CHEN Hua-tao, LlU Xiao-qing, ZHANG Hong-mei, YUAN Xing-xing, GU He-ping, CUl Xiao-yan, CHEN Xin

Institute of Industrial Crops, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, P.R.China

Abstract Salt stress is one of the major abiotic stresses affecting soybean growth. Genetic improvement for salt tolerance is an effective way to protect soybean yield under salt stress conditions. Successful improvement of salt tolerance in soybean relies on identifying genetic variation that confers tolerance in soybean germplasm and subsequently incorporating these genetic resources into cultivars. In this review, we summarize the progress in genetic diversity and genetics of salt tolerance in soybean, which includes identifying genetic diversity for salt tolerant germplasm; mapping QTLs conferring salt tolerance;map-based cloning; and conducting genome-wide association study (GWAS) analysis in soybean. Future research avenues are also discussed, including high throughput phenotyping technology, the CRISPR/Cas9 Genome-Editing System, and genomic selection technology for molecular breeding of salt tolerance.

Keywords: soybean, genetic variation, heredity, gene identification, salinity improvement

Salt stress is an important abiotic stress on soybean, a major oil crop, that could have strong negative effects on crop growth. Saline and alkaline soils were estimated at 397 and 434 million ha, respectively, based on the FAO/UNESCO soil map of the world from 1970 to 1980. It was reported that 19.5% of irrigated land and 2.1% of dryland were affected by salt stress (FAO/UNESCO soil map,http://www.fao.org). Soybean (Glycine max) is generally regarded as a salt-sensitive crop compared to other major crops (Munns and Tester 2008). Soybean yields are negatively affected when the soil salinity exceeds 5 dS m–1(Ashraf 1994; Phang et al. 2008). Toxicity occurs when Cl–and Na+ions are absorbed and accumulated at high concentrations in the soybean plant. The seed yields were reduced 50% when the electrical conductivity of the saturation extract of soil, electrical conductivity of the saturation extract (ECe), was 9 millimhos cm–1(Abel and MacKenzie 1964). Papiernik et al. (2005) reported that increasing salinity stress led to 40% yield reduction in soybean. In addition, high salinity can result in soybean plant death (Pathan et al. 2007; Phang et al. 2008). Based on the importance of soybean production and problems caused by salinity, improving the salt tolerance in soybean is becoming a major breeding target around the world.

1. Genetic diversity of salt tolerance in soybean

Multiple soybean germplasms have been used to testsalt tolerance, and rich genetic variation conferring salt tolerance has been observed in soybean (Table 1). Salt tolerance was evaluated on 15 soybean cultivars in 1981,and 10 cultivars were identified as salt tolerant (Parker et al.1983). Furthermore, Parker et al. (1983) found that the average Cl-content in the leaves of susceptible cultivars was 18 times higher than that in the tolerant cultivars and the sensitive soybean cultivars had 37% less yield than the tolerant cultivars. Yang and Blanchar (1993) identified 19 soybean cultivars as Cl-excluders from 60 lines with 12 from each maturity group (MG) from II to VI. In 2014,two additional soybean cultivars were identified from 257 lines (Zhang et al. 2014). Up to 151 germplasms, including Lee, Lee 68, and S-100, were identified as salt-tolerant,whereas 413 germplasms were designated as susceptible to salt (www.ars-grin.gov/npgs/searchgrin.html).

Variation in salt tolerance has also been observed in wild soybean. Two wild soybean accessions, BB52 and N23232, were collected from coastal and inland areas,and identified as tolerant under salt stress (Luo et al. 2005;Chen et al. 2013). Hamwieh and Xu (2008) reported a salttolerant wild soybean accession JWS156-1. PI483463, a wild soybean accession from America, was characterized as salt tolerant (Lee et al. 2009; Ha et al. 2013). Another wild soybean accession, W05, was characterized as having a high tolerance to salt (Qi et al. 2014). Do et al.(2016) determined that a total of 123 soybean accessions(117 cultivated soybean accessions and six wild soybean accessions) were salt tolerant based on hydroponic culture in the greenhouse. Xu et al. (2016) evaluated more than 600 soybean accessions, including wild soybean and cultivars of soybean, in a greenhouse condition. In these soybean accessions, several genotypes with high salt tolerance were identified, such as a Brazilian soybean cultivar FT-Abyara, a Chinese soybean cultivar Jindou 6,and a Japanese wild soybean accession JWS156-1. The high level of variation in soybean germplasm, including wild and cultivated soybean species, suggests that genetic improvement of salt tolerance is feasible.

2. Genetic studies on salt tolerance in soybean

The heredity of salt tolerance in soybean was previously analyzed as a quality trait using the classical genetics approach, and the gene symbols Ncl and ncl were proposed as the dominant for tolerance and the recessive for sensitivity,respectively (Abel 1969). Soybean roots of tolerant and sensitive plants absorb chloride ions in comparable amounts,but subsequent translocation of chloride higher in the plant is genetically controlled. Abel (1969) made crosses between parents with similar chloride accumulating capacity, and found that F2and F3progenies were similar to parents in chloride content. As a result, F2plants segregated from eight crosses in ratios of three chloride excluders to one includer. In tests of the F3progeny from these crosses, the F2excluder plants segregated in ratios of one excluder to two segregating,whereas the F2includer plants bred true. The progenies of BC1F1crosses segregated in a ratio of one excluder to one includer. These results suggested that chloride accumulation in the upper part of soybean is controlled by a single locus with exclusion being dominant (Ncl) and inclusion being recessive (ncl) (Abel 1969).

Shao et al. (1994) studied the inheritance of salt tolerance with crosses differing in salt-tolerant and -sensitive soybean varieties. Their results indicated that the F1, F2, and F3progenies were all salt tolerant when both parents were salttolerant, and that they were all sensitive when both parents were sensitive. When a salt-tolerant variety was crossed with a sensitive one, or vice versa, the F1was salt-tolerant;the F2segregated in a 3:1 ratio; the F3lines from salttolerant F2plants had a 1:2 segregating ratio; and lines from sensitive F2plants were all sensitive. In backcross progenies with sensitive varieties as the recurrent parents,the segregating ratio was always 1:1. Thus, the results indicated that the salt tolerance of soybean is controlled by one pair of dominant genes in the nucleus where salt tolerance is dominant to sensitiveness (Shao et al. 1994).The inheritance of salt tolerance in wild soybean wasstudied using the accession PI483463. Three soybean accession lines – PI483463, S-100, and Hutcheson were used to create crosses for testing salt tolerance inheritance and allelism of tolerance genes. F2plants from the cross of PI483463×Hutcheson segregated into a tolerant to sensitive ratio of 3:1. The F2:3lines segregated into a tolerant to segregating to sensitive ratio of 1:2:1. F2plants from the cross of PI483463×S-100 segregated into a tolerant to sensitive ratio of 15:1, indicating different genes from the two sources. Their results showed that Glycine soja line PI483463 had a single dominant gene for salt tolerance, which was different from the gene in G. max line S-100 (Lee et al. 2009). An F2:3population derived from a cross between Tiefeng 8 (tolerant) and 85-140 (sensitive)was also used to analyze the inheritance for salt tolerance in soybean. The F2:3population showed 1:2:1 segregation,indicating a single dominant gene for salt tolerance in soybean cultivar Tiefeng 8 (Guan et al. 2014a). Chen et al. (2011) investigated the inheritance of salt tolerance in the recombinant inbred line (RIL) population by the mixed major gene plus polygene inheritance model of quantitative traits, and showed that salt tolerance was controlled by three major genes in soybean (the F-3 model), and the heritability value of the major genes was 64.4%. It is well known that salt tolerance is a quantitative trait controlled by several genes in soybean, which is confirmed by the QTL mapping and GWAS results described below.

Table 1 Major salt tolerance germplasm accessions in soybean

3. QTL mapping associated with salt tolerance during seedling stage in soybean

The development of molecular markers and linkage/physical maps provide approaches to detect the quantitative trait loci (QTLs) associated with traits of interest in soybean.Recently, QTL mapping for salt tolerance was conducted during different soybean growth stages (but mainly focused on the seedling stage) by several research groups using different germplasms (Table 2). Lee et al. (2004) reported the first QTL mapping result for salt tolerance in soybean.A soybean F2:5population derived from the cross of S-100(salt-tolerant) and Tokyo (salt-sensitive) was used to detect the QTL associated with salt tolerance. First, RFLP markers were associated with the visual ratings of salt tolerance, and the target genes were mapped on linkage groups L and N.According to the information from RFLP, additional 32 SSR markers were selected to conduct further linkage analysis of the S-100×Tokyo F2:5population. As a result, a major QTL for salt tolerance was discovered between the SSR markers Sat_091 and Satt237 on linkage group N, accounting for 41 and 60% of the total genetic variation for salt tolerance in the field and greenhouse, respectively (Lee et al. 2004).

The RIL population with a molecular map consisting of 221 SSR markers was used to detect the QTLs associated with salt tolerance in soybean (Chen et al. 2008). Field and greenhouse experiments were conducted to evaluate 184 RILs derived from a cross between Kefeng 1 and Nannong 1138-2 for salt tolerance during the seedling stage.Subsequently, eight QTLs significantly associated with salt tolerance were identified using WinQTLCart software. A new QTL was identified between markers Sat_164 and Sat_358 on linkage group G in both the field and greenhouse experiments. Furthermore, a QTL named qppsN.1 was identified in the same location as a salt tolerance QTL previously reported in soybean (Lee et al. 2004).

Hamwieh and Xu (2008) used an F2population derived from a cross between Jackson (sensitive cultivar) and JWS156-1 (wild soybean accession) to detect salt tolerance QTL in hydroponic culture conditions. Results confirmed a major QTL associated with salt tolerance with a large dominant effect located on the soybean linkage group N,which accounted for 68.7% of the total phenotypic variance.The major salt tolerance QTL was further validated by three sets of salt-tolerant near-isogenic lines (NILs). NILs were developed by selfing the residual heterozygous soybean plants following marker-assisted selection using the SSR markers (Satt339, Satt237, GMES1100, Satt255, Sat_091,and Sat_304) around the salt-tolerant QTL located on linkage group N. The phenotypic evaluation of the NILs showed that the lines with the salt-tolerant QTL region had better growth phenotypes, compared to the lines without the salt-tolerant locus (Hamwieh et al. 2011). A subset of 66 iso-lines was developed and used for narrowing the mapping region of the salt-tolerant QTL located on linkage group N (chr. 03). The integration of recombination events and the salt response data indicated that the QTL is located in the region of an approximately 658 kb segment between SSR03_1335 at nucleotide 40505992 and SSR03_1359 at nucleotide 41164735 on chr. 03 (Ha et al. 2013). Guan et al.(2014a) mapped and validated a dominant salt tolerance gene located on chr. 03 in soybean cultivar Tiefeng 8.In their experiment, five new molecular markers were developed within the QTL mapping region. As a result, the salt tolerance QTL was mapped within 209 kb flanked by markers QS08064 and Barcsoyssr_3_1301 on chr. 03.

4. QTL mapping associated with salt tolerance during germination stage in soybean

It is well known that seed germination is critical for plant growth and life cycle completion (Wang et al. 2011). Seed germination is one of the plant growth stages that is highly susceptible to salt and is severely inhibited as salinity increases (Fredj et al. 2013). It is well established that soybean has different mechanisms for salt tolerance between germination and seedling stages (Pathan et al.2007; Phang et al. 2008).

Salt tolerance during seed germination is a critical determinant ofstable stand establishment under salt stress conditions. Therefore,identifying salt tolerance QTLs during the germination stage in soybean is crucial. Using a soybean RIL population, Kan et al.(2016) identified 11 salt tolerance QTLs during the germinationstage; these QTLs were located on six chromosomes for the three salt tolerance indices. These QTLs explained 4.49 to 25.94% of the phenotypic variation. The positive alleles of the 11 QTLs were confirmed from different parents of the RIL population. Further analyses were conducted with the 11 previously detected QTLs of salt tolerance using a natural population consisting of 196 soybean landraces. A total of 22 SSR loci associated with three salt tolerance indices were detected by association mapping. Interestingly, Kan et al. (2016) found that the SSR marker Sat_162 was closely linked to the co-localized QTLs at a site 792 811 bp from the gene Glyma08g12400.1, which was confirmed in response to salt stress during the germination stage in soybean. Zhang et al. (2014)analyzed the QTL for salt tolerance using epistatic association mapping during the germination stage in soybean. They detected a total of 83 QTL-by-environment interactions and one epistatic QTL using epistatic association mapping implemented with an empirical Bayes algorithm. Currently, the QTL information for salt tolerance at the soybean germination stage is limited by previously published reports. More work needs to be done to fully investigate QTLs conferring salt tolerance.

5. Fine mapping and positional cloning of the major salt tolerance QTL located on chr. 03

The major salt-tolerance QTL located on chr. 03 (linkage group N) was identified by several researchers using different soybean mapping populations. Subsequently, two research groups from China and Japan aimed to clone this major salt tolerance QTL to map salt tolerance QTL in soybean. Guan et al. (2014a) firstconstructed an RIL population, consisting of 367 RILs derived from a cross between Tiefeng 8 (salt-tolerant) and variety 85-140 (saltsensitive). Then they mapped this QTL between the two markers QS1101 and QS100011 and named it GmSALT3. Fine mapping of GmSALT3 was conducted using a new population. This new population consisting of 5 769 plants was derived from F5heterozygous lines between InDel markers QS1101 and QS100011. Then, GmSALT3 was mapped within a 17.5-kb region that had only one predicted gene: Glyma03g32900.Therefore, Glyma03g32900 was regarded as the candidate gene underlying the salt tolerance QTL GmSALT3 (Guan et al. 2014b).

Do et al. (2016) showed that the salt tolerance QTL(named Ncl) was mapped within a 58.8-kb region between the markers BARCSOYSSR_03_1342 and BARCSOYSSR_03_1338 on chr. 03. According to the reference soybean genome, there were seven predicted genes in this 58.8-kb region of the major salt tolerance QTL Ncl. Subsequently, a large segregation population consisting of 5 828 plants derived from F9residual heterozygous plants was used for screening recombination between BARCSOYSSR_03_1342 and BARCSOYSSR_03_1338.In total, 29 recombinants were identified with a homozygous genotype at one marker and a heterozygous genotype at the other marker. As a result, the salt tolerance QTL Ncl was mapped within a 16.6-kb region between the two markers SSR25.8 and CAPS42.4. Further analysis showed that there was only one gene, Glyma03g32900, within the 16.6 kb region. Furthermore, Qi et al. (2014) identified a novel ion transporter gene, GmCHX1, via a combination of de novo sequencing data and their previous germplasm resequencing data. The ion transporter gene GmCHX1 was confirmed as the causal locus of gene Glyma03g32900.Functional analysis of this salt locus was conducted by three research groups (designated GmSALT3, Ncl, and GmCHX1).

In addition, a 3.8-kb insertion of the Ty1/copia type retrotransposon fragment was identified to be responsible for the loss of function of the salt-tolerant gene Glyma03g23900(Guan et al. 2014b; Qi et al. 2014; Do et al. 2016; Kumawat et al. 2016). Abel (1969) first identified the dominant gene for salt tolerance using the classical genetics approach in soybean; he called it Ncl. Interestingly, based on the reference genome sequence information, the three genes- GmSALT3, GmCHX1, and Ncl- were located on the same locus as Glyma03g32900. Recent studies have indicated that the dominant gene Ncl described by Abel(1969) is the salt tolerance locus Glyma03g32900. To date, the salt tolerance gene Glyma03g32900 has been given three names (GmSALT3, GmCHX1, and Ncl), which is confusing to researchers. We propose that the gene be named Ncl, as first described by Abel (1969).

6. GWAS for salt tolerance in soybean

Benefitting from the advances of next-generation sequencing(NGS), genome-wide association studies (GWAS) were broadly used to detect the associated single nucleotide polymorphism (SNP) markers for many agronomy and resistance traits in soybean (Mamidi et al. 2011; Hao et al.2012; Hwang et al. 2014; Wen et al. 2014, 2015; Zhang et al.2015). Recently, GWAS was successfully performed to identify the SNP markers associated with salt tolerance in soybean.Kan et al. (2016) detected eight SNP-trait associations and 13 potential SNP-trait associations with salt tolerance during the seed germination stage by GWAS using a mixed linear model and TASSEL 4.0 software. Eight SNPs or potential SNPs were co-associated with salt tolerance in soybean.Based on the soybean genome database, nine candidate genes were located in or near the salt tolerance associated SNP marker region. Subsequently, five candidate genes Glyma08g12400.1, Glyma08g09730.1, Glyma18g47140.1,Glyma09g00460.1, and Glyma09g00490.3 were validated in response to salt stress during the soybean germinationstage. Based on phenotype and the SoySNP50K BeadChip database of 106 soybean lines, Patil et al. (2016) performed GWAS for salt tolerance using an expedited single-locus mixed model. The SNPs identified by GWAS pin-pointed a single and highly significant association for salt tolerance on chr. 03. In this associated region, GWAS found that the soybean salt tolerance gene Glyma03g32900 overlapped the associated locus (Patil et al. 2016). The whole-genome resequencing (WGRS) data of the 106 soybean lines were also used to detect the SNPs associated with salt tolerance during the seedling stage. Finally, a significant SNP within the salt tolerance gene Glyma03g32900 was identified and determined to explain up to 63% of the phenotypic variation.Genome-wide analysis showed that natural variation associated with the Glyma03g32900 gene has a major impact on salt tolerance in soybean. A total of 283 worldwide soybean germplasm lines and the SoySNP50K BeadChip database were used to perform the GWAS for salt tolerance(Zeng et al. 2017). A total of 45 SNPs representing nine genomic regions on nine chromosomes were significantly associated with salt tolerance based on GWAS analysis.Most of the SNPs significantly associated with salt tolerance were located within or near the major salt tolerance gene Glyma03g32900 on chr. 03. Moreover, seven putative novel QTLs represented by significant SNPs were identified for salt tolerance in soybean. GWAS has been a powerful tool for identifying all candidate loci conferring target traits at the whole genome level. For soybean salinity traits, more related QTLs/QTNs (quantitative trait nucleotide) contributing to salt tolerance will be identified and can be used to improve salinity through genomic selection strategies.

7. Future perspectives

At present, only experienced soybean researchers can screen for salt tolerance phenotyping. This screening involves a visual scoring method that can be used to screen a large population, however, it is inefficient and thus time-consuming. To improve soybean salinity tolerance,the next step is introducing tolerance genes such as Ncl(Glyma03g32900) into elite cultivars by molecular markerassisted genomic selection, and transgenic technologies;these techniques will produce a large population for selection of candidate plants. Then, the key question is to select the target soybean lines from the large population.In order to do this, new phenotyping technology needs to be developed, and this new phenotyping method will need to be high throughput and precise and cannot rely entirely on an experienced investigator of soybean salinity. High throughput phenotyping technology such as image-based phenotyping is needed to explore the genetic variation for salinity tolerance in soybean. Salinity response, measured as the effect of salt on growth rate at different developmental times, could explain genotypic variation for salinity tolerance in soybean. We have constructed an image-based phenotyping platform to enable quantitative, non-destructive assessment of the responses of soybean to salinity during the seedling stage under salt stress conditions. This highthroughput salinity phenotyping technology can be used to identify salt tolerance germplasms early on and can be applied as a selection tool in soybean breeding programs for salt tolerance.

Genomic resources can help to clone new QTLs and genomic selection can help in selection for salt tolerance,especially for minor and multiple genes. Genomic selection technology can be used to improve salt tolerance in soybean.Based on the developments of genome sequencing technology and efficient SNP genotyping technology, genomic selection has been used to explore the prediction accuracy of plant height and yield per plant in soybean (Ma et al. 2016).In practice, genomic selection is conducted in a soybean population (validation set) that is different from the reference soybean population (training set) in which the SNP marker effects were estimated based on kinship. After obtaining phenotypic information, SNP marker scores, and kinship or pedigree information, the effects for each SNP marker will be estimated in the training soybean set using certain statistical methods. Then, the genomic breeding value or genetic values of new soybean genotypes are predicted based only on each SNP marker effect.

CRISPR/Cas9 System is a powerful and precise tool for genome editing (Liu et al. 2017). The CRISPR/Cas9 Genome Editing System requires the co-expression of a Cas9 protein with a guide RNA vector expressed from the human U6 polymerase III promoter. A protospacer-adjacent motif (PAM, with the sequence NGG) occurs at the 3´ end of the Cas9 protein. After the guide RNA recognizes this target sequence, Cas9 unwinds the DNA to cleave both strands.Finally, the functional cassette, which is synthesized using a rescue donor vector, is inserted into the cleaved DNA.The precise genome modification by the CRISPR/Cas9 System avoids backcrossing, leading to a shorter breeding cycle than conventional breeding methods. In the future,a highly efficient transformation system for soybean will enable CRISPR/Cas9 to be used directly for improving salt tolerance in elite soybean varieties.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (31401407).