MLH1dn may cause random insertion, deletion, and substitution of bases at the whole-genome level. This potential defect remains to be investigated before applying the ePE5max system in practice. In summary, we demonstrate that the PEmax architecture and the epegRNA greatly improved prime editing efficiency in rice, and inhibition of cellular activity of MMR by co-expressing dominant negative OsMLH1 could synergize with the PEmax architecture and the epegRNA at some targets. We also demonstrate that the pegRNA scaffold-derived byproducts, which increased along with enhanced prime editing efficiency, could be eliminated by using the termination rule for design of pegRNAs. Our results suggest that the optimized PEs are now able to remove the main obstacle to broad applications of prime editing in rice. In addition, by using the optimized PE systems, we efficiently generated heritable TAP-IVS mutation in EPSPS in rice, which lays a solid foundation for rice non-transgenic glyphosate-resistance breeding. The optimized PE also facilitates overcoming the potential fitness cost associated with the TAP-IVS mutation by directed evolution of the TAP site of EPSPS. Author contributionsQ.-J.C., Y. Zhou, and X.-C.W. designed the research. Y.J., Y.C., D.Q., J.W., C.X., W.S., Z.C., and Y. Zhang performed the experiments. Q.-J.C., Y.J., Y.C., Y. Zhou, and X.-C.W. wrote the manuscript. Q.-J.C. supervised the project. Q.-J.C., Y. Zhou, and X.-C.W. designed the research. Y.J., Y.C., D.Q., J.W., C.X., W.S., Z.C., and Y. Zhang performed the experiments. Q.-J.C., Y.J., Y.C., Y. Zhou, and X.-C.W. wrote the manuscript. Q.-J.C. supervised the project. FundingThis work was supported by grants from the National Natural Science Foundation of China (grant nos. U19A2022 and 31872678 ) and the National Crop Breeding Fund (grant no. 2016YFD0101804 ).