Efficient generation of null-segregant parthenocarpic tomato by CRISPR/Cas9 editing CURRENT STATUS: POSTED

Background Genome editing technique is a powerful tool in plant the of crop plants. Currently, obtaining transgene-free mutant plants is an important issue. transgene-free mutant plants such as null-segregant, can be segregated out from selfing or back-crossed populations. However, such procedures are time consuming, and, therefore, the development of efficient systems to eliminate transgenes in mutant plants is required. Results CRISPR/Cas9 vectors targeting Sl IAA9 were introduced into the tomato cultivars, Ailsa Craig, Moneymaker, Super Roma and Rio Grande, via Agrobacterium transformation. Bi-allelic mutations were detected at ratios of 9–46% in CRISPR/Cas9-transgenic T0 shoots. Sequence analysis revealed that the bi-allelic mutations generated stop codons by frame-shift of SlIAA9 amino acid sequences, indicating that sliaa9 knockout mutants with parthenocarpic phenotypes were generated in the T0 generation with high efficiency in these cultivars. Subsequently, many null-segregant (transgene free) lines, as confirmed by PCR and Southern blot, were isolated rapidly in the T1 generation by self-pollination. Conclusions The method used in this study is effective in achieving rapid isolation of null-segregants with high efficiency, while adding important traits to various cultivars at the same time. Our developed system would further increase the advantages of using CRISPR/Cas9 for the production of novel crop varieties in the future. hybridizing bands in several T1 in PCR to T-DNA not detected hybridization bands for the LHCB gene used as a control were detected in all lines. These results indicate many null-segregants the can be efficiently isolated in various cultivars in the T1

in conventional breeding. One such new technology-genome editing-is now widely used as a powerful tool in plant genetic engineering to accelerate the rapid breeding of crop plants. One of the most convenient genome editing tools is the clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein-9 nuclease (Cas9) system, in which the Cas9 nuclease and a guide RNA (gRNA) form a complex, and induce a double-stranded DNA break (DSB) at the target site.
Plant genome editing tools are introduced into plant cells mainly by Agrobacterium-mediated transformation. When transgenes, including genome editing tools, are integrated into mutant genomes, specific regulations, such as the Cartagena Protocol on Biosafety, must be adhered to, both in terms of public acceptance and to facilitate the commercialization of new crop varieties generated by genome editing [9]. Several methods for transgene-free plant genome editing, including RNP (ribonucleoprotein) technology have recently been developed [10,11,12,13]. Null-segregants, i.e., transgene-free mutant plants, can be segregated out from selfing or back-crossed populations.
However, such procedures are time consuming, and, therefore, the development of efficient systems to eliminate transgenes in mutant plants is required.
Tomato (Solanum lycopersicum)-a model horticultural crop-is one of the most important vegetable and fruit crops worldwide. In wild-type tomato plants, fruiting can be induced by artificial pollination and application of exogenous auxin. Parthenocarpy, which allows fruit production without fertilization, is a hugely important agricultural trait, and has been produced by traditional breeding in limited species and cultivars. The fruiting of tomato is controlled by auxin signaling factors, tomato Auxin Responsive Factor 8 (SlARF8) and the auxin responsive transcription repressor INDOLE-3-ACETIC ACID 9 (SlIAA9) [14,15]. It has been shown that downregulation of SlIAA9 induces parthenocarpy [15,16].
Mutants of sliaa9 have been generated by ethyl methanesulfonate (EMS) treatment of several cultivars, such as Micro-Tom, Ailsa Craig, and M82 [17,18]. We previously reported the efficient generation of sliaa9 knockout mutants using the efficient CRISPR/Cas9 system [19]. To extend this technology to generate parthenocarpic tomatoes in the various genetic backgrounds of commercial cultivars, we used CRISPR/Cas9 to induce SlIAA9 mutations in multiple cultivars (Ailsa Craig, Moneymaker, Super Roma, and Rio Grande) simultaneously. The newly generated mutants exhibited parthenocarpic phenotypes, and null-segregants from these sliaa9 knockout lines were isolated systematically and rapidly in the T1 generation. The efficient generation of null-segregants of the mutant lines with the desired phenotype shown in this study is an important example of how useful traits can be generated rapidly in commercial cultivars by using NPBTs.
The 100% mutations determined as non-digested fragments in PCR-RFLP analysis (AC#52, MM#3, and RG#4 in Fig. 1A), which are presumed to be mutations introduced into almost all somatic cells, were isolated at a frequency of 23-46% in Ailsa Craig, Moneymaker, and Super Roma, and 9% in Rio Grande (Fig. 1B). The mutated sequences were then analyzed by Sanger sequencing of cloned PCR fragments from the target sites, amplified from shoots with 100% mutation from each of the various commercial cultivars (Fig. 1C, Additional file 1: Figure S2). The results revealed that the mutant lines have only one or two types of mutation, which were found at similar levels in all sequenced clones (AC#52, MM#3, SR#12, RG#4) (Fig. 1C, Additional file 1: Figure S2), suggesting that they are biallelic mutations. Stop codons generated by frame shifts were found downstream of each target 5 sequence (Fig. 1C, Additional file 1: Figure S2). These results confirm isolation of sliaa9 knockout mutants of commercial tomato cultivars in the T0 generation following efficient CRISPR/Cas9 editing.

sliaa9-crispr T0 mutant phenotypes
The phenotypes of sliaa9 T0 knockout mutants (sliaa9-crispr) in cultivars Super Roma, Moneymaker, and Rio Grande exhibited morphological changes in leaves, with similar tendency, i.e., the compound leaf shape was changed to a simple leaf with a longer petiole than that of wild-type ( Fig. 2A)-a phenotype previously seen in the leaves and petioles of Ailsa Craig [19]. The growth rates of these mutants were similar to that of wild-type ( Fig. 2A). Flower development was normal in all these mutants compared with wild-type plants (Fig. 2B); however, seedless fruits, or fruits with few seeds, were generated in the knockout mutants, suggesting a parthenocarpic phenotype (Fig. 2C, D). When mutations at the target locus were analyzed by PCR-RFLP analysis, 100% mutation was detected in fruits with parthenocarpic phenotypes (Additional file 1: Figure S3). We also counted fruit numbers in parthenocarpic phenotypes (Fig. 2E). In the wild-type, no fruits developed without pollination, whereas many seedless parthenocarpic fruits developed in the knockout mutants (Fig. 2E). There were no differences in fruit size between the mutants and wild-type (Additional file 1: Figure S4).
These results showed that parthenocarpic phenotypes were efficiently introduced into the various commercial cultivars in the T0 generation by the CRISPR/Cas9 vectors.

Mutation analysis in next generation plants
PCR-RFLP analysis of the T1 generation isolated by self-pollination from the T0 knockout lines showed that some of sliaa9-crispr T1 plant lines in Ailsa Craig, Moneymaker, and Rio Grande cultivars showed 100% somatic mutation (Fig. 3A). Sanger sequencing using cloned PCR fragments at the target site in the T1 lines detected single mutations, suggesting that these lines were homozygotes, and that they were segregated from the parental T0 generation (Figs. 1C, 3A, Additional file 1: Figure S5). Altered leaf morphology and parthenocarpic fruits were also detected in these T1 lines (Fig. 3B, C). We next analyzed a candidate targets for off-target effects (Additional file 1: Figure S6) in mature leaves from 6 T0 and T1 mutants by using amplicon sequence. The results showed that there were no off-target mutations (Additional file 1: Figure S6).

Isolation of null-segregants
Next, we examined null-segregants in the T1 generation. Since the latter were generated from T0 lines by self-pollination, the transgene, including the CRISPR/Cas9 cassette can be segregated out. To detect T-DNA insertion in the genomic DNA of sliaa9-crispr T1 progenies, the PCR primers shown in Additional file 1: Figure S1 were used to amplify the various T-DNA regions of the vector. Figure 4 shows results using the representative mutant lines, which indicate that no PCR bands from any T-DNA regions were detected in the several T1 lines tested (Fig. 4) lines, and 28.6% (2/7 plants) in RG#4 T1 lines (data not shown). Subsequently, a search for T-DNA regions was also performed by PCR in the T2 generation (MM#2-13 and MM#2-14) (Additional file 1: Figure S7). T-DNA was not detected in MM#2-13, a T1 progeny of MM#2 (data not shown), and PCR analysis showed that the T2 progenies of MM#2-13 also did not possess any T-DNA (Additional file 1: Figure S7). On the contrary, individuals with both the presence and absence of T-DNA segregated in the next generation of MM#2-14 (Additional file 1: Figure S7), which did harbor T-DNA in its genome (data not shown).
We then performed Southern blot analysis to detect T-DNA insertion and copy number in various progenies, including null-segregants. Genomic DNA from T0 and T1 progenies were digested by the restriction enzyme HpaI (Additional file 1: Figure S1), and T-DNA was detected using probes for gRNA and Cas9 (Fig. 5). Southern blot analysis suggested a single insertion of the transgene in several T0 and T1 lines. There were no hybridizing bands in several T1 progenies, in which PCR fragments corresponding to T-DNA were not detected (Figs. 4, 5), whereas hybridization bands for the LHCB gene used as a control were detected in all lines. These results indicate that many null-segregants lacking the CRISPR/Cas9 cassette can be efficiently isolated in various tomato cultivars in the T1 7 generation.

Discussion
In this study, parthenocarpic lines of various commercial tomato cultivars were generated by CRISPR/Cas9-mediated mutation of the SlIAA9 gene. These lines are null-segregants, which were generated rapidly by self-pollination of T0 plants. Bi-allelic mutants of SlIAA9 were generated in the T0 generation with high efficiency-a strong advantage in the rapid isolation of null-segregants in early generations. We have previously shown generation of a sliaa9 mutant in Micro-Tom using the same CRISPR/Cas9 vector with similarly high mutation efficiency [19], suggesting that the optimized vector can be used universally among tomato cultivars, although slightly lower mutation efficiency was detected in Rio Grande among the four cultivars used here. This might be due to tissue culture difficulties with Rio Grande; further improvement of experimental procedures, including transformation steps, would increase mutant isolation efficiency in this cultivar.
Down-regulation of SlIAA9 [15] and knockout mutants [16,19] in Micro-Tom and Ailsa Craig have been shown to result in parthenocarpy and morphological changes in leaves. sliaa9-crispr mutants in various commercial cultivars also clearly show parthenocarpy, even in the T0 generation, which is useful in crops with agricultural applications. Our results suggest that the function of SlIAA9 in fruit development is highly conserved among tomato varieties. Leaf morphological changes, however, showed slight differences among cultivars. The leaf morphology changes were more obvious in Moneymaker than in the other three cultivars; knockout plants in Moneymaker exhibited a longer petiole and shorter leaves than in the other cultivars. This might be caused by variations in SlIAA9 function among tomato cultivars. The leaf morphological changes in sliaa9-crispr mutants also affect tomato tree forms (Fig. 2). SlIAA9 downregulation in Ailsa Craig leads to changes in flower morphology [15]; however, we found no apparent differences between sliaa9-crispr mutant and wildtype flowers in Ailsa Craig (Fig. 2C). Further study might be needed to identify differences in mutant flowers between knockdown and knockout of SlIAA9 by CRISPR/Cas9; however, there is the possibility that expression levels of SlIAA9 might affect flower phenotype.
Agrobacterium-mediated transformation leads to T-DNA insertion into the genome, resulting in generation of a mutant line; however, null-segregant mutant lines generated by CRISPR/Cas9 can be developed through segregation of the T-DNA by self-pollination [20,21,22]. In this study, we isolated many null-segregants of sliaa9-crispr mutants in T1 progenies. Using conventional PCR-based analysis to test for various regions of the T-DNA in the progeny, no trace of T-DNA was found in the nullsegregants. However, it is difficult to eliminate the possibility that short regions of T-DNA have been inserted in the mutant genome that may escape detection by PCR-based analysis. In such cases, alternative technologies such as next generation sequencing might be required. The mutant lines we analyzed for off-target effects (Additional file 1: Figure S6) did not harbor any such small varied sequences that may have occurred upon T-DNA insertion (data not shown). In our PCR analysis, one

Conclusions
Generally, much time and labor are needed to isolate null-segregant mutants of homozygotes by back-crossing with wild-type. The method used in this study is effective in achieving rapid isolation of null-segregants with high efficiency, while adding important traits to various cultivars at the same time. While classical breeding requires a long time and much labor to produce new traits, the generation of new crop varieties that are otherwise difficult to cross-breed can be achieved using genome editing. Recent studies report transgene free systems as well as novel RNP methods and the optimization of transformation and tissue culture systems [10,11,13,23,24,25,26], which will further increase the advantages of using CRISPR/Cas9 for the production of novel crop varieties in the future.

Mutation Analyses in CRISPR/Cas9 Target Sites
Genomic DNA was isolated from tomato leaves and fruits using an SDS-based DNA extraction method.
Fragments, including the target sequence, were amplified PCR using PrimeSTAR GXL DNA Polymerase (TaKaRa, Japan). The recognition sequence of AccI is included at the predicted DSB position on the SlIAA9 target sequence. In PCR-RFLP, PCR fragments were digested with AccI (NEB, Japan) and analyzed by agarose-gel electrophoresis. For Sanger sequencing analysis, PCR fragments purified from agarose-gel were cloned by the Seamless ligation cloning extract (SLiCE) method [28] into cloning vector pNEB193 (NEB, Japan). All primers used for PCR are listed in Additional file 1: Table S1.

PCR for detection of null-segregant plants
Nine regions of the T-DNA in the CRISPR/Cas9 vector were selected and amplified using PrimeSTAR GXL polymerase (TaKaRa, Japan). Between 20 and 50 plants in the T1 generation and 8 plants from T2 generations were randomly selected for PCR-based detection of T-DNA regions. All primers used for PCR are listed in Additional file 1: Table S1. The amplicon sizes are also indicated in Additional file 1: Table S1. Technical replicates were performed three times for each line.

Southern blot analysis
Total DNA was extracted from leaves of WT and sliaa9-crispr mutants by a CTAB-based DNA isolation method. Briefly, tomato leaves were crushed after freezing with liquid nitrogen, and incubated in 2% CTAB solution. The lysates were extracted with chloroform: isoamyl alcohol (CI) (24:1), and then 1% CTAB solution was added to the water layer. Total DNAs were obtained by CsCl-EtOH precipitation.
Total DNA was digested with HpaI, which has a single restriction recognition site in the vector (Additional file 1: Figure S1), fractionated on 1% agarose gels, and transferred to nylon membrane (Zeta-Probe, Bio-Rad, USA) by capillary transfer. After UV cross-linking (UVP CL-1000 crosslinker, Analytik Jena, Upland, CA, USA), the membranes were hybridized with DIG-labelled probes for gRNA, Cas9, and LHCB (Solyc02g070970) genes. Probes were amplified using pEgP237-2A-GFP (gRNA and Cas9) or tomato genomic DNA (LHCB) using the primers listed in Additional file 1: Table S1, and labeled with DIG by using a DIG High Prime DNA Labeling and Detection Kit (Roche, Basel, Switzerland). Probe labeling, membrane hybridization, and detection were carried out according to the manufacturer's instructions. Hybridization signals were detected and imaged with PXi (SYNGENE, Bangalore, India).

Off-target analysis by deep sequencing using MiSeq
Deep sequencing was performed using MiSeq Reagent Kit v2 Nano (Illumina, Japan). Genomic DNAs, including the region of the CRISPR/Cas9 target sites and off-target sequences, were amplified by PCR.
One off-target candidate sequence (SL2.50 ch6 :26946923-26946946) was selected by Cas-OT to examine the mutation. PCR products were first separated by electrophoresis, then purified from the agarose gel using the Wizard SV Gel and PCR Clean-Up System (Promega, Japan) and used as templates for a second round of PCR. Second PCR primers were subjected to TruSeq (Illumina, Japan).
All primers used for PCR are listed in Additional file 1: Table S1. MiSeq data was analyzed using CLC Genomics Workbench software version 7.5.1 (CLC bio, Japan), mapped on the off-target candidate

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
YO is the guest editor of this journal. The authors declare that they have no competing interests.   Isolation of null-segregants in the T1 generation of the sliaa9-crispr mutants. Southern blot analysis was performed to isolate null-segregants of sliaa9-crispr mutants. DIG-labeled gRNA and Cas9 probes were used to detect T-DNA insertion in HpaI-digested genomic DNA of sliaa9-crispr mutants. The LHCB gene was used as an endogenous control.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download. Supplementary_materials_Hara_etal.pdf