The SC7-resistant soybean cultivar DN50 (Dongnong 50) used in this study was obtained from Northeast Agricultural University, Harbin. The SC7-susceptible soybean cultivar XQD (Xiaoqingdou) was obtained from the KeShan Branch of HeiLongJiang Academy of Agricultural Sciences, KeShan.

For SC7 inoculation, the seeds of DN50, XQD and near-isogenic inbred lines (NILs) of R_SMV-11 were planted in a greenhouse with a 16 h light/8 h dark photoperiod and maintained at 25^°C with 70% relative humidity. Soybean plants were inoculated with SMV strain SC7 following the methods described by Che et al. (2020). The control leaves were carried out with equivalent amounts of 0.01 M sodium phosphate buffer (PH 7.2–7.4). Foliar symptoms were monitored every three days after inoculation.

Primers were designed online using Primer 5 based on the Williams 82 reference genome. All primers used for fine-mapping and qRT-PCR assays for candidate genes are listed in Supplementary Table 1.

A single young leaf was collected from each plant at the V2 stage (one fully expanded trifoliate). Genomic DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method (Saghaimaroof et al., 1984). Total RNA was isolated from 1 g soybean leaves using TRIzol (Invitrogen, Shanghai, China) according to the manufacturer’s protocol. The extracted DNA and RNA were quantified using a NanoDrop 2000C ultra-micro spectrophotometer (Sunnyvale, CA, United States) and 1.5% agarose gel electrophoresis.

A cross was generated between the resistant cultivar DN50 and the susceptible XQD. After self-pollination of F₁ plants, 355 F₂ seeds were harvested. All individual F₂ plants were grown in a field in KeShan, China, under natural conditions and a bi-parental F₆ recombinant inbred lines (RIL) population was developed by single-seed descent. The F₆ RILs were planted in the field in KeShan, China, in 2018 for mapping of the R_SMV-11 locus.

Thirty highly resistant and 30 highly susceptible F₆ individuals were screened to generate the R- and S-bulks, respectively. DNA samples of the parental lines and the two bulks were subjected to whole-genome resequencing using the Illumina HiSeq X Ten platform, followed by standard paired-end 150-bp sequencing library construction. Primer sequences of the markers for mapping are listed in Supplementary Table 1. For fine-mapping, 11 markers between positions 1,632,020 and 4,642,090 were identified. Six recombinants were identified in the fine-mapping population using 11 markers, and the SC7 resistance phenotype of their progeny was evaluated to delimit the genomic interval containing R_smv-11. The genotypes of the R_smv-11 allele were analyzed by tagging marker M24 or M28.

cDNA was synthesized from total RNA using an Oligo (dT) 18 primer and PrimeScript 1st strand cDNA Synthesis Kit (Takara, Dalian, China). qRT-PCR analysis was performed to determine the transcript abundance of candidate genes using LightCycler 480 SYBR Green I Master (Roche, Mannheim, Germany) in a LightCycler 480 system (Roche, Mannheim, Germany). The soybean housekeeping gene Tubulin was used as the internal control. The relative transcript level of the target gene was calculated using the 2^–ΔΔCt method. Three biological replicates with three technical replicates each were performed. The expression data of candidate genes in different soybean tissues (leaf, stem, root, flower, seed, pod, and nodule) were obtained from the RNA-seq database (Machado et al., 2020). Primers used are listed in Supplementary Table 1.

To assess the variation in disease resistance between DN50 and XQD, we identified the phenotype of DN50 and XQD. The result domesticated soybean varieties DN50 and XQD, display clear differences in resistance to the SC7 viral strain in the field (Figures 1A–C). Under natural conditions, SC7 caused symptoms in up to 90% of leaves in XQD, but did not induce any symptoms in the leaves of DN50. Moreover, in the greenhouse, SC7-treated XQD exhibited enhanced rugosity, curling, and chlorosis symptoms compared with DN50 (Figures 1D,E). We also analyzed the relative accumulation of SC7 in top non-inoculated leaves at 21 days post inoculation (dpi) in the greenhouse. The accumulation of viral CP gene in DN50 was significantly lower than that in XQD (Figure 1F). Together, these results suggested that DN50 is more resistant than XQD to SC7.

To understand the molecular basis of SC7 resistance, we aimed to detect genetic differences between SC7-resistant and -susceptible plants. To this end, we first generated an F₂ population by crossing DN50 and XQD (Figure 2A). The F₂ population by four rounds of self-fertilization generated an F₆ population (Figure 2A). We selected two groups of F₆ plants reflecting segregation of the SC7 resistance trait (phenotype of F₆ populations revealed a 3:1 ratio of the resistance and susceptible to SC7): one group of 30 individuals showing resistance to SC7 and a second group containing 30 individuals susceptible to SC7 (Figure 2A).

Next, we sequenced the two bulked DNAs (R- and S-bulks), along with single DNA samples of the two parents (DN50 and XQD), on the Illumina HiSeq X Ten platform and then resequenced the genomes of the two parents at 30× coverage each and the two bulked DNAs at 20× coverage each (∼100 Gb data). After filtering, a total of 2,005,612 bi-allelic single-nucleotide polymorphisms (SNPs) and a total of 432,197 bi-allelic short insertions and deletions (InDels) were identified in the two parents, and 113,258 bi-allelic SNPs and 46,172 bi-allelic InDels were identified in the two bulked DNAs. We then aligned the resequencing data of the two bulked sample groups (R- and S-bulks) to the Williams 82 reference genome (Glycine max Wm82.a2.v2; Schmutz et al., 2010).

Using the genotype data from the parents and the two sample groups (R and S), we used the Euclidean distance (ED) algorithm to locate the SC7 resistance QTL locus. Only one QTL at the beginning of chromosome 11 was significantly associated with resistance to SC7; this QTL was named R_SMV-11 (Figures 2B,C). ED value analysis revealed that a physical region of 0.944 kb–3.418 Mb on chromosome 11 was possibly linked to SC7 resistance.

To more precisely map the R_SMV-11 gene within the previously identified candidate regions, we used the F₆ RIL population. We localized R_SMV-11 to a 207-kb region between markers M24 and M28 (Figure 3A), adjacent to the R_SMV-11 region mapped by BSA-seq.

To substantiate the relationship between the R_SMV-11 QTL and SC7 resistance, we compared the phenotypes of two F₆ near-isogenic lines (NILs) carrying either the functional R_SMV-11 allele (NIL-R_SMV-11) or the non-functional r_SMV-11 allele (NIL-r_SMV-11) (Figure 3B) in both the field and the greenhouse. The results showed that NIL-R_SMV-11 had greater resistance to SC7 in the field than NIL-r_SMV-11 (Figures 3C–E). Consistent with the field phenotypes, NIL-r_SMV-11 exhibited enhanced curling symptoms and chlorosis compared with NIL- R_SMV-11 in the greenhouse (Figures 3F,G). We also analyzed the relative biomass of SC7 in infected NIL-R_SMV-11 and NIL-r_SMV-11 soybean leaves at 21 days post infection (dpi) with SC7 in the greenhouse. The biomass of SC7 in NIL-R_SMV-11 was significantly lower than in NIL-r_SMV-11 (Figure 3H). These data confirm that the R_smv-11 allele could greatly enhance resistance to SC7, and that the R_smv-11 gene was in this candidate region.

In the 207-kb candidate mapping region, we found 25 genes (Figure 4A and Supplementary Table 2). We identified SNPs and InDels of all 25 genes in the two parents, and SNPs or InDels in 11 genes resulted in non-synonymous amino acid substitutions or frameshift mutation in the deduced protein sequences (Figure 4B). Of these 11 genes, only two (Glyma.11G028900 and Glyma.11G029900) harbored InDels predicted to cause frameshifts (Figure 4B and Supplementary Table 3).

To further screen candidate genes, we searched an RNA-seq database and retrieved the expression data for 11 candidate genes with DNA sequences differing between DN50 and XQD. We analyzed data for seven tissues: flower, leaves, pod, stem, nodules, seed, and root (Machado et al., 2020). Seven genes were constitutively expressed in all tissues: Glyma.11G028900, Glyma.11G029500, Glyma.11G029800, Glyma.11G030000, Glyma.11G030600, Glyma.11G031100, and Glyma.11G031000 (Figures 5A–G). Glyma.11G030900 was highly expressed in all tissues except the seeds (Figure 5H). The other three genes displayed low transcript abundance in most tissues; however, Glyma.11G029900 showed high expression in the stem, Glyma.11G030300 was highly expressed in roots, and Glyma.11G030800 was abundant in nodules and seeds (Figures 5I–K). Given that Glyma.11G029900, Glyma.11G030300, and Glyma.11G030800 were barely expressed in leaves, we excluded these three genes from this candidate region and subsequent analyses.

We next tested the expression of the eight candidate genes in response to SC7 or mock inoculation of DN50 leaves. Five of eight genes showed altered expression patterns following SC7 treatment as compared to mock inoculation: the expression levels of Glyma.11G028900 and Glyma.11G030600 were induced at 4 h post-inoculation; Glyma.11G029500, Glyma.11G030000, and Glyma.11G030900 were upregulated at 24 h post-inoculation; however, transcript levels of Glyma.11G029800, Glyma.11G031000, and Glyma.11G031100 did not change after SC7 inoculation (Figures 6A–H). These results suggested that Glyma.11G028900, Glyma.11G030600, Glyma.11G029500, Glyma.11G030000, and Glyma.11G030900 were the main candidate genes for the R_SMV-11 locus. Of these genes, only Glyma.11G028900 harbored a SNP and InDel variation in its coding sequence, resulting in a frameshift (Figure 2B). In addition, according to gene function annotation in the Phytozome database (Liu et al., 2016), Glyma.11G028900 corresponds to GmMATE68, a gene that encodes a multidrug and toxic compound extrusion (MATE) transporter. MATE transporters result in pleiotropic phenotypes, including enhanced plant disease resistance (Nawrath et al., 2002; Ishihara et al., 2008). Taken together, these results indicate that GmMATE68 is a likely candidate gene for the R_SMV-11 locus.