Generation of Transgene-Free Maize Male Sterile Lines Using the CRISPR/Cas9 System

Male sterility (MS) provides a useful breeding tool to harness hybrid vigor for hybrid seed production. It is necessary to generate new male sterile mutant lines for the development of hybrid seed production technology. The CRISPR/Cas9 technology is well suited for targeting genomes to generate male sterile mutants. In this study, we artificially synthesized Streptococcus pyogenes Cas9 gene with biased codons of maize. A CRISPR/Cas9 vector targeting the MS8 gene of maize was constructed and transformed into maize using an Agrobacterium-mediated method, and eight T0 independent transgenic lines were generated. Sequencing results showed that MS8 genes in these T0 transgenic lines were not mutated. However, we detected mutations in the MS8 gene in F1 and F2 progenies of the transgenic line H17. A potential off-target site sequence which had a single nucleotide that was different from the target was also mutated in the F2 progeny of the transgenic line H17. Mutation in the MS8 gene and the male sterile phenotype could be stably inherited by the next generation in a Mendelian fashion. Transgene-free ms8 male sterile plants were obtained by screening the F2 generation of male sterile plants, and the MS phenotype could be introduced into other elite inbred lines for hybrid production.


INTRODUCTION
Male sterility (MS) provides a useful breeding tool to harness hybrid vigor for hybrid seed production, saving labor, and time. MS can be classified into genic male sterility (GMS) and cytoplasmic male sterility (CMS), based on the inheritance pattern. CMS is usually caused by the mutation of a mitochondrial gene that leads to the defective function of plant mitochondria. According to the molecular mechanism of MS, CMS is classified into three types: CMS-T, CMS-C, and CMS-S (Kim and Zhang, 2018). CMS has been widely used in hybrid seed production in threeline systems. However, the application of CMS is limited because of few restorer line resources, instability of the CMS phenotype, and disease susceptibility of CMS-based hybrids (Weider et al., 2009).
Genic male sterile plants are usually generated by mutations of some nuclear gene, and these plants can be used for hybrid seed production in two-line systems. GMS is a more attractive alternative for hybrid breeding than CMS. In 1921, the first maize GMS mutant was identified (Eyster, 1921). Currently, more than 40 maize ms mutants have been reported. Male sterility 44 (MS44) gene was identified to encode a lipid transfer protein that is mainly expressed in the tapetum, and ms44 is a dominant mutant of the gene (Fox et al., 2017). Male sterility 45 (MS45) encodes a strictosidine synthase which plays an important role in the biosynthesis of terpenoid indole alkaloids, and it is required for the formation of the cell wall of male gametophytes (Cigan et al., 2001). Male sterility 32 (MS32) encodes a basic helix-loop-helix (bHLH) transcription factor that regulates the division and differentiation of anther cells (Moon et al., 2013). Male sterility 26 (MS26) encodes a cytochrome P450 monooxygenase, and it plays a role in fatty acid metabolism (Djukanovic et al., 2013). Male sterility 7 (MS7) was cloned by map-based cloning, and it was identified that the gene encodes a plant homeodomain finger transcription factor. The mutation of the MS7 gene led to abnormal programed cell death (PCD) of tapetal cells . Male Sterility 8 (MS8) encodes a putative ß-1,3-galactosyltransferase, and it mainly affects the meiotic stage of anther development . In the previous report, it was shown that the ms8 mutant plants have fewer branches, that they do not exert anthers, and that they have no viable pollen. The anthers of the ms8 mutant plants had more and shorter epidermal cells, while they also had less and larger tapetal cells (Wang et al., 2010). The ms45 mutant has been successfully used in a hybrid production platform called the seed production technology (SPT) (Wu et al., 2015). A multi-control sterility system based on the maize ms7 mutant has also been developed .
Maize male sterile mutants can also be generated by targeted mutation of candidate genes. Novel ms26 male sterile lines were generated by targeted mutagenesis of MS26 using the redesigned I-CreI homing endonuclease or the CRISPR/Cas9 (Djukanovic et al., 2013;Svitashev et al., 2015). Novel maize ms45 mutants have also been produced by targeted mutagenesis of the MS45 gene (Svitashev et al., 2015). Mutation of the ZmTMS5 gene using the CRISPR/Cas9 technology has generated maize tms5 male sterile mutants that are thermosensitive . In this study, we edited the maize MS8 gene using the CRISPR/Cas9 system, and we obtained novel transgene-free male sterile ms8 lines of maize. These mutant lines are valuable resources for maize hybrid seed production.

Construction of the sgRNA-Cas9 Expression Vector
The codons of Streptococcus pyogenes Cas9 (SpCas9) gene were optimized using maize biased codons (Liu and Scott, 2009). The codon-optimized gene was attached with two nuclear localization sequences (NLSs), MSERKRREKL and MISESLRKAIGKR, at the N-terminal and C-terminal ends, respectively. The Cas9 gene was synthesized and cloned into the pUC57 vector by Sangon Biotech (Shanghai, China) to generate the pUC57-Cas9 vector. Using pUC57-Cas9 plasmid as the template, the Cas9 fragment was amplified by PCR with the BsmBI and NcoI sites added to the N-terminal and with the BstEII site added to the C-terminal. The amplified Cas9 fragment was digested with BsmBI and BstEII, and it was inserted into the pCAMBIA3301 plasmid at the NcoI and BstEII sites to generate the pCAMBIA3301-Cas9 vector.
Maize U3 promoter sequence was also synthesized at Sangon Biotech (Shanghai, China), and it was ligated into the pUC57 vector to generate the pUC57-ZmU3 vector. The sequences of Cas9 and the single guide RNA (sgRNA) construct were shown in Supplementary File S1.

Agrobacterium-Mediated Maize Transformation
The pCAMBIA3301-Cas9-MS8ex2 vector was transformed into the Agrobacterium tumefaciens strain EHA105. Immature embryos from maize hybrid HiII were transformed according to the described method  with some modifications. Immature embryos about the size of 1.2 mm were isolated and suspended in liquid infection medium (Murashige and Skoog basal medium, 68.5 g L −1 sucrose, 36.0 g L −1 glucose, 100 µM acetosyringone, pH 5.2). The embryos were heated at 45 • C for 3 min, and then transferred to an A. tumefaciens suspension and incubated for 5 min. After inoculation, the embryos were transferred to solid cocultivation medium (half strength Murashige and Skoog basal medium, 20 g L −1 sucrose, 10 g L −1 glucose, 0.85 mg L −1 silver nitrate, 100 µM acetosyringone, 1.22 mg L −1 CuSO 4 , 8 g L −1 agar, pH 5.8) and were incubated in the dark at 23 • C for 3 days. The embryos were then transferred onto resting medium (Murashige and Skoog basal medium, 20 g L −1 sucrose, 10 g L −1 glucose, 1.5 mg L −1 2,4-Dichlorophenoxyacetic acid (2,4-D), 0.85 mg L −1 FIGURE 1 | The T-DNA region of the pCAMBIA3301-Cas9-MS8ex2 plasmid. The codon-optimized Streptococcus pyogenes Cas9 (SpCas9) gene was inserted into the pCAMBIA3301 plasmid at NcoI and BstEII sites after the 35S promoter. The ZmU3::sgRNA was inserted into the BamHI and HindIII sites. LB, T-DNA left border; polyA, 35S polyA terminator; bar, bar gene; 35S, CaMV 35S promoter; sgRNA, single guide RNA; ZmU3, Zea mays U3 promoter; Cas9, Streptococcus pyogenes Cas9 gene; Nos, Nos terminator; RB, T-DNA right border; BamHI, HindIII, NcoI, BstEII are restriction enzyme sites. The sequencing chromatograms of MS8 in maize inbred line B73, ms8-DelG/ms8-DelG, and ms8-InA/ms8-InA plants. T 0 transgenic line H17 plant was crossed with maize inbred line Zong31 to produce F 1 generation, which was then self-pollinated to produce F 2 progeny. Fragments flanking the target site were amplified by PCR using the genomic DNA of male sterile F 2 plants as the template, and the PCR products were directly sequenced. silver nitrate, 250 mg L −1 cefotaxime, 8 g L −1 agar, pH 5.8) and were cultured at 28 • C for 7 days. The embryos were moved to a selection medium (Murashige and Skoog basal medium, 20 g L −1 sucrose, 10 g L −1 glucose, 0.7 mg L −1 proline, 0.25 mg L −1 myoinositol, 1.5 mg L −1 2,4-D, 0.5 mg L −1 6-Benzylaminopurine (6-BA), 1.22 mg L −1 CuSO 4 , 250 mg L −1 cefotaxime, 1.5 mg L −1 bialaphos, 8 g L −1 agar, pH 5.8) and were maintained for 2 weeks under dim light (10 µmol m −2 s −1 ) at 28 • C. Then, bialaphos was increased to 3 mg L −1 in the selection medium for two rounds of 2-week selection. Resistant calli were placed in Murashige and Skoog medium containing 1 mg L −1 bialaphos for regeneration under fluorescent white light in a 16/8 h light/dark cycle. The regenerated shoots were transferred to Murashige and Skoog rooting medium. Two weeks later, the regenerated T 0 seedlings were transferred to soil and were grown in a greenhouse with 16/8 h light/dark cycle at 25-28 • C.

Maize Propagation
F 1 generation plants were produced by crossing the transgenic line H17 with the maize inbred line Zong31. These F 1 plants were self-pollinated to produce the F 2 generation. Backcrossing of the MS trait to maize inbred lines was performed by crossing transgene-free male sterile mutant plants ms8-6 or ms8-18 with maize inbred line Zheng58, respectively. The generated F 1 plants were then self-pollinated to produce the F 2 plants. The F 1 and F 2 plants were grown in the field in Beijing or Hainan Province, China.

PCR Analysis of the Transgenic Plants
Genomic DNA was extracted from the leaves of transgenic maize plants using the cetyltrimethylammonium bromide (CTAB) method (Murray and Thompson, 1980). For determining the mutations in the target gene, fragments flanking the target site were amplified by PCR using the genomic DNA as the template. For the transgenic male fertile F 1 generation plants, PCR products were cloned into pEASY R -T1 Simple Cloning vectors (TransGen, Beijing, China) and five randomly selected individual clones were sequenced, or the PCR products were directly sequenced. For the transgenic male sterile F 2 generation plants, PCR products were directly sequenced. For the determination of transgene-free male sterile plants, several fragments at the T-DNA region of the pCAMBIA3301-Cas9-MS8ex2 vector were amplified by PCR with the genomic DNA as the template. The

Phenotype Observation
At the flowering stage, the tassels of wild type (WT) and mutants were photographed in the field in Beijing, China. The spikelets were photographed in the laboratory, and the anthers were observed using an Olympus SZX7.
Inheritance Analysis of CRISPR-Mediated ms8 Genes

Knockout of MS8 Gene in the Progeny of Transgenic Maize Plants
After maize transformation, eight independent T 0 transgenic lines were generated. These transgenic lines were confirmed by PCR analysis of both Cas9 gene and bar gene (data not shown). To identify whether the MS8 genes in these transgenic lines were mutated, the targeted region of the MS8 gene was amplified and sequenced directly. Unexpectedly, sequencing results showed that none of these T 0 transgenic lines were mutated at the MS8 gene. These T 0 transgenic plants could not be self-pollinated to generate T 1 generation plants in the greenhouse; hence, they were crossed with maize inbred line Zong31 to produce F 1 generation, which was then self-pollinated to produce F 2 progeny. To determine the mutations in the target site in F 1 plants, fragments flanking the target site were amplified by PCR and were directly sequenced. The sequencing chromatograms of all the 12 F 1 plants had chaotic peaks after the target site (Supplementary Figure S1), indicating that there were hemizygous mutations in the MS8 gene at the target site in these plants. These PCR products were then cloned into pEASY R -T1 Simple Cloning vectors (TransGen, Beijing, China), and five randomly selected individual clones were sequenced. The sequencing results confirmed that two types of mutations occurred. One was named as ms8-DelG that had a guanine nucleotide deletion, and the other was named as ms8-InA that had an adenine nucleotide insertion (Supplementary Figure S1). Both of these two mutation types resulted in frame shift and a non-functional MS8 protein. The genotypes of these F 1 plants were shown in Supplementary Table S2. Among the 23 F 2 progeny plants from the ms8-DelG-1 line and the 27 F 2 progeny plants from the ms8-DelG-3 line (Supplementary Table S2), nine plants (ms8-1, ms8-2, ms8-3, ms8-4, ms8-5, ms8-6, ms8-7, ms8-8, and ms8-9) and six plants (ms8-19, ms8-20, ms8-21, ms8-22, ms8-23, and ms8-24) showed male sterile phenotype, respectively. Sequencing the ms8 gene in these male sterile plants identified the homozygous mutation in ms8-DelG (Figure 2). Among the 27 F 2 progeny FIGURE 3 | The phenotype of F 2 mutant and wild type plants. The phenotype of tassels, spikelets, and anthers of WT (A,D,G), ms8-DelG/ms8-DelG (B,E,H), and ms8-InA/ms8-InA (C,F,I). T 0 transgenic line H17 plant was crossed with maize inbred line Zong31 to produce F 1 generation, which was then self-pollinated to produce F 2 progeny. At the flowering stage, the WT tassel had anthers, and the anthers were full of pollen, whereas tassels from the two mutants did not exert anthers, and anthers were empty without any pollen. Bar = 1 cm in D-F; bar = 1 mm in G-I. Table S2), nine plants (ms8-10, ms8-11, ms8-12, ms8-13, ms8-14, ms8-15, ms8-16, ms8-17, and ms8-18) showed male sterile phenotype. Sequencing the ms8 gene in these male sterile plants identified the homozygous mutation in ms8-InA (Figure 2). We grew more F 2 progeny plants of these three lines to assess the ratio of male fertile and male sterile plants, and we found a 3:1 segregation between fertile and sterile plants ( Table 1). This result suggested that the male sterile phenotype was inherited in a Mendelian fashion.

plants obtained from line ms8-InA-2 (Supplementary
Similar to WT plants, new ms8 mutant plants have normal vegetative growth and female fertility. The phenotypes of these male sterile plants were significantly different from that of the WT at the flowering stage. The WT anthers were nontranslucent because the anthers were full of pollen grains (Figures 3A,D,G), whereas none of the mutant plants had exerted anthers (Figures 3B,C). Mature mutant anthers were empty and half-translucent when viewed on a light table, without any visible pollen grains (Figures 3E,F,H,I). At the late stage, the mutant anthers began senescing and shrinking; meanwhile, the mutant tassel became white and eventually withered.

Generation of Transgene-Free Male Sterile Plants
Because the MS phenotype was inherited in a Mendelian fashion, we speculated that transgene-free male sterile plants could be obtained in the progeny. To obtain transgene-free ms8 male sterile plants, we used PCR to screen 15 F 2 male sterile plants from the ms8-DelG-1 and ms8-DelG-3 lines. We also screened nine F 2 male sterile plants from the ms8-InA-2 line. Six pairs of primers were designed to amplify Cas9, sgRNA, and 35S::bar fragments ( Figure 4A). Results showed that transgene fragments existed in 11 F 2 mutant plants from the ms8-DelG-1 and ms8-DelG-3 lines, and seven F 2 mutant plants from the ms8-InA-2 line, respectively. However, the transgene fragment was eliminated in four F 2 male sterile plants from the ms8-DelG-1 and ms8-DelG-3 lines (ms8-6, ms8-20, ms8-22, and ms8-24) and in two F 2 male sterile plants from the ms8-InA-2 line (ms8-14 and ms8-18) (Figure 4B). These results suggested that mutations of the MS8 gene were fixed in these transgene-free plants.
To further investigate the inheritability of MS in these transgene-free mutants, we crossed two transgene-free male sterile plants ms8-6 and ms8-18, which belong to the ms8-DelG-1 and ms8-InA-2 lines, with maize elite inbred line Zheng58. All of the F 1 progeny plants were male fertile, whereas the F 2 population displayed a 3:1 segregation ratio between male fertile and male sterile plants ( Table 2). These results suggested that the MS phenotype could be introduced to other elite inbred lines for hybrid production.

Evaluation of Off-Target Mutations in Transgene-Free Male Sterile Mutants
Off-target mutations using the CRISPR/Cas9 system have been reported in many crops, including maize , rice (Endo et al., 2015), and soybean (Jacobs et al., 2015). To evaluate the possibility of off-target mutations induced by CRISPR/Cas9 in this study, we searched the potential off-target sites in the CRISPR-P website 1 . Nine potential off-target sites were found to have 95 or 85% identity with the target sequence. The fragments flanking four potential off-target sites were amplified from male sterile mutants and sequenced (Table 3). Sequencing results showed that mutations occurred in the potential off-target site sequence which had a single nucleotide that was different from the target. However, mutations did not occur in the sequences which had three nucleotides that were different from the target ( Table 3). For the potential off-target site sequences which had a single nucleotide that was different from the target, a cytosine nucleotide deletion allele was identified in some male sterile plants, but the genotypes were not associated with the male sterile phenotype (Supplementary Table S3).

DISCUSSION
Compared with CMS, GMS plant materials are much more useful for the hybrid SPT. Environment-sensitive GMS has been explored in two-line hybrid production system for wheat and rice (Whitford et al., 2013;Huang et al., 2014). Using GMS mutant ms45, DuPont Pioneer developed a hybrid production platform named SPT (Wu et al., 2015). Zhang et al. (2018) also developed a multi-control sterility system based on the ms7 mutant. Expanding male sterile germplasm resources is necessary for developing new SPT technology. The MS phenotype is usually controlled by the recessive mutation of nuclear gene; therefore, it is possible to generate male-sterile mutants using the CRISPR/Cas9 system. Several MS genes have been identified to control the development of pollen or anther. In this study, we chose male sterility 8 (MS8) as the targeted gene and generated novel ms8 mutant lines using the CRISPR/Cas9 system. The MS phenotype of the novel ms8 lines generated in this study could be introduced to other elite inbred lines for hybrid production, which is essential for the use of these mutants in producing hybrids. The new ms8 lines generated in this study showed a male sterile phenotype similar to the reported natural ms8 mutant in which the Mu transposon was inserted in the MS8 gene . If the CRISPR/Cas9 system works well, then there should be high editing efficiency in T 0 generation plants. In this study, we did not observe mutations in the MS8 gene in the T 0 plants. T 0 plants were crossed with maize inbred line Zong31 to generate F 1 plants, and the editing of MS8 gene in F 1 plants was observed. The main reason for the low mutation efficiency of CRISPR/Cas9 system might be that the sgRNA is not stable and has poor targeting efficiency. It has been shown that the low activity of 35S promoter in germline cells would lead to low editing efficiency (Fauser et al., 2014). Using meiotic promoter to drive the expression of Cas9 could increase the targeted mutagenesis efficiency (Eid et al., 2016;Mao et al., 2016). In this study, the Cas9 gene was driven by the 35S promoter and its activity might be low. To localize the Cas9 protein into the nucleus, we fused the amino acids MSERKRREKL and MISESLRKAICKR (Shieh et al., 1993) to the N-and C-terminal ends of the Cas9 protein. We speculate that these nuclear location sequences (NLS) might not completely transport the target Cas9 protein to the nucleus, which might be another reason for the low mutation efficiency.
Off-target events often occurred during the gene targeting process using the CRISPR/Cas9 system. To reduce off-target editing, several strategies such as SpCas9-HF1 variants (Kleinstiver et al., 2016) and CRISPR/Cas9 nickase have been used (Shen et al., 2014). In our study, off-target mutations occurred in sequence which had a single nucleotide that was different from the target. However, there were no mutations in the sequences which had three nucleotides that were different from the target sequence. In some cases, off-target mutations might lead to the defective growth of plants. In this study, the off-target mutations in the ms8 mutant did not lead to a negative effect on the mutant plants. Off-target mutations could also be eliminated by using the breeding methods of backcrossing or outcrossing.
Globally, there is a strict regulatory framework for transgenic crops. Using preassembled CRISPR/Cas9 ribonucleoproteins (RNPs) could generate transgene-free plants with edited genomes (Woo et al., 2015). CRISPR/Cas9 RNPs have also been successfully used to produce transgene-free genome-edited maize (Svitashev et al., 2016) and wheat (Liang et al., 2017). For the genome-edited crops, transgenes and off-targeted genes could also be eliminated by outcrossing or backcrossing. In general, the regulations for genome-edited crops without transgenes should be different from the traditional transgenic crops (Huang et al., 2016). In our study, transgenes were eliminated by backcrossing and transgene-free male sterile mutants were generated. These transgene-free male sterile mutants could be used for maize hybrid production systems without the limitation of regulatory frameworks for transgenic organisms. For application purposes, the MS phenotype of transgene-free mutants should be introduced into different elite maize lines by backcrossing.

CONCLUSION
In this study, a CRISPR/Cas9 vector targeting the MS8 gene of maize was constructed, and it was transformed into maize plants using an Agrobacterium-mediated method; consequently, we detected mutations in the MS8 gene in F 1 and F 2 progeny plants. Mutations in the MS8 gene and the male sterile phenotype could be stably inherited by the next generation in a Mendelian fashion. Transgene-free ms8 male sterile plants were obtained, and the MS phenotype could be introduced into other elite inbred lines for hybrid production.

AUTHOR CONTRIBUTIONS
RC, GW, and YJL designed the research and wrote the article. RC, QX, YL, JZ, and DR performed the research and analyzed the data.

ACKNOWLEDGMENTS
We thank the reviewers for the valuable advice to improve the manuscript.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2018.01180/ full#supplementary-material FIGURE S1 | Gene-targeting of MS8 gene in F 1 plants produced by crossing T 0 transgenic line H17 and inbred line Zong31. (A) The sequencing chromatograms of maize inbred line B73; (B,C) the sequencing chromatograms of ms8-DelG/MS8 plant; (D,E) The sequencing chromatograms of ms8-InA/MS8 plant. In (B,D), PCR products were directly sequenced. In (C,E), PCR products were cloned into pEASY R -T1 Simple Cloning vectors (TransGen, Beijing, China) and the randomly selected individual clone was sequenced. The PAM motif is underlined and in bold.