Enhanced rice salinity tolerance via CRISPR/Cas9-targeted mutagenesis of the OsRR22 gene

Salinity is one of the most important abiotic stress affecting the world rice production. The cultivation of salinity-tolerant cultivars is the most cost-effective and environmentally friendly approach for salinity control. In recent years, CRISPR/Cas9 systems have been widely used for target-site genome editing; however, their application for the improvement of elite rice cultivars has rarely been reported. Here, we report the improvement of the rice salinity tolerance by engineering a Cas9-OsRR22-gRNA expressing vector, targeting the OsRR22 gene in rice. Nine mutant plants were identified from 14 T0 transgenic plants. Sequencing showed that these plants had six mutation types at the target site, all of which were successfully transmitted to the next generations. Mutant plants without transferred DNA (T-DNA) were obtained via segregation in the T1 generations. Two T2 homozygous mutant lines were further examined for their salinity tolerance and agronomic traits. The results showed that, at the seedling stage, the salinity tolerance of T2 homozygous mutant lines was significantly enhanced compared to wild-type plants. Furthermore, no significantly different agronomic traits were found between T2 homozygous mutant lines and wild-type plants. Our results indicate CRISPR/ Cas9 as a useful approach to enhance the salinity tolerance of rice.


Introduction
The global crop production needs to double by 2050 to match the demands of the rapidly increasing population, changing diet, and increasing biofuel consumption (Ray et al. 2013). However, abiotic stress, which includes drought, salinity, low temperature, heat, flooding, and oxidative stress, severely limits the feasible yield increase, or even reduces crop production in large areas (Mahajan and Tuteja 2005). Among these abiotic stress, salinity poses one of the major threats to crop production since most crop plants cannot grow under a high concentration of salt (Munns and Tester 2008). Furthermore, soil salinity is hard to remove, which will cause a continuous decrease in crop production for many years. Over 400 million hectares of land throughout the world have been affected by salinity (http://www.plantstress.com/Articles/index.asp). Moreover, land affected by salt stress is arising due to various factors such as climate change, sea-level increases, and tsunamis (Kumar et al. 2013). Consequently, salinity remains a severe threat to the food supply.
Rice (Oryza sativa L.) is one of the most important food crops and forms the main staple food for more than half of the world's population. Since rice is a species originally grown in swamps and freshwater marshes, it is particularly sensitive to salt stress and rated as a particularly salt-sensitive crop (Dionisio-Sese and Tobita 1998; Kumar et al. 2013). Salinity is one of the major obstacles for rice production especially at the seedling stage (Lutts et al. 1995). Researching of rice salt tolerance is becoming increasingly urgent and improving the salt tolerance of rice has become an important breeding goal. Numerous salt tolerance quantitative trait loci were identified and few of them had been transferred into popular rice varieties via markerassisted selection (MAS) (Lang et al. 2011;Bimpong et al. 2016;Jing and Zhang. 2017). During the past two decades, many salt-related genes (SKC1, DST, OsRR22, OsHAL3, P5CS, SNAC2, and OsNAP) have been successfully cloned (Ren et al. 2005;Hu et al. 2008;Huang et al. 2009;Sun et al. 2009;Karthikeyan et al. 2011;Chen et al. 2014;Takagi et al. 2015). Among them, the OsRR22 gene encodes a 696-amino acid B-type response regulator transcription factor that is involved in both cytokinin signal transduction and metabolism; its loss of function has been reported to significantly increase salt tolerance (Takagi et al. 2015).
The CRISPR/Cas9 system is an accurate, convenient, and efficient genome-editing method developed during recent years (Shan et al. 2013). At present, the CRISPR/Cas9 system has been widely used for genome editing in major crops such as wheat Liang et al. 2017), maize (Svitashev et al. 2016;Zhu et al. 2016), and sorghum (Li et al. 2015;Cai et al. 2015). In rice, using CRISPR/Cas9 technology, many genes (OsPDS, OsERF922, OsHAK1, Badh2, and TMS5) have been knocked out and the expected phenotype was obtained Wang et al. 2016;Zhou et al. 2016;Nieves-Cordones et al. 2017;Shao et al. 2017). This system provides a new method for rice breeding. This study first reports the improvement of salinity tolerance via CRISPR/Cas9-targeted mutagenesis of the transcription factor OsRR22.

Plant growth conditions
The elite japonica rice cultivar WPB106 was bred from 'Huhan9/Huxiangjing//Huhan3/Huhan11' in our laboratory. All transgenic plants and WPB106 (wild type, WT) were grown in the greenhouse at 28-35°C, in Shanghai, or in fields at the station of the Shanghai Academy of Agricultural Sciences under normal growth conditions. For salinity stress at the seedling stage, seedlings of rice were cultivated in normal nutrient solution for 5 days after germination on a 96-well plate (Xia et al. 2017). They were placed in a growth chamber (14 h of daytime at 30°C and 10 h at night at 20°C with 70% relative humidity).

Vector construction
The Cas9 plant expression vector (pYLCRISPR/ Cas9Pubi-H) and the sgRNA expression vector (pYLgRNA) were provided by Prof. Yao-Guang Liu of the South China Agricultural University. The Cas9-OsRR22-gRNA expressing vector was constructed following previously described protocol (Ma et al. 2015a). Briefly, according to the design principles of the target sequences of the CRISPR/ Cas9 system, 19 to 20 bases upstream of the protospacer adjacent motif (PAM) were selected as candidate target sequence (Fig. 1 a). A BLAST search (http://blast.ncbi.nlm.nih.gov/Blast.cgi) of the target sequences (including PAM) against the rice genome was conducted to confirm their targeting specificity in the genome. The target sequence has a difference of at least two bases compared with similar non-target sequences within the PAM or PAM-proximal region. The gRNA expression cassette was synthesized via overlapping PCR. The target-specific sequence of gRNA (target OsRR22) was put at the 5′-end of the primers RR22-gRT+/RR22-OsU6aT-. Two PCR reactions were performed, using the plasmid pYLgRNA-OsU6a/LacZ as template. The first PCR was performed using the primer set U-F/RR22-OsU6aT-, and the second one used the primer set RR22-gRT+/gR-R (Table 1). The products of PCR 1 and 2 were used as templates for the third PCR reaction with the primer set U-GAL/Pgs-GAR to generate the full-length gRNA fragment (Table 1)  The expression of Cas9 is driven by the maize ubiquitin promoter (Pubi); the expression of the sgRNA scaffold is driven by the rice U6a small nuclear RNA promoter (OsU6a); the expression of hygromycin (HPT) is driven by two CaMV35S promoters (2 × 35S). Abbreviations: NLS, nuclear localization signal; Tnos, gene terminator; LB and RB, left border and right border, respectively. c Nucleotide sequences at the target site in the nine T 0 mutant rice plants. The recovered mutated alleles are shown below the wildtype sequence. The target site nucleotides are indicated with black capital letters. The PAM site is underlined. The red dashes indicate deleted nucleotides. The red capital letters indicate inserted or substituted nucleotides. The numbers on the right indicate the type of mutation and the number of nucleotides involved. Bi,^Bd,^and Bs^indicate insertion, deletion, and substitution of the indicated number of nucleotides, respectively; BWT^indicates wild-type a Cas9p expression cassette (Pubi::NLS::Cas9 p::NLS::Tnos) and a hygromycin resistance cassette (2×P35S::HPT::T35STnos) ( Fig. 1 b).

Agrobacterium-mediated rice transformation
The Cas9-OsRR22-gRNA expressing binary vector was introduced into the japonica rice WPB106 cultivar via the Agrobacterium-mediated transformation method described by Nishimura et al. (2006). Hygromycincontaining medium was used to select hygromycinresistant calli, and then, vigorously growing calli were transferred to regeneration media to generate green plants.

Identification of mutant transgenic plants
To determine the mutation at the target site, genomic DNA from the leaves of transgenic plants was extracted using a DNA Quick Plant System (TransGen Biotech, Beijing, China). Genomic DNA (50 ng) was used as template to perform PCR amplifi cation using PCR Mastermix (TIANGEN, Beijing, China). PCR was performed to amplify the genomic region containing the CRISPR/Cas9 target site, using specific primer pairs that surround the designed target site ( Table 1). The PCR products were directly sequenced or cloned into the pEASY-Blunt vector (TransGen Biotech, Beijing, China) and sequenced using the Sanger method. Mutations were identified by comparing the sequences of transgenic plants with those of WT plants. Mutations containing normal sequencing chromatograms were considered as homozygote mutations. Mutations containing superimposed sequencing chromatograms were considered heterozygous or bi-allelic mutations, which were decoded via degenerate sequence decoding (Ma et al. 2015b).
To identify T-DNA-free plants from T 1 , the plants were analyzed via PCR using HPT-specific and Cas9specific primers (Table 1) in combination with agarose gel electrophoresis. The pYLCRISPR/Cas9Pubi-H plasmids and the T 0 transgenic plants were selected as positive controls and WPB106 DNA and H 2 O were used as negative controls. HPT-and Cas9-negative plants were considered as T-DNA-free plants.

Greenhouse trials for salinity tolerance
To evaluate the salinity tolerance of plants at the seedling stage, a salt stress test was performed according to the method published by Takagi et al. (2015). We compared the salinity tolerance of 2-week-old WT and homozygous mutant plants in the greenhouse of the Shanghai Agrobiological Gene Center, Shanghai. Briefly, 2-week-old plants were treated with fresh groundwater and concentrations of 0.75% NaCl solution (pH = 7), respectively. After 2 weeks of treatment, the salinity tolerance was determined via plant height and shoot fresh and dry weights of 10 plants per line. Each line was replicated three times.

CRISPR/Cas9 design
To design a mutation specifically targeting the OsRR22 gene in rice, a 20-bp nucleotide sequence in the first exon of OsRR22 (GenBank Accession No BR000251.1) was chosen as the target site ( Fig. 1 a). The binary plasmid Cas9-OsRR22-gRNA ( Fig. 1 b) was constructed based on the CRISPR/Cas9 vector previously described by Ma et al. (2015a). The vector was used to transform the rice variety WPB106 via Agrobacterium-mediated transformation. Using site-specific PCR and Sanger sequencing, a total of nine WPB106 mutants were recovered from 14 T 0 hygromycinresistant transgenic WPB106 plants (64.3%). These nine mutants were detected and then subjected to zygosity analysis by cloning PCR products into the T vector for DNA sequencing. Sequence analyses detected two homozygous mutations, two heterozygous mutations, and five bi-allelic mutations. Based on allele mutation types, 44.4% of the mutations were nucleotide insertions, 11.1% of the mutations were nucleotide deletions, and 5.6% of the mutations were nucleotide substitutions; 11.1%, 5.6%, and 11.1% of the mutations were simultaneous nucleotide insertions and deletions, insertions and substitutions, and deletions and substitutions, respectively ( Table 2).
Transmission of CRISPR/Cas9-induced mutations to the T 1 generation To further understand the inheritance of the mutation, two homozygous (rr-2, rr-7), one heterozygous (rr-10), and one bi-allelic (rr-1) T 0 mutant plants (Fig. 1 c) were self-pollinated, and their progenies were genotyped at the target site. We randomly selected nine to 23 T 1 progenies derived from each T 0 plant for genotyping analysis ( Table 3). As expected, all of these T 0 putative homozygotes and their offspring had identical genotypes (rr-2 and rr-7), suggesting stable transmittance of the mutations in these homozygous mutant lines to the next generation. Bi-allelic mutations in T 0 plants were apparently transmitted to the T 1 generation following the Mendelian genetic law, indicating that the targeted mutations in T 0 plants were inherited normally. For example, the bi-allelic T 0 mutant plant rr-1 harbors two mutations (a 1-bp deletion (d1) and a 1-bp insertion (i1)); its T 1 progenies segregated in a ratio of 5 (d1):8 (d1/i1):4 (i1), which is consistent with the predicted Mendelian segregation (χ 2 = 0.176 < χ 2 0.05,2 = 5.99) ( Table 3). For the T 1 generation of heterozygous (rr-10), several new mutations (6 (d1/i1)) were generated. In combination, these results clearly demonstrated that CRISPR/Cas9induced gene mutations could be stably transmitted to subsequent generations.

Selection of T-DNA-free mutant rice lines
To obtain rice lines harboring the desired OsRR22 mutations without T-DNA of the construct Cas9-OsRR22, we conducted PCR amplification using the primer sets designed to amplify Cas9 and HPT sequences ( Table 1). The absence of transgenes was determined via negative PCR results of both Cas9 and HPT. T-DNA-free plants were found among most T 1 plants, with the proportion ranging from 20.0 to 33.3% (Table 3). These results indicated that T-DNA-free homozygous mutants could be acquired via segregation populations. We isolated two T-DNA-free homozygous mutant lines (rr-2-1 and rr7-4) in the T 1 generation to produce the T 2 population to identify the salinity-tolerant phenotypes, designated as WPB106-cas-1 and WPB106-cas-2.

Salinity tolerance was enhanced in OsRR22-induced mutations
To evaluate the salinity-tolerant phenotype of the obtained rice mutants, two homozygous mutant T2 lines (WPB106-cas-1 and WPB106-cas-2) with different allelic mutations and WT plants were treated with fresh groundwater and a concentration of 0.75% NaCl nutrition solution at the 2-week-old stage. Compared to WT, two mutant lines grew better than WT under this condition (Fig. 2 a). As measured after 2 weeks of treatment, the shoot fresh weight of WT was reduced by 50.3%, while WPB106-cas-1 and WPB106-cas-2 showed only 10.1% and 2.1% reduction in shoot fresh weight, compared to plants that were grown with fresh groundwater. Similarly, the shoot dry weight of WT had been reduced by 42.6%, whereas WPB106-cas-1 and WPB106-cas-2 showed only decreases of 12.9% and 12.3%. The 0.75% NaCl treatment also caused decreases of 31.8%, 20.3%, and 17.8% in plant height of WT, WPB106-cas-1, and WPB106-cas-2, respectively ( Fig. 2 b and Table S1). The significant difference analysis of the shoot fresh weight, shoot dry weight, and plant height indicated that two mutant lines were significantly different from WT plants. These results implicitly indicate that CRISPR/ Cas9-induced mutations in the OsRR22 gene enhanced the tolerance to salinity.

The main agronomic traits were not altered in rice mutants
To survey whether mutations in the OsRR22 gene affect other agronomic traits, we characterized two homozygous T 2 mutant lines by measuring their plant height, days to 50% flowering, no. of tillers per plant, no. of grains per panicle, spikelet fertility, 1000-seed weight, and yield per plant under normal field conditions. Student's t test showed that none of the T 2 mutant lines was significantly different from WT plants under normal growth conditions (Table 4). These results showed that CRISPR/Cas9-induced mutations in the OsRR22 gene did not significantly influence agronomic traits under normal field conditions.

Discussion
CRISPR/Cas9 is a new genome-editing technique, which is highly specific and efficient. So far, the CRISPR/Cas9 technology has been widely used to

Insertion and substitution
Deletion and substitution 55.6 (5/9) 22.2 (2/9) 22.2 (2/9) 11.1 (2/18) 44.4 (8/18) 5.6 (1/18) 11.1 (2/18) 5.6 (1/18) 11.1 (2/18) a Based on the number of each mutant genotype out of the total number of all mutant genotypes at the target site b Based on the number of each allele mutation type out of the total number of all allele mutation types at the target site Bi^and Bd^indicate insertion and deletion of the indicated number of nucleotides, respectively; Bd/i^indicates the simultaneous deletion and insertion of the indicated number of nucleotides. The numbers on the right indicate the type of mutation and the number of nucleotides involved WT, wild type; ND, not detected improve major crops, such as rape, corn, rice, and soybean (Bortesi and Fischer 2015). However, few studies reported the direct genome editing of elite rice cultivars with the CRISPR/Cas9 technology. The ERF transcription factor gene OsERF922 was mutated by CRISPR/ Cas9 to enhance the blast resistance of the rice variety Kuiku131 with normal phenotypes (Wang et al. 2016). Knockout of the thermo-sensitive genic male-sterile (TGMS) gene tms5 of 11 fertile elite cultivars produced TGMS lines with good agronomic characteristics (Zhou et al. 2016). WPB106, a water-saving and droughtresistant elite japonica cultivar (Luo 2010), has the advantages of drought resistance, early maturity, and cooking quality; however, it is very sensitive to salinity. To quickly improve its salt tolerance, we applied the CRISPR/Cas9 technology. In this study, we used the Cas9-OsRR22-gRNA expressing vector to knockout OsRR22 and achieved 64.3% mutant plants in T 0 transgenic plants. We obtained two homozygous mutant lines that harbor mutagenesis in OsRR22 without exogenous T-DNA. The evaluation of the salinity tolerance at the seedling stage showed that the salinity tolerance of T 2 homozygous mutant lines was significantly enhanced compared to that of WT plants. Furthermore, the result of field trials showed no significant difference between T 2 homozygous mutant lines and WT plants in the main  agronomic traits under normal field conditions. Our study provides a successful case for improving rice salinity tolerance via the CRISPR/Cas9 technology and thus demonstrated that OsRR22 has promising potential to accelerate the improvement of the salinity tolerance in rice breeding. Many genes involved with salinity tolerance have been identified in rice, such as OsNAC6, OsPP1a, OsTPS1, and OsNAP. Transgenic rice plants overexpressing these genes showed an improved tolerance to high salt stresses. However, transgenic plants generated by gene addition are subjected to rigorous genetically modified management. Breeding strategy using CRISPR/Cas9 technology knockdown of rice transcription factor has been demonstrated to be an alternative approach for genetic improvement of rice and avoiding transgenic issue. According to the reports, there are only a few genes acting as a negative regulator of salt tolerance. Although the DST knockdown mutant could effectively improve salt tolerance, it has a large change in agronomic traits, such as leaf width, the panicle number per plant, and the main panicle length (Huang et al. 2009). OsRR22 could significantly improve salt tolerance but no changes were found in other agronomic traits (Takagi et al. 2015). Through our experiments, improved salt tolerance in the OsRR22 knockout lines free of transgene has been verified. Similarly, there was no alteration in the agronomic traits under normal conditions, which has achieved our breeding goals. Salinity tolerance is usually related to drought tolerance. Interestingly, we also carried out the drought-tolerant identification of WT and two homozygous T 2 mutant lines at the seedling stage, and the results showed that there was no difference between WT and mutant lines under drought stress (Fig. S1).
In the present study, the Cas9-OsRR22-gRNAinduced mutagenic frequency of T 0 plants was 64.3% and the homozygous rate of T 0 mutant plants was 22.2%, which was similar to previous reported values in rice Wang et al. 2016;Zhou et al. 2016). Among six types of induced mutations in T 0 plants, single-nucleotide insertions were most frequently detected (up to 44.4%), which is consistent with a previous report ). In addition, allele mutations could be successfully transmitted to the next generations. Moreover, we observed new mutations within the T 1 offspring of rr-10, which are probably due to the continuous modification of WT alleles in Cas9-positive T 1 lines. T-DNA-free plants could be found in almost all T 1 segregation population. These results indicate a very convenient production of T-DNAfree homozygous mutation lines in the T 1 generation.
In conventional rice breeding, efforts to breed for salinity tolerance have been attempted. However, these usually required approximately one decade due to the lack of accurate screening techniques, lack of adequate resistance resources, and timeconsuming backcrossing procedure (Hoang et al. 2016). Compared to conventional breeding, the CRISPR/Cas9 technology offers the ability to shorten the breeding period and thus significantly reducing cost (Schaart et al. 2016). For example, this experiment showed that we only required 1 year to improve the salt tolerance of WPB106 via CRISPR/ Cas9 technology. Furthermore, the CRISPR/Cas9 technology is more accurate than conventional breeding, since it only creates mutations in the target gene without changing other genes. However, rarely, negative regulatory genes with the desired function and the requirement for a PAM (-NGG) sequence form limitations of the CRISPR/Cas9 system. Conventional breeding has the advantage to improve the complex trait, while the CRISPR/Cas9 technique has the advantage in the mutagenesis of key genes. Therefore, the present study indicates that combining the CRISPR/Cas9 technique with conventional rice breeding could become a very powerful new tool for crop improvement. provide a link to the Creative Commons license, and indicate if changes were made.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.