Title: Gene Targeting by Homology-Directed Repair in Rice Using a Geminivirus-Based CRISPR/Cas9 System
Abstract: Rice (Oryza sativa) is the staple food for more than half of the world's population. Technologies enabling precise and efficient DNA knock-in or replacement, hereinafter referred to as KI, have the potential to revolutionize the generation of crops by precision molecular breeding. Clustered regularly interspaced short palindromic repeats (CRISPR)-associated Cas9 (CRISPR/Cas9) has recently emerged as a promising genome editing tool allowing precise genomic manipulation in rice and other crops. However, due to the prevalence of non-homologous end joining (NHEJ) over homology-directed repair (HDR) in the repair of CRISPR/Cas9-induced double-strand breaks (DSBs), this genome editing tool has been mostly used to generate random insertions and deletions (Indels) in precise genomic locations in plants (Cong et al., 2013Cong L. Ran F.A. Cox D. Lin S. Barretto R. Habib N. Hsu P.D. Wu X. Jiang W. Marraffini L.A. et al.Multiplex genome engineering using CRISPR/Cas systems.Science. 2013; 339: 819-823Crossref PubMed Scopus (9981) Google Scholar, Feng et al., 2013Feng Z. Zhang B. Ding W. Liu X. Yang D.L. Wei P. Cao F. Zhu S. Zhang F. Mao Y. et al.Efficient genome editing in plants using a CRISPR/Cas system.Cell Res. 2013; 23: 1229-1232Crossref PubMed Scopus (684) Google Scholar, Miao et al., 2013Miao J. Guo D. Zhang J. Huang Q. Qin G. Zhang X. Wan J. Gu H. Qu L.J. Targeted mutagenesis in rice using CRISPR-Cas system.Cell Res. 2013; 23: 1233-1236Crossref PubMed Scopus (585) Google Scholar, Shan et al., 2013Shan Q. Wang Y. Li J. Zhang Y. Chen K. Liang Z. Zhang K. Liu J. Xi J.J. Qiu J.L. et al.Targeted genome modification of crop plants using a CRISPR-Cas system.Nat. Biotechnol. 2013; 31: 686-688Crossref PubMed Scopus (1249) Google Scholar, Ma et al., 2015Ma X. Zhang Q. Zhu Q. Liu W. Chen Y. Qiu R. Wang B. Yang Z. Li H. Lin Y. et al.A robust CRISPR/Cas9 system for convenient, high-efficiency multiplex genome editing in monocot and dicot plants.Mol. Plant. 2015; 8: 1274-1284Abstract Full Text Full Text PDF PubMed Scopus (1160) Google Scholar, Xie et al., 2015Xie K. Minkenberg B. Yang Y. Boosting CRISPR/Cas9 multiplex editing capability with the endogenous tRNA-processing system.Proc. Natl. Acad. Sci. USA. 2015; 112: 3570-3575Crossref PubMed Scopus (764) Google Scholar, Gao et al., 2016Gao X. Chen J. Dai X. Zhang D. Zhao Y. An effective strategy for reliably isolating heritable and Cas9-free Arabidopsis mutants generated by CRISPR/Cas9-mediated genome editing.Plant Physiol. 2016; 171: 1794-1800Crossref PubMed Scopus (153) Google Scholar). HDR-mediated DNA KI remains extremely challenging, partially due to the difficulty in delivering sufficient repair template, i.e., donor DNA. Here, by combining CRISPR/Cas9 to produce DSBs and geminiviral vectors to deliver abundant donor DNA into rice cells, we have achieved up to 19.4% targeted KI frequency in transgenic rice plants. Geminiviruses are a large family of plant viruses with single-stranded, circular DNA genomes (∼2.5–3.0 kb) and a wide host range, including monocotyledonous and dicotyledonous plants. Upon infection, they produce numerous replicons through rolling-circle replication, which may serve as repair templates during HDR (Hanley-Bowdoin et al., 2013Hanley-Bowdoin L. Bejarano E.R. Robertson D. Mansoor S. Geminiviruses: masters at redirecting and reprogramming plant processes.Nat. Rev. Microbiol. 2013; 11: 777-788Crossref PubMed Scopus (461) Google Scholar, Baltes et al., 2014Baltes N.J. Gil-Humanes J. Cermak T. Atkins P.A. Voytas D.F. DNA replicons for plant genome engineering.Plant Cell. 2014; 26: 151-163Crossref PubMed Scopus (329) Google Scholar). The bean yellow dwarf virus (BeYDV) has been used in tobacco and tomato for this purpose, increasing KI efficiency 10–100 times compared with conventional T-DNA delivery (Baltes et al., 2014Baltes N.J. Gil-Humanes J. Cermak T. Atkins P.A. Voytas D.F. DNA replicons for plant genome engineering.Plant Cell. 2014; 26: 151-163Crossref PubMed Scopus (329) Google Scholar, Cermak et al., 2015Cermak T. Baltes N.J. Cegan R. Zhang Y. Voytas D.F. High-frequency, precise modification of the tomato genome.Genome Biol. 2015; 16: 232Crossref PubMed Scopus (371) Google Scholar). The wheat dwarf virus (WDV) infects a variety of grasses, including most cereals, and has been previously used to express foreign proteins in wheat and maize cells (Suarez-Lopez and Gutierrez, 1997Suarez-Lopez P. Gutierrez C. DNA replication of wheat dwarf geminivirus vectors: effects of origin structure and size.Virology. 1997; 227: 389-399Crossref PubMed Scopus (19) Google Scholar). Recently, WDV was used to achieve high-efficiency gene targeting in hexaploid wheat calli, and it was also shown to express in rice protoplast (Gil-Humanes et al., 2017Gil-Humanes J. Wang Y. Liang Z. Shan Q. Ozuna C.V. Sánchez-León S. Baltes N.J. Starker C. Barro F. Gao C. et al.High-efficiency gene targeting in hexaploid wheat using DNA replicons and CRISPR/Cas9.Plant J. 2017; 89: 1251-1262Crossref PubMed Scopus (226) Google Scholar). However, no geminivirus-based genome editing system has been established for rice. Therefore, we chose WDV to develop an efficient KI system in this important crop and model plant. The coding region for the replication initiator protein (Rep/RepA) as well as the large intergenic region (LIR) and short intergenic region (SIR) of WDV were synthesized (Supplemental Figure 1) and cloned into the binary vector pCambia1300 together with GFP or GUS expression cassettes (35S-GFP-NOS or 35S-GUS-NOS) to produce the geminiviral reporter vectors WDV-GFP and WDV-GUS (Figure 1A and Supplemental Figure 2A). Rice calli (cv. Nipponbare) co-cultivated with Agrobacterium tumefaciens containing either WDV-GFP or WDV-GUS showed significantly stronger GFP or GUS signal compared with traditional T-DNA vectors (T-GFP and T-GUS) harboring the same 35S-GFP-NOS or 35S-GUS-NOS expression cassettes without the additional WDV sequences (Figure 1B and Supplemental Figure 2B). Once inside the host cells, sequences contained between the two LIRs were expected to circularize to form a replicon (Figure 1A and Supplemental Figure 2A). This circularization was confirmed by PCR using a pair of primers (Cycle-F and Cycle-R) facing opposite directions (Figure 1C). Time-course experiments showed strong signals up to 9 days after co-cultivation with an apparent peak at day 5–7 (Supplemental Figure 3). In order to determine the number of circular WDV-induced molecules per cell, protoplasts were isolated from calli 5 days after inoculation with Agrobacteria containing WDV-GFP or T-GFP, and fluorescent protoplasts were isolated by flow cytometry. GFP copy number per protoplast was calculated by qPCR using the endogenous locus Actin-1 (ACT1, LOC_Os03g50885) as internal reference. As shown in Figure 1D, copy numbers in rice cells containing WDV-based constructs were hundreds of times higher than those obtained using conventional binary vectors (589 ± 238 versus 0.78 ± 0.34, respectively), confirming their potential as efficient delivery tools for HDR purposes. To determine the feasibility of our KI approach, the GFP coding region was fused in-frame with the selection marker Neomycine phosphotransferase II (NPTII) using the self-cleaving peptide 2A (Kim et al., 2011Kim J.H. Lee S.R. Li L.H. Park H.J. Park J.H. Lee K.Y. Kim M.K. Shin B.A. Choi S.Y. High cleavage efficiency of a 2A peptide derived from porcine teschovirus-1 in human cell lines, zebrafish and mice.PLoS One. 2011; 6: e18556Crossref PubMed Scopus (836) Google Scholar) as a bridge (Supplemental Figure 4). The translation start codon "ATG" was removed from this fusion cassette in order to inhibit any undesired translation caused by random insertion of the T-DNA into the rice genome. Two loci, encoding ACT1 and glutathione S-transferase (GST, LOC_Os05g02530), respectively, were selected for targeted KI because of their high expression levels in rice calli. Successful incorporation of the GFP-2A-NPTII cassette into the intended targets should result in expression of the fusion fluorescent proteins (ACT1-GFP and GST-GFP), while self-cleaved NPTII should help to enrich the cells containing targeted KI during tissue culture. To facilitate homologous recombination, the GFP-2A-NPTII cassette was flanked by two homology arms (∼500 bp) for ACT1 or GST resulting in ∼2.6-kb donor DNA fragments, which were inserted into the WDV-GFP vector replacing the GFP expression cassette (designated as WDV2-ACT1 and WDV2-GST, respectively) (Figure 1E). Previous studies have shown that CRISPR/Cas9-induced DSBs at the target site greatly increase the KI frequency (Baltes et al., 2014Baltes N.J. Gil-Humanes J. Cermak T. Atkins P.A. Voytas D.F. DNA replicons for plant genome engineering.Plant Cell. 2014; 26: 151-163Crossref PubMed Scopus (329) Google Scholar, Cermak et al., 2015Cermak T. Baltes N.J. Cegan R. Zhang Y. Voytas D.F. High-frequency, precise modification of the tomato genome.Genome Biol. 2015; 16: 232Crossref PubMed Scopus (371) Google Scholar). We therefore designed CRISPR/Cas9 gRNA targets in the intended recombination sites for ACT1 and GST and verified their targeting efficiency in T0 transgenic rice (Supplemental Figure 5). The two U6-driven guide RNA (gRNA) expression cassettes (SG1 and SG2) were then incorporated into WDV2-ACT1 and WDV2-GST, respectively (Figure 1E). The cargo capacity of WDV in wheat is limited and replication efficiency decreases with increasing replicon sizes (Suarez-Lopez and Gutierrez, 1997Suarez-Lopez P. Gutierrez C. DNA replication of wheat dwarf geminivirus vectors: effects of origin structure and size.Virology. 1997; 227: 389-399Crossref PubMed Scopus (19) Google Scholar, Gil-Humanes et al., 2017Gil-Humanes J. Wang Y. Liang Z. Shan Q. Ozuna C.V. Sánchez-León S. Baltes N.J. Starker C. Barro F. Gao C. et al.High-efficiency gene targeting in hexaploid wheat using DNA replicons and CRISPR/Cas9.Plant J. 2017; 89: 1251-1262Crossref PubMed Scopus (226) Google Scholar). We studied the expression of the WDV system carrying DNA cassettes of different sizes in rice calli. Insertion of up to 3 kb DNA fragments between the LIR and SIR in WDV resulted in strong GFP expression, but GFP fluorescence dramatically decreased with larger fragments (Supplemental Figure 6). These results indicated that the WDV2-ACT1 (2.59 kb) and WDV2-GST (2.67 kb) cassettes should undergo efficient replication but cast doubts about the adequacy of the WDV system for Cas9 (4.2 kb) expression. Consequently, we used transgenic rice calli constitutively expressing Cas9 (Cas9+) to inoculate with Agrobacteria carrying WDV2-ACT1 or WDV2-GST constructs with or without gRNA expression cassettes. The conventional T-DNA vectors T-ACT1 and T-GST, which contain the same GFP-2A-NPTII and gRNA cassettes with homology arms but without the WDV elements, were used as controls (Supplemental Figure 7). The T0 transgenic plants generated using NPTII as selection marker were first analyzed by PCR using primers flanking the homology arms to identify successful KI events (Figure 1E and 1F), and then the junction sites of all KI events were further analyzed using Sanger sequencing (Figure 1G and Supplemental Figure 8). Twelve of 62 (19.4%) WDV2-ACT1-SG1 transgenic plants and six of 78 (7.7%) WDV2-GST-SG2 plants had incorporated the GFP-2A-NPTII cassette in the correct genomic locus; both of them were significantly higher than their T-DNA control (Figure 1H). Moreover, sequencing results showed that the insertion of the donor DNAs at their targeted sites was precise (Figure 1G). In agreement with previous reports (Baltes et al., 2014Baltes N.J. Gil-Humanes J. Cermak T. Atkins P.A. Voytas D.F. DNA replicons for plant genome engineering.Plant Cell. 2014; 26: 151-163Crossref PubMed Scopus (329) Google Scholar), the presence of gRNA, essential for the production of DSBs by Cas9, was critical to achieve targeted KI (Figure 1H). Sequencing results also showed that all KI plants were heterozygous with indels or large fragment deletions at their allelic sites, derived from NHEJ (Supplemental Figure 9). Consistent with the genotyping results, GFP fluorescence was observed in the ACT1 and GST KI calli (Supplemental Figure 10) as well as ACT1 KI seedlings (Figure 1I). Although we failed to observe GFP fluorescence in the GST KI seedlings, possibly due to the low expression level of GST in seedlings, western blot results indicated the presence of the expected GST-GFP fusion protein in the GST KI lines (Supplemental Figure 11). To overcome the limitation of using Cas9+ rice callus, Cas9 could be incorporated into the WDV binary constructs outside the WDV replicon (Baltes et al., 2014Baltes N.J. Gil-Humanes J. Cermak T. Atkins P.A. Voytas D.F. DNA replicons for plant genome engineering.Plant Cell. 2014; 26: 151-163Crossref PubMed Scopus (329) Google Scholar). For this purpose, we modified the WDV2 vectors to contain a Ubiquitin promoter::Cas9::NOS terminator cassette outside the WDV replicon boundaries (named WDV3) (Supplemental Figure 12) and used them to transform wild-type rice calli. A total of 123 transgenic lines were regenerated and genotyped by PCR and Sanger sequencing. The targeted KI frequency for ACT1 and GST was 8.5% and 4.7%, respectively (Figure 1H). Although the KI frequency was lower than using Cas9+ calli, these results demonstrate that it is possible to combine the WDV-derived donor DNA delivery system and CRISPR/Cas9 into a single vector to facilitate targeted KI applications in rice research and breeding. We found that the T-DNA elements of WDV3, i.e., Cas9 and Rep, were integrated into the genome of KI plants, and the circular form of replicons was also detected (Supplemental Figure 13), indicating that the WDV elements (LIR, SIR, and Rep) still function in the KI plants. Although we did not find obvious difference in phenotypes between T-DNA KI lines and WDV KI lines at the seedling stage, the effect of the Rep protein on the growth of KI plants needs further evaluation. A practical and efficient KI method is critical for the development of precision molecular breeding in rice. The WDV-derived targeted KI approach presented here provides a simple and efficient tool for the delivery of abundant donor DNA into rice cells as well as the generation of DSBs, resulting in efficiencies of up to 19.4%. Future work is needed to improve the delivery efficiency of WDV, and to adapt our method for knock-in of sequences into the 5′ or the middle of genes, since our method is currently limited by the necessity of using a selection marker for enrichment of cells with KI events in tissue culture. We anticipate that rapid advances in genome editing technologies as well as ongoing developments in DNA delivery methods will make the requirement of using a selection marker superfluous in the near future. This study was supported by the Chinese Academy of Sciences.