Identiﬁcation and Validation a Major QTL from “Sea Rice 86” Seedlings Conferred Salt Tolerance

: Saline stress severely a ﬀ ects rice ( Oryza sativa L.) growth and development and reduces crop yield. Therefore, developing salt-tolerant and high-yielding rice using quantitative trait loci (QTLs) and linkage markers is a priority for molecular breeding. Here, the indica rice Sea Rice 86 (SR86) seedlings showed higher tolerance than ordinary rice varieties in saline soil, and a dominant e ﬀ ect on salinity sensitivity was demonstrated by genetic analysis. We constructed bulked segregant analysis pools using F 2 populations from parents Dianjingyou 1 as the recipient and SR86 as the donor. We identiﬁed a 2.78 Mb region on chromosome 1 as the candidate region. Using simple sequence repeat markers and substitution analysis, we mapped the target region within 5.49 cM in the vicinity of markers RM8904–RM493. We speculated that this QTL, named qST1.1 , might contribute signiﬁcantly to the salt tolerance of SR86. The high salt tolerance of introgression lines obtained by marker assistant selection (MAS) conﬁrmed that the qST1.1 region was associated with salinity tolerance. This newly-discovered QTL will be helpful for the analysis of the salt-tolerant mechanism of rice and breeding high-quality rice varieties using MAS.


Introduction
Salinity is one of the major abiotic stresses that adversely effects plant growth and development, constraining crop yield [1,2]. At least 6% of the world's land is affected by salinity [3], and the yield of many rice varieties can be reduced by up to 50% in response to 50 mM NaCl. Hence, efforts to increase the salt tolerance of rice would aid in increasing the use of marginal saline-alkaline land and improving crop production [4,5]. Rice is considered as most susceptible to salt stress in the seedling, reproductive, and 2-3 leaf stages [6]. Previous investigations clarified that salinity influences rice growth in three phases: osmotic stress, ionic toxicity, and oxidative stress; and salt tolerance is a comprehensive quantitative trait involving both genetic characteristics and physiological responses [7][8][9].
Emerging new technologies, combining high-throughput sequencing technology with traditional technologies such as RNA-sequencing (RNA-seq), alternative splicing analysis, mapping by sequencing (MBS), bulked segregant analysis (BSA), whole-genome resequencing, and quantitative trait locus (QTL) mapping, have facilitated the analysis of similarly complex characters [10]. Marker assisted selection (MAS) based on QTL location is a generally accepted technique for breeding new salt-tolerant rice varieties [11]. However, it is difficult to assess salt tolerance by evaluating physiological traits. A DJY1 and donor SR86; the overlapped segment substitution lines (BC 2 F 3 ) were derived from a cross between recipient (DJY1, YJ26, YJ29, YJ32) and donor SR86. The introgression line 2018H3T22 (T22) was from BC 2 F 5 and was homozygous for the SR86 genotype of qST1.1. T22 was bred as follows: BC 2 F 3 plants (SR86 × YJ26) with good salt tolerance were transplanted to the field; some of the plants in BC 2 F 4 (collected from BC 2 F 3 ) were tested for salt tolerance and used for genetic linkage analysis with SSR markers; the plants that had good salt tolerance and were homozygous for the SR86 genotype of qST1.1 were transplanted; and the BC 2 F 5 plants contained the homozygous target fragment were served as the introgression line. Both hybridization and propagation were carried out at Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, located in Menglun, Jinghong, Yunnan Province, China, with planting twice per year.

Phenotype Measurements
Phenotype identification of all the populations and the two parents was conducted in greenhouses (natural illumination, 25-28 • C). Seeds were germinated at 30 • C. After germination, they were grown in plastic seedling pots with 2 cm plant spacing and normal seedling management (red soil: organic matter: coconut bran = 1:2:1). When the plants grew to the trifoliate stage, the soil was dried under natural conditions (but without drought stress), then treated with 1.2% NaCl (w/v) solution and the water level was kept at 1 cm depth above the soil during the treatment. Finally, the phenotype of the populations and parents were identified when they were significantly different in appearance (after 8-10 d of salt treatment, usually). According to the salt damage degree classification standard developed by the International Rice Research Institute (IRRI) (Table S1), all the lines or individuals of populations were assigned into five grades as 1, 3, 5, 7, and 9 corresponding to resistance levels: extremely resistant, resistant, medium resistant, medium sensitive, and sensitive, respectively.

DNA-BSA Analysis
Plants of the F 2 generation and the parents SR86 and DJY1 were treated with 1.2% NaCl solution at the trifoliate stage. The healthy leaves of 100 individuals in F 2 belonging to grade 7 or 9 at day 8, 100 individuals belonging to grade 1 or 3 at day 12, and the two parents were collected and quickly frozen with liquid nitrogen then stored at -80 • C .The genomic DNA of all samples was extracted using the cetyltrimethylammonium ammonium bromide (CTAB) method [25]. Four mixed DNA pools (R01-DJY1, R02-SR86, R03-salt-tolerant pool, R04-salt-sensitive pool) were constructed for bulked segregant analysis (BSA). The BSA sequence and analysis were performed by Biomarker Biotechnology Co., Ltd. (Beijing, China).

Genotype Identification and QTL Analysis
Genomic DNA was extracted from fresh leaves as described above. Primers of SSR markers were selected from those reported by McCouch et al. and Ganie et al. [26,27]. A total of 36 SSR markers covering the region 7.9-12.26 Mb on chromosome 1 were used for testing parental polymorphisms and mapping genotypes; 9 markers with polymorphisms among the population were screened to proceed the QTL mapping using 186 F 2 (SR86 × DJY1) individuals (Table S2). PCR reactions (20 µL) contained: 10 µL 2 × Taq Master Mix (Novoprotein Scientific Inc.), 1 µL forward primer (10 µmol/µL), 1 µL reverse primer (10 µmol/µL), 2 µL DNA samples, and 6 µL ddH 2 O. The PCR reactions involved: 95 • C for 5 mins, and then 26 cycles of 30 s at 95 • C, 30 s at 60 • C, and 30 s at 72 • C, followed by 72 • C for 2 mins. The PCR products were detected by polyacrylamide gel electrophoresis (PAGE) and stained with silver.

Measurements of Agronomic Traits
Normal sowing and germinating were done for seeds, and when the seedlings grew to four leaves stage, the seedlings were transplanted to the field with six rows per plot and 25 × 20 cm between plants. The agronomic traits of the plants in the middle of the middle rows were measured when plants grew to maturity under normal field management. At least 10 individuals of each line were randomly Agronomy 2020, 10, 410 4 of 12 selected for measurement. Plant height (cm) (awns excluded), the panicle length(cm), number of filled grains per panicle, number of primary branches, number of secondary branches, effective tiller number, 1000-grain weight (g), grain length (mm), grain width (mm), grain length-to-width ratio, and grain yield per plant (g) were measured with units per plant and averaged. The experiments were conducted three times independently.

Statistical Analysis
QTL IciMapping software (version 3.2) was used to construct a genetic linkage map based on the genotypes and phenotypes. SPSS (version 18) in International Business Machines (IBM) was used to analyze data of agronomic traits in the introgression lines and two parents. Student's t-tests were used for analysis, and the threshold of p < 0.05 was considered as statistically significant, and values of p < 0.01 as highly significant.

Phenotypes and Genetics Analysis of Salt Tolerance in SR86
SR86 is a popular variety for salt-tolerant rice breeding. We compared the salt tolerance of SR86 with four common rice varieties grown in saline soil. SR86 showed higher salt tolerance compared with DJY1, YJ26, YJ29, and YJ32 under increasing concentrations of NaCl in the soil (Table 1 and Table  S1). SR86 was classified as salt resistant and graded as 1-3 with survival for more than 15 d, while the other varieties survived less than 10 d during the trifoliate stage in soil with 1%, 1.2%, and 1.5% NaCl. The significant difference of the phenotypes was clearly visible, and we chose 1.2% NaCl for further analysis. We also performed the above-mentioned experiments with the same treatments under hydroponic conditions; surprisingly, there was no difference between SR86 and the other varieties ( Figure 1A and Figure S1), which might imply a different mechanism of salt tolerance between growth in soil and in hydroponic solutions.
To carry out genetic analysis of salt tolerance in SR86, we hybridized SR86 with salt-sensitive variety DJY1 to obtain an F 1 population. The F 1 generation showed stronger salt tolerance than DJY1 after 8 d under 1.2% NaCl treatment ( Figure 1A), and we speculated that the salt tolerance of SR86 was effectively dominant.
To further assess the characteristics of phenotypical variation in salt tolerance of SR86, the phenotype classification was performed in the F 2 population according to the degree of salt damage under 1.2% NaCl treatment. The F 2 population showed different degrees of damage among 893 individuals after salt treatment for 8 d and there were significantly more salt-tolerant plants (448 individuals scored as grades 1-3) than salt-sensitive plants (166 individuals scored as grades 7-9). Thus, the salt tolerance showed dominance or partial dominance, which was consistent with the F 1 population phenotype ( Figure 1B, C and Table S3).

A New Potential QTL (qST1.1) Involved in the Salt Tolerance of SR86
We selected individuals with extreme phenotypes as two mixed pools (salt-sensitive pool and salt-tolerant pool) from the F2 population for BSA. Four mixed DNA pools (R01-parent DJY1, R02parent SR86, R03-salt-tolerant pool, R04-salt-sensitive pool) were constructed and used for DNA sequencing. A total of 71.07 Gbp clean data were filtered from the raw data: quantification of data showed Q20 ≥ 91.87% and Q30 ≥ 85.7% and the GC content was 41.91%-43.43% across all four mixed samples. The reading depth of both parental pools, R01 and R02, was 13 × because of the small number of samples, while the reading depths of the two offspring pools, R03 and R04, were 72 × and 63 × , respectively (Table S4 and S5). When the pooled samples were aligned to the reference genome, we identified respectively in the four pools (R01-R04) about 0.27 million, 2.10 million, 2.23 million, and 2.22 million single nucleotide polymorphisms (SNPs); more than 7500, 43,800, 47,300, and 47,300 indels in genome; 9228 genes, 37,395 genes, 38,923 genes, and 38,890 genes with non-synonymous SNPs; and 3023 genes, 15,223 genes, 16,225 genes, and 16,226 genes with indels (Tables S6-S8). The

A New Potential QTL (qST1.1) Involved in the Salt Tolerance of SR86
We selected individuals with extreme phenotypes as two mixed pools (salt-sensitive pool and salt-tolerant pool) from the F 2 population for BSA. Four mixed DNA pools (R01-parent DJY1, R02-parent SR86, R03-salt-tolerant pool, R04-salt-sensitive pool) were constructed and used for DNA sequencing. A total of 71.07 Gbp clean data were filtered from the raw data: quantification of data showed Q20 ≥ 91.87% and Q30 ≥ 85.7% and the GC content was 41.91%-43.43% across all four mixed samples. The reading depth of both parental pools, R01 and R02, was 13 × because of the small number of samples, while the reading depths of the two offspring pools, R03 and R04, were 72 × and 63 × , respectively (Tables S4 and S5). When the pooled samples were aligned to the reference genome, we identified respectively in the four pools (R01-R04) about 0.27 million, 2.10 million, 2.23 million, and 2.22 million single nucleotide polymorphisms (SNPs); more than 7500, 43,800, 47,300, and 47,300 indels in genome; 9228 genes, 37,395 genes, 38,923 genes, and 38,890 genes with non-synonymous SNPs; and 3023 genes, 15,223 genes, 16,225 genes, and 16,226 genes with indels (Tables S6-S8). The SNP-index between R03 and R04 was calculated and visualized ( Figure 2A and Figure S2). Two adjacent intervals were found on chromosome 1 that exceeded the threshold value 0.99 (Figure 2A and Table 2). Since the distance between the two sections was short, we considered these two sections as one candidate region at around 9.46-12.24 Mb, which contained 448 open reading frames (ORFs). This result indicated that the genes responsible for salt tolerance in SR86 might be located within this 2.78 Mb region on chromosome 1. . &, *, and # represent the plant number, genetic position (cM) of markers on the chromosome, and phenotype, respectively. W386 was crossed from SR86 and DJY1; W401 was crossed from SR86 and YJ26; W402 was crossed from SR86 and YJ29; W404 was crossed from SR86 and YJ32; SR86 was the donor. Start: the start position of the associated region; End: the end position of the associated region; Size: the size of associated region (Mb); Gene number: the number of genes in the associated region. of markers on the chromosome, and phenotype, respectively. W386 was crossed from SR86 and DJY1; W401 was crossed from SR86 and YJ26; W402 was crossed from SR86 and YJ29; W404 was crossed from SR86 and YJ32; SR86 was the donor. Start: the start position of the associated region; End: the end position of the associated region; Size: the size of associated region (Mb); Gene number: the number of genes in the associated region.
To further narrow the candidate region, 36 SSR markers were selected to screen for polymorphism in the region of 7.45-12.39 Mb. Among them, nine SSR markers that showed polymorphism were used to perform genetic linkage analysis and QTL mapping of 186 individuals in the F 2 population, with the following parameters: SR86 was R (salt-resistant), DJY1 was S (salt-sensitive); grades 1 and 3 were assigned as R, grades 7 and 9 were assigned as S. From the analysis, a QTL was identified within a 5.49 cM region flanked by markers RM8094 and RM493 and extending from 11.24 to 12.26 Mb on chromosome 1. The salt tolerant-related region was designated as qST1.1 and showed a LOD score of 14.8014 ( Figure 2B); the phenotypic variance explained (PVE) by this QTL was 62.6%, which suggested that qST1.1 might play a key role in salt tolerance in SR86. Additionally, the additive effect value and dominant effect value of the enhancing alleles for salt tolerance were derived from SR86 (Table S9) indicating that qST1.1 had a positive effect on increasing the salt tolerance of rice. Furthermore, qST1.1 was located in the region of Saltol, where there was a key salt-resistant gene SKC1 [12,16]. However, whether SKC1 was important in determining salt tolerance of SR86 remains to be verified.

qST1.1 Validation by Substitution Experiments and Introgression Line T22
To evaluate the accuracy and precision of the target region qST1.1 that we identified by QTL mapping, segment substitution lines were constructed and used to analyze the salt tolerance related regions within qST1.1. Four BC 2 F 3 populations (W386, W401, W402, and W404) generated using the donor SR86 and recipients DJY1, YJ26, YJ29, and YJ32, respectively, were screened for recombined homozygotes with or without the qST1.1 region. From the 400 introgression lines, 15 homozygous substitution lines with different length overlapping segments were screened for phenotype investigation ( Figure 2C). The phenotypes showed that the lines harbored with the introgression region between RM8094 and RM493 conferred salt tolerance; in contrast, the other recombined lines that lacked the RM8094-RM493 region were sensitive to salt. By the substitution mapping, we narrowed down and mapped qST1.1 to the interval between flanking markers RM8094 and RM493 with a length of 5.49 cM ( Figure 2B). This substitution experiment confirmed the consistency and accuracy of our QTL mapping, and so we inferred that qST1.1 was likely associated with the salt tolerance of SR86 in saline soil conditions. We aimed to utilize the introgression lines carrying qST1.1 from SR86 in breeding practice for salt tolerance. The BC 2 introgression line T22, derived from crossing SR86 with YJ26, was evaluated for 12 agronomic traits including plant height, yield per plant, and seed setting rate, etc. (Table 3). It was notable that the grain number per panicle and the yield per plant of T22 were significantly higher than those of the parents, although the seed setting rate and 1000-grain weight of T22 were less than YJ26. Notably, the lower plant height of T22 (100.52 cm) was suitable for lodging resistance compared with SR86 (163.22 cm), and this feature might contribute to the increased yield in T22. Importantly, the salt resistance of T22 was similar to, or even better than, SR86 (Figure 3). These data proved that T22 did have a better performance than its parents and offered the prospect of being useful for breeding new rice varieties with improved salt tolerance. We developed further salt-tolerant QTL introgression lines in four different genetic backgrounds (DJY1, YJ26, YJ29, YJ32 as recipients); seven recombinants harboring qST1.1 showed salt resistance ( Figure 2B and Table S10). These salt-tolerant QTL introgression lines might also be useful in breeding for salt tolerance with evaluation under saline soil conditions. The asterisk * and ** indicate significant difference from control at p < 0.05 and p < 0.01, respectively; the T22 column was marked as the comparison result with YJ26.
substitution lines with different length overlapping segments were screened for phenotype investigation ( Figure 2C). The phenotypes showed that the lines harbored with the introgression region between RM8094 and RM493 conferred salt tolerance; in contrast, the other recombined lines that lacked the RM8094-RM493 region were sensitive to salt. By the substitution mapping, we narrowed down and mapped qST1.1 to the interval between flanking markers RM8094 and RM493 with a length of 5.49 cM ( Figure 2B). This substitution experiment confirmed the consistency and accuracy of our QTL mapping, and so we inferred that qST1.1 was likely associated with the salt tolerance of SR86 in saline soil conditions. We aimed to utilize the introgression lines carrying qST1.1 from SR86 in breeding practice for salt tolerance. The BC2 introgression line T22, derived from crossing SR86 with YJ26, was evaluated for 12 agronomic traits including plant height, yield per plant, and seed setting rate, etc. (Table 3). It was notable that the grain number per panicle and the yield per plant of T22 were significantly higher than those of the parents, although the seed setting rate and 1000-grain weight of T22 were less than YJ26. Notably, the lower plant height of T22 (100.52 cm) was suitable for lodging resistance compared with SR86 (163.22 cm), and this feature might contribute to the increased yield in T22. Importantly, the salt resistance of T22 was similar to, or even better than, SR86 (Figure 3). These data proved that T22 did have a better performance than its parents and offered the prospect of being useful for breeding new rice varieties with improved salt tolerance. We developed further salt-tolerant QTL introgression lines in four different genetic backgrounds (DJY1, YJ26, YJ29, YJ32 as recipients); seven recombinants harboring qST1.1 showed salt resistance ( Figure 2B and Table S10). These salt-tolerant QTL introgression lines might also be useful in breeding for salt tolerance with evaluation under saline soil conditions. . Figure 3. Identification of salt tolerance in IL-T22. It shows the salt resistance of T22 was similar or even better than SR86 compared to YJ26 after being treated with 1.2% NaCl solution for 8 d.

Discussion
Improving the salt tolerance of rice is valuable for increasing crop yield. Salt tolerance is a quantitative trait, and a complex set of genetic and physiological mechanisms have been reported as related to salt tolerance in rice, including regulation of osmotic potential, ion transport capacity, antioxidant ability, and hormone-related pathways [22]. Various kinds of genes and QTLs have been shown to contribute to resisting salt stress under experimental conditions. However, few of them work in practice to improve the salt tolerance of rice in the field, and there remains a need to explore QTLs and genes that might make a greater contribution under saline conditions [28,29].
SR86, an ancient indica subspecies that is phylogenetically close to the divergence point of the major rice varietals, has been shown to have strong salt tolerance [23]. Surprisingly, only a few studies on SR86 have been conducted, and QTL mapping or analysis of the salt tolerance mechanism in SR86 has not been reported. Most of the previous studies on rice salt tolerance were conducted in hydroponic solutions to ensure a homogeneous medium for the salt treatment; however, in the current study, the salt-response differences between SR86 and salt-sensitive varieties were not observed under hydroponic conditions ( Figure S1). We simulated saline soil conditions based on the natural growth environment of SR86, and observed a different phenotype between SR86 and salt-sensitive varieties. In this study, SR86 seedlings were treated with 1.2% NaCl solutions applied to the soil. The electrical conductivity of the water in the soil was measured and showed a relatively high NaCl content at Agronomy 2020, 10, 410 9 of 12 about 0.95% to 1.12% ( Figure S3), which indicated that SR86 has a high salt tolerance compared with other rice varieties. The different performance of SR86 between soil and hydroponic conditions might be attributed to the complex soil environment, such as the soil microorganisms, root exudates, and interaction between the plants and soil microenvironment [30,31].
The identified QTL was mapped using BSA and found to be located within two candidate regions, one around 9.46-12.13 Mb and the other one around 12. 21-12.24 Mb. The QTL, which was named as qST1.1, was located between the markers RM8094 and RM493 in the 11.24-12.26 Mb regions on chromosome 1, with an interval of 5.49 cM between flanking markers. This qST1.1 was located within the QTL "Saltol", which was reported as contributing more than 70% of the variation in salt uptake in a population, with an LOD >14.5 [32]. "Saltol" has proven to be an effective salt tolerance region and was successfully applied in breeding. Field experiments proved that "Saltol" significantly improved the salt tolerance of rice at the seedling stage and maintained stable yield under salt stress [28,33]. Thomson et al. also found some QTL segments related to salt tolerance at the seedling stage in this chromosome region, and some segments overlapped with SKC1, possibly encoding the same gene (OsHKT1;5) [13]. SKC1, as a major member of the high-affinity K + transporter (HKT), functions as a Na + -selective transporter and was mapped using populations derived from a cross between a salt-tolerant indica variety, Nona Bokra, and a susceptible elite japonica variety, Koshihikari [16]. Coincidentally, SKC1 was also located in qST1.1, and SR86 had the same amino acid sequence as Nona Bokra ( Figure S4). The balance of cations, such as Na + , K + , and Ca 2+ , and their transport ability played key roles in the salt tolerance response of rice [31,34,35]. Whether the ions transport affects salt tolerance in SR86 remains to be further explored.
Numerous potential salt-responsive genes are located within qST1.1, which contains 198 ORFs, including 34 retrotransposon proteins, 26 transposon proteins, 57 expressed proteins, 9 hypothetical proteins, 7 peroxidase precursors, 5 transcription factors (TFs), namely OsMADS89 (LOC_Os01g18440), OsWRKY9 (LOC_Os01g18584), OsSPL12 (LOC_Os01g18850), OsMYB78L (LOC_Os01g19330), and OsMYBL (LOC_Os01g19970), and other genes such as PIF1 family genes, and protein kinase genes. Transcription factors play an essential role in rice development and responses to biotic and abiotic stresses [36][37][38][39]; these five TFs have been reported to respond to salt stress in rice to varying degrees [40,41]. The reactive oxygen species signaling pathway mediated by peroxidase was also critical, as it evoked a cascade of responses related to stress tolerance [42][43][44]. The aggregation of related genes in the same chromosome region might be a key feature underlying the salt tolerance of SR86.

Conclusions
In this study, we combined bulked segregant analysis and QTL mapping using parents SR86 (salt-tolerant), DJY1 (salt-sensitive), and their F 2 population. We identified a QTL designated as qST1.1 that played a significant role in salt tolerance in SR86 grown under soil conditions. This QTL was further validated by homozygous substitution lines and introgression lines. This qST1.1 not only improves the ability of plants to resist salt stress, but also does not affect yield; it therefore shows potential as a genetic resource for breeding salt-tolerant rice using MAS. We further applied salt stress using high concentrations of NaCl in soil, and the saline soil environment was simulated to avoid undue disturbance of normal growth conditions. The QTL identified by this method will be more effective for future breeding practices for salt-tolerance. This study will not only help to better understand the genes and mechanisms of salt tolerance in rice, but also provides new insights that will underpin breeding strategies for salt-tolerant rice.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4395/10/3/410/s1, Figure S1: The phenotype of SR86, T22, DJY1, and YJ26 in different salt concentrations under hydroponic conditions. NaCl dissolved and diluted in 1/4 Hoagland solution, Figure S2: SNP-index distributed on chromosomes. (a) R03-salt-tolerant pool, (b) R04-salt-sensitive pool, and (c) R03-R04 pool. The black arrow points to the BSA location range, Figure S3: Linear regression graph of NaCl concentration and electrical conductivity. The blue scattered points mean different concentrations of NaCl solutions and the corresponding electrical conductivity, and the blue dotted line is the trendline. The yellow triangles represent the sample solution, the sample was diluted 5 times and determined, Figure S4: Sequence alignment of SKC1 (OsHKT1;5) protein in SR86, Nona Bokra, DJY1, and YJ26. The dark blue shows the different sites, Table S1: Grading standard of salt tolerance on the relative salt damage rate, Table S2: SSR marks tested in this study, Table S3: The phenotypic frequency statistics of F 2 population under 1.2% NaCl treatment, Table S4: Data quantity and quality of whole-genome resequencing, Table S5: Sequencing depth and coverage statistics of whole-genome resequencing, Table S6: SNP statistics of whole-genome resequencing, Table S7: Indel statistics of whole genome and coding regions, Table S8: Classification and statistics of differential genes produced by various mutations, Table S9: The qST1.1 affected salt tolerance in F 2 at the seedling stage, Table S10: Different length homozygous overlapping segment lines screened from BC 2 F 3 .