Identification of major quantitative trait loci and candidate genes for seed weight in soybean

Abstract

Four major quantitative trait loci for 100-seed weight were identified in a soybean RIL population under five environments, and the most likely candidate genes underlying these loci were identified. Seed weight is an important target of soybean breeding. However, the genes underlying the major quantitative trait loci (QTL) controlling seed weight remain largely unknown. In this study, a soybean population of 300 recombinant inbred lines (RILs) derived from a cross between PI595843 (PI) and WH was used to map the QTL and identify candidate genes for seed weight. The RIL population was genotyped through whole genome resequencing, and phenotyped for 100-seed weight under five environments. A total of 38 QTL were detected, and four major QTL, each explained at least 10% of the variation in 100-seed weight, were identified. Six candidate genes within these four major QTL regions were identified by analyses of their tissue expression patterns, gene annotations, and differential gene expression levels in soybean seeds during four developmental stages between two parental lines. Further sequence variation analyses revealed a C to T substitution in the first exon of the Glyma.19G143300, resulting in an amino acid change between PI and WH, and thus leading to a different predicted kinase domain, which might affect its protein function. Glyma.19G143300 is highly expressed in soybean seeds and encodes a leucine-rich repeat receptor-like protein kinase (LRR-RLK). Its predicted protein has typical domains of LRR-RLK family, and phylogenetic analyses reveled its similarity with the known LRR-RLK protein XIAO (LOC_Os04g48760), which is involved in controlling seed size. The major QTL and candidate genes identified in this study provide useful information for molecular breeding of new soybean cultivars with desirable seed weight.

Fig. 1

Seed traits of the two parental soybean accessions PI and WH. a Seed morphology of PI and WH. Scale bar, 1 cm. b Statistical analysis of the 100-seed weight of PI and WH. c Seed length of PI and WH. Scale bar, 1 cm. d Statistical analysis of the seed length of PI and WH. e Seed width of PI and WH. Scale bar, 1 cm. f Statistical analysis of the seed width of PI and WH. The photograph and phenotypic data of 100-seed weight, seed length and seed width were obtained under 2019DT environment. All data and error bars in charts represent mean ± standard deviation of three replications (n = 100 × 3 for 100-seed weight; n = 10 × 3 for seed length and seed width). Student’s t-tests (two-tail) were used to compare the significant differences between PI and WH

Fig. 2

Genotyping map and genetic map constructed from resequencing data of the NJPW-RIL population. a The genotype of 4702 bins based on the recombination breakpoints identified in 300 NJPW-RILs derived from the cross of PI and WH. Each horizontal line represents a single RIL across 20 soybean chromosomes. Red and blue bars represent the parental genotypes of PI and WH, respectively. b Distribution and genetic distance of bin markers on 20 soybean chromosomes in the NJPW-RIL population. The horizontal black lines on each chromosome represent bin markers

Fig. 3

The quantitative trait loci (QTL) for 100-seed weight identified in the NJPW-RIL population under multiple environments. a 2014LH, b 2015DT, c 2015JP, d 2018DT, e 2019DT and f MEAN represent the environments of 2014Liuhe, 2015Dangtu, 2015Jiangpu, 2018Dangtu, 2019Dangtu, and the mean value of 100-seed weight across five environments, respectively. LOD, logarithm of odds; the horizontal dotted lines represent LOD thresholds calculated from 1000-permutation tests (significance level of 0.05) by using the CIM model in WinQTLCart2.5 Software, which were 3.60, 3.50, 3.50, 3.50, 3.70 and 3.60 for 2014LH, 2015DT, 2015JP, 2018DT, 2019DT and MEAN (the mean value of 100-seed weight value across five environments), respectively

Fig. 4

Relative expression levels of six candidate genes in the seeds of two parental soybean accessions PI and WH at different developmental stages. Relative expression levels of six candidate genes, including Glyma.19G143300 a Glyma.19G182400 b Glyma.20G053200 c Glyma.20G055900 d Glyma.20G062700 e and Glyma.20G081600 f, in the seeds of two parental lines PI (larger seed) and WH (smaller seed) at four developmental stages of 10, 20, 30, and 40 DAF (days after flowering). GmUKN1 (Glyma.12G02500) was used as an internal control. The data represent the mean ± standard deviation (n = 3 × 3 = 9). * and ** represent significant difference in the relative expression level between PI and WH at 0.05 and 0.01 level, respectively; ns, not significant (Student’s t-test, two-tail)

Fig. 5

Sequence analyses of Glyma.19G143300 and its predicted protein structure. a Polymorphisms in the coding region of Glyma.19G143300 between the two parental lines of soybean RIL population and the reference genome sequence of Williams 82. b The amino acid change of S (serine) to F (phenylalanine) due to the SNP polymorphism in the coding region of Glyma.19G143300 as shown in a. c, d The predicted protein structure of Glyma.19G143300 in PI and WH, respectively. The first gray boxes represent LRRNT_2 domains (leucine-rich repeats at the N terminus), the green boxes represent LRR (tandem leucine-rich repeats) domains, the blue boxes represent transmembrane regions, and the boxes at the end represent the kinase domains of Pkinase_Tyr domain in c (gray box) and STYKc domain in d (orange box). e, f The three-dimensional structure of Glyma.19G143300 protein in PI and WH, respectively. The white arrows indicate the difference between PI and WH. g Phylogenic tree of Glyma.19G143300 and the known leucine-rich repeat receptor-like kinase (LRR-RLK) proteins. The tree was constructed using MEGA version 6.0. The numbers on the branches indicate the 1000 bootstrap values. Scale bar unit, divergence distance. The figure was generated using the full-length amino acid sequences of the proteins, including AT3G19700 and AT4G39400 from Arabidopsis thaliana, LOC_Os04g48760 and LOC_Os09g12240 from Oryza sativa, GRMZM2G149051 from Zea mays and Glyma.19G143300 from Glycine max

Fig. 6

Sequence and allelic variation in Glyma.19g143300 among soybean recombinant inbred lines (RILs), the two parents and 186 soybean accessions. a Sequence variation in the coding region of Glyma.19g143300 from 60 RILs (with 30 largest and 30 smallest 100-seed weight), the two parental lines of PI and WH, and Williams 82 (W82). The position of the sequence variation is relative to the start codon (ATG), which is shown on the top. 100-SW, 100-seed weight. The RILs were named with PW + number, for example, PW233 represent a RIL derived from the cross of PI × WH. b Boxplot of 100-seed weight for two groups of soybean RILs carrying two different CDS types of Glyma.19g143300, in the 60 RILs with extreme 100-seed weight. The phenotypic data of 100-seed weight were the mean value across 5 environments. c Boxplot of 100-seed weight for two groups of soybean accessions carrying two different CDS types of Glyma.19g143300. The sequence variations of Glyma.19g143300 and phenotypic data of 100-seed weight were downloaded from the database (http://www.mbkbase.org/soybean). Statistical significance of the difference between two groups was determined by two-sided Wilcoxon test. The center bold line represents the median; box edges indicate the upper and lower quantiles; whiskers show the 1.5 × interquartile range and points indicate outliers