Genomic and transcriptomic analysis reveal molecular basis of salinity tolerance in a novel strong salt-tolerant rice landrace Changmaogu

Background Salt stress is an important factor that limits rice yield. We identified a novel, strongly salt tolerant rice landrace called Changmaogu (CMG) collected from a coastal beach of Zhanjiang, Guangdong Province, China. The salt tolerance of CMG was much better than that of the international recognized salt tolerant rice cultivar Pokkali in the germination and seedling stages. Results To understand the molecular basis of salt tolerance in CMG, we performed BSA-seq for two extreme bulks derived from the cross between CMG and a cultivar sensitive to salt, Zhefu802. Transcriptomic sequencing was conducted for CMG at the germination and young seedling stages. Six candidate regions for salt tolerance were mapped on Chromosome 1 by BSA-seq using the extreme populations. Based on the polymorphisms identified between both parents, we detected 32 genes containing nonsynonymous coding single nucleotide polymorphisms (SNPs) and frameshift mutations in the open reading frame (ORF) regions. With transcriptomic sequencing, we detected a large number of differentially expressed genes (DEGs) at the germination and seedling stages under salt stress. KEGG analysis indicated two of 69 DEGs shared at the germination and seedling stages were significantly enriched in the pathway of carotenoid biosynthesis. Of the 169 overlapping DEGs among three sample points at the seedling stage, 13 and six DEGs were clustered into the pathways of ABA signal transduction and carotenoid biosynthesis, respectively. Of the 32 genes carrying sequence variation, only OsPP2C8 (Os01g0656200) was differentially expressed in the young seedling stage under salt stress and also showed sequence polymorphism in the ORFs between CMG and Zhefu802. Conclusion OsPP2C8 was identified as the target candidate gene for salinity tolerance in the seedling stage. This provides an important genetic resource for the breeding of novel salt tolerant rice cultivars.

Salinity is one of the most common abiotic stresses limiting crop production. Except for in coastal areas, improper irrigation and the use of poor-quality water aggravate the salinization of arid and semi-arid soil in inland areas (Rahman et al., 2016). Rice (Oryza sativa L.) is a staple food crop that is salt sensitive in both young seedling and reproductive stages (Kumar et al., 2013). Identifying rice germplasm that is salt tolerant and breeding rice cultivars that are salt tolerant are the most economic and effective methods for the reduction of rice yield loss caused by salinity.
Rice salinity tolerance is a complex trait controlled by quantitative trait loci (QTLs) (Roy et al., 2011;Wang et al., 2013) and also shows different physiological mechanisms (Li and Xu, 2007). Although many salt tolerant QTLs have been detected in different rice lines (Koyama et al., 2001;Ammar et al., 2009;Pandit et al., 2010;Gong et al., 1999;Kumar et al., 2015), only several major salt tolerant rice QTLs or genes such as qSKC1 (Lin et al., 2004;Ren et al., 2005), qSNC7 (Lin et al., 2004), Saltol (Thomson et al., 2010), and OsRR22 (Takagi et al., 2015) have been identified by genomic methods. A major QTL, qSKC1, encodes a sodium transporter in rice under salt stress (Ren et al., 2005). OsRR22 encodes a B-type response regulator protein that acts as a transcription factor regulating genes involved in osmotic responses and/or ion transport between parenchyma cells and vascular tissue cells of roots (Takagi et al., 2015).
Next generation sequencing (NGS) technologies contribute to discovering genome-wide genetic variation and genotyping in a highly efficient way (Huang and Han, 2013). The relatively low cost of sequencing enables the use of genome and transcriptomic sequencing to map some agronomic traits, especially quantitative traits (Varshney et al., 2014b;Pandey et al., 2017). The bulk segregant analysis (BSA) method is more effective in rapidly locating candidate genomic regions that underlie the target genes based on whole genomic resequencing for the extreme bulks and both parents (Takagi et al., 2013). Several transcriptomic studies have identified numerous differentially expressed genes in salinity-tolerant rice varieties compared to salinitysensitive rice varieties (Kawasaki et al., 2001;Walia et al., 2005;Cotsaftis et al., 2011;Wang et al., 2016;Shankar et al., 2016). These differentially expressed genes (DEGs) are generally associated with stress signaling, ion transport, transcription regulation, and some specific metabolic processes (Kawasaki et al., 2001;Walia et al., 2005;Cotsaftis et al., 2011;Kumar et al., 2013). However, current studies of the tolerance of rice to salinity mainly focus on the classic salt-tolerant rice cultivar, Pokkali, and its derived line, FL478. We identified a new salttolerant landrace called Changmaogu (CMG), which shows much stronger tolerance to salinity than Pokkali at the germination and young seedling stages (Fig. 1a).
In this study, we performed BSA sequencing for two extreme bulks and transcriptomic sequencing for CMG in the germination and young seedling stages under salt stress. This research aimed to identify the salt tolerant candidate genes in the seedling stage and provide insight into the molecular basis of salinity tolerance in CMG.

Evaluation of Salinity Tolerance of CMG and Construction of the Extreme Mapping Population for BSA-Seq
The rice landrace Changmaogu (CMG), collected from a coastal beach of Zhanjiang, Guangdong Province, China, showed strong tolerance to the salinity of sea water (the total salt concentration: 370 mM) (Fig. 1a). To investigate the salt tolerance of CMG and identify the optimal salt concentration, the germination rates of CMG, Pokkali, and Zhefu802 were scored under six different salt (NaCl) concentrations (85 mM, 100 mM, 120 mM, 140 mM, 150 mM, and 170 mM). The germination rates of the three rice lines gradually decreased with increasing salt concentration. Overall, CMG showed a much stronger tolerance to salinity than Pokkali and Zhefu802 under all six salt concentrations (Fig. 1b). The germination rate of CMG was similar to that of Pokkali and much greater than that of Zhefu802 in the 85 mM NaCl treatment. In the 100 mM NaCl treatment, the germination rate of CMG was higher than that of Pokkali. The germination rate of CMG varied based on the NaCl concentration; the germination rates were 69%, 48.67%, and 18.7% in the 100, 120, and 140 mM NaCl treatments, respectively. Hence, the 120 mM NaCl treatment was considered the optimal concentration. The germination rate of CMG, Pokkali, and Zhefu802 were 48.67%, 17.33%, and 4.67%, respectively, in the 120 mM NaCl solution (Fig. 1b).
To map the genes related to salinity tolerance in CMG, we constructed an F 2 population derived from the cross between the strongly tolerant cultivar CMG and the sensitive cultivar Zhefu802. This F 2 population was used to identify the degree of salinity tolerance in the young seedling stage using F 2:3 seeds in the 120 mM NaCl treatment. We identified the salinity tolerance of more than 1000 F 2:3 seeds and detected 241 and 68 F 2:3 lines showing no germination and more than 40% germination rate, respectively. We continued to cultivate the 68 F 2:3 seeds for 25 days after salt treatment and selected 30 extremely tolerant individuals showing good growth for 25 days after salt stress and 30 individuals sensitive to salinity to construct the tolerant and sensitive salinity bulks for BSA-seq (Fig. 1c).

Measurement of Abscisic Acid (ABA)
Abscisic acid (ABA) is a plant hormone that regulates plant growth and development and is rapidly increased by abiotic stresses such as drought and salinity (Mahajan and Tuteja, 2005). High salinity increases ABA (Kumar et al., 2013). To investigate whether the salt tolerance of CMG is associated with ABA, we tested the content of ABA in the roots and leaves under both normal growth conditions and salt stress at the germination and young seedling stage in CMG and Zhefu802. At the germination stage, the ABA levels in CMG and Zhefu802 increased under salt stress, compares with normal growth condition. But the ABA levels in CMG were equivalent to those in Zhefu802 under salt stress and normal growth condition, respectively. Compared with under normal growth condition, the ABA levels in both extreme bulks did not significantly change under salt stress. But the ABA levels (about 185 μg/L) in extreme salt tolerant bulk were remarkably higher than those (about 154 μg/L) in salt sensitive bulk under salt stress and normal growth condition (Fig. 1d). At the young seedling stage, compared with normal growth conditions, the ABA levels in roots and leaves increased under salt stress in both CMG and Zhefu802, and the increase of ABA in CMG was significantly higher than that in Zhefu802 (Fig. 1e). In roots of CMG, the ABA content instantly reached the highest level (162.70 μg/L) at the first sampling point (SSI) and then gradually reduced at SSII (145.80 μg/L) and SSIII (136.76 μg/L). In leaves of CMG, the ABA level gradually increased at three sampling points, SSI (164.80 μg/L), SSII (169.43 μg/L), and SSIII (172.16 μg/L).
However, Zhufu802 showed the different changing trend of ABA content in roots and leaves, compared with CMG. Namely, the ABA level gradually increased in roots (120.54 μg/L, 124.09 μg/L and 131.27 μg/L) and decreased in leaves (148.59 μg/L, 144.96 μg/L and 141.66 μg/L) at the sampling points from SSI to SSIII.

Sequencing and Mapping of Reads to the 'Nipponbare' Reference Genome
Whole genome re-sequencing data were generated for both parents and the extreme salinity tolerant bulk (T_ bulk) and salinity sensitive bulk (S_bulk) ( Table 1). A total of 42.15 million clean reads were generated for the salinity tolerant parent, CMG, 35.79 million clean reads for the salt sensitive parent, Zhefu802, 71.88 million clean reads for S_bulk, and 72.86 million clean reads for T_bulk. The obtained sequencing data were 12.63 Gb for CMG, 10.72 Gb for Zhefu802, 21.54 Gbfor S_bulk, and 21.83Gb for T_bulk, and more than 96% clean reads were mapped to the Nipponbare reference genome. The Fig. 1 The identification of salt tolerance and the assay of abscisic acid (ABA) in Zhefu802 and Changmaogu (CMG). a: The seedlings of CMG, Zhefu802, and Pokkali when the dry seeds were immersed in sea water (total salt:2%) at room temperature for 30 days; b: The germination tendencies of CMG, Zhefu802, and Pokkali at different NaCl concentrations; c: The germination rates of CMG, Zhefu802, and Pokkali in 0.12 mol/L NaCl solution; d: the ABA levels of both parents and two bulks at germination stage; e: the ABA levels in roots and leaves of CMG at different sampling points under salt stress rates of high quality (Q30) bases were more than 91% (91.82% for CMG, 93.66% for Zhefu802, 92.37% for S_ bulk, and 92.78% for T_bulk).

Candidate Regions for Salinity Tolerance by Bulked Segregant Analysis (BSA) Based on Genomic Resequencing
In total, we detected 1,043,738 SNPs, including 96,226 nonsynonymous coding SNPs and 253,340 small InDels, between both parent lines. Between the two extreme bulks, we detected 331,923 SNPs including 32,239 nonsynonymous coding SNPs and 82,091 small InDels (Tables 2 and 3; Fig. 2a,b). An association analysis of salinity tolerance and polymorphic markers was performed using Euclidean Distance (ED) and SNP/InDel-index methods. We detected seven and six related candidate regions covering 3.66 Mb and 10.16 Mb based on SNPindex and ED, respectively (Additional file 1: Table S1). Similarly, according to polymorphic InDels, we obtained 10 and nine candidate regions covering 3.63 Mb and 9.58 Mb based on InDel-index and ED, respectively (Additional file 1: Table S2). We overlapped the results based on the two methods, identifying six and six candidate regions covering 3.66 Mb and 2.6 Mb, respectively, based on the threshold value of the confidence interval of 0.8914 of Δ(SNP/InDel-index) at the 95% significant level ( Fig. 2c;Additional file 1: Table S3). Finally, with further analyses, we identified that the six overlapped regions as the most likely candidate regions for salinity tolerance at the seedling stage. These six candidate regions were located on chromosome 1 and covered a 2.6 Mb region containing 430 genes (Table 4). Based on the sequence polymorphism in ORFs of 430 genes between CMG and Zhefu802, we found that 23 genes carried nonsynonymous coding SNPs and eight genes contained frame shift mutations (Table 5).

Transcriptome Sequencing for CMG under Salt Stress at Germination and Seedling Stages
To understand the differentially expressed genes responding to salt stress, we performed transcriptomic sequencing under the conditions of normal growth and salt stress (120 mM NaCl solution). We randomly selected 24 DEGs for real time PCR (RT-PCR) to validate the differential expression (Additional file 1: Table S4). The RT-PCR results were in good agreement with the transcriptomic sequencing results (Table 6). We investigated the number of upexpressed and downexpressed genes at germination and the seedling stage and found that the number of downregulated genes was greater than that of upregulated genes at the germination stage. The number of upregulated genes was greater than that of downregulated genes at the seedling stage, especially at the first sampling point (SSI) (Fig. 3a). Similarly, the number of differentially expressed transcription factors (TFs) had a similar tendency at the germination and seedling stages (Fig. 3b).

Screening the Target Candidate Genes for Salinity Tolerance
In 31 possible candidate genes including 23 genes containing nonsynonymous coding SNPs and eight frameshift coding genes (Table 5), nonsynonymous and frameshift mutations simultaneously occurred in Os01g0712250 and Os01g0655700. One gene (Os01g0655400) encodes a putative transposon protein and 13 genes encode hypothetical proteins. We investigated the expression of the remaining 17 genes using transcriptomic sequencing. Only Os01g0656200 (OsPP2C8), encoding a protein phosphatase 2C family protein (PP2C), had significantly increased expression levels at the young seedling stage; the expression levels were not altered at the germination stage under the salt treatment. Compared with normal growth conditions, the expression of other genes did not change in the salt treatment. We also compared the sequence of Os01g0656200 (OsPP2C8) in CMG and Zhefu802 according to BSA sequencing, and we detected one SNP (A in Zhefu802 to G in CMG) in the CDS region of OsPP2C8causing the alternation of the corresponding amino acid (Thr in Zhefu802 to Ala in CMG) (Fig. 6a). We also developed the corresponding primers for the variant site and validated the variant site by sequencing and aligning PCR products from CMG (Fig. 6b). Hence, OsPP2C8 was thought to be the most likely candidate gene for conferring salinity tolerance in the seedling stage in CMG.

Discussion
Rice is considered relatively tolerant to salt at germination and sensitive to salinity at the young seedling and reproductive stages (Heenanet al., 1988;Khatun et al., 1995;Zeng et al., 2001). Rice salinity tolerance is also thought to be a quantitative trait. Conventional QTL analysis is a laborious and time-consuming process because of the requirement of genotyping and phenotyping a large number of individuals derived from a biparental cross (Lim et al., 2014). Whole genomic sequencing assisted BSA has been widely used for the analysis of quantitative traits controlled by a few major genes and is thought to be less useful for the identification of minor QTLs . In this study, we performed a BSA analysis based on whole genome sequencing and transcriptome sequencing of rice subjected to salt stress to identify the major QTLs for salinity tolerance in the rice landrace CMG, which has a strong tolerance to salinity.

Mapping Candidate Regions for Salinity Tolerance in CMG
In the whole genome sequencing analysis, we detected a large number of SNPs and InDelInDels between CMG and Zhufu802 and obtained six candidate regions for salinity tolerance in CMG. In these six regions, three regions are small (approximately 20 kb; Chr1: 26110000-26,130,000, 26,160,000-26,180,000, and 36,140,000-36, 160,000) and contained two to three putative genes, respectively. Os01g0648000 (located in the region from 26,110,000-26,130,000 on Chr1) encodes a putative potassium channel protein that has been previously shown to be associated with salt stress (Fuchs et al., 2005); however, it was not differentially expressed in our transcriptase analysis. Another two candidate regions are large (0.81 Mb spanning 26,490,000-27300000and 1.19 Mb from 29, 440,000-30,630,000 on Chr1). In the candidate region from 26,490,000-27,300,000, we detected a sequence variant in the CDS region of Os01g0656200, the expression of which was upregulated in the seedling stage under the salt stress treatment. In agreement with these results, a previous study also showed that Os01g0656200 was associated with salt tolerance in different salt tolerant rice lines (Walia et al., 2005;Pandit et al., 2010). Os01g0656200 (OsPP2C8) encodes a type of protein phosphatase 2C family protein (PP2C). PP2C was known to be involved in abscisic acid signal transduction in higher plants (Sheen, 1998). Three PP2Cs (AtPP2CG1, AtPP2CA, and OsPP2C51) were found to enhance salinity tolerance in an abscisic acid-dependent mannerin Arabidopsis and promoted seed germination in rice Cui et al., 2013;Bhatnagar et al., 2017). We detected a high level of ABA in the leaves and roots of CMG under salt stress in the seedling stage and the expression of OsPP2C8 was significantly up-regulated at the three sampling points in the seedling stage. CMG shows strong salinity tolerance in the germination and seedling stages, and the expression of this gene was significantly up-regulated in the seedling stages under salt stress. Hence, OsPP2C8 is also considered the most likely candidate gene for salt tolerance in CMG.

Plant Hormone and Salinity Tolerance
Plant hormones are thought to be the most important endogenous substances for regulating various physiological responses that lead to adaptation to salinity (Pearce et al., 2001b). Yang et al. (2014) assayed the levels of gibberellin, cytokinin, auxin, and abscisic acid under salt stress in tomato, and the study indicated that ABA played a major role in tomato salt tolerance. The exogenous application of ABA was found to offset the effects of osmotic and ion stress from salt stress conditions in commonbean (Khadri et al., 2007), wheat (Gurmani et al., 2013), and rice (Sripinyowanich et al., 2013) by reducing the sodium concentration and improving osmotic adjustment. In this study, we also found that the ABA level largely increased in the leaves and roots, even after approximately 30 min under salinity stress. Overlapping DEGs among three seedling sampling points were enriched in the pathways of carotenoid biosynthesis and ABA signal transduction, and two DEGs shared in the germination and seedling stages were also clustered into carotenoid biosynthesis. Duan et al. (2012) reported that the genes responsive to salt stress in the tomato line 'Moneymaker' and its wild genotype PI365967 (tolerant to salt) were involved in the carotenoid and ABA biosynthetic pathway. Similar results were also obtained in a close relative of Arabidopsis, Thellungiella, which showed strong tolerance to salt (Wong et al., 2006).
Combined with the above studies, these results likely imply that similar mechanisms in the adaptation to salinity stress apply for both cultivated plants and wild species, namely, rapid accumulation of ABA upon salinity stress condition set off the ABA-dependent signal transduction pathway to activate downstream target genes to respond to salt stress.
The transcriptome analysis under salt stress indicated that only two genes OsNCED3 (Os03g0645900) and Os03g0645966, a hypothetical gene of 69 DEGs were significantly clustered into the pathway of carotenoid biosynthesis. Interestingly, OsNCED3 was down-expressed in the germination stage and upexpressed at three sampling points in the seedling stage. Overexpression of rice OsNCED3 increased the accumulation of ABA, reduced relative water loss, and delayed seed germination in Arabidopsis (Hwang et al., 2010). In the sequence variants of the two genes in the BSA-seq analysis, we found that there were three nonsynonymous SNPs in the CDS regions. It is possible that down-expression of OsNCED3 decreases the level of ABA to promote seed germination in CMG in the germination stage. In addition, no shared DEGs among the germination and seedling stages were identified in plant hormone signaling pathways, which seems to indicate that the mechanism of salinity tolerance in the germination stage might be different from that in the seedling stage. Hence, OsNCED3 could be invovled into salinity tolerance in CMG at the germination stage, but it was not mapped into the candidate region based on the mapping population of CMG/Zhefu802. The enrichment analysis of overlapped DEGs between the germination and seedling stages. a: Sixty-nine DEGs were enriched in biological processes between the germination and seedling stages; b: 69 enriched DEGs overlapped in the metabolic pathways between the germination and seedling stages; c: 164 DEGs were enriched in biological processes that were shared among three sampling points at the seedling stage; b: 164 DEGs were enriched in metabolic pathways that were shared among three sampling points at the seedling stage Among 13 DEGs enriched in the ABA signal transduction pathway in the seedling stage, one (Os02g0766700), two (Os03g0297600 and Os05g0473101), and 10 genes belong to the ABF (ABRE binding factors), PYR/PYL, and PP2C families, respectively. Os02g0766700 encodes a bZIP transcription factor. Previous studies have indicated that bZIP transcription factors play crucial roles in the ABA signaling pathway in plants (Amir Hossain et al., 2010;Liu et al., 2014). The PYR/PYL family has been identified as ABA receptors Ludwików, 2015). In the model of ABA signaling in plants, PP2Cs first interact with SnRK2s to form a reversible regulatory module in the manner of ABA-independence (Umezawa et al., 2009). The PYR-PP2C-SnRK2 complex is the primary framework for ABA signaling and phosphorylates downstream substrates, including bZIP transcription factors to activate ABA-responsive gene expression. We detected a sequence variant in the CDS region of OsPP2C8 from CMG, Zhefu802, and Nipponbare to alter the corresponding amino acid, which could cause the activation of the downstream genes in the pathway of ABA signaling to increase the salinity tolerance of CMG.

Materials
A rice salinity tolerant landrace, Changmangu (CMG), was collected from a coastal beach in Zhanjiang city, Guangdong Province, China. Zhefu802 is a rice cultivar that is sensitive to salinity.
Approximately 2000 plants in the F 2 population were derived from a cross between CMG and Zhefu802. Their leaves were harvested for BSA-seq to map the salinity tolerant genes. The seeds from the F 2 plants (F 2:3 ) were also collected to identify salinity tolerance. The seeds of more than 1000 F 2:3 lines from the F 2 population and two parents were immersed in 120 mM NaCl solution for 4 days to investigate germination rate at 25°C in the artificial climate incubator (MGC-450HP). The F 2:3 lines with more than 40% germination rate continued to be cultivated under salt stress for 21 days. Thirty F 2:3 lines growing well were selected and their corresponding F 2 plant leaves were used to construct the salt tolerant bulk.

Bulked Segregant Analysis by Pooled Sequencing
BSA-seq was used to identify the genes regulating the tolerance to salinity in the above mentioned F 2 population. We selected 30 extremely tolerant and 30 F 2 plants sensitive to salinity to create extreme sample pools. Genomic DNA was extracted using a modified CTAB (Hexadecyltrimethylammoniumbromide) method and purified by chloroform: phenol (1:1) (Chen and Ronald 1999). The DNA quality was checked using an Agilent bioanalyzer 1000 (Agilent Technologies, Singapore). Library preparation was performed according to the manufacturer's protocol. Genomic re-sequencing was conducted to generate paired-end 100-base (PE100) reads using the Illumina Hiseq 2000 platform (Illumina Technologies), which was conducted by Biomarker (China). Clean reads were aligned to reference genome sequences of the Japonica rice Nipponbare genome (http://ftp.ensemblgenomes.org/pub/release-24/plants/fasta/oryza_sativa/ dna/Oryza_sativa.IRGSP-1.0.24.dna.toplevel.fa.gz) using BWA software (Li and Durbin, 2009). SNPs and small InDelInDels were detected using GATK software (Mc-Kenna et al., 2010). The tool of Mark Duplicate in Picard (http://sourceforge.net/projects/picard/) was used to eliminate PCR duplication to increase SNP/InDel-calling accuracy. SNP/InDel-index was calculated for all the SNP/InDel positions. We excluded SNP/InDel positions with multiple genotypes and read depth < 4 from the two bulk sequences. The association analysis was conducted by Euclidean Distance (ED) and SNP/InDelindex, respectively (Hill et al., 2013;Fekih et al., 2013). The overlapped regions based on the above two methods were considered candidate regions for salinity tolerance.

Sample Preparation and Transcriptome Sequencing
Dry seeds from CMG were immersed in 120 mMNaCl solution and sterile water for 4 days. The young buds and roots were harvested at germination and rapidly stored in liquid nitrogen for transcriptomic sequencing. Thirty-day-old seedlings in nutrient solution were treated with 120 mM NaCl solution for 0 min, 30 min, 3 h, and 24 h; approximately 2 g of roots per treatment were collected and rapidly stored in liquid nitrogen for transcriptomic sequencing of the seedling stage. Total RNA samples were extracted using the TRIzol reagent (Invitrogen) and then treated by RNase-freeDNase I (Takara) to remove genomic DNA. mRNA libraries were created according to the standardprotocols provided by Illumina. The mRNA quality including mRNA concentration and fragment size was tested using Qubit2.0 and Agilent 2100.mRNA was enriched using Dynabeads oligo (dT) (Dynal; Invitrogen) and fragmented using fragmentationbuffer. Double-stranded cDNAs were produced using reversetranscriptase (Superscript II; Invitrogen) and random hexamerprimers and further purified using AMPure XP beads. The purified double-stranded cDNA samples were enriched by PCR to construct the final cDNA libraries for sequencing using Hiseq 2500 (150 bp paired ends) by Biomarker (China). All raw-sequence reads data were uploaded to NCBI SequenceRead Archive (SRA, http://www.ncbi.nlm.nih.gov/Traces/sra) with accession numbers SRP143635.
Clean reads were also aligned toreference genome sequences of the Japonica rice Nipponbare genome using TopHat2 . Gene expression differences among different sampling points were detected using the EBSeq package (v1.10.1). Flod change≥2 and False Discovery Rate (FDR) < 0.01 were set to act as the standard for screening the DEGs. Functional classification of DEGs, including Gene Ontology and KEGG pathways, were analyzed using the GOseq R package (Release2.12) and KOBAS software (v2.0).

Real-Time PCR Confirmation of DEGs
A total of 30 DEGs were randomly selected to confirm the transcriptomic sequencing results using real time PCR (RT-PCR). The corresponding sequences of these genes were obtained from the rice genome sequence database (Rap-db). These primers were designed according to the CDS sequences of the corresponding genes using Primer3 software (http://frodo.wi.mit.edu/) ( Table  S1). The Osactin1 gene was used as the internal control. Total RNAs were isolated from thesame samples for transcriptomic sequencing using the TRIzol reagent (Invitrogen) for RT-PCR. First-strand cDNA was synthesized from 1 mg of DNase I-treated RNA samples in a 20 μl reaction solution with the oligo (dT) primer, using a Rever TraAce-akit (TOYOBO). Standard RT-qPCR was performed using SYBR Green SuperMix (Bio-Rad) on a CFX96 Real Time System (BioRad).

Assays of abscisic Acid Level
Thirty-day-old seedlings of Zhefu802 and CMG were treated in 120 mM NaCl solution for 0 min (mock), 30 min (SSI), 3 h (SSII), and 24 h (SSIII). Ten young seedlings were prepared at every sampling point to assay the level of ABA. The ABA content was scored using the plant abscisic acidELISA kits (AndyGene) according to the manufacturer's protocol.

Conclusions
In this study, we mapped six candidate regions for salt tolerance on chromosome 1 based on BSA-seq using extreme populations and identified 32 candidate genes according to the sequence polymorphism in the regions of promoters and ORFs between CMG and Zhefu802. The transcriptome analysis under the conditions of salt stress and normal growth identified numerous DEGs at the germination and young seedling stages. We investigated the expression of the 32 candidate genes and found that OsPP2C8 (Os01g0656200) was differentially expressed in the seedling stage under salt stress. Hence, OsPP2C8 was identified as the target candidate gene for salinity tolerance in the seedling stage, which will provide an important genetic resource for salt tolerant rice breeding.