Decades of Genetic Research on Soybean mosaic virus Resistance in Soybean

This review summarizes the history and current state of the known genetic basis for soybean resistance to Soybean mosaic virus (SMV), and examines how the integration of molecular markers has been utilized in breeding for crop improvement. SVM causes yield loss and seed quality reduction in soybean based on the SMV strain and the host genotype. Understanding the molecular underpinnings of SMV–soybean interactions and the genes conferring resistance to SMV has been a focus of intense research interest for decades. Soybean reactions are classified into three main responses: resistant, necrotic, or susceptible. Significant progress has been achieved that has greatly increased the understanding of soybean germplasm diversity, differential reactions to SMV strains, genotype–strain interactions, genes/alleles conferring specific reactions, and interactions among resistance genes and alleles. Many studies that aimed to uncover the physical position of resistance genes have been published in recent decades, collectively proposing different candidate genes. The studies on SMV resistance loci revealed that the resistance genes are mainly distributed on three chromosomes. Resistance has been pyramided in various combinations for durable resistance to SMV strains. The causative genes are still elusive despite early successes in identifying resistance alleles in soybean; however, a gene at the Rsv4 locus has been well validated.

The history of soybean (Glycine max) symptoms caused by Soybean mosaic virus (SMV) starts in the U.S. as early as 1915, when it was first observed and documented in Western scientific literature by Clinton (1916) [1]. Several years later, Gardner and Kendrick (1921) [2] described the symptoms in detail as mosaic, dark green leaf areas with misshapen and stunted leaflets, and the discoloration of the seed coat (mottling) caused by the bleeding of hila pigment into other regions. Other symptoms included local and/or systemic necrosis [3]. Since then, many scientific studies describing the principles of SMV infection The virus continuously adapts and evolves, producing new isolates that overcome host resistance. Several isolates that overcame resistance have been identified over the past two decades in South Korea and Brazil. SMV-G3A and G5H were initially reported by Cho et al. (1983) [43] as mutated strains from G3 and G5, respectively. The G3A mutant induced mosaic symptoms in Marshall, Ogden, and 'Jangbaekkong'. The G5H mutant caused necrosis in 'Suweon 97' and a mosaic reaction in Jangbaekkong. Mosaic symptoms were later observed by Kim and Lee (1991) [73] in Marshall and 'Paldalkong' after infection by G5DH. In Brazil, the SMV isolate SMV-95-1 was classified as a member of SMV-G5 [74]. In South Korea, an evolved SMV strain G7H overcame genetic resistance [75], and a new resistance-breaking isolate L-RB from the G2 pathotype was identified in Canada [23]. In recent years, the genetic analysis of SMV diversity in Iran revealed a new SMV mutation in a P3 protein, overcoming Rsv4 resistance in the soybean cultivar 'V94-5152' [76]. A novel resistance-breaking SMV-R strain was found to be common in the central and southern provinces of China [77]. These reports indicated that gene-mediated resistance in soybean is not durable due to high selection pressure from the constantly evolving SMV isolates generated by viral RNA polymerase, which lacks the proof-reading ability [4]. The effective control of new and aggressive strains requires significant efforts to manage and limit disease prevalence in soybeans [78]. Therefore, disease management and new sources of resistance are essential in plant breeding and crop improvement [46,52].

Molecular Mechanisms of SMV Infection
The molecular interactions between SMV and the plant host are complex, and many mechanisms are still unknown. The virus is released directly into the host cell via the mechanical damage of soybean tissue and its infection action appears to be similar to other Potyviruses [78,79]. Upon entering a susceptible plant cell, the coat protein (CP) is removed (virion encapsidation) and the genetic information is translated. The RNA genome is a direct template for translation using the cap-independent internal ribosome entry site (IRES) for initiation [78]. Two products of translation are produced as precursors to functional proteins: a long polyprotein as a result of the translation of the entire genome, and a short polyprotein P3N-PIPO produced via a ribosomal frameshift. After translation, polyproteins are subjected to proteolytic processing by three self-encoded proteases to yield mature proteins [78][79][80].
The SMV infection cycle consists of several major steps: the entry, uncoating, translation, replication, and cell-to-cell and long-distance movement. Shortly after translation, the viral genome is replicated by NIb, an RNA-dependent RNA polymerase (RdRp), in association with cytoplasmic membranes that create a specific micro-environment to protect the viral genome from silencing [78,80,81]. The copied molecules are coated (virion assembly) and some of the viral particles move into the neighbor cells through plasmodesmata (PD). The intercellular passage requires the assistance of the CI-formed, cone-shaped structures that can increase the plasmodesmata size exclusion limit (SEL). The formation of the CI conical structures that anchor to and extend through the PD is mediated by the P3N-PIPO protein [82]. In addition to CI, CP, and P3N-PIPO, several other viral proteins, including HC-Pro and VPg, may also play a role in viral transport through plasmodesmata [82][83][84]. Long-distance movement occurs when the virus spreads through the vascular system and can infect cells located far from the initial infection point (systemic infection) [80,84,85]. However, this phloem-dependent movement is poorly understood.
SMV resistance-mediated signaling pathways in soybean remain largely unknown. Babu et al. (2008) [86] used microarray technology to detect expression changes of the 'Williams 82' (PI 518671) SMV-susceptible (rsv) genome infected by the SMV-G2 strain. Many genes that were related to hormone metabolism, cell wall biogenesis, chloroplast functions, and photosynthesis were significantly downregulated at 14 days post-inoculation. The genes involved in defense were upregulated at the late stages, suggesting that the response to SMV was delayed and the plant could not combat the infection. In addition, in research conducted by Chen et al. (2016) [87], small RNAs, degradome, and transcriptome sequencing analyses were used to identify multiple differentially expressed genes and microRNAs in Williams 82 infected by three SMV isolates and PI 96983 (resistant or necrotic to SMV; Rsv1) by SMV-G7. The knock-down of the eukaryotic translation initiation factor 5A (eIF5A) gene diminished necrosis and enhanced viral accumulation, suggesting its essential role in the SMV-G7-induced and Rsv1-mediated signaling pathway. The status of the soybean-SMV interactions has recently been summarized by Liu et al. (2016) [79].

Symptom Induction by SMV
The symptoms induced by SMV depend on many factors, including the host genotype, virus strain, plant age at infection, and environment [88]. Genetic resistance to SMV seems to be the most efficient strategy to control the disease [5,19,89]. The first and most important step in producing soybeans with SMV resistance is to identify resistant germplasm and then to study the molecular mechanisms [90,91]. Individual cultivar reactions to SMV strains are classified into three main responses: resistant (R, symptomless); necrotic (N, systemic necrosis); or susceptible (S, mosaic) ( Figure 1) [13,31].
inoculation. The genes involved in defense were upregulated at the late stages, suggesting that the response to SMV was delayed and the plant could not combat the infection. In addition, in research conducted by Chen et al. (2016) [87], small RNAs, degradome, and transcriptome sequencing analyses were used to identify multiple differentially expressed genes and microRNAs in Williams 82 infected by three SMV isolates and PI 96983 (resistant or necrotic to SMV; Rsv1) by SMV-G7. The knock-down of the eukaryotic translation initiation factor 5A (eIF5A) gene diminished necrosis and enhanced viral accumulation, suggesting its essential role in the SMV-G7-induced and Rsv1-mediated signaling pathway. The status of the soybean-SMV interactions has recently been summarized by Liu et al. (2016) [79].

Symptom Induction by SMV
The symptoms induced by SMV depend on many factors, including the host genotype, virus strain, plant age at infection, and environment [88]. Genetic resistance to SMV seems to be the most efficient strategy to control the disease [5,19,89]. The first and most important step in producing soybeans with SMV resistance is to identify resistant germplasm and then to study the molecular mechanisms [90,91]. Individual cultivar reactions to SMV strains are classified into three main responses: resistant (R, symptomless); necrotic (N, systemic necrosis); or susceptible (S, mosaic) ( Figure 1) [13,31]. Resistant soybeans have no disease symptoms and are undistinguishable from noninfected plants. A host plant is resistant if it can block viral replication and movement (cell-to-cell or long-distance) and, therefore, cannot be detected visually or by using the enzyme-linked immunosorbent assay (ELISA) test [92]. A resistant reaction restricts the Resistant soybeans have no disease symptoms and are undistinguishable from noninfected plants. A host plant is resistant if it can block viral replication and movement (cell-to-cell or long-distance) and, therefore, cannot be detected visually or by using the enzyme-linked immunosorbent assay (ELISA) test [92]. A resistant reaction restricts the virus to the infection site and prevents disease dispersion. In a resistant reaction, the pathogen is described as avirulent, and the virus-host interaction is incompatible [80].
Necrosis is a hypersensitive (HR) protective system activated in response to SMV. In general, the necrotic reaction causes necrotic spots on inoculated leaves, the yellow and brown discoloration of upper leaves and veins, stunting of the entire plant, browning of stems and petioles, defoliation, and plant death as a system of reduction in disease within the crop [56,81,93]. The signaling pathway causing the necrotic reaction remains largely unknown. Some soybean accessions develop necrotic symptoms in a short time after infection, leading to plant death in the V2 developmental stage [56]. In necrotic plants, viral replication and cell-to-cell or long-distance movement occur at a lower level. The necrotic Viruses 2022, 14, 1122 6 of 29 reaction can be classified into three subgroups: local necrosis, where virus proliferation is restricted to the initially infected cell (necrotic spots on the leaves, stems, and petioles); systemic necrosis (necrotic spots on upper leaves, browning of leaf veins, petioles, and stems); and stem-tip necrosis, when the reaction is not restricted to the initial infection site (bud blight, stunting of the plants, defoliation, and plant death) [40,81]. Necrosis is associated with environmental temperature and the heterozygous state of resistance alleles; however, local and systemic necrosis may also be observed when a homozygous resistance gene interacts with specific SMV strains [31,32,34,53,93,94]. High temperatures exceeding 37 • C tend to mask the expression of mosaic symptoms in both homozygous and heterozygous plants [93]. There has been disagreement among researchers in terms of the classification of necrotic responses as either susceptible or resistant. Some researchers consider necrosis as a susceptible reaction due to significant yield losses or plant death [95][96][97][98]. However, results from many genetic studies suggest that necrotic plants should be classified as resistant when evaluating segregating populations due to the phenotypic expression of heterozygotes [32,94,99]. In addition, the necrotic reaction is always associated with the presence of resistance genes/alleles interacting with specific SMV strains. Therefore, necrotic plants should be classified with the resistant class as they contain resistance genes regardless of the symptom expression [94,99].
Susceptible soybean genotypes develop a characteristic stunted growth and mosaic leaf pattern, leading to curling leaves, twisting leaf edges downward, and seed-coat mottling [100,101]. In susceptible plants, a substantial yield reduction has been reported when SMV infection occurs at the early stages of plant development [6,35]. A host plant is considered fully susceptible if the virus can successfully complete replication and cell-to-cell and long-distance movement [102]. When the disease symptoms are present, the pathogen is described as virulent, and the virus-host reaction is compatible [80]. There are two types of susceptible phenotypes distinguished as early and late mosaics. Early mosaic is the most common response observed as soon as 5-7 days after inoculation. Late mosaic has a 2-3 weeks delay in symptom expression [26]. Mosaic symptoms are more severe at lower temperatures, whereas higher temperatures alter stem-tip necrosis to mosaic symptoms or mask the mosaic symptoms [93,103,104].

SMV Resistance Genes
Early genetic studies on host resistance to SMV were conducted in Japan [96]; however, there has been significant progress since the establishment of the SMV classification system by Cho and Goodman (1979) [13] for SMV resistance genes and their alleles [5,32,34,37,56,60,[66][67][68]70,105,106]. Soybean resistance to SMV is controlled by four nuclear genes, and there is no significant cytoplasmic effect on the SMV reaction as determined by reciprocal crosses [31,53,65].
The identification of resistance genes (R-genes) in a plant genome is based on inheritance and allelism studies [37,54,55,58,64,66]. Recombinant inbred line (RIL) populations, developed by crossing a resistant genotype and a susceptible genotype, e.g., 'Essex' (PI 548667) or 'Lee 68' (PI 559369), are the most common type of a bi-parental mapping population used to determine the source, location, and number of genetic resistance genes. On the other hand, an allelism test crosses the resistant genotype in question with a set of several differential resistant genotypes [32]. Observed segregations of F 2 and F 2:3 lines inoculated with SMV have been used for testing goodness-of-fit to the expected genetic ratios by the Chi-square test (χ 2 ). The presence of recombinant plants indicates that the resistant parent in question carries different alleles for SMV resistance; otherwise, genes in both parents are allelic at the same locus if there is no segregation in the progeny. If a genotype in question turns out to be allelic to a reported locus, it is necessary to compare the pattern of reaction symptoms of the newly identified allele with those of known alleles to determine if the SMV reaction of the new allele is unique. Good examples of such a genetic strategy were tested in 'Tousan 140' [66], 'Zao 18' [68], 'Raiden' (PI 360844) [55], Suweon 97 [54], and 'J05' [67]. The results can be confirmed through molecular studies by applying genetic markers such as simple sequence repeat (SSR) or single-nucleotide polymorphism (SNP) that are positioned within or nearby the gene of interest [67,69].
The classical genetic analysis for virus resistance is based on the gene-for-gene hypothesis established by Flor (1955) [107]. Typically, a single dominant resistance gene of the host plant specifically interacts with the product of a single dominant avirulence gene in the pathogen. A gene-for-gene relationship exists when the presence of a gene in one population is contingent on the continued presence of a gene in another population, and when the interaction between the two genes leads to a single phenotypic expression by which the presence or absence of the relevant gene in either organism may be recognized [108]. There are different perceptions on the interaction of SMV and soybean. In China, resistance to different SMV strains was presumed to be conditioned by different genes, and there was no pleiotropic effect on a resistance gene [109]; however, in the U.S., SMV resistance genes were shown to be pleiotropic. Therefore, two different reactions to separate SMV strains may be controlled by the same host gene [110].
To date, four independent loci for SMV resistance, Rsv1, Rsv3, Rsv4, and Rsv5, have been identified [53,59,63,111] (Table 1). Most resistant soybean genotypes carry a single dominant gene [34,47,106]; however, some genotypes contain two genes for resistance [60,[66][67][68]70,71]. One landrace '8101', collected from South Korea, was identified to carry three SMV R-genes [72]. Pyramiding resistance genes is a technique commonly used in breeding programs to control plant pathogens. However, this was not the case for SMV due to concerns about the possible occurrence of a superior SMV isolate with the ability to overcome all R genes. Recent molecular analyses revealed that the resistance breakdown mechanisms for different resistance genes are diverse and simultaneous mutations in multi-viral proteins are required for SMV to gain virulence to multiple R genes [22,112]. Therefore, the chance for the occurrence of a superior SMV isolate that can overcome all resistance genes is extremely low, and pyramiding R genes may be a sound option. Indeed, the pyramiding of three R-genes using marker-assisted selection (MAS) has been reported, and potential lines may be released to support breeding programs [61,64,89,113].
The first allele, Rsv, was discovered by Kiihl and Hartwig (1979) [53] in PI 96983, collected from South Korea and reassigned later as Rsv1 by Chen et al. (1991) [31]. PI 96983 was crossed to a susceptible cultivar, and an F 2 population indicated a dominant nature of Rsv1. The Rsv1 allele in PI 96983 showed greater effectiveness in suppressing virus development when evaluated by grafting [53]. Buffalo exhibited the same reaction as PI 96983 to different strains. Isogenic lines derived from 'Williams (6)' × Buffalo exhibited two types of reactions: resistant to G1 and G5, and resistant to G1 but susceptible to G5. Williams was reported to be susceptible to all SMV strains [117,118]. Resistance to SMV-G1 and G7 strains in Buffalo was under monogenic control at the Rsv1 locus [119]; therefore, it was proposed that the resistance gene in Buffalo may be a different allele at the Rsv1 locus [120,121]. In the present SMV classification system, Buffalo was replaced by PI 96983, which carries a single dominant allele at the same locus [52].
The Ogden soybean cultivar carries an allele designated as Rsv1-t, previously known as rsv t , and it shows resistance to SMV strains G1, G2, and G4-G6, but necrosis when inoculated with G3 and G7 [31,47,53,66,69,71]. The ratio of necrotic to resistant plants was 1:1 in the F 2 populations from Ogden × susceptible genotype [31,53], indicating the incomplete dominant nature of Rsv1-t. The heterozygous plants at Rsv1-t permitted greater virus development than the homozygotes. Some plants with heterozygous alleles displayed necrosis and were grouped with resistant classes [53].
The soybean cultivar York carries the Rsv1-y allele, and it is resistant to the less virulent strains G1-G3 and susceptible or necrotic to the more virulent strains G4-G7 [13,31,33]. Resistance to SMV in York is controlled by a single dominant gene [33], which was later designated as Rsv1-y [31]. When York was crossed to a susceptible genotype, nearly onefourth of the plants observed in the F 2 populations were necrotic. Additionally, there was a low level (about 0.6-1.3%) of necrotic tissue, but no susceptible lines were detected in the crosses York × PI 96983, York × Marshall, and York × Ogden [31]. Silva et al. (2004) [122] reported the soybean cultivar FT-10 to carry an allele at the Rsv1 locus, probably originating from Davis, and a gene symbol of Rsv1-d was assigned. According to its pedigree, Rsv1-d should be the same as Rsv1-y, because Davis and York showed the same symptoms to G1 to G7 and presumably carry the same resistance gene [13,111] derived from 'Dorman' (PI 548653) [65]. The Rsv1-y in York was later proved to be a separate locus, although tightly linked to Rsv1 (2.2 cM), and assigned a new locus Rsv5 [65]. Chen et al. (1991) [31] confirmed that the genes in Marshall and Kwanggyo are allelic to the Rsv1 locus and designated them as Rsv1-m and Rsv1-k, respectively. Marshall contains the single dominant allele Rsv1-m, and confers resistance to strains G1, G4, and G5, and necrosis to the rest of the strains [31,34]. At the homozygous state, the Rsv1-m in Marshall confers complete resistance or necrosis to SMV strains, but Rsv1-k in Kwanggyo displays a mixture of symptoms of resistance and delayed necrosis to some strains. This type of necrosis is characterized by systemic lesions and stem browning, which is different from typical stem-tip necrosis ( Figure 1). Rsv1-k confers resistance to SMV-G1 through G4, and necrosis to SMV-G5 through G7. When Kwanggyo was crossed to a susceptible genotype, there was an equal number of necrotic and resistant plants in the F 2 population indicating the partial dominance of Rsv1-k [31]. Additionally, the viral strain could not be recovered from the necrotic plants because viral replication and movement were restricted by the death of cells surrounding the infection site [31,123].
The Raiden cultivar contains a single Rsv1-r allele and displays resistance to SMV-G1 through G4 and G7, but necrosis to SMV-G5 and G6 [55,111]. Raiden and 'L88-8431', a Williams back-crossed (BC 5 ) isogenic line with SMV resistance derived from Raiden, were crossed to a susceptible soybean cultivar. The F 2 populations and F 2:3 progenies from both crosses did not show any segregation for susceptibility and proved that the resistance to SMV in Raiden and 'L88-8431' was controlled by a single dominant gene allelic to Rsv1 [55].
LR1 contains the Rsv1-s allele and confers resistance to SMV-G1 through G4 and G7, and systemic necrosis to SMV-G5 and G6 [34,37]. The LR1 line, derived from Essex × PI 486355, was crossed with Lee 68, Ogden, and York, and results showed a single dominant gene at the Rsv-1 locus. The Rsv1-s confers partial resistance, and its expression is genedosage dependent, where homozygosity conferred resistance and heterozygosity conferred systemic necrosis [37].
PI 507389, released as the soybean cultivar 'Tousan 50', contains the Rsv1-n allelic to Rsv1 and does not show any resistance, but shows necrosis when infected with G1, G2, G5, and G6 [56]. The necrosis allele Rsv1-n is co-dominant with the allele for susceptibility, and the heterozygotes showed a mixed phenotype of necrosis and a susceptible reaction (N/S). The resistance allele became recessive to the susceptible allele as the mixed N/S reaction shifted to susceptible at a later stage in response to more virulent strains [56,124].
Suweon 97 contains the most valuable Rsv1-h allele conferring resistance to all SMV strains G1-G7 and G7A [54]. Inheritance and allelic studies showed that the Rsv1-h allele displays complete dominance at the Rsv1 locus [54].
Recently, a single dominant allele Rsv1-c has been discovered in 'Corsica' (PI 559931) that brings different patterns of responses to SMV strains than other genotypes with known genes. The unique pattern is characterized by early resistance (ER) at the seedling stage to G2, G5, and G7 strains, but susceptibility to G1, G3, and G6 [57].
In summary, the alleles at the Rsv1 locus are diverse, abundant, and widely distributed [34,57,111]. Among the ten alleles of the Rsv1 locus, Rsv1-y and Rsv1-k are most frequent in diverse germplasm [106,125]. The alleles of the Rsv1 locus exhibit partial or complete dominance and confer resistance to some, but not all, SMV strains. The alleles at the Rsv1 locus can be grouped into several distinct classes: (A) alleles conferring resistance to most or all SMV strains, such as Rsv1-h in Suweon 97 and Rsv1 in PI 96983; (B) alleles conferring both resistance and necrosis such as Rsv1-t, Rsv1-m, Rsv1-k, and Rsv1-s alleles in Ogden, Marshall, Kwanggyo, Raiden, and LR1, respectively; (C) alleles conferring resistance or a mosaic reaction as Rsv1-y in York and Rsv1-c in Corsica; and (D) alleles conferring necrosis and a mosaic reaction as Rsv1-n in PI 507389. Genetic studies on Rsv1 demonstrated that alleles with necrotic symptoms to specific SMV strains in the homozygous state are dominant to alleles resistant or susceptible to the same strain, and alleles resistant in the homozygous state often exhibit necrosis when they occur in a heterozygote with a susceptible allele [32]. The necrotic allele (Rsv1-n) is co-dominant with the allele for susceptibility (rsv).
Later, four allele-derived SNPs were found in N11PF fragments amplified from the SCAR primer and validated in two F 2 populations of PI 96983 × Lee 68, and 'Myeongjunamulkong' × Lee 68 [127]. Two other SNP markers were developed from the L20a fragment, a Toll and Interleukin-1 Receptor (TIR)-Nucleotide-Binding Site (NBS) genomic sequence and mapped close to Rsv1 [130]. In another study, Rsv1 in the J05 soybean was mapped using the F 2 population derived from a cross of J05 (R) × Essex (S). The Rsv1 locus was flanked by Sat_154 and Satt510 with 0.5 and 2.3 cM, respectively [69]. Hayes et al. (2004) [130] reported six clones on chr. 13 (LG F) around the Rsv1 locus associated with SMV resistance, and three of them were sequenced (GenBank Accession No. AY518517-AY518519). In a population derived from PI 96983 × Lee 68, one of the six clones, the 3gG2 gene, was co-segregating with Rsv1 and the distance between 3gG2 and Rsv1 was 0 cM; hence, 3gG2 (Wm82.a2.v1: Glyma.13g190800; Gm13: 30,426,359-30,430,201) became a strong candidate gene for the Rsv1 locus. This 3390 bp gene encodes a protein product similar to non-N-terminal (non-TIR) NBS-LRR, common for disease resistance genes. In addition to PI 96983, Marshall (Rsv1-m) and Ogden (Rsv1-t) were reported to contain the 3gG2 gene [69,130].

Discovery and Rejection of Rsv2 Locus
Based on a genetic analysis, the soybean line 'OX670' was identified as carrying a new SMV resistance gene independent of Rsv1. The 15R:1S segregation ratio observed in the F 2 population of OX670 (R) × 'L78-379' (S) indicated that the gene in OX670 is independent of the Rsv1 locus, and, therefore, the Rsv2 symbol was assigned to SMV resistance in OX670 [145]. However, it was later presumed that this novel resistance gene was derived from Raiden [55,146]. Based on allelism tests of Raiden and L88-8431 (Williams isogenic line carrying the R-gene from Raiden), SMV resistance was confirmed to be at the Rsv1 locus with partial dominance, and named as the Rsv1-r allele [55]. Later, it was documented that OX670 carries the two SMV resistance loci Rsv1 and Rsv3, derived from Raiden and 'Harosoy' (PI 548573), respectively; therefore, the Rsv2 in Raiden and its ancestors do not exist [5,60]. In another study, the reaction pattern to different SMV strains and genetic modes conferred by the gene from Raiden were similar to those conferred by the Rsv1-s allele in the breeding line LR1 derived from PI 486355 [37,110]. PI 486355 was shown to contain two resistance genes: Rsv1-s and Rsv4. Both alleles, Rsv1-r in Raiden and Rsv1-s in LR1, confer resistance to SMV-G1 through G4 and G7, but necrosis to G5 and G6 ( Table 1). The SMV reaction pattern and the genetic behavior of Rsv1-r are like those of other Rsv1 alleles. It is highly likely that the Rsv1-r in Raiden and Rsv1-s in PI 486355 originated from the same source. The Rsv1-r allele is gene-dosage dependent, because it confers resistance in the homozygous state and systemic necrosis in heterozygotes infected by SMV-G7 [37,55,110].
Rsv3 was first identified as an independent and dominant gene in OX686, conditioning stem-tip necrosis to SMV-G1 [103]. The OX686 line was derived from a 'Columbia' × Harosoy cross. Columbia was presumed to carry Rsv3 and Rsv4, conferring resistance to all SMV strains [71], while Harosoy carries resistance to SMV-G5 to G7 at the Rsv3 locus, but is susceptible to G1 through to G4. Therefore, it is most likely that Rsv3 in OX686 was derived from Columbia rather than Harosoy. The Rsv3 locus in Harosoy is partially dominant, giving rise to a necrotic reaction in heterozygous plants infected by SMV-G7 [60]. The pedigree analysis of Harosoy and Hardee indicated that the resistance gene Rsv3 in both genotypes was derived from Chinese accessions, but their relationship is unknown. They showed the same reaction pattern to SMV strains, but different genetic behavior, being allelic at the same locus [60]. Buss et al. (1999) [58] reported that L29, a derivative of Hardee, exhibited resistance to G5, G6, and G7, but susceptibility to G1 through to G4 (Table 1). Inheritance and allelism studies showed that L29 carries a new single, completely dominant allele at the Rsv3 locus. Based on the same strategy, PI 61944 was reported to contain a different, single, dominant Rsv3 allele with a distinct symptom pattern of necrosis to SMV-G1 and resistance to G7 [61].   [62] reported two new alleles of the Rsv3 locus, Rsv3-c in South Korean PI 399091, and Rsv3-h in a Chinese land race PI 61947. The Rsv3-c gives a susceptible reaction to G1, G2, and G6 strains, but resistance to G3, G5, and G7, whereas Rsv3-h confers necrosis to G1 and G2, and resistance to G3 to G7. The CI protein is found to be the effector of Rsv3-mediated resistance [18,147,148].

Mapping of the Rsv3 Locus
The Rsv3 gene was mapped on chr. 14 (LG B2) in two crosses of L29 (R) × Lee 68 (S), and Tousan 140 (R) × Lee 68 (S) using data collected from the F 2 generations. Primarily, the Rsv3 gene was flanked by A519F/R at 0.9 cM and M3Satt at 0.8 cM (Table 2). Additionally, the SSR marker Satt063 was mapped at the same side of A519 with a distance of 6.5 cM to Rsv3 [138]. Later, the locus was mapped in the J05 soybean cultivar using an F 2 population derived from a cross with Essex (S). Two SRR markers, Sat_424 and Satt726, were closely linked with Rsv3 with a distance of 1.5 and 2.0 cM, respectively [69] (Table 2).

Rsv4 Locus Confers Resistance to Most or All SMV Strains
Four alleles at the Rsv4 locus have been identified in V94-5152, PI 88788, and 'Beeson' (PI 548510) (Rsv4-b) [5,57,63,71,160]. The Rsv4 alleles confer complete dominant resistance to all SMV strains, but in some cases may show early resistance (ER) at the seedling stage and the expression of delayed mild susceptibility with mosaic symptoms at a late stage [5,63]. Unlike Rsv1 and Rsv3, no hypersensitive reaction has been observed, suggesting that the Rsv4 locus may have unique molecular mechanisms [5,71,113,156]. Due to a wide spectrum of resistance to SMV strains, there is great interest in pyramiding the Rsv4 locus with Rsv1 and Rsv3 loci as a strategy to defend against infections caused by multiple SMV strains [32].
The first Rsv4 allele was found in PI 486355 (Rsv1Rsv4) [37]. Both genes were isolated in F 3:4 -derived lines from the cross of PI 486355 (R) × Essex (S). The 'D26' soybean line, later named 'LR2', was advanced to F 6:7 and released as V94-5152. V94-5152 carries a single Rsv4 gene conferring resistance to all strains [37,63,70]. Later, another allele at the Rsv4 locus was identified in PI 88788 [5]. This resistance was confirmed in crosses PI 88788 × LR2 and PI 88788 × V94-5152, and reported to be controlled by a single dominant Rsv4 gene. Shakiba et al. (2011Shakiba et al. ( , 2013 [57,160] proposed a new Rsv4-b allele in the soybean cultivar Beeson, displaying a unique resistance pattern when compared with other Rsv4 alleles: early resistance to strains G1, G2, and G6, resistance to G5 and G7, and susceptibility to G3 (Table 1). Recently, inheritance and allelic studies revealed that resistance to SMV in PI 438307 is controlled by a single dominant Rsv4-v allele. PI 438307 is resistant to SMV-G1 through to G6, and resistant at seedling stages to SMV-G7 [64]. The P3 protein was found to be the effector of Rsv4-mediated resistance [22,161].

Novel Locus Rsv5
Rsv1-y in the York soybean has been classified as an allele of the Rsv1 locus; however, several phenomena recently raised a question of whether the Rsv1-y allele belongs to the Rsv1 or if it is a distinct, but tightly linked, locus. According to the study by Shi et al. (2008b) [131], a PCR-based marker Rsv1-f/r for the detection of the Rsv1 candidate gene 3gG2 (Wm82.a2.v1: Glyma.13g190800), completely linked to Rsv1, could amplify a specific sequence from soybean genotypes carrying various Rsv1 alleles, except for Rsv1-y. Recently, Yang et al. (2013) [133] concluded that there might be one or two dominant R-genes tightly flanking the Rsv1 locus by performing a cross of PI 96983 (R) × Nannong 1138-2 (S). The potential Rsc-pm gene confers resistance to the Chinese SMV strains SC3, SC6, and SC17, and was positioned between BARCSOYSSR_13_1128 and BARCSOYSSR_13_1136. The other gene, Rsc-ps, confers resistance to SMV-SC7, and was reported between BARCSOYSSR_13_1140 and BARCSOYSSR_13_1155 ( Table 2). The Rsv1 locus is located at the genomic region on the long arm of chr. 13 (LG F), where multiple R genes have previously been reported [121,130] and is tightly linked to a cluster of genes containing an NBS-LRR (www.soybase.org, accessed on 18 May 2022). This area of the chromosome is extremely complex, conferring resistance not only to SMV, but also to soybean aphids [163] and other plant pathogens, e.g., Phytophthora [164] and Fusarium [165]. The Rsv1 locus seems to be complex itself, with the possibility of having a variety of at least ten different copies of the same gene. In a recent study by Klepadlo et al. (2017b) [65], two alleles of the Rsv1 locus were analyzed by crossing PI 96983 (Rsv1) and York (Rsv1-y) to determine whether Rsv1 and Rsv1-y belong to the same locus. The occurrence of segregating and susceptible lines indicated tight linkage between the two SMV R-genes, with an estimated genetic distance of 2.2 cM. A new symbol of Rsv5 was assigned for the resistance gene in York to replace the original Rsv1-y allele name [65]. This locus separation helps explain why York is susceptible to G5-7, a rare pattern for Rsv1 alleles. The P3 protein has been considered as a possible effector of Rsv5-mediated resistance [21].

Gene Combinations Reduce Vulnerability of SMV Infection
The inheritance of SMV resistance has been extensively studied for many years. In over 80% of reported cultivars, resistance is controlled by a single gene [31,37,94,146,166]. Only a few soybean accessions carry two resistance genes in different combinations (Rsv1Rsv3, Rsv1Rsv4, or Rsv3Rsv4) that interact in a complementary fashion, giving resistance to SMV-G1 through G7 (Table 1). Up to now, only one soybean genotype '8101' has been found to carry three resistance genes, giving rise to resistance to all SMV strains [72].
Eight genotypes, namely, OX670, 'Hourei', Tousan 140, J05, 'Jindou 1', Zao 18, PI 486355, and Columbia, carry two resistance genes in various combinations of Rsv1, Rsv3, and Rsv4 [60,[66][67][68][69]71]. The presence of two genes for SMV resistance substantially reduces plant vulnerability to infection by various strains/isolates [70,89]. The Rsv1 allele in OX670 is derived from Raiden and the Rsv3 allele from Harosoy [60]. The Rsv1 in Tousan 140 was similar to Rsv1-y for the same resistance reaction to G1 and G3, local necrotic lesion to G2, and susceptibility to G5 through G7, and Rsv3 for the same resistance to G5-G7 and susceptibility to G1-G3 [66]. The Rsv1 and Rsv3 in Hourei were similar to the two genes in Tousan 140 that condition similar reactions to G1 and G7, but sources of the two genes are not known. Rsv1 allele in Tousan 140 gives resistance to G1 and G7 strains, whereas Rsv3 allele confers resistance to G7, but susceptibility to G1 [66]. Rsv1 in Zao 18 is similar to Rsv1-y (Rsv5) conditioning resistance to G1 and susceptibility to G7, and Rsv3 conferring resistance to G7 and susceptibility to G1 [68]. The soybean cultivar J05 was also identified carrying Rsv1 and Rsv3 [67,69], but it is not known whether or not the two genes are new alleles at Rsv1 or Rsv3. The digenic combination of Rsv1 and Rsv3 provides a wide potential for SMV resistance. Jindou, a soybean cultivar developed in China, is resistant to SMV-G1 through to G7, and was identified to contain resistance genes of Rsv1-y (Rsv5) and Rsv3 [167].
Only one soybean accession, PI 486355, was reported to carry Rsv1-s and Rsv4 [37,63,70]. PI 486355, originally named SS74185 and collected in South Korea, is resistant to SMV G1 through to G7 in the US and 23 SMV isolates from China [50]. Rsv1-s confers resistance with partial dominance to G1-G4 and G7, but necrosis to G5 and G6, which is similar to the reaction pattern of Rsv1-r in Raiden. In contrast, Rsv4 exhibited complete dominance and confers resistance to all SMV strains (G1-G7). These two genes were separated in two different F-3:4 derived lines, LR1 and LR2, from a cross between PI 486355 (R) and Essex (S) [37].
Columbia contains two dominant genes, Rsv3 and Rsv4, and displays resistance to all SMV strains [59,71]. In populations of Columbia × Lee 68, a unique response to SMV-G1 has been designated to be resistant at the early seedling stage with delayed mild mosaic symptoms approximately three weeks after initial inoculation. The same populations remained resistant to SMV-G7. The Rsv3 in Columbia and OX686 is the same gene that confers resistance to G7, but necrosis to G1. The Rsv4 gene in Columbia confers resistance to G7, and Rsv3 and Rsv4 interact in a complementary fashion and provide complete resistance to G1-G3 and G5-G7, and stem-tip necrosis to G4 [71].
A significant genetic source of SMV resistance was also mapped on chr. 6 (LG C2). Yang and Gai (2011) [175] crossed 'RN-9' (R) × '7605' (S) to study the resistance to the Chinese SMV-SC15 strain. Results indicated that a single dominant gene, designated as Rsc15, conferred the SMV resistance, and it was located between Sat_213 and Sat_286 with distances of 8.0 and 6.6 cM, respectively. The candidate gene Glyma.18g154900 in RN-9 encodes the cytochrome protein GmCYB5 targeting the P3 protein of SMV to inhibit its proliferation (Luan et al., 2019) [183]. This novel resistance gene has never been found in germplasm infected by American SMV strains. Thus, this gene has not been found or reported in the US germplasm.

Genetic Resources and Breeding for SMV Resistance
A set of NILs is a powerful genetic resource for the identification of genes or QTLs associated with a trait of interest. Each NIL carries a genomic segment associated with the trait of interest from a donor parent and is incorporated into the genomic background of the recurrent parent. Constructing a set of NILs requires a bi-parental hybridization followed by repetitive backcrossing and an extensive molecular study [184]. Two sets of isogenic lines for SMV resistance alleles were constructed (Table 5). Williams isogenic lines (L-series) were developed by Richard L. Bernard, University of Illinois, and were derived from backcrossing Williams (S) with 10 different resistant lines resulting in isogenic lines carrying different alleles of the Rsv1 and Rsv3 loci [118]. Essex isogenic lines (V-series) carrying different alleles of Rsv1, Rsv3, and Rsv4 loci were developed by Glenn Buss, Dep. of Crop and Soil Environmental Sciences, Virginia Polytechnic Institute and State University. Essex isogenic lines showed resistant, necrotic, or susceptible reactions when infected by the same SMV strain from G1 to G7. For example, infection by SMV-G1 provided resistance in 'V94-3971' and 'V97-9003', necrosis in 'V262', and susceptibility in 'V229'. In this case, induced symptoms did not depend on the virus strain, but instead on the host genotype [63,113,118]. In contrast, a specific genotype could exhibit multiple symptoms when infected with different strains. For example, V94-3971 shows resistance to G1 but necrosis to G7, V229 exhibits susceptibility to G1 but resistance to G7, and V262 is necrotic to G1 but susceptible to G7. These isolines with different SMV resistance alleles in the same genetic backgrounds could allow for future investigations on specific resistance gene-SMV strain interactions. These isolines may also be useful in future studies of the gene-for-gene concept in the soybean-SMV system. Several gene-pyramided lines were developed utilizing MAS for the incorporation of SMV resistance genes at different loci to develop cultivars with durable resistance [89]. Saghai Maroof et al. (2008) [113] pyramided Rsv1, Rsv3, and Rsv4 using three different Essex isogenic lines, each carrying a single R-gene. Two-gene and three-gene isogenic lines with combinations of Rsv1Rsv3, Rsv1Rsv4, and Rsv1Rsv3Rsv4 acted in a complementary manner conferring resistance against all SMV strains, whereas isogenic lines of Rsv3Rsv4 displayed a late susceptible reaction to the selected SMV strains not observed in the donor sources. Shi et al. (2009) [89] also pyramided the same genes from the cross J05 (Rsv1Rsv3) × V94-5152 (Rsv4) using eight markers: Sat_154, Satt510, and Rsv1-f/r for Rsv1; Satt560 and Satt063 for Rsv3; and Satt266, AI856415, and AI856415-g for Rsv4. Five F 4:5 lines were identified to be homozygous for all eight marker alleles and presumably carried all three SMV resistance genes that potentially provide multiple and durable resistance to SMV. Cervantes (2012) [185] performed the cross PI 96983 (Rsv1) × Columbia (Rsv3Rsv4) and developed 11 pyramided lines with three resistance genes homozygous and heterozygous on each locus. Three SSR markers, Satt510, Sct_064, and Satt296, were mainly used for screening Rsv1, Rsv3, and Rsv4 loci. Three SMV resistance genes, Rsc4, Rsc8, and Rsc14Q, were identified and mapped on soybean chromosomes 14, 2, and 13 from Dabaima, Kefeng 1, and Qihuang 1 cultivars, respectively [186]. The soybean cultivar Nannong 1138-2 is widely grown in the Yangtze River valley of China. The crosses were performed between (Qihuang 1 × Kefeng 1) and (Dabaima × Nannong 1138-2). Ten SSR markers linked to these three resistance loci were used for gene pyramiding, and the progenies were evaluated by inoculations with 21 SMV strains from China. Five F7 homozygous pyramided families exhibited resistance to all 21 strains [186].
SMV resistance was also successfully introgressed into high-yielding backgrounds in breeding programs. Harosoy has been used as a donor of resistance to Japanese strains SMV-C and D, resulting in resistant varieties such as 'Fukuibuki'. Harosoy harbors the Rsv3 gene conferring resistance to US SMV-G5 through to G7, and it was suggested that Rsv3 confers resistance to the Japanese SMV-C and D. This resistance was introgressed by the recurrent backcrossing of Fukuibuki (R) × 'Ohsuzu' (S, high yield) with the assistance of MAS. Three years of field trials showed that the SVM-resistant breeding line 'Tohoku 169' was equivalent to Ohsuzu for important agronomic traits, such as seed size, maturity date, and yield [187].
Recently, there have been several transgenic soybean lines with SMV resistance that have been developed by overexpressing resistance-related genes (GmKR3 and GmAKT2) or expressing viral nucleic acid sequences of genetic elements (P1, CP, and HC-pro genes) for pathogen-derived resistance (PRD) [188][189][190][191][192][193][194]. Additionally, an RNAi technique has been successfully used to develop transgenic soybean lines with SMV resistance [189,[195][196][197]. Specifically, a recent study conducted by Gao et al. (2019) [197] used a soybean endogenous gene, eIF4E, which is a major susceptibility factor for some RNA viruses in the RNAi technique. It confirmed that transgenic soybean lines with the silencing of the eIF4E gene had an enhanced and broad resistance to SMV-SC3, SC7, SC15, SC18, and SMV-R. In addition, transgenic soybean lines expressing P3 and HC-Pro genes also demonstrated a strong and stable resistance to SMV-SC3, SC7, SC15, SC18, and SMV-R, with a significant yield potential [195,196].

Presence and Future of SMV Resistance Research
The soybean is one of the most important commodity crops grown on a large scale around the world. Given the historical perspective of research on the physiological and genetic characteristics of disease caused by SMV, it is fundamental to look at the achievements of the past century, from the disease discovery in 1915, followed by the beginning of genetic studies in 1979, to its present status of genetic mapping and gene pyramiding. Great progress has been achieved in the past 35 years in our understanding of genetics and the molecular basis of resistance mechanisms. These discoveries have been advanced considerably by the whole-genome shotgun sequencing of Williams 82 [198]. During the earlier years of genetic studies, we gained much information on SMV resistance genes and their correlations with visible symptoms. Based on distinct reaction patterns of a given genotype to specific SMV strains, we can predict the host resistance gene or alleles (Table 1). For example, with resistance to SMV-G1 and mosaic or necrosis to G7, a given soybean accession may carry an R-gene at the Rsv1 locus. In contrast, with a G1-susceptible and G7-resistant reaction, a genotype in question may carry an Rsv3 allele. However, if a given accession shows resistance to all SMV strains, it may carry one of several gene possibilities, such as Rsv1-h, Rsv4, Rsv1Rsv3, Rsv1Rsv4, Rsv3Rsv4, or Rsv1Rsv3Rsv4, and in such cases, allelism and inheritance studies are necessary for the identification of resistance and the confirmation of allelomorphic relationships.
The identification of SMV resistance in diverse soybean accessions is crucial for breeding and production purposes, and the discovery of new resistance genes is likely to continue to provide effective SMV resistance to a broad and ever-changing range of SMV isolates. Although SMV resistance loci have been reported in many soybean genotypes, most of the modern commercial cultivars are susceptible to SMV, particularly to more virulent strains [34,125]. The use of genetic resistance is the primary method of controlling the disease, and with the emergence of new resistance-breaking SMV isolates, the search for new sources of host resistance is necessary to avoid potential vulnerability. Isogenic lines carrying different alleles at the same locus and mixing them up as a synthetic variety based on the distribution changes of SMV strains can be an effective approach to control multiple viral strains. Isogenic and pyramiding lines are also good materials for genetic analysis, gene mapping, gene expression, and cloning.
None of the SMV resistance genes have been cloned yet; however, mapped and fine-mapped regions pinpoint candidate genes that could be transformed soon. Genetic mapping studies have provided information on the approximate location of the genes. The Rsv1 and Rsv5 loci are difficult to map due to the presence of many tightly linked genes in the same genomic region that confer resistance to several other diseases, requiring new techniques to detect and map single genes. The identification of the exact locus is likely to begin a new era of detailed biology research, especially proteomics, to gain knowledge about molecular mechanisms that underlie soybean resistance to SMV. Fine-mapping, together with the availability of the complete soybean genome sequence, is likely to make the molecular cloning of these resistance genes possible, which could largely accelerate breeding programs for the control of SMV using these natural resistance sources.
Recent studies suggest that novel resistance to viral pathogens may be developed through advanced biotechnology [199]. Like other positive-sense RNA viruses, SMV requires many host factors to establish its infection in soybeans. The silencing or mutation of a host factor gene may generate SMV resistance. The recently developed precise genome-editing technology could make this approach accessible to the soybean industry and consumers as it becomes possible to develop transgene-free resistance to SMV. Target host factor genes can be precisely deleted through genetic transformation, followed by transgene removal via traditional breeding. This promise propels ongoing research to characterize molecular SMV-soybean interactions and identify host factors that are essential for SMV infection.
In summary, multiple SMV strains and isolates occur worldwide, causing various reactions in soybeans, including resistance, necrosis, and susceptibility. Resistance is a typical immune reaction without virus entry, multiplication, and movement, resulting in no symptoms of infection or disease. Necrosis could be localized or systemic with limited viral replication and movement, leading to very low levels of detectable virus. Necrosis is characterized as local necrotic spots, vine necrosis, stem browning, defoliation, and stem-tip necrosis, often leading to plant death. Susceptible reactions show typical mosaic, vein clearing, and leaf curling symptoms, with maximum viral replication and movement. SMV strains and isolates are classified into strain groups or pathotypes with different virulence and symptomology on host genotypes with different resistance genes and alleles. There are several classification systems in different countries, where different SMV strains/isolates and differential genotypes are used. Therefore, research is needed to compare and unify different classifications around the world. Genetic resources for SMV resistance are readily available and quite diverse. Several major resistance loci and many alleles have been identified, all of which reside in four chromosomal regions, namely, chrs. 2 (LG D1b), 6 (LG C2), 13 (LG F), and 14 (LG B2). Rsv1 is the most common, with the most diverse alleles in soybean germplasm, predominantly found in the US gene pool. Rsv1 alleles confer resistance with incomplete dominance to less virulent strains but necrotic or susceptible reactions to more virulent strains. In contrast, Rsv3 alleles confer resistance to more virulent strains but susceptible reactions to less virulent strains. Rsv4 confers resistance with complete dominance to all SMV strains only at the early stage. Rsv1-h is the single, most effective gene for SMV resistance. However, multiple allelic combinations such as Rsv1Rsv3, Rsv1Rsv4, and Rsv1Rsv3Rsv4 are available, although rare, naturally in soybean germplasm, and can be achieved through gene pyramiding. Such gene combinations could provide multiple and durable resistance and prevent resistance from breaking down due to the genetic mutation of existing strains or the emergence of new isolates. Rsv5 is a newly discovered locus and tightly linked to Rsv1 in a common region on chromosome 13, where there are many disease resistance genes. The novel gene for SMV resistance on chr. 6 (LG C2) deserves more attention in future research. As with the strain differential systems in different countries, various resistance genes reported from different countries should be tested for allelism or differentiated through molecular approaches. To date, all the SMV resistance genes worldwide to date likely belong to five genetic loci: Rsv1 and Rsv5 on chr. 13 (LG F), Rsv3 on chr. 14 (LG B2), Rsv4 on chr. 2 (LG D1b), and the novel locus Rsc15 on chr. 6 (LG C2). Molecular markers, particularly gene-specific markers, are very useful in the identification, differentiation, and selection of specific resistance genes/alleles of interest in germplasm screening, gene-pyramiding, or breeding and selection for SMV resistance. As the virus constantly evolves with changes of virulence and pathogenicity and possibilities of breaking resistance in the host, collaboration is needed between geneticists and virologists to understand the molecular mechanisms of interactions between the virus and host in preparation for strain mutation and resistance breaking. Breeders need to stay alert and leverage genetic information and technology advancements to take advantage of genetic resources and molecular tools in their breeding effort for resistance, as multiple, durable, and preventable resistance is the ultimate goal.