Natural Viration And Functional Analysis Of Soybean Genes Underlying Soybean Seed Oil Traits

Soybean is an important oilseed crop,which contributes to about 61%of the world oilseed production and almost 28%of the total global vegetable oil consumption in 2018.Soybean oil consists of five prominent fatty acids that comprise 98.4%of the total oil composition:palmitic acid(16:0,PA),stearic acid(18:0,ST),oleic acid(18:1,OL),linoleic acid(18:2,LI),and linolenic acid(18:3,LN).Soybean can provide humans with high quality vegetable oil because of enrichment in unsaturated fatty acids.China is the center of soybean origin in the world and has abundant soybean germplasm resources.The soybean seed oil content has been significantly increased after domestication and improvement,while the genetic basis underlying such a significant change is largely unknown.The seed oil content and fatty acid composition are quantitative traits which are controlled by multiple genes and required the cooperation of multiple pathways.Therefore,studying the molecular basis and candidate genes underlying soybean domestication,improvement,and seed oil accumulation would be useful to further improve seed oil traits.In this study,187 soybean landraces and 195 released cultivars,which originated from six soybean eco-regions in China,were grown in three years and their seed oil content and fatty acid composition were determined.Genome-wide high-density SNP markers were used for population genetic analysis and association study.As a result,we identified elite soybean germplasm with high seed oil content and/or high OL content,as well as candidate loci that were selected during soybean improvement and also significantly associated with oil traits.We then used gene expression analysis,plant transgenic technology and population genetics to reveal a sucrose efflux transporter gene,GmSWEET39(Sugars Will Eventually be Exported Transporter),controlling soybean seed oil content and under selection.In addition,genome-wide association study(GWAS)and KEGG pathway analysis were employed to identify candidate loci and candidate genes that were significantly associated with seed oil content and five fatty acid contents.The main results obtained in this study are as follows:1.Phenotypic variation in seed oil traits among Chinese cultivated soybeans and identification of elite soybean germplasmsNear-infrared and gas chromatograph methods were used to determine the seed oil content and fatty acid contents of 187 landraces and 195 released cultivars grown in 2012,2014 and 2015,which originated from different regions of China.The variation range of seed oil content in 382 accessions is 15.80%?24.43%while that for oleic acid content is 14.94%?38.45%.The average seed oil content of landraces and released cultivars was 19.77%and 20.93%,respectively,reflecting that the average seed oil content has been improved in released cultivars over landraces during soybean breeding history.Correlation analysis showed a significant positive correlation between seed oil content and OL content(r=0.19,P<0.01),but a significant negative correlation with PA content(r=-0.2,P<0.01)and LN content(r=-0.55,P<0.01).Based on the average phenotypic data in three years,we identified 15 soybean accessions with high seed oil content(?23%)and 6 soybean accessions with high OL content(?33%),and most of these selected germplasms are released cultivars.2.Selective sweeps related to seed oil content during soybean improvementTo identify the improvement-associated loci,71293 SNP markers with MAF>0.05 were employed to analyze genome-wide FST(fixation index)and ROD(reduction of diversity)in 187 landraces and 195 released cultivars.Following the criteria of FST?0.19 and ROD?0.84(corresponding to top 2%sites in the whole-genome),we identified 47 SNPs subjected to selective sweeps.We further compared the physical locations between these 47 SNPs and the QTLs of seed oil content recorded in the Soybase database,and found that 25 SNPs locate within the previously mapped QTL regions.These significant loci could be the selective sweeps related to seed oil content during soybean improvement.3.GmSWEET39 is a candidate gene affecting seed oil content during soybean improvementLots of oil content-related genes are specifically expressed in seeds.In order to identify the key candidate genes in above identified loci,the 864 genes within the LD decay distance(786 kb)of these 25 SNPs were examined for their expression levels in different soybean tissues,using the RNA-seq data from Soybase.Five genes,Glyma.08g116300,Glyma.09g185500,Glyma.15g049200,Glyma.08g088000 and Glyma.15g050800,are highly expressed in soybean seeds.Then we employed qRT-PCR analyses to compare the transcript abundance of these five genes in soybean seeds at 10,20,30 and 40 days after flowering(DAF),using two soybean accessions differing in oil content.We found that only one gene,Glyma.15g049200(GmSWEET39),showed much higher relative expression in multiple stages of developing seeds with high oil content than that with low oil content.Next,regional association study of the seed oil content was performed with SNPs(MAF?0.05)within the 10-Mb region surrounding GmSWEET39 in 382 Chinese cultivated soybean accessions.Six SNPs showed significant(FDR?0.1)association with total seed oil content.Two of them,Gm15₃852076 and Gm15₃852306,not only locate 4.5 kb downstream of the candidate gene GmSWEET39 but also locate within the previously confirmed seed oil QTL regions(cqSeed oil-007 and cqSeed oil-010).In addition,these two SNPs exhibited strong LD with GmSWEET39(located in the same LD block).We found soybean accessions carrying the Gm15₃852076-G or Gm15₃852306-C allele have significantly higher average seed oil content than those with the Gm15₃852076-T or Gm15₃852306-A allele,and the frequencies of Gm15₃852076-G and Gm15₃852306-C alleles are higher in released cultivars than landraces,indicating that these two loci subjected to selection in this region associated with seed oil content.SWEET genes were predicted to encode a bidirectional sugar transporter.Considering that seed oil biosynthesis is directly affected by the accumulation of sucrose in sink tissues,therefore,GmSWEET39 was selected as a candidate gene controlling seed oil content during soybean improvement.4.Functional verification and superior alleles of GmSWEET39The tissue expression patterns of GmSWEET39 were confirmed using samples of different tissues from a soybean accession Kexin4(22.84%seed oil content).We found that GmSWEET39 was highly expressed in soybean seeds and leaves,with the highest level in seeds at 30 DAF.Next,we investigated the expression levels of GmSWEET39 in the seeds at 30 DAF among a representative population composed of 80 accessions with high,moderate and low seed-oil-content,which were selected from 382 cultivated soybean accessions.Further analysis showed that the relative expression level of GmSWEET39 was positively correlated with seed oil content(r=0.63,P=4.35 e-10).Subcellular localization in tobacco leaves and Arabidopsis protoplasts showed that GmSWEET39 protein was expressed on plasma membrane.The approximately 2.4 kb region of GmSWEET39 including its 1.6 kb promoter and full length of CDS in 80 accessions were sequenced.Based on the polymorphic sites within GmSWEET39,the promoter region and CDS region were classified into three types(Prol-Pro3)and two types(CDS1-CDS2),respectively.A total of six haplotypes(HaplHap6)were identified by combining three promoter types and two CDS types.We then analyzed the effects of GmSWEET39 haplotypes,promoter and CDS types on soybean seed oil content and gene relative expression.As a result,accessions with Hap6(Pro3+CDS2),Pro3 or CDS2 had higher expression levels of GmSWEET39 and average seed oil content than those containing the other alleles.These results suggest that Hap6(Pro3+CDS2),Pro3 and CDS2 are the potential superior alleles of GmSWEET39 that might improve soybean seed oil content.To further verify the superior alleles of GmSWEET39,we overexpressed GmSWEET39CDS1 and GmSWEET39CDS2 using CaMV 35S promoter in Arabidopsis,respectively.Five homozygous Arabidopsis lines containing 35S:GmSWEET39CDS1 and 10 homozygous lines containing 35S:GmSWEET39CDS2 were obtained at T3 generation.Gas chromatographic analysis revealed that the fatty acid content in mature seeds of GmSWEET39CDS2 overexpressing lines increased by 33%on average.However,no significant difference was observed between GmSWEET39CDS1 overexpressing lines and control lines.Next,we found that the fatty acid content in soybean hairy roots of 35S:GmSWEET39CDS2 and GmSWEET39Pro3:GmSWEET39CDS2(GmSWEET39Hap6)was significantly higher than that in the roots containing empty vector and significantly higher than that in the roots overexpressed 35S:GmSWEET39CDS1.In addition,accessions with GmSWEET39Pro3,GmSWEET39CDS2 or GmSWEET39Hap6 had higher seed oil content than those containing the other alleles among a soybean population of 155 recombiant inbred lines.These results confirmed that GmSWEET39CDS2 or the combination GmSWEET39Pro3 and GmSWEET39CDS2 are the superior alleles of GmSWEET39 that can improve soybean seed oil content.We further sequenced the GsSWEET39 gene from 41 wild soybean accessions,and found 22 haplotypes,19 types of promoter and five types of CDS,indicating that SWEET39 gene has more variation in wild soybean than that in cultivated soybean.Furthermore,the frequencies of superior SWEET39 alleles in wild soybean are significantly lower than those in cultivated soybean,which suggests that SWEET39 gene could be selected during domestication from wild soybean.5.Other genetic loci and candidate genes underlying soybean seed oil and fatty acid contentsTo identify the seed oil traits-associated genetic loci and candidate genes,the BLUP(best linear unbiased prediction)values,which were calculated from seed oil and fatty acid contents in 382 soybean accessions in three years,and 71293 SNP markers with MAF?0.05 were employed to perform GWAS.Following the threshold of FDR?0.1,a total of 12,12 and 7 SNPs were found significantly associated with seed oil content,LI content and LN content,respectively.Next,the candidate genes within the LD region of these significant association SNPs were identified using gene annotation and metabolic pathway analysis.As a result,a total of 12,10 and 2 candidate genes related to oil traits were identified in the regions of 38.47 Mb-38.56 Mb on chromosome 5,6.89 Mb-7.08 Mb on chromosome 8,and 5.37 Mb-5.43 Mb on chromosome 12,respectively.Further functional studies of these candidate genes could provide useful information for improving soybean seed oil traits.