Linkage and association mapping of wild soybean (Glycine soja)seeds germinating under salt stress

SHI Mei-qi ,LIAO Xi-liang ,YE Qian ,ZHANG Wei ,LI Ya-kai ,Javaid Akhter BHAT ,KAN Gui-zhen,YU De-yue,2

1 National Key Laboratory of Crop Genetics and Germplasm Enhancement,National Center for Soybean Improvement,Nanjing Agricultural University,Nanjing 210095,P.R.China

2 Innovative Center of Molecular Genetics and Evolution,School of Life Sciences,Guangzhou University,Guangzhou 510405,P.R.China

Abstract Salinity threatens soybean germination,growth and production.The germination stage is a key period in the life of soybean.Wild soybean contains many genes related to stress resistance that are valuable resources for the genetic improvement of soybean.To identify the genetic loci of wild soybean that are active during seed germination under salt stress,two populations,a soybean interspecific hybrid population comprising 142 lines and a natural population comprising 121 wild soybean accessions,were screened for three germination-related traits in this study.By using single-nucleotide polymorphism (SNP) markers with three salt tolerance indices,25 quantitative trait loci (QTLs),21 significant SNPs (-log10(P)≥4.0) and 24 potential SNPs (3.5＜-log10(P)＜4.0) were detected by linkage mapping and a genome-wide association study (GWAS) in two environments.The key genetic region was identified based on these SNPs and QTLs.According to the gene functional annotations of the W05 genome and salt-induced gene expression qRT-PCR analysis,GsAKR1 was selected as a candidate gene that responded to salt stress at the germination stage in the wild soybean.These results could contribute to determining the genetic networks of salt tolerance in wild soybean and will be helpful for molecular marker-assisted selection in the breeding of salt-tolerant soybean.

Keywords: salt tolerance,wild soybean,QTLs,GWAS,GsAKR1

1.Introduction

Soil salinity is one of the major abiotic stresses that restricts plant growth and reduces agricultural productivity worldwide (Acquaah 2015).Approximately 60 million ha of the world’s agricultural land area is affected by salt as a result of improper irrigation and fertilization,and the situation is likely to become progressively worse by 2050 (Yamaguchi and Blumwald 2005;Waniet al.2016;Doet al.2019;Xiaet al.2019;Zhanget al.2020).Soybean [Glycinemax(L.) Merr.],which is rich in protein and oil,has long been an important source of nutrients for humans and animals.Previous studies have shown that the agronomic traits of soybean,including seed germination,plant growth,biomass,and final yield after the soybean growth cycle,are severely affected by salt stress (Pathanet al.2007;Phanget al.2008).Salt tolerance is an inheritable quantitative trait that is affected by many genetic and nongenetic factors (Pathanet al.2007;DeRose-Wilson and Gaut 2011;Ashraf and Foolad 2013;Longet al.2013).

Seed germination is a critical stage during the plant growth cycle and plays an important role in ensuring the survival of new seedlings and increasing yields in saline soils (Wanget al.2011).Recently,much progress has been made in understanding soybean salt tolerance at the seedling stagevialinkage mapping,and some major quantitative trait loci (QTLs) for salt tolerance have been identified in many studies (Guanet al.2014;Qiet al.2014;Wanget al.2016;Zhanget al.2019;Jiaet al.2020).Several QTLs underlying salt stress at the soybean germination stage have been identifiedvialinkage and association mapping (Qiuet al.2011;Zhanget al.2014,2019;Kanet al.2015,2016).For example,Zhanget al.(2019) used 184 recombinant inbred lines(RILs) to map salt tolerance-related QTLs and identified the candidate geneGmCDF1by combining genomewide association study results.Fewer studies have been reported on the genetic mechanisms underlying salt tolerance at the germination stage than at the seedling stage (Kanet al.2015;Zhanget al.2019).Therefore,determining the response mechanism of salt tolerance at the germination stage and applying molecular markerassisted techniques will be very helpful for the genetic improvement of salt-tolerant soybean.

The plant materials used in these previous studies mostly comprised cultivated soybean materials instead of wild soybean materials.Compared with its wild counterpart,cultivated soybean has much lower genetic diversity because of bottlenecks and human selection(Hytenet al.2006).Wild soybean is the immediate progenitor of domesticated soybean,and without reproductive isolation,it is actually interfertile (Zhuang 1999;Andersson and Vicente 2010).Generally,wild soybean exhibits high allelic diversity.Interesting agronomic traits from wild soybean,such as salt tolerance,have been effectively applied to improve cultivated soybean genotypes (Leeet al.2009;Andersson and Vicente 2010;Tuyenet al.2010;Qiet al.2014).Recently,more studies on wild soybean have been conducted,and they all have achieved good results.A novel salt tolerance gene,Glysoja01g005509,was identified in wild soybeanviawhole-genome sequencing involving QTL mapping of a core panel of 96 RILs derived from reciprocal crosses of W05 and C08;and at the same time,a recombinant bin map was constructed (Qiet al.2014).In addition,the soybean salt tolerance alleleNc12was identified in wild soybean PI483463 (Leeet al.2009),and a significant QTL controlling alkalinity tolerance was detected in linkage group (LG) D2 by using RIL and residual heterozygous line(RHL) populations derived from crosses between Jackson and wild soybean JWS156-1 (Tuyenet al.2010,2013).Shenet al.(2018) found an R2R3-MYB transcription factor from wild soybean,GsMYB15,that increasedHelicoverpa armigeraresistance and salt tolerance in transgenicArabidopsis(Shenet al.2018).Overexpression ofGsPRX9(Jinet al.2019) orGsERD15B(Jinet al.2020)enhanced salt tolerance in soybean.

In the present study,we combined linkage analysis with a genome-wide association study (GWAS),and the major loci and genes governing salt tolerance in wild soybean at the seed germination stage were identifiedviaassociation and linkage mapping.These loci and genes can be used to genetically improve soybean salt tolerance.

2.Materials and methods

2.1.Plant materials and salt tolerance evaluation

A segregating soybean population comprising 142 F2:8RILs derived from a cross between wild soybean (Jiangpu wild soybean-5) and cultivated soybean (Nannong 06-17) was used in this study.The RILs were subsequently used to construct a high-density genetic map and a QTL map for salt tolerance at the germination stage.A total of 121 wild soybean accessions for the GWAS were collected from southern China to northeastern China.The field trials were conducted at the Jiangpu Agronomic Experimental Station of Nanjing Agricultural University(32°12′N,118°37′48′′E),Nanjing,China,in the summers of 2013,2014 and 2015.The plant materials were planted under a completely randomized block design and replicated three times,with 100 cm×100 cm hill plots and five plants per plot.The RIL population seeds from 2014 to 2015 and the wild soybean seeds from 2013 to 2014 were collected at maturity to evaluate the phenotypes related to germination under salt stress.

Before germination,the seeds were first dried.After treatment with 15 mL NaCl solution (150 mmol L-1) or pure water (0 mmol L-1),50 uniform seeds were then placed in sterilized Petri dishes (9 cm in diameter) containing two sheets of filter paper.The seeds were subsequently incubated in the dark at (25±1)°C.After 24 h,we counted the number of germinated seeds with the radicle and plumule lengths equivalent to the seed length or half of the seed length,respectively.We subsequently rinsed the seeds and placed them into new dishes containing filter paper before adding either 5 mL NaCl solution(150 mmol L-1) or pure water (0 mmol L-1).These experimental procedures were performed daily for the next 6 days.Three replications were included,and the mean values were used for data analysis (Zhanget al.2019).

The three germination traits evaluated were the germination index (GI=Σ(Gt/Dt),where Dt represents t days of seed germination and Gt represents the cumulative number of germinated seeds on day t),germination rate (GR (%)=(Number of germinated seeds/Total seed number used in the test)×100) (Wanget al.2011),and germination potential (GP(%)=(Number of germinated seeds on day 3/Total number of experimental seeds)×100) (Zhanget al.2019).Salt tolerance (ST) was defined as the ratios of the three germination-related traits(GI,GR and GP) under salt conditions to the same traits under no-salt conditions (Longet al.2013).

2.2.Phenotypic data analysis

Statistical analysis of all the phenotypic data was performed with SAS/STAT 9.4 Software (SAS Institute Inc.,Cary,NC,USA).Descriptive statistics were determined,and correlation analyses were performed using the mean values of all the phenotypic data from the traits of 121 wild soybean accessions and 142 F2:8RILs across two environments.Analysis of variance (ANOVA)was performed for all the traitsviathe PROC GLM procedure of SAS,and Pearson’s correlations between the traits were calculated from the mean valuesviathe PROC CORR procedure of SAS.Theh2of salt tolerancerelated traits was estimated by fitting the following formula in the R software package “l(fā)me4” (Merket al.2012):whereis the genetic variance,is the interaction of genotype with locations,is error variance,ris the number of replications,andnis the number of locations.The estimates ofandwere obtained from an analysis of variance (ANOVA)assuming the locations were random.

2.3.Genetic map assembly and QTL mapping

Single-nucleotide polymorphism (SNP) genotyping was performed at Michigan State University on an Illumina platform (Illumina,Inc.,San Diego,CA,USA).The Illumina Infinium SoySNP6K BeadChip genotyping array was used to construct the genetic linkage map.The SNPbased map was constructed with JoinMap 4.1 (Van Ooijen 2011) using a logarithm of odds (LOD) score of 3.0 to 10 to confirm the order of markers,and the recombination values were converted to genetic distances (cM) using the Kosambi mapping function (Akondet al.2013).

QTL mapping and effect estimation were performedviaWinQTL Cart v2.5 (http://statgen.ncsu.edu/qtlcart/),with model 6 used for composite interval mapping (CIM).The control marker,window size and walk speed were set to 5,10 and 1 cM,respectively.An optimum threshold was obtained when the permutation times and significance were set at 1 000 and 0.05,respectively (Churchill and Doerge 1994).Significant QTLs were identified according to LOD peaks that exceeded the optimum threshold,while potential QTLs were defined as QTLs associated with LOD peaks that did not exceed the optimum threshold but were higher than a threshold of 2.0 (Kinget al.2013).Confidence intervals were set as the map interval corresponding to a“LOD peak”-1 decline on both sides of the LOD peak.

2.4.SNP genotyping and association analysis

A total of239 658 SNPs with a minor allele frequency(MAF)＞5% and a missing rate＜20% from the NJAU 355K SoySNP array were used in a GWAS of salt tolerance traits for the 121 wild soybean accessions.The linkage disequilibrium (LD) decay rate,defined as the chromosomal distance where the LD decays to half of its maximum value,was 80 kb in the 121 wild soybean accessions.GWAS analysis was performed using a compressed mixed linear model (Zhanget al.2010).All the work was conducted with the GAPIT version 2.0 package (Lipkaet al.2012).We considered -log10P≥4.0 as the significance threshold(Yanget al.2014),and SNPs significantly associated with salt tolerance were identified according to this threshold.Additionally,SNPs of interest that were associated with salt tolerance indices (3.5＜-log10P＜4.0) were defined as potential SNPs (Kanet al.2015,2016).

2.5.Quantitative real-time PCR (qRT-PCR) analysis

After imbibition for 24,48,72,and 96 h,the germinated seeds of the wild soybean extreme materials (Njau-W093 and Njau-W101) were used as cDNA templates,and the constitutively expressed soybean geneTubulin(GenBank accession number: AY907703) was used as the reference gene to determine the expression levels of the two candidate genes.The primers used for qRT-PCR of the nine genes are shown in Table 1.An ABI 7500 system was used for qRT-PCR,according to the instructions of the RealUniversal Colour PreMix (SYBR Green) (FP201;TIANGEN BIOTECH,Beijing,China).

Table 1 Sequences of the specific primers for qRT-PCR

The PCR program was as follows: 50 cycles of predenaturation at 95°C for 15 min,denaturation at 95°C for 10 s,and annealing and extension at 60°C for 32 s.

The amplification curve was analyzed after PCR,and the relative expression of the target gene was calculated based on themethod (Livak and Schmittgen 2001),where ΔΔCt=(CtTarget-CtTubulin)genotype-(CtTarget-CtTubulin)calibrator.

2.6.DNA extraction and whole-genome resequencing (WGRS)

The genomic DNA of each individual was extracted from soybean seeds using the TIANGEN Plant Tissue DNA kit (TIANGEN DP320).After the DNA samples were delivered,a quality control test was carried out on the specimens using nanodrop for the DNA quantity estimation,and the qualified DNA (＞3 μg;concentration ＞ 30 ng μL-1;OD260/OD280=1.80-2.00) was used for further study.

Illumina paired-end sequencing (PE150) was used for the WGRS of soybean genotypes.After quantification by TBS380,paired-end libraries of the two lines were sequenced by Shanghai Biozeron Biothchology Co.,Ltd.(Shanghai,China) with the Illumina HiSeq PE 2×151 bp read length.The soybean reference genomes Gmax_W82 and Gmax_ZH13 were obtained from NCBI,and the gene annotation information (including functional annotation and gene family information) was downloaded from SoyBase(https://soybase.org/data/public/Glycine_max/).

2.7.Variation detection and annotation

The raw paired end reads were trimmed and quality controlled by Trimmomatic with default parameters(http://www.usadellab.org/cms/uploads/supplementary/Trimmomatic).The high-quality sequencing reads were aligned to the soybean reference genome sequence using BWA Software (http://bio-bwa.sourceforge.net/) with “bwa mem” mode.After removing PCR-duplication reads with SAMtools Software (http://samtools.sourceforge.net/),the sequencing depth and coverage were calculated based on the alignments by custom Perl scripts.A valid BAM file was used to detect SNPs and short InDels by the GATK“HaplotypeCaller” function (http://www.broadinstitute.org/gatk/).The variant call format (VCF) files were then generated by quality filtering (Variant Filtration with parameters: QD＜2.0||FS＞60.0||MQ＜40.0||SOR＞10.0),and the VCF files were filtered with VCFtools (version 0.1.11;parameters: --minQ 20 --minDP 4).When the gene annotation file of the reference genome was provided,the annotation of detected variations was performed by ANNOVAR (http://www.openbioinformatics.org/annovar/),including SNP (synonymous or nonsynonymous mutations of SNPs) and InDel.Structural variations (SVs) were identified by BreakDancer (http://breakdancer.sourceforge.net)

3.Results

3.1.Phenotypic variation and correlation analysis

For the three salt tolerance indices: i.e.,the ratio of the germination index under salt conditions to the germination index under no salt conditions (ST-GI),the ratio of the germination potential under salt conditions to the germination potential under no salt conditions (ST-GP)and the ratio of the germination rate under salt conditions to the germination rate under no salt conditions (ST-GR),the means,standard deviations,ranges and heritability(h2) of the 142 RILs and 121 wild soybean accessions are shown in Table 2.The mean values of ST-GI,STGR and ST-GP in the RIL population were 0.44,0.58 and 0.44,and their values varied from 0.03 to 0.97,0.05 to 1.00,and 0.00 to 1.07,respectively;while in the natural population,the means were 0.33,0.44 and 0.32,and the values ranged from 0.03 to 0.80,0.04 to 0.121,and 0.00 to 1.02,respectively.ANOVA indicated a significant difference (P＜0.001) between the three salt tolerance indices for each individual material of the two populations.In addition,significant (P＜0.001 andP＜0.01) genetic variations for ST-GI and ST-GR were found in the two populations across the two environments.By contrast,no difference was found for ST-GP in either population across the two environments.

Table 2 Descriptive statistics,ANOVA and heritability (h2) of salt tolerance indices based on the means of the traits in 142 recombinant inbred lines (RILs) and 121 wild soybean accessions

In RILs,h2values of the three salt tolerance index traits (ST-GI,ST-GR and ST-GP) were 94.81,93.50 and95.51%,respectively (vs.72.00,78.52 and 78.59% in wild soybean).Theh2values of the three salt tolerance indices in cultivated soybean were higher than those in wild soybean,indicating that salt tolerance-related traits in wild soybean were more sensitive to the environment than those in cultivated soybean.

The phenotypic segregation for ST-GI,ST-GR and ST-GP in the two populations approximately fit a normal continuous distribution (Fig.1),indicating these three salt tolerance indices are controlled by multiple genes.Pearson’s correlations of the salt tolerance indices were analyzed based on the means of the two populations(Table 3).Significant positive correlations (P＜0.001) were found among the three salt indices (ST-GI,ST-GR and ST-GP).In addition,the correlations among the three salt indices (ST-GI,ST-GR and ST-GP) were similar for the two populations.

Fig.1 Frequency distributions of three salt tolerance indices (ST-GI,salt tolerance (ST) is defined as the ratio of germination index under salt conditions and germination index under no salt conditions;ST-GR,ST is defined as the ratio of germination rate under salt conditions and germination rate under no salt conditions;ST-GP,ST is defined as the ratio of germination potential under salt conditions and germination potential under no salt conditions) of a wild soybean natural population (A) and a soybean interspecific recombinant inbred lines (RILs) population (B) based on the means of the traits.

Table 3 Phenotypic correlations between three salt tolerance indices based on the means of the traits in 142 recombinant inbred lines (RILs) and 121 wild soybean accessions1)

3.2.Genetic map construction

A total of5 361 SNPs were produced in the RIL population.A genetic map was constructed using JoinMap 4.1 Software based on deleting missing SNP data plus heterozygosity and no proper segregation among the RILs.A total of 1 018 SNP markers were distributed across the 20 LGs (Table 4).The basic information of the LGs is presented in Appendix A.The map covers approximately 2 177.33 cM,and the average distance of each marker is 2.14 cM.The lengths of the LGs ranged from 40.05 cM (H) to 184.34 cM (J),the number of SNP markers on each LG ranged from 7 to 103,and the average number was 50.9.Thus,the number and density of SNP markers in the LGs were not uniform.

Table 4 Distribution of SNPs and their properties on the linkage groups (LGs) and chromosomes (Chr.) of soybean based on the ‘Jiangpu wild soybean-5’ by ‘Nannong 06-17’ recombinant inbred lines (RILs) population genetic map

3.3.QTLs for the three salt tolerance indices

The QTLs were correlated with salt tolerance at the germination stage across three different environments:2014,2015 and both years combined.Thus,a total of 23 QTLs,including seven for ST-GI,eight for ST-GR,and eight for ST-GP,were mapped over seven chromosomes.These 23 QTLs,with LOD values ranging from 2.08 to 6.32,explained 5.00 to 16.94% of the phenotypic variation.Among all the QTLs,the additive effect of nine QTLs was positive (all of which were from Jiangpu wild soybean-5),while the additive effect of the others was negative (from Nannong 06-17).A summary of all the detected QTLs is presented in Table 5.

Table 5 QTLs for salt tolerance indices based on the mean traits of three replications in 142 recombinant inbred lines (RILs)

Table 5 (Continued from preceding page)

For the three salt tolerance indices,eight QTLs (two for ST-GI,three for ST-GR,and three for ST-GP) were detected in 2014.One of the two QTLs for ST-GI,one of the three QTLs for ST-GR and two of the three QTLs for ST-GP were significantly associated with salt tolerance.In 2015,11 QTLs (three for ST-GI,three for ST-GR,and five for ST-GP) were detected in the RIL population.One of the three QTLs for ST-GI and three of the five QTLs for ST-GP were significantly associated with salt tolerance.Based on the average data for the RIL population across the two environments,a total of nine QTLs were identified:three for ST-GI,three for ST-GR,and three for ST-GP.One of the three QTLs for ST-GI,one of the three QTLs for ST-GR and two of the three QTLs for ST-GP were significantly associated with salt tolerance.

In addition,five QTLs were detected across the two environments:qST-GI-10,qST-GR-11-1,qSTGP-8,qSTGP-9,andqST-GP-10-1,and several markers that control two or more salt tolerance traits were identified.For example,three pairs of QTLs (qST-GI-9andqSTGP-9,qST-GI-10andqST-GP-10,as well asqST-GR-11andqST-GP-11) and two sets of three QTLs (qST-GI-5,qST-GR-5andqST-GP-5as well asqST-GI-8,qST-GR-8,andqST-GP-8) were colocalized.

Some QTLs related to salt tolerance were located on chromosomes 5,10 and 11,and showed largely overlapping confidence intervals (CIs) that were considered as the same QTLs.These QTLs included the following sets:qSTGI-5-1,qSTGI-5-2andqSTGI-5-3;qSTGR-5-1andqSTGR-5-2;qSTGR-11-1andqSTGR-11-2;qSTGP-5-1,qSTGP-5-2andqSTGP-5-3;andqSTGP-10-1andqSTGP-10-2.

3.4.SNPs associated with salt tolerance traits identified by GWAS

A total of 21 significant SNPs (-log10P≥4.0) and 24 potential SNPs (3.5＜-log10P＜4.0) were identified in the natural wild soybean population across the three environments (Table 6).These SNPs,which showed phenotypic variation ranging from 14.11 to 37.34%,were distributed mainly on chromosomes 2,3,10,18,and 19,and the greatest number of associations was found on chromosome 10.Of these significant SNPs,19 were associated with only a single salt tolerance trait,while two were associated with two traits.

In 2013,AX-93898882 and AX-93659405 were found to be significantly associated with the ST-GI and ST-GP,respectively,and four potential SNPs (AX-93867730,AX-93781777,AX-94181780,and AX-93898882) were identified in the wild soybean population.In 2014,nine SNPs (six significantly associated with the ST-GI,two(AX-93689000 and AX-93905792) significantly associated with the ST-GR,and one (AX-93745705) significantly associated with the ST-GP) and seven potential SNPs were detected in this population.Based on the average data of the natural wild soybean population across the two environments,10 significant SNPs and nine potential SNPs were identified (Table 6).Among them,four SNPs(AX-94015122,AX-94031902,AX-93855738 and AX-93895778) were associated with both ST-GI and ST-GR,and AX-93952841 was associated with ST-GR and ST-GP.Moreover,AX-93867730,AX-93987448,AX-93689000,AX-94077990 and AX-94114773 were detected in two or three environments,and nine SNPs were co-associated with two or three salt tolerance indices.

Table 6 Significant SNPs (-log10P≥4.0) and suggestive SNPs (3.5＜-log10P＜4.0) for salt tolerance indices in 121 soybean accessions

Table 6 (Continued from preceding page)

Table 7 Haplotypes of 142 recombinant inbred lines (RILs)based on two markers (ss247255001 and ss247257642)

Furthermore,we found that six SNPs (including four out of 21 significant SNPs and two out of 24 potential SNPs) were located on chromosome 19,forming a small cluster flanked by the SNP markers AX-93895778 and AX-93899012 (-log10P≥4.0) with a physical position of 40 652 185-49 890 446 bp.On chromosome 10,we detected three of 21 significant SNPs and four of 24 potential SNPs,distributed in two regions: one region ranged from 10 305 629 (AX-93776990)-10 855 397 (AX-94073665) bp,and the other ranged from 38 184 512 (AX-94077990)-39 275 128 (AX-93781777) bp.

3.5.Prediction of candidate genes and expression pattern analysis in soybean

Among 23 QTLs,three QTLs (qST-GI-10,qST-GP-10-1andqST-GP-10-2) were associated with salt tolerance in the present study and were colocalized in regions with overlapping intervals.Additionally,they were consistently identified in the two environments and in both environments combined,strongly supporting their candidate gene identification.Combining the above QTLs with those of our association analysis,we identified an SNP located in the same region in which these QTLs are located.The SNP AX-93781777 (locus10:39 275 093-39 275 163),located on chromosome 10,was significantly associated withqSTGI-10(in 2014 and 2015) andqSTGP-10(qSTGP-10-1in 2014 and2015,andqSTGP-10-2in the combined environment)mapping in the RIL population.In addition,the SNP AX-93781777 was associated with both ST-GR (2013) and ST-GP (2013) in the wild soybean accessions,the SNP AX-93781777 was downstream of the QTLs,and the physical distance between them was approximately 305 kb.Therefore,the 28 possible candidate genes within this corresponding interval were searched within the W05 genome.The functional annotations of these 28 genes are listed in Appendix B.Among them,seven genes had no biological function annotation.We then compared the similarities of amino acid sequences among the 21 genes that were functionally annotated and their corresponding genes in cultivated soybean.The amino acid sequences of nine of the 21 genes were 100% similar to their corresponding sequences.The remaining 12 genes had similarities ranging from 58.79 to 99.84%.The names of the corresponding genes (based on Wm82.a2.v1)and their functional annotations are given in Appendix C.These results indicated that the potential functions of these 21 genes could be partially predicted based on the expression patterns of their corresponding genes in cultivated soybean.

To further elucidate the potential functions of the 21 genesviafunctional annotations,the expression patterns of the corresponding genes in cultivated soybean were searched within publicly available RNA sequencing (RNAseq) data,although expression data were missing for two genes.Three genes were weakly expressed in all tissues,whileGlyma.10g163300(Glysoja.10g0027264)was highly expressed in all tissues,especially in young leaves (Fig.2).Glyma.10g164100(Glysoja.10G027271)was expressed in nearly all tissues,but not in nodules.Notably,nine of the genes were highly expressed in the root,and were greatly affected by salt stress due to their direct contact with the soil.Thus,we focused on these nine genes in our further analyses.

Fig.2 Tissue expression profiles of 19 genes based on RNA-seq data (downloaded from the SoyBase website,http://www.soybase.org/soyseq).DAF,day after flower.

To determine whether these nine genes responded to salt stress at the germination stage,the induced expression of each of the nine genes was analyzed using germinated seeds of the wild soybean extreme materials Njau-W093 and Njau-W101.Based on the real-time PCR results,GsAKR1(Glysoja.10G027248) was chosen as the candidate gene.Its salt stress-induced expression pattern is shown in Fig.3.The results indicate thatGsAKR1responds to salt stress at the germination stage;in the two extreme accessions,the expression patterns significantly differed.After exposure to salt stress,the expression ofGsAKR1in Njau-W093 was downregulated during the first 24 h but then upregulated at 48 h,after which the expression returned to the initial level.In Njau-W101,the expression significantly changed until 96 h.Overall,the expression pattern ofGsAKR1was considerably altered after treatment with 150 mmol L-1NaCl.This expression pattern will be validated in additional soybean accessions in future studies.

Fig.3 Expression patterns of GsAKR1 (Glysoja.10G27248)in response to salt stress at the germination stage.The X-axis represents 24,48,72,and 96 h after seed germination under the 0 and 150 mmol L-1 NaCl treatments.The Y-axis represents the fold change (2-ΔΔCt) of the expression in response to NaCl.

Based on these results,GsAKR1was selected as a possible candidate gene involved in the salt response at the germination stage.

3.6.Analysis of introgression lines with significant phenotypic differences based on the candidate interval

By considering the phenotypic data of the three salt tolerance indices from three different environments,we identified two soybean lines,NJAU-C121 and NJAU-C135,revealing the significant phenotypic differences for salt tolerance from the 142 RILs.NJAU-C121 was identified as the salt tolerant line,while NJAU-C135 was identified as the salt sensitive line.In the candidate interval that we identifiedvialinkage analysis and GWAS,two markers(ss247255001 and ss247257642) were present in this region.Based on these two molecular markers,the 142 RILs were classified into five groups (Table 7).In this classification,the NJAU-C121 and NJAU-C135 genotypes belonged to two different groups among the five groups,suggesting their association in the regulation of salt tolerance in soybean.

In addition,WGRS was carried out for these two lines,and the SNP density distribution within the soybean genome is shown in Appendix D.After mapping against the soybean Gmax_W82 reference genome,we observed that NJAU-C121 possessed more SNP variations in the candidate interval.Being a salt tolerant line,NJAU-C121 has more SNPs suggesting that it is more diverse from Gmax_W82 and might have retained better alleles(Appendix D).Furthermore,we found 289 different SNPs between the two extreme materials in the 305 kb fragment between the SNP AX-93781777 and the QTL marker ss247255001 on chr.10.The complete mapping results are shown in Appendix E.Among these SNPs,17 nonsynonymous SNPs were located in the exons of the 38 genes.In addition,21 of the total SNPs led to synonymous mutations (Appendix E).Generally speaking,when a gene coding region is mutated,the protein encoded by it may change accordingly,resulting in the differences of plant phenotypes,which is also the genetic basis of QTL mapping.We found that an SNP caused the non-synonymous mutation in the exon ofGlyma.10G161700,the corresponding gene ofGsAKR1in cultivated soybean,suggesting thatGsAKR1is a potential candidate gene for salt tolerance in soybean.

We also compared the whole genomes of NJAU-C121 and NJAU-C135 with the soybean Gmax_ZH13 reference genome,and 12 more SNPs were found in the candidate interval.Interestingly,one of these SNPs was found in the exon ofSoyZH13_10G147500,the corresponding gene ofGsAKR1in Zhonghuang 13.It is also noteworthy that in the two soybean reference genomes,the nonsynonymous SNP showed different genotypes in the two lines,and NJAU-C121 was homozygous at this site while NJAU-C135 was heterozygous.

4.Discussion

4.1.Phenotypic variation of salt tolerance-related traits in the two populations

Many genomic and genetic studies of crop plants have used unique germplasm resources,such as wild accessions,to identify important loci and genes(Andersson and Vicente 2010;Huet al.2020).Wild soybean is a valuable genetic resource for improving soybean cultivation because of its rich collection and high genomic diversity (Hamwieh and Xu 2008).

In this study,we used two populations,one with 142 RILs and one with 121 wild soybean accessions,to detect key loci.ANOVA,h2and the descriptive statistics of three traits (ST-GI,ST-GR and ST-GP) were identified for the two populations in three environments (Table 2).The means of these three traits were higher in the 142 RILs than in the wild accessions.Theh2value of ST-GP was the highest among the three traits in both populations (95.51 and 78.59% in RILs and wild accessions,respectively).Interestingly,theh2values of the three salt tolerance indices were also higher in cultivated soybean,suggesting that these three traits are relatively stable in cultivated soybean and not particularly affected by environmental factors.

4.2.New markers related to salt tolerance-related traits through linkage and association analysis

SNP markers with high amounts of variation or diversity have been widely used to construct dense maps.In this study,a linkage map with 1 018 SNP markers,a map length of 2 177.33 cM and an average marker distance of 2.14,was constructed using a soybean RIL population,which was beneficial for the genetic dissection of soybean salt tolerance indices.

In modern breeding programs,QTL mapping and GWAS have increasingly become the important first steps in marker-assisted selection and gene identification (Price 2006;Abdurakhmonov and Abdukarimov 2008;Soto-Cerda and Cloutier 2012).In the present study,linkage mapping and GWAS were used together to detect key regions and identify candidate genes.Linkage mapping can produce more precise positioning and a wider range of genetic variations (Karikariet al.2019).Thus,a total of 25 QTLs for the three salt tolerance indices clustering on five chromosomes (5,8,9,10,and 11) were mapped in the RIL population.Although the mapping results were similar to those of previous studies (Chenet al.2008;Zhanget al.2014;Kanet al.2015),several novel QTLs related to salt tolerance were identified in this study:qST-GI-5,qST-GR-5,qST-GP-5,qST-GI-11,qSTGR-11,andqST-GR-16.The marker ss247255001 was detected in two traits and two environments (qSTGI-10andqSTGP-10-1).Interestingly,qSTGP-10-1andqSTGP-10-2were considered as the same QTLs because of their largely overlapping confidence intervals (CIs).Some salt tolerance-related QTLs,root character-related QTLs and drought tolerance-related QTLs were found in similar positions on chr.10 (Carpentieri-Pipoloet al.2012;Lianget al.2014;Zhanget al.2014).In addition,three sets of QTLs were also considered to be the same QTLs:qSTGI-5-1,qSTGI-5-2andqSTGI-5-3;qSTGR-5-1andqSTGR-5-2;as well asqSTGP-5-1,qSTGP-5-2andqSTGP-5-3(Table 5).The physical positions ofqSTGI-5,qSTGR-5,andqSTGP-5were similar.Lianget al.(2014)found one QTL for the lateral root number (LRN) and one QTL for the shoot weight (SW) located in the nearby region on chr.5 (Lianget al.2014).LRN and SW are two root characteristics,and it has been reported that high resistance to adversity stress is associated with early and fast root growth,long main roots,and more extensive lateral roots (Sunet al.1996;Lianget al.2014).Both drought stress and salt stress are osmotic stresses that can cause water loss in plants and make it difficult for plants to absorb water,indicating that these two stresses have very similar effects on the cells.

Because of the broad genetic base of natural populations,the association mapping approach is used to detect and characterize quantitative traits(Abdurakhmonov and Abdukarimov 2008;Soto-Cerda and Cloutier 2012).In a previous study using 1 142 SNPs,a GWAS analysis of 191 natural populations for three germination-related traits under high salt conditions determined 13 suggestive SNPs associated with salt tolerance on chrs.2,3,6,8,9,12,13,14,17 and 18(Kanet al.2015).In the present study,the use of GWAS analysis for detecting salt tolerance in wild soybean identified 21 significant SNPs and 24 potential SNPs located on chrs.2,3,5,6,7,10,13,14,16,17,18,19,and 20 in the wild soybean population across three environments.Among the SNPs associated with salt tolerance,the significant SNP AX-93698081 on chr.3 associated with ST-GI in our study was 148 kb away from those detected by Kanet al.(2015).The SNP AX-94119046 on chr.14 associated with ST-GI (significantly)and ST-GR was 2.4 Mb away from those detected by Kanet al.(2015).In addition,the significant SNP AX-94031902 on chr.6 was associated with ST-GI and STGR,and it was far from those detected by Kanet al.(2015).The physical positions of many SNPs found in our study were less than 5 Mb away from those detected by Zenget al.(2017) at the vegetative stage (Zenget al.2017).

Overall,a similar genetic region on chr.10 was detected in different years and in different populations,indicating that these salt tolerance-related loci are conserved and play an important role in the salt tolerance of soybean.This region is worthy of further study.

4.3.Candidate genes for salt tolerance

Comparing these mapping results in the two populations,we found multiple loci within the same genetic region.The SNP (AX-93781777) located on chr.10 was associated with ST-GR and ST-GP,and it was detected in a wild soybean population in 2013 and associated withqST-GI-10andqST-GP-10in the RIL population.The possible candidate genes within the region surrounding the significant SNP AX-93781777 and QTLs were queried.When the gene functional annotation and expression profiles were considered,GsAKR1was predicted to be a candidate gene for salt tolerance and was mapped onto chr.10.The tissue expression pattern based on public information resources showed thatGsAKR1is expressed in all tissues and at relatively high levels in the roots.According to the real-time PCR results,GsAKR1can respond to salt stress rapidly and significantly at the germination stage.To further verify the candidate geneGsAKR1,we performed whole genome resequencing on the contrasting soybean genotypes of NJAU-C121 and NJAU-C135,and compared the differences for the SNPs and InDels in the above candidate interval.We found that non-synonymous mutations occurred in the exon ofGlyma.10G161700(GsAKR1),possibly leading to amino acid mutation and in turn leading to phenotypic differences.GsAKR1encodes a putative aldo-keto reductase 1,and AKRs comprise a superfamily of NAD(P)-linked oxidoreductase superfamily proteins.Plant AKRs may be actively involved in detoxifying stressinduced reactive carbonyls (Senguptaet al.2015).In celery,mannose-6-phosphate reductase (M6PR),a plant AKR member,was identified and overexpressed inArabidopsisthaliana.The resulting transgenic lines presented enhanced salt tolerance by inducing mannitol and glucosyl-mannitol dimer biosynthesis (Zhifang and Loescher 2003),indicating that plant AKRs play an important role in providing abiotic stress tolerance.Additionally,together with the results of induced expression experiments,the results showed thatGsAKR1plays an important role in the salt tolerance of wild soybean.Hence,GsAKR1can be considered a potential candidate gene for improving soybean salt tolerance,although further investigation is required.

Wild soybean is an important germplasm resource for studying stress resistance genes.The use of wild soybean to identify salt tolerance genes has been reported in previous studies.For example,Qiet al.(2014) identified a novel gene (Glysoja01g005509),namedGmCHX1,and Fenget al.(2020) determined thatGsERF7expression in wild soybean roots was responsive to salt.Similar to other crop species,the narrow genetic base of cultivated soybean is a bottleneck for breeding high-yield varieties (Concibidoet al.2003;Li 2008).An important goal of genomic and genetic studies of crop species is to identify important loci and genes by taking advantage of unique germplasm resources,such as wild accessions.The introduction of genes from wild species by advanced backcross methods has been successfully employed to identify yield-enhancing genes in both tomato and rice (Tanksley and McCouch 1997).Because of its abundance of resources and high genomic diversity,wild soybean is obviously a valuable genetic resource for improving cultivated soybean.Annual wild soybean has many flowers and pods,a high reproductive coefficient,and high genetic diversity (Zhuang 1999;Andersson and Vicente 2010);so its use could constitute an effective method to broaden the cultivar soybean genetic base (Tanksley and McCouch 1997).In addition,because cultivated soybean was domesticated from wild soybean at least 5 000 years ago (Hytenet al.2006;Leeet al.2011),no barriers to sexual reproduction exist between them.Wild soybean may be a potential genetic resource for improving cultivated soybean because the former can be used to detect essential loci that cannot be found in the latter because of the absence of variation,ultimately improving cultivated soybean in the future.

5.Conclusion

In soybean,plants subjected to salt stress at the germination stage displayed a significantly reduced crop yield.We used two populations to assess the indices of three salt-related traits in their germination.Using linkage and association mapping,we detected the saltrelated candidate interval and identifiedGsAKR1as the candidate gene in this key region based on both the tissue and induced expression analysis.These results provide genetic and molecular evidence and resources for studying the response mechanism of soybean salt tolerance and further improving cultivated soybean.

Acknowledgements

This work was supported by the Natural Science Foundation of Jiangsu Province,China (BK20191313),the Fundamental Research Funds for the Central Universities,China (KYZ201705) and the Bioinformatics Center of Nanjing Agricultural University,China.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm