CRISPR/Cas9-mediated homologous recombination in tobacco
- 1.1k Downloads
Co-transformation of multiple T-DNA in a binary vector enabled CRISPR/Cas9-mediated HR in tobacco. HR occurred in a limited region around the gRNA target site.
In this study, CRISPR/Cas9-mediated homologous recombination (HR) in tobacco (Nicotiana tabacum L. ‘SR-1’) was achieved using binary vectors comprising two (T1–T2) or three (T1–T2–T3) independent T-DNA regions. For HR donor with the tobacco acetolactate synthase gene, SuRB, T-DNA1 contained ΔSuRBW568L, which lacked the N-terminus region of SuRB and was created by three nucleotide substitutions (ATG to GCT; W568L), leading to herbicide chlorsulfuron (Cs) resistance, flanked by the hygromycin (Hm)-resistant gene. T-DNA2 consisted of the hSpCas9 gene and two gRNA inserts targeting SuRB and An2. For the 2nd HR donor with the tobacco An2 gene encoding a MYB transcription factor involved in anthocyanin biosynthesis, T-DNA3 had a 35S promoter-driven An2 gene lacking the 3rd exon resulting in anthocyanin accumulation after successful HR. After selecting for Hm and Cs resistance from among the 7462 Agrobacterium-inoculated explants, 77 independent lines were obtained. Among them, the ATG to GCT substitution of endogenous SuRB was detected in eight T1–T2-derived lines and two T1–T2–T3-derived lines. Of these mutations, four T1–T2-derived lines were bi-allelic. All the HR events occurred across the endogenous SuRB and 5′ homology arm of the randomly integrated T-DNA1. HR of the SuRB paralog, SuRA, was also found in one of the T1–T2-derived lines. Sequence analysis of its SuRA-targeted region indicated that the HR occurred in a limited (< 153 bp) region around the gRNA target site. Even though some T1–T2–T3-derived lines introduced three different T-DNAs and modified the An2 gRNA target site, no signs of HR in the endogenous An2 could be observed.
KeywordsHomologous recombination CRISPR/Cas9 T-DNA Tobacco
Clustered regularly interspaced short palindromic repeats
Double-strand breaks. HMA, heteroduplex mobility assay
Genome editing is a genetic modification technique that involves introduction of targeted DNA double-strand breaks (DSBs) using artificial endonucleases. Nuclease-induced DSBs are mainly repaired by two different pathways: nonhomologous end-joining (NHEJ) and homology-directed repair (HDR) (Steinert et al. 2016). NHEJ generates small insertion/deletion mutations leading to disruption of the translational reading frame of a coding sequence or the binding sites of trans-acting factors in promoters. HDR-mediated repair can be used to introduce desired sequences by homologous recombination (HR) across the target locus with donor DNA (Sander and Joung 2014). In recent years, a new genome editing technique using CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9) derived from the bacteria Streptococcus pyogenes has been developed (Jinek et al. 2012; Sander and Joung 2014). Studies using CRISPR/Cas9 have been reported not only in animals and microorganisms, but also in plants (Bortesi and Fischer 2015). CRISPR/Cas9-mediated genome editing via NHEJ has been performed in several plant species, such as Arabidopsis (Feng et al. 2013, 2014; Jiang et al. 2014), tobacco (Nekrasov et al. 2013; Gao et al. 2015), and rice (Shan et al. 2013; Miao et al. 2013), many of which produced knockout individuals with a loss of function. Multiple gene modifications through NHEJ using multiple guide RNAs (gRNAs) have been reported in Arabidopsis (Li et al. 2013; Xing et al. 2014), rice (Zhang et al. 2014; Endo et al. 2015; Xie et al. 2015), and wheat (Wang et al. 2014).
CRISPR/Cas9 is a promising technology for gene targeting (GT) in plants because HR is not the fundamental repair mechanism in plants and introducing DSBs at the target locus can greatly enhance the GT frequency (Steinert et al. 2016). Till date, very few studies have reported CRISPR/Cas9-mediated GT or HR (Ma et al. 2017). Li et al. (2013) first reported CRISPR/Cas9-mediated replacement of PDS gene in tobacco (Nicotiana benthamiana) protoplasts. CRISPR/Cas9-mediated gene replacements were also performed by particle bombardment in rice (Svitashev et al. 2015) and soybean (Li et al. 2015). This strategy and the geminiviral system (Čermák et al. 2015) are expected to provide a higher number of copies of the donor template for HR (Zhao et al. 2016), compared with that obtained in Agrobacterium-mediated transformation. These methodologies have advantages and disadvantages (e.g., copy number, usability, necessity of equipment), and Agrobacterium-mediated transformation has also been applied for CRISPR/Cas9-mediated GT studies. Targeted integration of kanamycin resistance cassette flanked by homologous sequence was accomplished in Arabidopsis ADH1 locus (Schiml et al. 2014). In this study, Cas9 gene, AtU6 promoter-driven ADH1 gRNA, and GT donor cassette with the gRNA target site at both ends were placed in the same T-DNA region, and two independent transgenic Arabidopsis lines containing the GT event were identified from among 1400 seedlings. In addition, replacement of the Arabidopsis TFL gene by co-transformation with dual-gRNA/Cas9 vector and TFL donor vector was also achieved with a frequency of 0.8% (Zhao et al. 2016). In rice, gene targeting efficiency of ALS gene was increased by suppression of DNA ligase 4, while Cas9 expression did not significantly affect GT efficiency (Endo et al. 2016).
In this study, CRISPR/Cas9-mediated HR in tobacco (Nicotiana tabacum) was performed. Even though CRISPR/Cas9-mediated genome editing via NHEJ has been reported in tobacco (Gao et al. 2015; Mercx et al. 2016, 2017), no reports on CRISPR/Cas9-mediated HR exist. To introduce both the Cas9–gRNA expression cassette and the HR donor cassette in the host cells separately, Agrobacterium-harboring binary vectors having multiple T-DNA regions were used for transformation. Two genes were chosen as the targets of HR. One of the targets, the tobacco acetolactate synthase (ALS/AHAS) gene, SuRB has been well-characterized (Lee et al. 1988). Given that mutations in the ALS gene confer resistance to herbicides, a number of studies have used this gene for GT experiments (Beetham et al. 1999; Endo et al. 2006, 2007, 2016; Kochevenko and Willmitzer 2003; Lee et al. 1990; Okuzaki and Toriyama 2004; Saika et al. 2015; Sun et al. 2016; Svitashev et al. 2015; Townsend et al. 2009; Zhang et al. 2013; Zhu et al. 1999). The other HR target gene used is the tobacco anthocyanin 2 (An2) encoding a R2R3 MYB transcription factor involved in tobacco anthocyanin synthesis (Pattanaik et al. 2010). Given that ectopic overexpression of An2 gene enforces anthocyanin accumulation (Pattanaik et al. 2010), a 35S promoter-driven An2 gene without the 3rd exon was used as an HR donor. This type of knock-in system was previously demonstrated in tomato using a geminivirus-mediated CRISPR/Cas9 system (Čermák et al. 2015). Using the multiple T-DNA vector harboring these HR donors, this study examined the occurrence of CRISPR/Cas9-mediated HR in tobacco.
Materials and methods
Binary vectors, pGII–T1–T2 and pGII–T1–T2–T3, were derived from pGreenII (Hellens et al. 2000; Fig. 1). For T-DNA2 construction, the plasmid pX330-U6-Chimeric_BB-CBh-hSpCas9 (Addgene #42230; Cong et al. 2013) and the binary vector pER8 (Zuo et al. 2001) were used to create the G10–90 promoter-driven hSpCas9 with a pea rbcS E9 terminator (Coruzzi et al. 1984). Arabidopsis U6–26 promoter-driven gRNA was constructed according to Feng et al. (2013). Oligonucleotides used for gRNA constructs are listed in Table S1. The 35S promoter-driven HPT (hygromycin phosphotransferase) with a NOS terminator cassette was derived from the binary vector pIG121Hm (Hiei et al. 1994). Modification of SuRB gene was performed using the PrimeSTAR mutagenesis basal kit (Takara Bio, Otsu, Japan). These binary vectors and a helper plasmid pSoup (Hellens et al. 2000) were introduced into Agrobacterium (Rhizobium radiobacter) strain EHA105 (Hood et al. 1986) by the freeze–thaw method (Höfgen and Willmitzer 1988). A binary vector pSMAB-T1 derived from pSMAB704 (Igasaki et al. 2002) containing the same T-DNA1 construct as those of pGII–T1–T2 and pGII–T1–T2–T3 was introduced into the strain EHA101 (Hood et al. 1986) also, by the freeze–thaw method.
Agrobacterium transformation of tobacco
Genetic transformation of tobacco (Nicotiana tabacum L. ‘SR-1’) was performed by following the Agrobacterium-mediated transformation based on the leaf disk procedure (Horsch et al. 1985) and selected against 50 mg L− 1 hygromycin and 100 nM chlorsulfuron. The selection cultures were maintained by subculturing the explants on 0.8% (w/v) agar-solidified Murashige and Skoog (MS; Murashige and Skoog 1962) medium containing 3% (w/v) sucrose, 1 mg L− 1 6-benzylaminopurine (BA), and 0.1 mg L− 1 of 1-naphthaleneacetic acid (NAA) every 2–3 weeks. In vitro culturing was carried out at 25 °C, with a 16-h light and 8-h dark period. For further selection, the cultures were exposed to 100 nM bispyribac-sodium or 200 nM chlorsulfuron as well.
PCR and Southern blot analyses
For PCR analysis, genomic DNA was extracted from in vitro-grown calli or shoots according to Edwards et al. (1991). Primers used for the analysis are listed in Table S2. For the Southern blot analysis, genomic DNA was isolated from samples of in vitro-grown calli or shoots using a Nucleon PhytoPure DNA extraction kit (GE Healthcare, Piscataway, NJ) and following the supplier’s instructions. PvuII-digested genomic DNA (5-µg aliquots) were electrophoresed on a 0.8% (w/v) agarose gel, blotted onto nylon membrane, and fixed by ultraviolet (UV) irradiation. The blots were hybridized with digoxigenin (DIG)-labeled (Roche Diagnostics, Mannheim, Germany) SuRB probe following the supplier’s instructions. For the SuRB probe, a 449-bp PCR fragment was amplified with the following primers: 5′-AACCAAATGTGGGGTTGAAA-3′ and 5′-CAAAATTGCTTGGCCTCATT-3′.
Heteroduplex mobility assay (HMA)
HMA (Delwart et al. 1993) was performed for the SuRB, SuRA, and An2 genes. Primers listed in Table S2 were designed to amplify the regions containing CRISPR/Cas9 target sites. PCR products were electrophoresed on 10% non-denaturing polyacrylamide gels that were later stained with ethidium bromide. Gel images were captured with a FLA7000 (GE Healthcare) laser scanner.
RNA was extracted according to a previously described method of Chomczynski and Sacchi (2006). Reverse transcription was carried out at 37 °C using a MultiScribe reverse transcriptase (Thermo Fisher Scientific, Waltham, MA) and an oligo-dT primer (Takara Bio). PCR was performed at an annealing temperature of 60 °C and using the following primers: 5′-CGCCCACCCTTGGAAACTCCG-3′ with 5′-AGCCTTATAGAACCGATCCTCCAGC-3′ for T-DNA1 (ΔSuRBW568L), and 5′-CTATTCTCCGCTTTGGACTTGGCA-3′ with 5′-AGGACCTCAGGACAACGGAAACG-3′ for NtACT9 (Cortleven et al. 2009; GenBank X69885).
Vectors for CRISPR/Cas9-mediated HR
To evaluate the effect of CRISPR/Cas9-mediated HR in tobacco, binary vectors having multiple T-DNA regions were constructed. The vector pGII–T1–T2 has two T-DNA regions (T-DNA1 and T-DNA2). The T-DNA1 consists of tobacco SuRB genomic locus interlaid with a 35S promoter-driven hygromycin resistance gene (35S-HPT) for the HR donor (Fig. 1). This SuRB (designated as ΔSuRBW568L thereafter) is truncated at its 5′-coding region (nucleotides 1-602) and modified by three nucleotide substitutions (ATG to GCT) conferring the W568L substitution and consequently generating a PvuII recognition site (CAGCTG). The W568L mutation is known to confer herbicide resistance (Townsend et al. 2009). The lengths of its homology arms are 1422 and 1415 bp. The T-DNA2 consists of G10–90 promoter (Ishige et al. 1999)-driven hSpCas9 gene (Cong et al. 2013) and two sets of Arabidopsis U6 promoter-driven gRNA (U6-gRNA1 and U6-gRNA2). The gRNA1 and gRNA2 are specific for SuRB and tobacco An2 genes, respectively. Given that the ATG to GCT substitution overlaps with the SuRB gRNA target site, Cas9 does not cleave the ΔSuRBW568L. In addition to the T-DNA1 and T-DNA2, T-DNA3 was inserted in the pGII–T1–T2–T3 vector (Fig. 1). The T-DNA3 consists of the genomic locus of An2 interlaid with a 35S promoter for the HR donor. This An2 (designated as ΔAn2 thereafter) donor lacks the 3rd exon, which comprises 133 amino acids out of the 220 amino acids of the An2 protein, to which a 35S promoter was inserted upstream of the start codon of the An2 gene. Along with the 35S insertion, a 358 bp of 5′ upstream region from the start codon containing the TATA-box and the An2 gRNA target site was deleted, rendering Cas9 unable to cleave ΔAn2. The lengths of its homology arms are 1201 and 1286 bp. HR of both homology arms to the tobacco genomic regions is required to achieve a high expression of an intact An2 protein, which elevates anthocyanin accumulation (Pattanaik et al. 2010). To confirm the effects of CRISPR/Cas9-mediated DSBs on HR, a vector pSMAB-T1 containing only the T-DNA1 was also used as a control.
Agrobacterium-mediated transformation of tobacco for selection of Hm- and Cs-resistant calli
Agrobacterium strains harboring the above-mentioned binary vectors were used in the leaf disk transformation protocol in tobacco. When only hygromycin was used for selection, a large number of shoots and calli were obtained irrespective of the vector used (Fig. 2a). Contrarily, the frequencies of resistant calli varied depending on the vector when cultured on a selection medium containing both hygromycin and chlorsulfuron. Table 1 shows the result of the selection culture. In the pSMAB-T1 vector, only one hygromycin and chlorsulfuron-resistant (designated as HmR/CsR thereafter) callus was obtained from 2784 explants with a frequency of 0.036%. Higher frequencies (0.51 and 1.40%) of HmR/CsR calli were obtained using the pGII–T1–T2 and pGII–T1–T2–T3 vectors, respectively. Consequently, 16 (#1 to #16) and 61 (#1 to #61) independent HmR/CsR calli lines were obtained using the pGII–T1–T2 (designated as T1–T2 thereafter) and pGII–T1–T2–T3 (designated as T1–T2–T3 thereafter) vectors, respectively, and used for subsequent experiments.
Summary of the transgene integration in the Hm and Cs-resistant calli lines obtained
BsR callid (%)
Genotyping of the HmR/CsR lines
To determine the integration of T-DNA1, T-DNA2, and/or T-DNA3 in the genome of the HmR/CsR lines obtained, genomic PCR was performed. For each T-DNA region, two primer sets listed in Table S2 were used. In the case of HPT gene within T-DNA1, all the 77 HmR/CsR lines were positive for PCR amplification (Fig. 2b; Tables 1, S3, S4). It is consistent with the fact that these lines were resistant to hygromycin. Another T-DNA1 region, ΔSuRBW568L was also detected in 70 out of the 77 lines (Tables S3, S4). In the case of T-DNA2, hSpCas9 and gRNA regions were analyzed. Frequencies of the T-DNA2 integration in the genome of HmR/CsR lines derived from T1–T2 and T1–T2–T3 were very different. In the T1–T2 lines, 14 out of 16 (87.5%) lines showed integration of both hSpCas9 and gRNA regions (Fig. 2c; Tables 1, S3). However, only six lines were detected by amplifying either region in the T1–T2–T3-derived HmR/CsR lines (Fig. 2c; Tables 1, S4). Of the six lines, four lines showed integration of both hSpCas9 and gRNA regions, whereas the remaining two lines were detected to have one of two regions (Table S4). In the case of T-DNA3, two primer sets were used to detect integration of the ΔAn2 region. One primer set is able to detect both, the endogenous An2 and the ΔAn2 genes by amplifying the 981- and 1491-bp-sized fragments, respectively (Table S2). The results of the PCR amplification are shown in Fig. 2d. As expected, all of the HmR/CsR lines derived from T1–T2 had only the endogenous An2-specific fragment (Fig. 2d; indicated by asterisk), whereas 37 out of 61 (60.7%) T1–T2–T3 lines had the ΔAn2-specific fragment, in addition to the An2-specific fragment (Fig. 2d; Tables 1, S4). The other primer set specifically detects the 35S-ΔAn2 region as a 380-bp fragment (Table S2), which provided similar results (Table S4).
Analysis of HR of endogenous SuRB in the HmR/CsR lines
To confirm whether HR of endogenous SuRB occurred in the HmR/CsR lines, PCR–RFLP was performed. The PCR amplification was performed using the endogenous SuRB-specific primers that do not overlap with the ΔSuRBW568L region in T-DNA1 (Fig. 3a). All the amplified products matched the expected size (3462 bp) of the endogenous SuRB with no trace of the 5688-bp fragment from the expected HR of the two homology arms of T-DNA1 and SuRB locus (Table S2, Fig. 3a). If HR occurred with 5′ homology arm of the ΔSuRBW568L region, a new PvuII restriction site was generated from the three nucleotide substitutions (CAATGG to CAGCTG; Fig. 1a). To confirm this, the PCR-amplified fragments were restriction digested by PvuII restriction enzyme, which is expected to generate approximately 0.5, 1, and 2 kb bands after the modification (Fig. 3a). PCR–RFLP fragments of the unmodified SuRB locus are 1 and 2.5 kb in size. In the T1–T2-derived HmR/CsR lines, 4 (#2, #6, #8, #9) out of the 16 (25%) lines showed 0.5, 1, and 2 kb bands (Fig. 3b, Table S3), indicating that these lines are a result of bi-allelic modification. The other four (#3, #5, #10, #16) lines showed 0.5, 1, 2 and 2.5 kb bands, indicating mono-allelic modification of these lines. On the contrary, only 2 out of 61 (3.3%) T1–T2–T3-derived HmR/CsR lines and none (0 out of 1) of the pSMAB-T1-derived HmR/CsR lines had the three nucleotide substitutions (Table S4).
The nucleotide substitutions were further confirmed by the Southern blot analysis in the T1–T2-derived HmR/CsR lines. Genomic DNA of the seven PCR–RFLP-positive samples and one negative sample (#11) as a control were restriction digested with PvuII and probed with a 449-bp fragment of the genomic SuRB locus containing no overlapping homologous region of T-DNA1 (Fig. 3a). Consistent with the results of the PCR–RFLP, these seven samples had a signal at approximately 3.6-kb band, indicating that a new PvuII site was generated (Fig. 3a, c). The non-transgenic and line #11 sample had a signal only of a 4.1-kb band (see Fig. 3a) and the mono-allelic lines (#3, #5, #10) had both, the 3.6- and 4.1-kb bands (Fig. 3c).
Sequence analysis of the endogenous SuRB gRNA target site was conducted by the above-mentioned endogenous SuRB-specific PCR. Consistent with the results of PCR–RFLP and Southern blot analysis, only the three nucleotide substitutions (i.e. generating a new PvuII site) were detected from the expected bi-allelic T1–T2 lines (#2, #6, #8, #9; Fig. 3d). In the mono-allelic T1–T2 lines (#3, #5, #10, #16), one-nucleotide insertion (#3) and deletions (#5, #10, #16) were detected in addition to the three nucleotide substitutions (#10, #16). Additionally, modification of the endogenous SuRB gRNA target site was also analyzed by HMA (positions of the primers are shown in Fig. S1a). A subset of the T-DNA2-introduced lines (#4, #5, #6, #10, #11, #12, #16 in T1–T2 and #1, #7, #43, #46 in T1–T2–T3) showed distinct retarded bands (Fig. S1b, Tables S3, S4).
Analysis of HR of endogenous SuRA in the HmR/CsR lines
In addition to SuRB, the tobacco genome contains SuRA gene (Lee et al. 1990), which shares the SuRB gRNA target site with 100% sequence similarity. Therefore, it is possible to induce CRISPR/Cas9-mediated HR across one or two homology arms of T-DNA1 and the endogenous SuRA locus (Fig. 4a). In fact, HMA of the endogenous SuRA including the gRNA target site showed retarded bands in 13 out of 16 (81%) T1–T2-derived HmR/CsR lines (Fig. S1c, Table S3). PCR fragments of the endogenous SuRA in the T1–T2 line #1 that were restriction digested with PvuII showed approximately 0.17, 0.5, 0.67, and 1.4 kb bands, suggesting that a new PvuII site was mono-allelically generated by the three nucleotide substitutions (Fig. 4a, b). Two T1–T2 lines (#7 and #12) contained unexpected bands instead of (#7) or in addition to (#12) the 0.67-kb band. In contrast, no T1–T2–T3-derived lines were detected with the three nucleotide substitutions of SuRA (Table S4). To confirm the modification of SuRA, sequence analysis around the gRNA target site was performed in the above-mentioned three T1–T2 lines (#1, #7, #12). As expected, the three nucleotide substitutions (generating the PvuII site) and a two-nucleotide deletion were detected in the #1 line (Fig. 4c). In the #7 and #12 lines, insertions of a DNA fragment (123 and 295 bp, respectively) were detected with the bi-allelic and mono-allelic lines, respectively. These fragments were derived from the tobacco genome and the insertions introduced termination codons (Fig. 4c; indicated by box). In the #12 line, a two-nucleotide deletion event was also detected.
To investigate the extent of replacement of the endogenous SuRA by HR with the ΔSuRBW568L region in T-DNA1 (Fig. S2a), the #1 sequence was compared with the SuRA and SuRB sequences. As shown in Fig. S2b, there are four and three SNPs upstream and downstream of the gRNA target site, respectively, between SuRA and SuRB within the 321-bp region analyzed (Fig. S2b; indicated by boxes). Consequently, all of the SNPs of the #1 sequence were associated with SuRA.
Different levels of resistance to Bs and Cs and mRNA expressions of ectopic ΔSuRB W568L in the HmR/CsR lines
Sensitivity to another ALS inhibitor, bispyribac-sodium (Bs) in the HmR/CsR lines was tested. Consequently, 14 out of 16 (87.5%) T1–T2-derived lines exhibited resistance to Bs (Fig. S3, Tables 1, S3). In the case of T1–T2–T3-derived lines, 3 out of 61 (4.9%) showed resistance to Bs (Fig. S3, Tables 1, S4), and 6 other lines exhibited moderate resistance to Bs (Table S4; indicated as “− +”). Selection of the HmR/CsR lines at a higher concentration (200 nM) of Cs also showed similar trends to the selection for Bs (Tables S3, S4).
To explore the possible reasons for the Cs resistance apart from the W568L and W571L mutations of the endogenous SuRB and SuRA genes, respectively, ectopic mRNA expression of the ΔSuRBW568L region in T-DNA1 was analyzed by RT-PCR. The results reveal that the ΔSuRBW568L transcript was detected in a subset of the HmR/CsR lines (Fig. 4d). The ectopic ΔSuRBW568L expression was more frequently seen in the T1–T2–T3-derived lines than in the T1–T2-derived lines.
Analysis of HR of endogenous An2 in the HmR/CsR lines
Modification of the An2 gRNA target site (Fig. 5a) in the T1–T2-derived HmR/CsR lines (#1 to #11) was analyzed by HMA. Nine out of 11 lines showed retarded bands of the amplified fragments of the endogenous An2 region around the target site (Fig. 5b). Sequence analysis of the An2 target region in the six lines (#1 to #6) showed various mutation including insertions and deletions (Fig. 5c).
Among the T1–T2–T3-derived HmR/CsR lines, both T-DNA2 and T-DNA3 were detected in lines #1, #7, #26, and #43 (Table S4). HMA of the An2 gRNA target region in the four lines indicated that lines #1 and #7 had retarded bands (Fig. S4). However, despite the efforts to find evidence of HR across the endogenous An2 and ΔAn2 of T-DNA3, no trace of HR could be obtained by the HR-specific PCR in these four lines.
Effect of CRISPR/Cas9 on HR of the SuRB and SuRA gRNA target sites
This study showed that HR of tobacco SuRB or SuRA genes was enhanced by CRISPR/Cas9-mediated DNA cleavage. As it is inconceivable that CRISPR/Cas9-mediated NHEJ or spontaneous mutation caused the ATG to GCT substitution at the target sites in SuRB or SuRA, HR was considered to occur if the three nucleotide substitutions were detected in the HmR/CsR lines. Even though one HmR/CsR line was obtained after transformation with the pSMAB-T1 vector, the ATG to GCT substitution did not occur. This result suggests that there were no HR events in the 2784 explants developed after transformation with pSMAB-T1. A previous study using Agrobacterium-mediated GT in Arabidopsis estimated that one GT event of ALS locus occurred for every 3120 T-DNA integration events (Endo et al. 2006). Compared with the findings from studies on Arabidopsis, the frequency of HR in the present study seems to be low, which may be due to the use of hygromycin and chlorsulfuron for selection of HR of both arms. It is possible that the different frequencies of HR reflect the different types of cells (i.e. egg cells vs. somatic cells) used for transformation. In the T1–T2 lines, from 3115 explants that were transformed, nine lines (0.29%) represented the ATG to GCT substitution, and four lines underwent a bi-allelic mutation. Considering the fact that the generation of HmR calli (Fig. 2a, upper panel) was not any different among the three vectors used in this study, we infer that differences in the vector backbone (e.g., pSMAB vs. pGreen) do not affect the transformation efficiency. These results also indicate that the latter vector additionally containing the CRISPR/Cas9 expression cassette is able to stimulate HR. This is inconsistent with a previous finding that CRISPR/Cas9 does not affect GT in rice (Endo et al. 2016). It may be due to the differences in the T-DNA delivery systems between the two studies; in the rice study, Cas9 and donor/gRNA cassettes were independently transformed. In this study with the T1–T2–T3 lines, however, only two out of 61 (3.3%) HmR/CsR lines had the ATG to GCT substitution. These results were observed because the frequency of T-DNA2 integration (9.8%) in the T1–T2–T3 lines was extremely lower than that recorded in the T1–T2 lines (87.5%, Tables 1, S3, S4). Actually, all of the ATG to GCT substituted T1–T2 lines were introduced via T-DNA2 (Table S3). Although integration of two T-DNA regions within a binary vector is possible (Komari et al. 1996), integration of three T-DNA regions within a vector is unknown. The T1–T2–T3 vector harboring the three T-DNA regions may suffer impaired T-DNA2 delivery. Exceptionally, the ATG to GCT substituted T1–T2–T3 line #46 did not show T-DNA2 integration. These results suggest that a stable Cas9/gRNA expression is generally required for HR and that transient Cas9/gRNA expression may rarely lead to HR.
Features of HR of SuRB and SuRA loci
Although the HPT gene was detected in all the HmR/CsR lines analyzed, no lines exhibited HR across the two homology arms of T-DNA1 and endogenous SuRB locus. Thereby, once the T-DNA1 had been randomly integrated into the tobacco genome, HR likely occurred between the integrated T-DNA1 and endogenous SuRB or SuRA loci. Results of the Southern blot analysis (Fig. 3c) also support this notion. An additional band (approximately 6 kb) was detected in T1–T2 line #3 in the Southern blot analysis, which may be due to ectopic gene targeting (Hanin et al. 2001; Endo et al. 2006), or unexpected HR between the endogenous SuRB or SuRA loci.
T1–T2 line #1 had ATG to GCT substitution in the endogenous SuRA. This line provides information about the HR region across ΔSuRBW568L and endogenous SuRA because there are some SNPs between the donor and recipient sequences. As a result, the HR occurred within a 153-bp region around the gRNA target site. In a previous GT study of ALS locus in rice, a subset of the selected lines of a callus was mutated at one (W548L) of the two (W548L, S627I) sites of GT donor (Endo et al. 2016), suggesting that the HR occurred within a limited range (approx. < 400 bp from the gRNA target sites). GT using donors of a short length of chimeric RNA/DNA oligonucleotides is known (Beetham et al. 1999; Kochevenko and Willmitzer 2003; Okuzaki and Toriyama 2004; Zhu et al. 1999). We did not find evidence of HR of both arms; therefore, HR within a limited region may be more likely to occur from CRIPSR/Cas9-mediated HR.
Alternative mechanisms of resistance to Bs and Cs in the HmR/CsR lines
We observed differences in resistance to Bs and Cs in the HmR/CsR lines. All the lines with ATG to GCT substitution exhibited resistance to 100 nM Bs and 200 nM Cs, suggesting that mutations of the W568L of SuRB and the W571L of SuRA confer these levels of resistance. Interestingly, some of the T1–T2 lines, however, did not contain W568L or W571L mutations, yet they exhibited resistance to 100 nM Bs and 200 nM Cs. Among these, lines #7 and #12 had an insertion of the tobacco genomic DNA fragment at the SuRA target site. These lines produced truncated SuRA proteins (WT:667AA, #7:583AA, #12:580AA) that contain the insertion-derived short C-terminal sequences (#7:13AA, #12:11AA) and W571 mutations (#7:W571F, #12:W571Y). These mutant SuRA proteins may contribute to resistance against the Bs and Cs. Even though CRISPR/Cas9-induced mutations were predominantly 1-bp insertion and short deletions in seven Arabidopsis genes, larger (21–100 bp) insertions are also found, but with a rare frequency (Feng et al. 2014). Accordingly, we concluded that the insertion events in #7 and #12 may be rare events and were likely salvaged by the Hm and Cs selection.
Conversely, many of the T1–T2–T3 lines were sensitive or moderately resistant to 100 nM Bs or 200 nM Cs. One possible explanation for the weaker resistance to Bs and Cs could be the ectopic ΔSuRBW568L expression frequently found in the T1–T2–T3 lines, which may confer moderate resistance by acquiring the start codon of the ΔSuRBW568L ORF.
Modification of endogenous An2 in the HmR/CsR lines
As shown in Fig. 5, modification of the An2 target site was evident in most of the T1–T2 lines analyzed. Even though no retarded band of HMA was found in line #5 (Fig. 5b), insertion and deletion were observed when the sequence was analyzed. This result suggests that most of the lines harboring T-DNA2 underwent CRISPR/Cas9-mediated modification of the An2 target site. Interestingly, even though T-DNA2 was not detected in the line #7, a retarded band of HMA was detected in this line. This result again implies that transient Cas9/gRNA expression can modify the An2 target site, and this assumption is supported by the findings of a previous study on biolistic-mediated transient expression of CRISPR/Cas9 in wheat (Zhang et al. 2016).
We did not detect expected HR across endogenous An2 and ΔAn2 of the T-DNA3 in the T1–T2–T3-derived HmR/CsR lines #1, #7, #26, and #43, which contain both the T-DNA2 and T-DNA3. As HR of both the homology arms is necessary to achieve a high An2 expression, we believe the frequency of the successful HR may be too low to detect even in the lines containing active Cas9/gRNA at the An2 target site. This inference is consistent with the present result of HR on SuRB or SuRA, all of which were HR of the 5′ arm.
Author contribution statement
KiM conceived and designed the research. AH, IS, KK, and KiM performed the experiments. KiM and AH analyzed the data. All authors contributed reagents, materials, and analysis tools. KiM, YI, and NK wrote the manuscript. All authors read and approved the manuscript.
The authors would like to thank Dr. Roger Hellens (Queensland University of Technology), Dr. Philip Mullineaux (University of Essex) and Dr. Mark Smedley (John Innes Centre) for the pGreen vector system, Dr. Feng Zhang (Massachusetts Institute of Technology) for the pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmid, Dr. Nam-Hai Chua (Rockefeller University) for the pER8 vector, and Dr. Hiroaki Ichikawa (NIAS, Japan) for the pSMAB704 vector. We also thank Editage (http://www.editage.jp) for English language editing.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- Delwart EL, Shpaer EG, Louwagie J, McCutchan FE, Grez M, Rübsamen-Waigmann H, Mullins JI (1993) Genetic relationships determined by a DNA heteroduplex mobility assay: analysis of HIV-1 env genes. Science 19:1261–1275Google Scholar
- Feng Z, Mao Y, Xu N, Zhang B, Wei P, Yang DL, Wang Z, Zhang Z, Zheng R, Yang L, Zeng L, Liu X, Zhu JK (2014) Multi-generation analysis reveals the inheritance, specificity, and patterns of CRISPR/Cas-induced gene modifications in Arabidopsis. Proc Natl Acad Sci USA 111:4632–4637CrossRefPubMedGoogle Scholar
- Igasaki T, Ishida Y, Mohri T, Ichikawa H, Shinohara K (2002) Transformation of Populus alba and direct selection of transformants with the herbicide bialaphos. Bull FFPRI 1:235–240Google Scholar
- Mercx S, Smargiasso N, Chaumont F, Pauw ED, Boutry M, Navarre C (2017) Inactivation of the β(1,2)-xylosyltransferase and the α(1,3)-fucosyltransferase genes in Nicotiana tabacum BY-2 cells by a multiplex CRISPR/Cas9 strategy results in glycoproteins without plant-specific glycans. Front Plant Sci 8:403CrossRefPubMedPubMedCentralGoogle Scholar