Soil salinization due to excessive application of agrochemicals and poor irrigation practices seriously reduces crop yield, and is an increasing threat to the sustainable development of the world food supply (Munns and Tester, 2008; Zhu, 2016). Soybean, which is rich in oil and protein, is not only used as food but also as an industrial raw material for animal feed. As a moderately salt-tolerant crop plant, soybean yield can be reduced by 40% under salt stress (Papiernik et al., 2005). Salt stress can inhibit soybean seed germination and seedling growth. It is therefore necessary to identify salt tolerance loci for future molecular breeding of salt-tolerant soybean.

Soybean germplasm vary in salt tolerance, which provides an opportunity to study the genetic control of this trait. Early research showed that F₂ populations of chloride includer×excluder varieties segregated in ratios of three non-necrotic plants (low chloride content) to one necrotic plant (high chloride content), indicating a single gene governs the inheritance of salt tolerance. The gene symbol Ncl was proposed for the dominant Cl⁻ excluder allele in soybean (Abel, 1969). Decades later, a codominant random amplified polymorphic DNA (RAPD) marker was reported to be tightly linked with the salt-tolerance gene (Guo et al., 2000). A great deal of research has focused on identifying salt tolerance quantitative trait loci (QTLs)/genes in recent decades. The major salt tolerance QTL was first mapped on linkage group N (LG N) in a segregating population of S-100 (salt tolerant)×Tokyo (salt sensitive; Lee et al., 2004). The salt tolerance locus on LG N was proven to be conserved in both cultivated and wild soybean (Lee et al., 2004; Hamwieh and Xu, 2008; Tuyen et al., 2010; Hamwieh et al., 2011; Ha et al., 2013; Guan et al., 2014a). The candidate gene Glyma03g32900 (Wm82.a1.v1.1) associated with soybean salt tolerance in this interval was identified by resequencing (GmCHX1) and map-based cloning (GmSALT3; Guan et al., 2014a; Qi et al., 2014), and was also designated as the chloride excluder locus Ncl (Do et al., 2016). GmSALT3 encodes an endoplasmic reticulum-localized protein in the Cation/Proton Antiporter (CPA2) family of transporters, and is dominantly expressed in root phloem and xylem (Guan et al., 2014b). Through a combination of physiological and genetic approaches, we demonstrated that GmSALT3 promoted soybean salt tolerance via restricting Na⁺ loading and conducting Cl⁻ retranslocation from the shoot, thus maintaining high salt tolerance under saline conditions (Liu et al., 2016; Qu et al., 2021). Sequence insertion and deletion cause premature stop codon in some GmSALT3 alleles, and amino acid changes or variation at an intron splice donor site in other alleles contribute to functional loss of GmSALT3 (Guan et al., 2014b). Other structural variation or coding sequence changes have been observed in diverse soybean germplasm, especially wild soybeans, and kompetitive allele specific PCR (KASP) and PCR-based markers have been developed to screen salt-tolerant soybean accessions (Patil et al., 2016; Lee et al., 2018). Sequencing, SNP-based KASP, and PCR-based electrophoresis assays show a high prediction rate (>95%) in identification of salt-tolerant genotypes, indicating that Glyma03g32900 is necessary for salt tolerance at the seedling stage (Guan et al., 2014a; Do et al., 2016; Lee et al., 2018). However, we found that a Chinese soybean landrace, Peking, contained a salt-tolerant allele of GmSALT3 but was salt sensitive (Guan et al., 2014b). The result was confirmed by Patil et al. (2016), who also found another nine lines that did not show the expected correlation between haplotype and phenotype. This raises the possibility that other genes or modifiers may affect salt tolerance in Peking. Salt tolerance QTLs in soybean located on several chromosomes other than Chr. 03 have been identified through genetic analysis and genome-wide association studies (GWAS; Chen et al., 2008; Zeng et al., 2017b). Compared to wild soybean, the genetic diversity of cultivated soybean decreased dramatically due to bottlenecks during domestication and human selection (Lam et al., 2010). Recently, a salt tolerance-related gene, GsERD15B, was cloned from a panel of 182 wild soybean lines using GWAS, and a 7-bp insertion in the promoter region appeared to be the functional polymorphism (Jin et al., 2021). An inositol polyphosphate 5-phosphatase gene (Gs5PTase8) from wild soybean was shown to increase salt tolerance in plants (Jia et al., 2020). Therefore, wild soybean germplasm are valuable resources for mining novel salt tolerance genes. We here generated two segregating F_2:3 populations using salt-tolerant wild soybean NY36-87 as the male parent and two salt-sensitive cultivars (Zhonghuang39 and Peking) as the female parents. This was intended to (1) identify salt tolerance loci/genes in NY36-87, and (2) explore the genetic factors underlying salt sensitivity in the soybean landrace Peking, which contains a tolerant haplotype of GmSALT3.

Over the past half century, scientists have focused on genetic analysis and begun to identify genes that control different traits associated with salt tolerance (Abel and Mackenzie, 1964; Abel, 1969; Shao et al., 1994). Because the traits used in different studies have varied, most salt tolerance gene mapping was performed as QTL using bi-parental populations, natural populations, or a combination of both. Leaf chloride content was considered to be an indicator of soybean response to salinity, and high chloride toxicity to be related to leaf necrosis (Abel and Mackenzie, 1964). Thus, leaf chloride accumulation and necrosis were used to study the genetic factors of salt tolerance in soybean, and the two traits were controlled by a single dominant gene in different genetic backgrounds (Abel, 1969; Shao et al., 1994). Using the salt tolerance rating score that classifies based on leaf scorching symptoms, a major QTL that could explain most of the phenotype variation in greenhouse and field environments was mapped on soybean LG N (Chr. 03; Lee et al., 2004). Nine QTLs were identified in a recombinant inbred line (RIL) population when the traits measured were tolerance rating, survival time, and survival ratio. The locus qppsN.1, tightly linked with Satt237, was located in the same region as the one mapped on LG N (Chen et al., 2008). The major QTL on Chr. 03 (LG N) was independently proven to be a conserved locus (Hamwieh and Xu, 2008; Hamwieh et al., 2011; Ha et al., 2013; Guan et al., 2014a). These studies led to cloning of the candidate gene, Glyma03g32900, under this dominant locus (Qi et al., 2014; Guan et al., 2014b; Do et al., 2016). Glyma03g32900 affects Na⁺ and Cl⁻ accumulation in soybean shoot under salt stress (Do et al., 2016; Liu et al., 2016). Recently, the association of Glyma03g32900 with Na⁺ and Cl⁻ accumulation in leaf was verified by genetic mapping of QTLs related to leaf Na⁺ and Cl⁻ contents (Do et al., 2018). In this study, using a population derived from the salt-sensitive cultivar Zhonghuang39 and the salt-tolerant wild soybean NY36-87, we mapped a dominant tolerance gene on Chr. 03 to a 98-kb interval, which included GmSALT3 (Glyma03g32900; Figure 2A). This indicated that the tolerance gene Glyma03g32900 is conserved in salt-tolerant germplasm.

The availability of high-throughput genome resequencing and SNP arrays promoted the identification of salt tolerance-related genes from diverse soybean germplasm (Zeng et al., 2017a; Jin et al., 2021). Two tolerant QTLs were mapped on Chr. 13 and Chr. 15, with Glyma.13g161800 and Glyma.15g091600 proposed as the candidates (Zeng et al., 2017b). A novel dominant locus related to leaf sodium content was mapped on Chr. 13, and the functional allele was from salt-sensitive parent Williams 82 (Do et al., 2018). These results showed that novel loci underlie salt tolerance in diverse soybean germplasm, even in salt-sensitive accessions. A similar result was reported related to soybean seed weight; the PP2C-1 allele from the small-seeded wild soybean, but not from big-seeded cultivars, is responsible for seed weight promotion (Lu et al., 2017). In association analysis of Glyma03g32900 functional markers with salt tolerance, several soybean accessions that may carry novel salt tolerance loci other than Glyma03g32900 were found (Guan et al., 2014b; Patil et al., 2016; Lee et al., 2018). Peking is one such accession that is salt sensitive but contains the tolerant allele of GmSALT3. In the F₂ population of Peking×NY36-87, the segregation of salt-tolerant and salt-sensitive genotypes fit a simple Mendelian ratio, suggesting that a dominant gene confers the salt tolerance in NY36-87. A novel salt tolerance-related locus on Chr. 18, which we called GmSALT18, was mapped to a 241-kb region (Figure 3B). There are 29 gene annotations based on the assembly of Williams 82 (Glyma.Wm82.a2.v1).² Three genes, including Glyma.18g026200 and Glyma.18g026500, which may relate to salt stress (Gong et al., 2001; Bouchabke-Coussa et al., 2008), and a predicted K⁺/H⁺-antiporter, Glyma.18g027900, were located in this region. In addition, potential genomic structure variations between Glyma.18g026700 and Glyma.18g026900 were observed in assemblies of soybean ZH13, wild soybean W05 and PI 483463 (Supplementary Figure 2). Given the facts that variation in GmSALT18 of Peking may be a minor allele, and possible genomic structure variation, further study of fine mapping and genome sequencing of Peking and NY36-87 could be useful for successful identification of the candidate gene.

The discrepancy between genotype and phenotype for a particular gene, such as GmSALT3, may be due to the phenotype sorting method making it difficult to distinguish between moderately tolerant or sensitive germplasm (Patil et al., 2016). In this study, we classified salt tolerance as a binary (normal growth of plants=salt tolerant, plant death=salt sensitive) rather than rating salt tolerance on a 1–5 scale (Figure 1); this provided a more exact phenotype for gene mapping. Using a RIL population of Kefeng No. 1 (salt sensitive)×Nannong1138-2 (salt tolerant), a major QTL (qtrG.1) for salt tolerance ratings was previously located on Chr. 18 flanked by Sat_164 and Sat_358 (Chen et al., 2008), which is 9Mb from GmSALT18. This suggests that GmSALT18 may represent a novel salt tolerance locus in wild soybean. Considering that the expression patterns of GmSALT3 in Peking and NY36-87 were similar (Figure 5B), it is unlikely that GmSALT18 plays a role in regulation of GmSALT3. Peking accumulated more Na⁺ and K⁺ in leaves compared to salt tolerant germplasm (Figure 4), as was observed in salt-sensitive cultivars 85–140 and NIL-S, which contain the sensitive GmSALT3 allele (Guan et al., 2014b; Qu et al., 2021). The mechanism underlying Na⁺ accumulation in Peking thus requires further study.

In summary, through mapping of salt tolerance genes using two F_2:3 populations, we identified the salt tolerance gene GmSALT3 in the population Zhonghuang39×NY36-87. In the population Peking×NY36-87, a novel locus, GmSALT18, was found to be responsible for the difference in salinity tolerance between Peking and NY36-87. Further research is needed to clone the corresponding gene underlying GmSALT18 and demonstrate the genetic relationship between GmSALT3 and GmSALT18.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

RG, RC, and LQ designed the research. XG, JJ, and LY performed the salt tolerance phenotype evaluation. LY and XG performed salt stress and gene expression. XG performed fine mapping of the salt tolerance genes and dada analysis. XG, RG, and LQ wrote the manuscript. All authors contributed to the article and approved the submitted version.

This research was funded by the Natural Science Foundation of China (31830066) and Central Public-interest Scientific Institution Basal Research Fund (S2021ZD02).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.