Rsv1 was originally identified in the soybean line PI 96983 (Kiihl and Hartwig, 1979), and it confers extreme resistance (ER) to SMV-G1 through G6 but not to SMV-G7 (Chen et al., 1991; Hajimorad and Hill, 2001; Table 1). Multiple Rsv1 alleles including Rsv1-y, Rsv1-m, Rsv1-t, Rsv1-k, and Rsv1-r have been identified from different soybean cultivars with differential reactions to SMV G1–G7 strains (Chen et al., 2001). Rsv1 was initially mapped to soybean linkage group F on chromosome 13 (Yu et al., 1994) and two classes of NBS-LRR sequences (classes b and j) were identified in this resistance gene cluster (Yu et al., 1996). A large family of homologous sequences of the class j Nucleotide biinding site-leucine rich repeat (NBS-LRRs) clustered at or near the Rsv1 locus (Jeong et al., 2001; Gore et al., 2002; Peñuela et al., 2002). Six candidate genes (1eG30, 5gG3, 3gG2, 1eG15, 6gG9, and 1gG4) in PI96983 were mapped to a tightly clustered region near Rsv1, three of them (3gG2, 5gG3, and 6gG9) were completely cloned and sequenced (GenBank accession no. AY518517–AY518519). Among the three genes, 3gG2 was found to be a strong candidate for Rsv1 (Hayes et al., 2004). When 3gG2, 5gG3, and 6gG93 were simultaneously silenced using Bean pod mottle virus -induced gene silencing (BPMV-VIGS), the Rsv1-mediated resistance was compromised, confirming that one or more of these three genes is indeed the Rsv1 (Zhang et al., 2012). Because, the sequence identities of these three R genes are extremely high along the entire cDNAs, it is impossible to differentiate which one(s) is Rsv1.

Several studies indicate that two or more related non-TIR-NBS-LRR gene products are likely involved in the allelic response of several Rsv1-containing lines to SMV (Hayes et al., 2004; Wen et al., 2013; Yang et al., 2013). Wen et al. (2013) generated two soybean lines, L800 and L943, derived from crosses between PI96983 (Rsv1) and Lee68 (rsv1) with distinct recombination events within the Rsv1 locus. The L800 line contains a single PI96983-derived member 3gG2, confers ER against SMV-N (an avirulent isolate of G2 strain). In contrast, the line L943 lacks 3gG2, but contains a suite of five other NBS-LRR genes allows limited replication of SMV-N at the inoculation site. Domain swapping experiments between SMV-N and SMV-G7/SMV-G7d demonstrate that at least two distinct resistance genes at the Rsv1 locus, probably belonging to the NBS-LRR class, mediate recognition of HC-Pro and P3, respectively (Khatabi et al., 2013; Yang et al., 2013).

Rsv3 was originated from “L29,” a ‘Williams’ isoline derived from Hardee (Bernard et al., 1991; Gunduz et al., 2000). The diverse soybean cultivars carrying Rsv3 alleles condition resistance to SMV G5 through G7, but not G1 through G4 (Jeong et al., 2002; Table 1). Rsv3 locus was firstly mapped between markers A519F/R and M3Satt on MLG B2 (chromosome 14) by Jeong et al. (2002), and was subsequently mapped on MLG-B2 with a distance of 1.5 cM from Sat_424 and 2.0 cM from Satt726 (Shi et al., 2008). The 154 kbp interval encompassing Rsv3 contains a family of closely related coiled-coil NBS-LRR (CC-NBS-LRR) genes, implying that the Rsv3 gene most likely encodes a member of this gene family (Suh et al., 2011).

Rsv4 confer resistance to all 7 SMV strains (Chen et al., 1993; Ma et al., 1995). It was identified in soybean cultivars V94-5152 and mapped to a 0.4 cM interval between the proximal marker Rat2 and the distal marker S6ac, in a ∼94 kb haplotype block on soybean chromosome 2 (MLG D1b++W) (Hayes et al., 2000; Saghai Maroof et al., 2010; Ilut et al., 2016). A haplotype phylogenetic analysis of this region suggests that the Rsv4 locus in G. max is recently introgressed from G. soja (Ilut et al., 2016). Interestingly, this interval did not contain any NB-LRR type R genes. Instead, several genes encoding predicted transcription factors and unknown proteins are present within the region, suggesting that Rsv4 most likely belongs to a novel class of resistance gene (Hwang et al., 2006; Ilut et al., 2016).

SMV isolates are classified into seven strains (G1–G7) based on phenotypic reactions on a set of differential soybean cultivars (Cho and Goodman, 1982). The modification of avirulence factors of plant viruses by one or more amino acid substitutions can convert avirulence to virulence on hosts containing resistance genes and therefore, can be used as an approach to determine the avirulence factor(s) of a specific resistant gene.

Rsv1, a single dominant resistance gene in soybean PI 96983 (Rsv1), confers ER against SMV-G1 through G6 but not to SMV-G7 (Chen et al., 1991; Hajimorad and Hill, 2001; Table 1). SMV-N (an avirulent isolate of strain G2) elicits ER whereas strain SMV-G7 provokes a lethal systemic hypersensitive response (LSHR) (Hajimorad et al., 2003; Hayes et al., 2004). SMV-G7d, an evolved variant of SMV-G7 from lab, induces systemic mosaic (Hajimorad et al., 2003). Serial passages of a large population of the progeny in PI 96983 resulted in emergence of a mutant population (vSMV-G7d), which can evade Rsv1-mediated recognition and the putative amino acid changes that potentially responsible for the mutant phenotype is initially tentatively narrowed down to HC-Pro, coat protein, PI proteinase or P3 (Hajimorad et al., 2003; Seo et al., 2011) and was later mapped to P3 through domain swapping between the pSMV-G7 and pSMV-G7d (Hajimorad et al., 2005). The amino acids 823, 953, and 1112 of the SMV-G7d are critical in evading of Rsv1-mediated recognition (Hajimorad et al., 2005, 2006). By generating a series of chimeras between SMV-G7 and SMV-N in combination with site-directed mutagenesis, Eggenberger et al. (2008) and Hajimorad et al. (2008) independently showed that gain of virulence on Rsv1-genotype soybean by an avirulent SMV strains requires concurrent mutations in both P3 and HC-Pro and HC-Pro complementation of P3 is essential for SMV virulence on Rsv1-genotype soybean (Table 1). A key virulence determinant of SMV on Rsv1-genotype soybeans that resides at polyprotein codon 947 overlaps both P3 and a PIPO-encoded codon. This raises the question of whether PIPO or P3 is the virulence factor. Wen et al. (2011) confirmed that amino acid changes in P3, and not the overlapping PIPO-encoded protein, which is embedded in the P3 cistron, determine virulence of SMV on Rsv1-genotype soybean. Chowda-Reddy et al. (2011b) constructed a chimeric infectious clone of G7, in which the N-ternimal part of CI was swapped with the corresponding part of G2. Compared with wildtype G7, this chimeric strain lost virulence on Rsv1-genotype plant but gained infectivity on Rsv3-genotype plant, indicating an essential role of CI for breaking down both Rsv1- and Rsv3-mediated resistance (Chowda-Reddy et al., 2011b). Together, it appears that P3, HC-Pro and possibly CI are virulent determinants for Rsv1-mediated resistance (Table 1).

It has been proven that cytoplasmic inclusion cistron (CI) of SMV serves as a virulence and symptom determinant on Rsv3-genotype soybean and a single amino acid substitution in CI was found to be responsible for gain or loss of elicitor function of CI (Seo et al., 2009; Zhang et al., 2009a). Analyses of the chimeras by exchanging fragments between avirulent SMV-G7 and the virulent SMV-N showed that both the N- and C-terminal regions of the CI cistron are required for Rsv3-mediated resistance and the N-terminal region of CI is also involved in severe symptom induction in soybean (Zhang et al., 2009a). In addition to CI, P3 has also been reported to play an essential role in virulence determination on Rsv3-mediated resistance (Chowda-Reddy et al., 2011a,b; Table 1).

Gain of virulence analysis on soybean genotypes containing Rsv4 genes showed that virulence on Rsv4 carrying cultivars was consistently associated with Q1033K and G1054R substitutions within P3 cistron, indicating that P3 is the SMV virulence determinant on Rsv4 and one single nucleotide mutation in the P3 protein is sufficient to compromise its elicitor function (Chowda-Reddy et al., 2011b; Khatabi et al., 2012; Wang et al., 2015). However, the sites involved in the virulence of SMV on Rsv4-genotype soybean vary among strains (Wang et al., 2015).

It is clear now that P3 plays essential roles in virulence determination on Rsv1, Rsv3, and Rsv4 resistant loci, while CI is required for virolence on Rsv1 and Rsv3 genotype soybean plants (Chowda-Reddy et al., 2011a,b). These results imply that avirulent proteins from SMV might interact with the soybean R gene products at a converged point. This evolved interactions sometimes could give SMV advantage in breaking resistance conferred by different SMV resistant genes simply by mutations within a single viral protein. On the other hand, since multiple proteins are involved in virulence on different resistant loci, concurrent mutantions in multiple proteins of SMV are required to evade the resistance conferred by different SMV resistant genes. The likelihood of such naturally occuurred concurrent mutations in different viral proteins is low. Therefore, integration of all three SMV resistant genes in a single elite soybean cultivar may provide long-lasting resistance to SMV in soybean breeding practice (Chowda-Reddy et al., 2011b).

The key components in R gene mediated disease resistant signaling pathway have been identified in model plant Arabidopsis, among which, RAR1 (Required for Mla 12 Resistance), SGT1(Suppressor of G2 Allele of Skp1) and HSP90 (Heat Shock Protein 90) are the most important ones (Belkhadir et al., 2004). Using BPMV-VIGS, it has been shown that Rsv1-mediated ER against SMV in soybean requires RAR1 and SGT1 but not GmHSP90, suggesting although soybean defense signaling pathways recruit structurally conserved components, they have distinct requirements for specific proteins (Fu et al., 2009). However, Zhang et al. (2012) showed that silencing GmHSP90 using BPMV-VIGS compromised Rsv1-mediated resistance. In addition, silencing GmEDR1 (Enhanced Disease Resistance 1), GmEDS1 (Enhanced Disease Susceptibility 1), GmHSP90, GmJAR1 (Jasmonic Responsive 1), GmPAD4 (Phytoalexin Deficient 4), and two genes encoding WRKY transcription factors (WRKY6 and WRKY 30), all of which are involved in defense pathways in model plant Arabidopsis, Rsv1-mediated ER was also compromised (Table 2). These results suggest that the host components required for R gene-mediated resistant signaling pathways are conserved across plant species.

Mitogen-activated protein kinase (MAPK) cascades play important roles in disease resistance (Meng and Zhang, 2013). The function of MAPK signaling pathways in disease resistance was investigated in soybean using BPMV-VIGS (Liu et al., 2011, 2014, 2015). Among the plants silenced for multiple genes in MAPK pathway, the plants silenced for the GmMAPK4 and GmMAPK6 homologs displayed strong phenotypes of activated defense responses (Liu et al., 2011, 2014). Consistent with the activated defense response phenotypes, these plants were more resistant to SMV compared with vector control plants (Liu et al., 2011, 2014), indicating that both genes play critical negative roles in basal resistance or PAMP-triggered immunity (PTI) in soybean. The constitutively activated defense responses has been reported for mpk4 mutant in Arabidopsis (Petersen et al., 2000) and the positive role of MPK6 in defense responses is well-documented (Meng and Zhang, 2013). However, the negative role of MAKP6 homologs has not been reported previously (Liu et al., 2014), indicating that both conserved and distinct functions of MAPK signaling pathways in immunity are observed between Arabidopsis and soybean.

Numerous host factors participate in defense responses in plants. Identification of these factors may facilitate rationale design of novel resistant strategies. Recently, it has been shown that silencing GmHSP40.1, a soybean nuclear-localized type-III DnaJ domain-containing HSP40, results in increased infectivity of SMV, indicating a positive role of GmHSP40.1 in basal resistance (Liu and Whitham, 2013). A subset of type 2C protein phophatase (PP2C) gene family, which participate ABA signaling pathway, is specifically up-regulated during Rsv3-mediated resistance (Seo et al., 2014). Synchronized overexpression of GmPP2C3a using SMV-G7H vector inhibits virus cell-to-cell movement mediated by callose deposition in an ABA signaling-dependent manner, indicating that GmPP2C3a functions as a key regulator of Rsv3-mediated resistance (Seo et al., 2014). An ortholog of Arabidopsis K⁺ weak channel encoding gene AKT2, was significantly induced by SMV inoculation in the SMV highly resistant genotype, but not in the susceptible genotype (Zhou et al., 2014). Overexpression of GmAKT2 not only significantly increased K⁺ concentrations in young leaves but also significantly enhanced the resistance against SMV, indicating alteration of K⁺ transporter expression could be a novel molecular approach for enhancing SMV resistance in soybean (Zhou et al., 2014). Molybdenum cofactor (Moco) is required for the activities of Moco-dependant enzymes. Cofactor for nitrate reductase and xanthine dehydrogenase (Cnx1) is known to be involved in the biosynthesis of Moco in plants. Soybean plants transformed with Cnx1 enhanced the enzyme activities of nitrate reductase (NR) and aldehydeoxidase (AO) and resulted in an enhanced resistance against various strains of SMV (Zhou et al., 2015). The differentially expressed genes in Rsv1 genotype in response to G7 infection have been identified (Chen et al., 2016a). Knocking down one of the identified genes, the eukaryotic translation initiation factor 5A (eIF5A), diminished the LSHR and enhanced viral accumulation, suggesting an essential role of eIF5A in the Rsv1-mediated LSHR signaling pathway. Eukaryotic elongation factor 1A (eEF1A) is a well-known host factor in viral pathogenesis. Recently, Luan et al. (2016) showed that silencing GmeEF1A inhibits accumulation of SMV and P3 protein of SMV interacts with GmeEF1A to facilitate its nuclear localization and therefore, promotes SMV pathogenicity.

Small RNAs play a fundamental role in anti-viral defense. Three miRNAs, miR160, miR393 and miR1510, which have been previously shown to be involved in disease resistance in other plant species, have been identified as SMV-inducible miRNAs through small RNA sequencing approach (Yin et al., 2013), implying that these three miRNAs might play roles in SMV resistance. Chen et al. (2015) recently showed that the expression of miRNA168 gene is specifically highly induced only in G7-infected PI96983 (incompatible interaction) but not in G2- and G7-infected Williams 82 (compatible interactions). Overexpression of miR168 results in cleavage of miR168-mediated AGO1 mRNA and severely repression of AGO1 protein accumulation (Chen et al., 2015). Silencing SGS3, an essential component in RNA silencing, suppressed AGO1 siRNA, partially recovers the repressed AGO1 protein, and alleviates LSHR severity in G7-infected Rsv1 soybean (Chen et al., 2015). These results strongly suggest that miRNA pathway is involved in G7 infection of Rsv1 soybean, and LSHR is associated with repression of AGO1.

Chen et al. (2016b) recently performed small RNA (sRNA)-seq, degradome-seq and as well as a genome-wide transcriptome analysis to profile the global gene and miRNA expression in soybean in response to three different SMV isolates. The SMV responsive miRNAs and their potential cleavage targets were identified and subsequently validated by degradome-seq analysis, leading to the establishment of complex miRNA-mRNA regulatory networks. The information generated in this study provides insights into molecular interactions between SMV and soybean and offer candidate miRNAs and their targets for further elucidation of the SMV infection process.

The multiple soybean cultivars transformed with an RNA interference (RNAi) construct targeted for SMV HC-Pro displayed a significantly enhanced resistance against SMV (Gao et al., 2015). Soybean plants transformed with a single RNAi construct expressing separate short hairpins or inverted repeat (IR) (150 bp) derived from three different viruses (SMV, Alfalfa mosaic virus, and Bean pod mottle virus) confer robust systemic resistance to these viruses (Zhang et al., 2011). This strategy makes it easy to incorporate additional short IRs in the transgene, thus expanding the spectrum of virus resistance. As the cases in the other plant species, these studies demonstrate that RNA silencing is obviously the most effective approach for SMV resistance.

Bean pod mottle virus -induced gene silencing system has been proven successful in gene function studies in soybean (Zhang et al., 2009b, 2010; Liu et al., 2015). There are four GmMAPK4 homologs that can be divided into two paralogous groups (Liu et al., 2011). The sequence identities of ORFs within the groups are greater than 96%, whereas the identities between the groups are 88.7% (Liu et al., 2015). The BPMV-VIGS construct used for silencing GmMAPK4 by Liu et al. (2011) actually can silence all four of the isoforms simultaneously. When only one parologous group was silenced by using construct targeted for the 3′ UTR (the sequence identity of the 3′ UTRs between the two parologous groups is less than 50%), the activated defense response was not observed, indicating that silencing the four GmMAPK4 isoforms simultaneously is necessary for activating defense responses in soybean. Using the same approach, it has been differentiated that GmSGT1-2 but not GmSGT1-1 is required for the Rsv1-mediated ER against SMV (Fu et al., 2009). Thus, VIGS is currently the most powerful tool in overcoming the gene redundancy in soybean.