Detection of rare nematode resistance Rhg1 haplotypes in Glycine soja and a novel Rhg1 α‐SNAP

This study pursued the hypothesis that wild plant germplasm accessions carrying alleles of interest can be identified using available single nucleotide polymorphism (SNP) genotypes for particular alleles of other (unlinked) genes that contribute to the trait of interest. The soybean cyst nematode (SCN, Heterodera glycines [HG]) resistance locus Rhg1 is widely used in farmed soybean [Glycine max (L.) Merr.]. The two known resistance‐conferring haplotypes, rhg1‐a and rhg1‐b, typically contain three or seven to 10 tandemly duplicated Rhg1 segments, respectively. Each Rhg1 repeat carries four genes, including Glyma.18G022500, which encodes unusual isoforms of the vesicle‐trafficking chaperone α‐SNAP. Using SoySNP50K data for NSFRAN07 allele presence, we discovered a new Rhg1 haplotype, rhg1‐ds, in six accessions of wild soybean, Glycine soja Siebold & Zucc. (0.5% of the ∼1,100 G. soja accessions in the USDA collection). The α‐SNAP encoded by rhg1‐ds is unique at an important site of amino acid variation and shares with the rhg1‐a and rhg1‐b α‐SNAP proteins the traits of cytotoxicity and altered N‐ethylmaleimide sensitive factor (NSF) protein interaction. Copy number assays indicate three repeats of rhg1‐ds. G. soja PI 507613 and PI 507623 exhibit resistance to HG type 2.5.7 SCN populations, in part because of contributions from other loci. In a segregating F2 population, rhg1‐b and rhg1‐ds made statistically indistinguishable contributions to resistance to a partially virulent HG type 2.5.7 SCN population. Hence, the unusual multigene copy number variation Rhg1 haplotype was present but rare in ancestral G. soja and was present in accessions that offer multiple traits for SCN resistance breeding. The accessions were initially identified for study based on an unlinked SNP.

and improvement (Tuberosa, 2012). Recent developments, such as single nucleotide polymorphism (SNP) genotyping and next generation sequencing have made it possible to prescreen germplasm for desirable traits through correlation with sequence information (Poland et al., 2012). In some cases, this approach can be extended to wild relatives of crop plants. Novel alleles of loci of agronomic importance may then be subjected to functional characterization. Alleles of interest can be moved into elite varieties via sexual crosses or transgenic approaches.
Soybean [Glycine max (L.) Merr.] is one of the world's most important legume crops, providing a major source of vegetable oil, protein meal for animal feed, and potential sources of renewable energy (http://soystats.com/) (Graham & Vance, 2003;Stacey, 2010. Soybean cyst nematode (SCN) (Heterodera glycines [HG]) is the most economically damaging pathogen of soybean, routinely causing upwards of US$1 billion each year in the United States alone (Davis et al., 2008;Koenning & Wrather, 2010;Mitchum, 2016). Soybean cyst nematode eggs can remain viable for years in cysts, which are recalcitrant to many environmental or chemical conditions, making control of established SCN populations difficult . Resistant varieties and crop rotation are the major methods of SCN control. Glycine soja is the wild ancestor of cultivated soybean, and the more diverse gene pool in G. soja species vs. domestic soybean offers a source of novel alleles or genes for traits of agronomic interest including SCN resistance (Hyten et al., 2006;Liu et al., 2020. The strong-effect quantitative trait locus (QTL) Rhg1 (Resistance to Heterodera glycines 1) is the most used source of SCN resistance, with the rhg1-b haplotype from PI 88788 present in the vast majority of SCN-resistant soybean grown in the United States. Concibido et al., 2004;Donald et al., 2006. Intriguingly, Rhg1 is a tandemly repeated block of four genes, and the number of repeats varies from one in susceptible soybean varieties or three in most rhg1-a haplotypes to nine or 10 copies in rhg1-b haplotypes (Cook et al., 2012). None of the genes in the repeated block are reminiscent of canonical defense or disease resistance genes. Both silencing and overexpression experiments have established a role in SCN resistance for three of the four genes in the Rhg1 locus, one of which (Glyma.18G022500) encodes an α-SNAP (alpha-soluble Nethylmaleimide sensitive factor (NSF) attachment protein) (Cook et al., 2012).
The α-SNAP is a functionally conserved eukaryotic protein that interacts in multimeric complexes with both NSF and soluble NSF attachment protein receptor (SNARE) proteins to mediate vesicular trafficking (Jahn & Scheller, 2006). In particular, α-SNAP and NSF cooperate to promote vesicle trafficking through their disassembly for recycling of the bundled v-and t-SNARE complexes that form during vesicle fusion (Jahn & Scheller, 2006). In soybean, we recently discovered

Core Ideas
• Germplasm carrying useful alleles can be identified using SNP data for other genes. • Glycine soja accessions with potentially valuable SCN resistance were identified. • A novel variant of the Rhg1 α-SNAP SCN resistance protein was identified and characterized. • The unusual Rhg1 multigene copy number variation structure arose prior to the domestication of soybean.
that Rhg1-encoded α-SNAPs are unusual in that they bind poorly to wild-type NSF (Bayless et al., 2016). In Nicotiana benthamiana Domin transient assays, expression of rhg1-a or rhg1-b α-SNAPs disrupts vesicle trafficking and is cytotoxic, eventually causing cell death (Bayless et al., 2016(Bayless et al., , 2018. Furthermore, these aberrant α-SNAPs accumulate approximately tenfold in syncytial cells as a response to SCN, suggesting a role of vesicle trafficking efficiency in the outcome of SCNsoybean interactions (Bayless et al., 2016(Bayless et al., , 2019. Soybean haplotypes containing three Rhg1 repeats (low-copy [LC] haplotypes; rhg1-a; Peking-type or Hartwig-type) encode a distinct α-SNAP protein while those haplotypes containing nine to 10 Rhg1 repeats (high-copy [HC] haplotypes; rhg1b; PI 88788-type) encode a second distinct α-SNAP protein (Cook et al., 2014). To maximize SCN resistance, varieties carrying LC rhg1-a haplotypes require presence of additional QTL at chromosome 11 (associated with a loss-of-function intron retention allele at a chromosome 11 α-SNAP gene) and at chromosome 8 (Rhg4, encoding a serine hydroxymethyltransferase) (Bayless et al., 2018;Lakhssassi et al., 2017;Liu et al., 2012. These and other studies (Lakhssassi et al., 2020;Liu et al., 2017;Patil et al., 2019;Yu et al., 2016) have identified additional attributes shared by or distinct between the rhg1-a vs. rhg1-b haplotypes. Linkage disequilibrium has been observed between the soybean chromosome 18 Rhg1 locus and a chromosome 7 locus, with segregation distortion observed only in genotypes carrying an SCN resistance-associated Rhg1 allele (Webb et al., 1995;Kopisch-Obuch & Diers, 2005;Vuong et al., 2015). Our group recently found that this distortion is attributable to a unique NSF allele that is encoded at Glyma.07G195900 on chromosome 7 (Bayless et al., 2018). Termed NSF RAN07 (for Rhg1-associated NSF on chromosome 7), this allele is present in 11% of 19,645 soybean accessions in the USDA collection but remarkably is present in all soybean accessions and segregating progeny that are homozygous for the SCN resistanceassociated rhg1-a or rhg1-b haplotypes. At the apparent α-SNAP-NSF binding interface, the encoded NSF RAN07 protein carries atypical amino acids at sites that become proximal to the unusual amino acids that distinguish α-SNAP Rhg1 LC and α-SNAP Rhg1 HC. The NSF RAN07 protein exhibits higher affinity than wild-type NSF for binding those resistance-associated α-SNAP proteins and enables the viability of soybeans carrying Rhg1-encoded SCN resistance (Bayless et al., 2018).
The gradual evolution of SCN populations toward HG types (or races) that partially or largely overcome the SCN resistance in commercially grown soybean varieties (McCarville et al., 2017; has motivated searches for new SCN resistance sources. Specific G. soja accessions have already provided new sources of SCN resistance that show promise (Brzostowski et al., 2017;Usovsky et al., 2020;Yu & Diers, 2017). In a recent study by Lee et al. (2015b), wholegenome sequencing revealed but did not further investigate a tandem duplication of Rhg1 in G. soja accession Jidong 5 (W06), suggesting that resistance-conferring Rhg1 haplotypes may have arisen prior to the divergence of G. max and G. soja. Further research on Rhg1 duplications in G. soja accessions can provide insights into the evolution of SCN resistance.
In the present study, we used a SNP marking the physically unlinked but genetically associated NSF RAN07 allele to discover and examine diversity of Rhg1 in G. soja germplasm available through the USDA collection. We identified infrequent G. soja accessions containing tandem repeat copies of the Rhg1 locus that represent an apparent progenitor of soybean rhg1-a and rhg1-b. We found that the α-SNAP encoded at G. soja multicopy Rhg1 loci is unique, yet carried structural and functional similarities to the rhg1-a and rhg1-b α-SNAPs. Partially as a result of contributions from other loci, specific G. soja accessions carrying NSF RAN07 and rhg1-ds were found to also carry strong resistance to problematic HG type 2.5.7 SCN.

DNA extractions and oligonucleotide primers
Soybean genomic DNA was extracted from young cotyledons from respective G. soja or G. max accessions using standard CTAB extractions as in Keim et al. (1988) with modifications from Cook et al. (2012). The oligonucleotide primers used for polymerase chain reaction (PCR) and other assays are listed in Supplemental Table S3.

Copy number variation assays
Copy number variation assays were performed essentially as in Lee et al., 2015b. In brief, 80 ng of extracted genomic DNA was subjected to qPCR using primers directed against the bridge junction as in Figure 1 or Glyma.18G022800 using SolisBiodyne 5× Firepol (SolisBiodyne; Cat. No. 08-36-00001). Copy number was estimated using the ΔCt method where ΔCt is the difference between Glyma.18G022800 and bridge amplifications. For TaqMan assays, experiments were performed essentially as in Lee et al. (2016).

Sequence variant detection
Flow cell data were demultiplexed and Illumina adapter sequences trimmed by the University of Wisconsin Biotechnology Center facility. Universal sequences were trimmed using Cutadapt (v2.8; Martin, 2011 Danecek et al., 2011) was used to pull out variants within these regions and filter for a depth of three and variant quality value of 50. A fasta sequence file was obtained for each sample using the FastaAlternateReferenceMaker function within GATK. These fasta files were then imported to SplitsTree for network analysis (Huson & Bryant, 2006).

Phylogenetic analysis and phylogenetic trees
To determine the relationship between G. soja and G. max or among G. soja, we aligned all the SNPs available in the SoySNP50K database (Song et al., 2013(Song et al., , 2015. The program SplitsTree (v4.16.1) was used to perform the alignment and construct the network (Huson & Bryant, 2006). Uninformative sites were excluded in the network construction. Rooted phylogenetic trees were built using MEGAX (Kumar et al., 2018;Stretcher et al., 2020). The rooted phylogeny for the Rhg1 locus was performed using DNA sequences described in the preceding section, corresponding to bp Gm18:1,636,000-1,659,000 of G. max genome assembly version Glyma.Wm82.a2.

RNA isolation and cDNA synthesis
Total RNA was extracted from expanding soybean trifoliates using either the DirectZol RNA miniprep plus kit (Zymo Research) or the RNeasy mini kit (Qiagen) using manufacturer's instructions. RNA samples were DNase treated, and total RNA was quantified using spectrophotometry. Integrity of RNA was checked by visualization of rRNA bands on a 1.2% agarose gel. cDNA synthesis was performed using the iScript cDNA synthesis kit (Bio-Rad) with ∼1 μg RNA as input. The Peking soybean accession used in these studies was PI 548402.

Vector construction
The G. soja α-SNAP alleles (Glyma.18G022500) from each of the six multicopy Rhg1 G. soja, and the NSF (Glyma.07G195900), SHMT Rhg4 (Rhg4; Glyma.08G11490), α-SNAP Ch11 (Glyma.11G234500), and 5′ NSF open reading frame (ORF) from G. soja accession PI 507623, were amplified from generated cDNA (for α-SNAP Rhg1 , NSF and Rhg4), or genomic DNA (for α-SNAP Ch11 and 5′ NSF), using Kapa HiFi polymerase (Roche). They were placed directly under the control of the soybean ubiquitin promoter and nopaline synthase terminator in pBlueScript using Gibson assembly (Gibson et al., 2009) and subsequently sequence verified. For the α-SNAP Rhg1 and the NSF, the promoter-ORF-terminator expression cassette was digested with XbaI and PstI or NotI and SalI (New England Biosciences), respectively, and the fragments were purified using the Zymoclean large fragment DNA recovery kit (Zymo Research). Purified DNA fragments were then ligated into the pSM101 binary vector (Melito et al., 2010) using T4 DNA ligase (New England Biosciences). For sequencing of the bridge junction, the NSF Ch07 amplicon and the last exon of α-SNAP Ch11 ,the G. soja Rhg1 bridge junction, NSF Ch07 , or α-SNAP Ch11 amplicon was amplified from genomic DNA using GoTaq green (Promega) and cloned into pGEM-T Easy using ligation per manufacturers recommendations (Promega). Constructs were verified by colony PCR and diagnostic digest before being sequenced by Sanger sequencing.

Transient Agrobacterium expression in Nicotiana benthamiana
Agrobacterium tumefaciens strain GV3101 (pMP90) containing the indicated expression cassettes was infiltrated using a blunt tip tuberculin syringe at an OD 600 = 0.8 into young leaves of N. benthamiana plants. The GV3101 cultures were grown overnight at ∼28˚C in media containing 25 μg ml −1 kanamycin and 100 μg ml −1 rifampicin and induced for ∼3 h in 10 mM MES (pH = 5.6), 10 mM MgCl 2 and 100 mM acetosyringone prior to leaf infiltration. N. benthamiana plants were grown at 25˚C with a photoperiod of 16 h light at 100 μE m −2 s −1 and 8 h dark.

Antibodies and immunoblotting
Generation and validation of rabbit polyclonal antibodies raised against Rhg1 α-SNAP LC , Rhg1 α-SNAP HC , and α-SNAP WT was previously described in Bayless et al. (2016). Tissue preparation and immunoblots were performed essentially as in Bayless et al. (2016Bayless et al. ( , 2018. Bradford assays were performed on each leaf or root extract supernatant and equal amount of OD 595 total protein were loaded onto SDS-PAGE gels. Immunoblots for α-SNAP and NSF were incubated overnight at 4˚C in 5% nonfat dry milk TBS-T (50 mM Tris, 150 mM NaCl, 0.05% Tween 20) at 1:1000. Secondary horseradish peroxidase conjugated goat antirabbit IgG (Sigma) was added at 1:10,000 and incubated for 1 h at room temperature on a platform shaker followed by five washes with TBS-T. Chemiluminescence signal detection was performed with SuperSignal West Dura chemiluminescent substrate (Thermo Scientific) and developed using a Chemi Doc MP chemiluminescent imager (Bio-Rad).
For antibody sensitivity assays, purified recombinant α-SNAP was serially fivefold diluted to concentrations of 800, 160, and 32 pg. The proteins were then loaded onto an SDS-PAGE gel and immunoblots performed as above. To confirm loading of gel lanes, the gel was stained using ProteoSilver Kit (Sigma) according to manufacturer's instruction.

Recombinant proteins
A vector encoding the ORF of the G. soja ortholog of Glyma.18G02250 from PI 507623 was generated, and recombinant protein expression and purification were performed as in Bayless et al. (2016). In brief, the ORF was cloned into the pRham-N-His-SUMO expression vector, protein was purified using PerfectPro Ni-NTA resin as per manufacturer's instruction, His and SUMO tags were cleaved by incubating the dialyzed protein with 1 unit per 100 μg protein of SUMO express protease (Lucigen) and removed by rebinding to the Ni-NTA column, and the recombinant protein was eluted in SNAP freeze-down buffer (50 mM Tris, 50 mM NaCl, 10% (w/v) glycerol, 0.5 mM TCEP, pH = 8). Both purity and quantity of protein were evaluated by Coomassie blue staining relative to BSA standards.

In vitro binding assays
In vitro binding assays were performed as previously described (Bayless et al., 2016). In brief, 20 μg of the specified recombinant α-SNAP was added to the bottom of a 1.5 mL polypropylene tube, unbound α-SNAP was removed with wash buffer, and 20 μg of either NSF or NSF RAN07 in NSF binding buffer (20 mM HEPES, 2 mM EDTA, 100 mM KCl, 500 μM ATP, 1 mM DTT, 1% (w/v) PEG 4000) was added and incubated on ice for 10 min. The solution was then removed and excess NSF was removed by washing twice with NSF binding buffer. Samples were subsequently boiled in 1× SDS loading buffer and separated on a 10% Bis-Tris SDS-PAGE gel. To detect bound NSF, the gel was stained using the Pro-teoSilver Kit (Sigma) according to manufacturer's instruction and quantified relative to α-SNAP abundance by densitometric analysis using ImageJ.

Nematode resistance tests
The nematode resistance tests were performed according to Niblack et al. (2009) with modifications described by Yu et al. (2016). In brief, the genotypes LD10-10198 and PI 507623 were crossed in a greenhouse. The F 1 seeds were generated and grown to maturity to produce an F 2 population. Ninetysix F 2 plants from the cross were evaluated in a resistance bioassay.
In the first SCN bioassay, seed of the SCN indicator lines (Niblack et al., 2009), the G. soja accession, and the susceptible check 'Lee 74′ were germinated on germination paper. After 3 d, individual seedlings were transplanted into PVC tubes filled with steam-sterilized sandy soil and packed in plastic crocks. Each tube was infested with ∼2,000 SCN eggs derived from either HG type 0 or HG type 2.5.7 nematode populations. The experimental unit was a tube with an individual plant and the tubes were randomized using a completely randomized design. The test with each isolate was done separately and there were three replicates of the indicator lines and six replicates of the PI. After 30 d, cysts were collected by washing roots over a 250 μm sieve. The cysts were then counted, and the female index (FI) was determined using the following equation: FI (%) = (number of female cysts on an entry) / (average number of female cysts on susceptible control 'Lee 74′) × 100. The F 2 population was evaluated for resistance to a HG type 2.5.7 population according to the procedures listed above. In this test, each F 2 plant was placed in a separate tube and 10 plants of the susceptible control Lee 74 was included. DNA was isolated from a leaf sample taken from each F 2 plant according to Yu et al. (2016) and tested with the genetic marker Satt309 according to Kim et al. (2010). A single factor analysis of variance was completed to test for an association between the genetic marker and FI.

A small number of Glycine soja accessions contain a multicopy Rhg1 locus
As an entry point to identify G. soja accessions that potentially carry resistance alleles at Rhg1, we searched for accessions that carry genetic markers associated with the G. max NSF RAN07 allele at Glyma.07G195900. NSF RAN07 is essential for the viability of soybean plants carrying the cyst nematode resistance Rhg1 haplotypes rhg1-a and rhg1-b (Bayless et al., 2018). The SoySNP50K Infinium BeadChip had previously been used to generate genetic data for over 19,000 domesticated soybean and over 1,000 wild soybean accessions from 84 countries, testing over 47,000 SNP loci in each accession (Song et al., 2013(Song et al., , 2015. In the absence of strongly predictive Rhg1 markers in the SoySNP50K dataset, we used SNP ss715597431 that is specific for the R 4 Q polymorphism in the encoded NSF RAN07 protein that was recently shown to be necessary for the viability of G. max that carry multicopy Rhg1 haplotypes (Bayless et al., 2018). When we queried the SoySNP50K dataset for accessions annotated as G. soja that contain this SNP, only 21 lines had the corresponding polymorphism, representing ∼1.7% of the ∼1,100 G. soja annotated accessions (Supplemental Table S1). We hypothesized that some of these accessions might also contain Rhg1 haplotypes that confer SCN resistance.
The Rhg1 locus was then examined in the 21 selected G. soja accessions. Prior work has indicated that Rhg1-mediated resistance in G. max is driven substantially by the presence of three to 10 tandem duplicate copies of the >30 kb Rhg1 locus, in so-called LC (rhg1-a) and HC (rhg1-b) genotypes (Cook et  noted a G. soja accession (W06/Jidong5) with three copies of the Rhg1 locus. The vast majority of soybean accessions in the USDA collection, including the cultivar Williams 82, are SCN-susceptible and carry the single-copy Rhg1 haplotype (Bayless et al., 2019;Cook et al., 2012;Lee et al., 2015b). To test the hypothesis that the R 4 Q-containing G. soja also carry multicopy Rhg1 loci, we developed Rhg1 repeat junction PCR primers as in Cook et al. (2012) but specific for G. soja. Primer pairs were first designed (Figure 1a) to amplify the 5′ and 3′ terminal portions of Rhg1, to confirm amplification of Rhg1, and allow DNA sequencing of the G. soja products. A PCR assay was then used to test for the Rhg1 repeat junction by pairing one outward-directed PCR primer for each terminus of the Rhg1 repeat, which should only give a product if the 3′ terminus of one Rhg1 repeat segment is adjacent to the 5′ terminus of the next Rhg1 repeat. Screening of the 21 G. soja accessions revealed that six contain a duplicated Rhg1 (Figure 1b). Use of a different oligonucleotide pair, and a limited set of changes to the template preparation and PCR conditions, produced the same results regarding presence of the repeat junction PCR product in these accessions (Supplemental Figure S1a).

3.2
The Glycine soja multicopy Rhg1 locus shares structural characteristics with Glycine max multicopy Rhg1 haplotypes The Rhg1 repeat junction was previously shown to be conserved between HC and LC SCN-resistant G. max genotypes, suggesting a shared lineage (Cook et al., 2012(Cook et al., , 2014Lee et al., 2015b). To test whether these multicopy Rhg1 G. soja loci also share that lineage, we cloned and sequenced the repeat junction from the six repeat-positive genotypes shown in Figure 1b. Those G. soja accessions were found to carry a Rhg1 repeat junction that is identical to that of the rhg1a soybean genotype Peking and the rhg1-b soybean accession PI 88788 ( Figure 1c). As previously reported, the Rhg1 repeat junction is not present in the single-copy Rhg1 of the G. max reference genome for Williams 82 (Figure 1c). The above observations support the hypothesis that all known multicopy Rhg1 types in annual Glycine species arose from a shared event. Because of the differences from G. max multicopy Rhg1 haplotypes we describe later, and to follow upon established use of the allele terms rhg1-a, rhg1-b, and rhg1-c for previously described soybean haplotypes, we named this G. soja multicopy Rhg1 haplotype 'rhg1-ds', short for rhg1-d (soja).
As it is possible that G. soja accessions carrying the rhg1-ds haplotype arose through hybridization events, we performed unrooted phylogenetic analyses (Huson & Bryant, 2006). Phylogenetic analyses were first conducted using genome-wide SNP data from the SoySNP50K dataset (Song et al., 2013(Song et al., , 2015, to assess the overall relatedness of accessions carrying NSF RAN07 and rhg1-ds, to each other and to a more broadly representative set of G. soja accessions. The comparison set was a relatively random set of G. soja accessions chosen because they have been included in recent publications from other research groups (Figure 2a; Supplemental Figure  S2a). At the whole-genome scale, G. soja accessions carrying an NSF RAN07 -encoding allele are dispersed throughout the phylogeny (Figure 2a). The rare rhg1-ds-containing G. soja accessions were more narrowly clustered, suggesting a shared derivation within G. soja. Unsurprisingly, individual rhg1-ds-containing G. soja also can be closely related to G. soja that do not carry rhg1-ds (Figure 2a). At a genome-wide level the rhg1-ds-containing G. soja group separates from G. max that carry rhg1-a or rhg1-b, suggesting that rhg1-ds arose independent of these haplotypes rather than from a recent hybridization with G. max that carry rhg1-a or rhg1-b (Supplemental Figures S2a and S2b).
To study the relatedness of the rhg1-ds locus to the rhg1-a and rhg1-b loci, we used a targeted approach to extract and sequence genomic DNA from Rhg1 resulting in >70% coverage of the rhg1-ds loci from PI 507613, PI 407287, and PI 378695A. These accessions were chosen because they are relatively distant from each other in the phylogenetic analysis ( Figure 2a). The data were compared with previously determined Rhg1 sequences from G. max and G. soja. Figure 2b shows that rhg1-ds forms a clade with rhg1-a and rhg1-b that is distinct from the more common wild-type (single-copy, SCN-susceptible) haplotypes. The Rhg1 loci from the susceptible G. max and G. soja are more similar to each other than to rhg1-a, rhg1-b, or rhg1-ds. Although the shared origin of soybean rhg1-a and rhg1-b and G. soja rhg1-ds was already implied from their identical Rhg1 repeat junction sequences (Figure 1c), the phylogeny shown in Figure 2b further demonstrates their apparent shared derivation. The data also suggest that rhg1-ds arose prior to the split between rhg1-a and rhg1-b.
We selected two rhg1-ds G. soja accessions, PI 507613 and PI 507623 (Figure 2a), to characterize in further detail. They were collected in 1983 as wild plants, ∼80 km from each other, in central Japan. We also continued to study rhg1-dscontaining PI 342434 because, although presently annotated in the USDA germplasm collection as a G. max, it carries leaf shape and plant architecture traits that are intermediate between G. max and G. soja and contains a genome-wide SNP pattern that clusters with G. soja accessions (Figure 2a). PI 342434 is also from Japan (donated to USDA in 1969) and was originally annotated as a Glycine ussuriensis (Regel & Maack) 'tsurumame' edible wild soybean. We sought to determine the Rhg1 copy number of these three accessions using both a genomic DNA qPCR method and a copy number variation TaqMan assay as in Lee et al. (2015bLee et al. ( , 2016. Our implementation of the TaqMan assay (example shown in F I G U R E 2 Multicopy Resistance to Heterodera glycines 1 (Rhg1)-containing Glycine soja group separately from multicopy Rhg1-containing G. max at the levels of whole genome and Rhg1 locus. (a) Unrooted phylogeny assessing relatedness of whole-genome single nucleotide polymorphism (SNP) signatures of a representative set of G. soja accessions. Accessions containing rhg1-ds and NSF RAN07 (red) were analyzed alongside those containing NSF RAN07 but lacking the multicopy rhg1-ds (black) and a representative semirandom set of non-rhg1-ds non-NSF RAN07 G. soja accessions (green). Note that G. max accessions carrying rhg1-a or rhg1-b also carry NSF RAN07 . (b) Rooted phylogeny generated using genomic sequences across 23 kb of the Rhg1 locus from the denoted accessions comparing relatedness of the rhg1-a (light blue), rhg1-b (dark blue), and rhg1-ds (red) loci as well as the Rhg1 loci from accession W05 (a G. soja carrying a single-copy Rhg1; green) and G. max cv. Williams 82 (orange) Supplemental Figure S1b) was unsuccessful, as we obtained erratic data even for controls, but those assays did indicate a rhg1-ds copy much lower than in rhg1-b cultivar Fayette and similar to that of rhg1-a cultivar Forrest. Genomic qPCR assays ( Figure 1d) were more reproducible and also included domesticated soybean cultivars of known Rhg1 copy number 1, 3, and 10 as controls. The tested G. soja and G. max had approximately the same number of copies of Rhg1 as Peking (three copies; Figure 1d). This is consistent with the hypothesis that after multicopy Rhg1 haplotypes became established, higher soybean Rhg1 copy number soybeans such as rhg1-b with nine or 10 copies may have been a trait derived under positive selection during breeding (Lee et al., 2015b).

3.3
Glycine soja multicopy Rhg1 encodes a distinct α-SNAP Investigation of the G. soja rhg1-ds locus revealed that it encodes a novel Rhg1 α-SNAP variant (Figure 3a). The contributions of Rhg1 α-SNAP proteins to SCN resistance depend heavily on the presence of altered sets of C-terminal amino acid residues at sites that are otherwise highly conserved across multicellular eukaryotes (Bayless et al., 2016;Cook et al., 2014). Distinct sets of C-terminal amino acids are carried by each of the Glyma.18G022500 Rhg1 protein products: α-SNAP Rhg1 LC (from LC, Peking-type rhg1-a) and α-SNAP Rhg1 HC (from HC, PI 88788-type rhg1-b) (Cook et al., 2014; Figure 3a). The multicopy G. soja rhg1-ds haplotypes and the product from G. max PI 342434 encoded an identical α-SNAP that diverged from the susceptible G. max and G. soja reference genomes (wild-type Wm82 and G. soja W05, and PI 483463, respectively), and from α-SNAP Rhg1 LC and α-SNAP Rhg1 HC (Figure 3a). To differentiate this protein from the much more common wild-type α-SNAP Rhg1 protein that is produced in G. soja accessions that carry single-copy Rhg1 haplotypes, and in adherence with the original naming convention for Rhg1-encoded α-SNAP proteins (Bayless et al., 2016), we termed this protein 'α-SNAP Rhg1 Gsm', for G. sojamulticopy. The single-copy Rhg1 G. soja accessions studied because they carried the NSF RAN07 R 4 Q polymorphism were all found to encode an α-SNAP The Plant Genome Immunoblots detecting levels of endogenous N. benthamiana NSF or α-SNAP (α-SNAP-WT), as well as recombinant Glycine α-SNAPs delivered via A. tumefaciens as in (b). Same sample was used for entire columns in (c) but probed with three separate antibodies. Leaf tissue was harvested ∼60 h after infiltration; Ponceau S staining indicates relative levels of total protein in each sample amino acid sequence identical to that encoded in the SCNsusceptible Williams 82 soybean reference genome. Like its counterparts in established SCN-resistant soybean varieties, α-SNAP Rhg1 Gsm is polymorphic at amino acid residues predicted to form electrostatic interactions with NSF (Bayless et al., 2018). No other amino acid polymorphisms are present between the α-SNAP of tested Rhg1 single copy G. soja and α-SNAP Rhg1 Gsm. Curiously, some of the variant (nonwild-type) amino acid residues of α-SNAP Rhg1 Gsm at the Cterminus are the same as in α-SNAP Rhg1 LC while others are the same as in α-SNAP Rhg1 HC. These shared residues again suggest, as with the shared Rhg1 repeat junction and SNPbased phylogeny, that there is a close evolutionary relationship between SCN-resistant G. max Rhg1 haplotypes and the G. soja rhg1-ds haplotype.
α-SNAP Rhg1 Gsm was subsequently tested for two recently discovered resistance-associated functions. Transient expression of Rhg1-resistance type α-SNAPs in N. benthamiana is cytotoxic and induces cell death by 6-7 d after infiltration with A. tumefaciens carrying this construct, with multiple lines of evidence suggesting that this phenotype is due to disruption of normal α-SNAP-NSF function (Bayless et al., 2016(Bayless et al., , 2018. In the present experiments, introduction of a cDNA encoding α-SNAP Rhg1 Gsm driven by a constitutive promoter also caused obvious cell death in N. benthamiana leaves by 6 d after infiltration (Figure 3b). Expression of wild-type α-SNAP or an empty vector control did not cause any cell death, while expression of α-SNAP Rhg1 LC or α-SNAP Rhg1 HC caused death similar to that caused by α-SNAP Rhg1 Gsm (Figure 3b). Expression of all of the α-SNAPs was confirmed using previously described custom antibodies (Figure 3c; Bayless et al., 2016Bayless et al., , 2018. Concurrent with cell death, overexpression of resistancetype Rhg1 α-SNAPs has been found to cause elevated accumulation of endogenous NSF in N. benthamiana-an apparent cellular feedback response to disrupted NSF function (Barszczewski et al., 2008;Bayless et al., 2016;Naydenov et al., 2011;Zhao et al., 2007). Expression of α-SNAP Rhg1 Gsm caused similar accumulation of NSF, well above that observed in control leaves expressing either the α-SNAP from Williams 82 or the empty vector (Figure 3c). These cytotoxicity and NSF accumulation results suggest that, like the widely used PI 88788-type and Peking-type Rhg1 α-SNAPs, the G. soja multicopy α-SNAP Rhg1 Gsm disrupts normal α-SNAP-NSF function because of its C-terminal polymorphisms.

Glycine soja Rhg1 α-SNAP interacts weakly with wild-type NSF Ch07
In light of the above results, we investigated the interaction between α-SNAP Rhg1 Gsm and the two NSF types, wild-type NSF (NSF Ch07 ) and G. soja-encoded NSF RAN07 . We first validated the presence of NSF RAN07 in the six G. soja rhg1ds accessions using primers specific either to the NSF RAN07 or NSF Ch07 and NSF Ch13 (Glyma.13G180100) as a positive control. As predicted by the initial SoySNP 50K screen of G. soja germplasm, all six of the tested accessions had NSF RAN07 amplicons (Supplemental Figure S3). Interestingly, PI 507614B and PI 407287 gave amplicons for both NSF RAN07 and NSF Ch07 across plant samples, whereas the other accessions only gave amplicons to NSF RAN07 (Supplemental Figure S3). Subsequent sequencing of PCR products from PI 507614B or PI 407287, generated using primers specific to the 5′ portion of Glyma.07G195900 (5′ UTR and first intron), only showed sequences analogous to NSF RAN07 . This suggests that an additional NSF Ch07 -like locus may be adjacent to sequences in the genome not captured by our primer set. Cloning of Glyma.07G195900 for each of the six rhg1-ds G. soja accessions as well as PI 342434 G. max revealed a DNA sequence identical to that of the NSF RAN07 from known resistant-type soybeans. Sequencing of a PCR product carrying the 5′ portion of PI 468396B, previously characterized by from Zhang et al. (2017a), which does not have the R 4 Q polymorphism-associated SNP, showed an NSF Ch07 allele. This further indicates the accuracy of the above primer set. We subsequently determined the sequence of the NSF Ch07 amplicons in PI 507614B and PI 407287, by cloning the amplicon into a pGEM-T plasmid and sequencing. Intronic polymorphisms revealed that the amplicons derived from PI 507614B and PI 407287 have a sequence similar to the NSF Ch07 from Williams 82. This suggests that there is an NSF Ch07 allele in these accessions, although perhaps at a chromosomal location that was not amplified from the above primer set.
We then proceeded to investigate physical interaction between α-SNAP Rhg1 Gsm protein and the two NSF types. Using an Escherichia coli expression system, we produced recombinant α-SNAP Rhg1 Gsm protein and NSF Ch07 and NSF RAN07 proteins, as well as α-SNAP Rhg1 LC and α-SNAP Rhg1 HC from resistant G. max varieties, and α-SNAP Rhg1 WT from single-copy susceptible Williams 82. We then performed in vitro binding assays between the different NSFs and α-SNAPs. Relative to single-copy susceptible Williams 82, the resistance-associated G. max α-SNAPs interacted poorly with wild-type NSF Ch07 (Figure 4a), recapitulating prior observations (Bayless et al., 2016(Bayless et al., , 2018. The G. soja α-SNAP Rhg1 Gsm also exhibited deficient interaction with wild-type NSF Ch07 (Figure 4a). This suggests that, as occurs with the resistance-associated soybean Rhg1 α-SNAPs, presence of the G.soja α-SNAP Rhg1 Gsm is likely to disrupt the normal cellular vesicle trafficking that requires efficient α-SNAP-NSF interaction in the absence of NSF RAN07 . These results are quantified in Figure 4b. We also investigated if α-SNAP Rhg1 Gsm interacts better with NSF RAN07 than with NSF Ch07 . Consistent with previous findings regarding α-SNAP Rhg1 LC and α-SNAP Rhg1 HC (Bayless et al., 2018),  there was marked increase in the binding of α-SNAP Rhg1 Gsm to NSF RAN07 relative to binding with wild-type NSF Ch07 (Figures 4a and 4b).

3.5
Glycine soja with multicopy rhg1-ds do not carry the rhg1-a α-SNAP copia retrotransposon, chromosome 11 α-SNAP truncation, or resistance-associated Rhg4 The additional genetic features known to most prominently associate with some Rhg1-mediated SCN resistance are the rhg1-a α-SNAP copia (RAC) retrotransposon, the chromosome 11 α-SNAP intron-retention allele that encodes a nonfunctional truncated α-SNAP, and resistance-associated alleles of Rhg4 that encode a serine hydroxymethyltransferase (Bayless et al., 2018(Bayless et al., , 2019Lakhssassi et al., 2017;Liu et al., 2012). We examined the six rhg1-ds G. soja accessions (and PI 342434 G. max) for these genetic features. None of these seven accessions had a SNP signature of RAC integration (ss715606985 G to A; Supplemental Table S1) or any PCR-amplification features of the 4.8 kb RAC insertion at Glyma.18G022500 ( Figure S4b). They also did not carry short sequence repeats that could be a hallmark of transposon excision at a possible excised RAC site in rhg1-ds.
The genomic region that in some soybean accessions encodes an intron-retention allele at the Glyma.11G234500 α-SNAP Ch11 gene was also sequenced. We first examined SoySNP50K data for the SNP associated with the intron retention allele ss715606985 C to T but noted that for many G. soja the nucleotide was uncalled at this SNP position. By PCR and sequencing, we determined that all seven of the rhg1ds accessions carried an allele lacking the intron retention SNP (Supplemental Figure S5). When antibodies against α-SNAP WT were used in protein immunoblots, we discovered that PI 507613, but not PI 507623, had a low apparent level of α-SNAP WT protein in both roots and shoots (Figure 5a; Supplemental Figure S6a). This low level of α-SNAP WT is analogous to previous experiments with rhg1-a soybean lines that carry the α-SNAP intron-retention allele on chromosome 11 (Bayless et al., 2018). DNA sequencing revealed that while the encoded C-terminus of α-SNAP Ch11 in PI 507623 is similar to Williams82, the C-terminus of α-SNAP Ch11 in PI 507613 carries a single amino acid polymorphism (E 285 V; Figure 5b). The apparent low levels of α-SNAP WT protein in PI 507613 may be due to reduced recognition by the antibody used, or to other regulatory mechanisms, but apparently are not due to a Glyma.11G234500 intron retention allele.
We also performed immunoblots with leaf and root extracts of PI 507613 and PI 507623 using other antibodies to detect resistance-associated α-SNAP Rhg1 LC and α-SNAP Rhg1 HC. For tissue from developing trifoliates, immunoblots revealed a low basal expression of α-SNAP Rhg1 Gsm in G. soja PI showing the E 285 V polymorphism, relative to other rhg1-ds-containing G. soja (PI 507623), single-copy G. soja (PI 468396B), and G. max cultivars Williams 82 and Forrest 507613 and PI 507623 relative to the Forrest (rhg1-a) control (Supplemental Figure S6a). However, in rhg1-ds roots we observed more α-SNAP Rhg1 Gsm protein, at levels analogous to Forrest soybean roots (Figure 5a). Control blots of purified α-SNAP Rhg1 Gsm protein showed antibody recognition at sensitivities similar to that for α-SNAP Rhg1 HC or LC, against which the antibodies were raised (Supplemental Figure S6b).
Rhg4 encodes a variant serine hydroxymethyltransferase in the allele (not genetically linked to Rhg1) that plant breeders pair with rhg1-a to achieve useful SCN resistance in domesticated soybean (Liu et al., 2012;Patil et al., 2019;Yu et al., 2016). Sequencing of Rhg4 cDNA clones from the six rhg1-ds G. soja accessions (as well as soybean Rhg4 transcripts from SCN-susceptible cultivar Essex and SCN-resistant Peking as controls) revealed absence of the P 130 R and N 358 Y polymorphisms associated with Peking-type resistance, suggesting that the six rhg1-ds G. soja accessions carry a susceptibletype (wild-type) Rhg4.

3.6
The Glycine soja PI 507613 and PI 507623 are resistant to virulent nematode populations In light of the above observations about rhg1-ds, including observations about the absence of a resistance-associated T A B L E 1 Relative soybean cyst nematode (SCN) cyst production on select Glycine soja accessions and Heterodera glycines (HG) type indicator G. max lines for two nematode populations Rhg4 and chromosome 11 α-SNAP alleles, we sought to determine the SCN resistance of select rhg1-ds G. soja accessions. The rhg1-ds G. soja lines PI 507613, PI 507623, and the G. soja-like PI 342434 were tested with HG type 0 and HG type 2.5.7 nematode populations. We compared the FIs (relative number of cysts formed) to those for G. max HG type indicator lines and the G. soja PI 468916, which is the source of the SCN resistance QTL cqSCN-006 and cqSCN-007 (Wang et al., 2001) (Table 1). As might be predicted because of the absence of resistance-associated Rhg4 or chromosome 11 α-SNAP alleles, PI 342434 was only able to confer moderate resistance to an HG type 0 SCN population. Intriguingly, PI 342434 was not able to confer substantial resistance to an HG type 2.5.7 SCN population. Even more intriguing, PI 507613 and PI 507623 were both able to confer strong resistance to HG type 0 and HG type 2.5.7 nematode populations, with levels comparable to PI 468916 (Table 1). The above findings, taken as a whole, do not demonstrate causation but do suggest that the novel rhg1-ds haplotype is as functional as rhg1-a for SCN resistance against HG type 0 SCN. The resistance of PI 507613 and PI 507623 to the HG type 2.5.7 SCN population (that PI 342434 was susceptible to) indicates that in PI 507613 and PI 507623, loci other than rhg1-ds also contribute to SCN resistance against economically important HG type 2.5.7 SCN populations. A summary of these data may be found in Supplemental  Table S2.

rhg1-b and rhg1-ds made statistically similar partial-resistance contributions against an HG type 2.5.7 SCN population
To test the effect of the novel rhg1-ds allele on resistance relative to the rhg1-b allele, a F 2 population from a cross of G. max LD10-10198 (rhg1-b resistance) × G. soja PI 507623 (rhg1-ds) was tested for SCN resistance. Each plant in the population was evaluated for resistance in a greenhouse using an HG type 2.5.7 SCN population against which rhg1-b soybean exhibit partial resistance, and each plant was tested with the marker Satt309 that maps within 1 cM of Rhg1 (Kim et al., 2010). One month after the plants were inoculated, the susceptible check Lee 74 averaged 219 SCN females, while the 96 F 2 plants averaged 120 females. Within the F 2 population, no significant association was observed between the Satt309 allele present and level of SCN resistance [single factor ANOVA Pr(>F) = 0.34]. Hence rhg1-ds was associated with partial resistance-no distinction in their contribution to resistance against the tested HG type 2.5.7 SCN population was detected between rhg1-b and the rhg1-ds haplotype derived from PI 507623. This experiment, together with that of Table 1 (above), also provides further evidence that there are loci other than Rhg1 in PI 507623 that contribute to the observed high-level resistance of PI 507623 to the virulent HG type 2.5.7 SCN population that was used.

DISCUSSION
Crop wild relatives are an appealing source for novel traits of agronomic interest, and this is evident in domesticated soybean. Recent resequencing of G. soja along with landrace and elite G. max varieties revealed that many rare alleles were lost from G. soja during its domestication to form G. max (Hyten et al., 2006;Liu et al., 2020). Many of these lost traits are of agronomic importance. Quantitative trait loci associated with improved yield, salt tolerance, alkaline salt tolerance, soybean aphid resistance, foxglove aphid resistance, and SCN resistance have all been found to be present in G. soja but absent in G. max (Concibido et al., 2003;Hesler et al., 2013;Kabelka et al., 2005Kabelka et al., , 2006Kim et al., 2013;Lee et al., 2009Lee et al., , 2015aWang et al., 2001;Yu & Diers, 2017). With respect to SCN, G. soja PI 468916 has served as a donor for the QTL cqSCN-006 and cqSCN-007 that each contribute modest resistance to a highly virulent SCN population but provide substantial and synergistic resistance even to HG type 1.2.3.5.6.7 SCN when combined in G. max with rhg1-b and a chromosome 10 QTL from PI 567516C (Brzostowski et al., 2017). Here, we used a SNP marker associated with NSF RAN07 to select G. soja accessions to screen for multicopy Rhg1 haplotypes, leading to discovery of the rhg1-ds haplotype that encodes a unique α-SNAP with SCN resistance-associated functional features. The hypothesis was that plant germplasm accessions that are strong candidates to carry alleles of interest can be identified using available SNP genotypes for particular alleles of other (unlinked) genes that also contribute to the trait of interest. Our study highlights the power of deploying new findings regarding the genetic architecture of traits of interest (such as the requirement for NSF RAN07 in Rhg1-mediated SCN resistance), along with available data such as the SoySNP50K data for ∼20,000 Glycine accessions, to prescreen germplasm and identify a manageable number of accessions for functional analysis. Recent next-generation sequencing work has further identified structural and nucleotide variants within domesticated soybean (Torkamaneh et al., 2019). The same study also identified loss of function alleles within soybean germplasm, potentially functioning as a sequence-catalogued mutant library analogous to those available in Arabidopsis thaliana (L.) Heynh. If genetic features of potential interest are known, there are increasingly powerful resources available for in silico germplasm prescreens. A potential limitation of using the NSF RAN07 SNP for in silico germplasm screens is that it may not fully capture the diversity of Rhg1-containing G. soja. There may be lines that do not share with soybean (Bayless et al., 2018) the requirement for NSF RAN07 for Rhg1-mediated SCN resistance. In the present study and in previous work, NSF RAN07 was found to be present in all multicopy Rhg1 accessions but also in a minority of the much larger pool of single-copy Rhg1 accessions (Bayless et al., 2018), suggesting that NSF RAN07 is necessary for multicopy Rhg1 resistance while demonstrating that NSF RAN07 can be present in the absence of multicopy Rhg1 resistance. Some accessions expressing NSF RAN07 but lacking multicopy Rhg1 might exhibit elevated SCN resistance, but that trend has not been observed in the documented SCN resistance of soybean accessions that express NSF RAN07 but lack multicopy Rhg1.
We and others have identified SoySNP 50K (or KASP) SNPs associated with other resistance-associated genetic elements that might be used during in silico germplasm screens for possible SCN resistance. In addition to NSF RAN07 , these include the chromosome 11 intron retention α-SNAP knockout, the RAC (rhg1-a-alpha SNAP copia) retrotransposon in the first intron of rhg1-a α-SNAP genes, and SNPs specific for SCN resistance-associated alleles of Rhg4 (Bayless et al., 2018(Bayless et al., , 2019Matsye et al., 2012;Shi et al., 2015;Tran et al., 2019). These might be used to select G. soja accessions that can subsequently be tested for the presence of useful alleles at other loci that contribute to SCN resistance whose presence is not adequately tested by the SoySNP50K dataset. More direct tests using markers not on the SoySNP50K chip could also be carried out de novo on large numbers of germplasm accessions (albeit at appreciable new expense), for example to detect Rhg1 tandem repeat junctions, novel C-terminal sequences in α-SNAP proteins, or new Rhg4 alleles.
The present discovery of PI 507613 as an accession of interest for SCN resistance provides an extended example of the above hypothesis, that useful alleles at multiple loci controlling a trait are likely to co-occur in individual accessions. Phenotypic screening of USDA accessions for SCN resistance apparently missed PI 507613, which was instead flagged for further study by our NSF RAN07 genotypic screen. PI 507613 would have been passed over in a screen for RAC but would be positive in a screen for the Rhg1 tandem repeat junction. PI 507613 would also be negative for Rhg4 or the chromosome 11 intron-retention α-SNAP, but we did discover another α-SNAP with unusual C-terminal amino acids in this accession. The example gains more interest when considering that, in addition to finding the novel rhg1-ds, we found that PI507613 carries the valuable trait of SCN resistance to HG type 2.5.7 SCN (that apparently is not dependent on traditional rhg1a/Rhg4/chromosome 11 α-SNAP genotypes). This PI 507613 SCN resistance is dependent on more loci than just rhg1-ds. We postulate that it was much more likely to identify such an accession once we focused on G. soja accessions that carry one or more SCN resistance alleles of interest. A productive target for future studies will be to map and identify the full complement of other loci that contribute to resistance to HG type 2.5.7 SCN in PI 507613, PI 507623, and possibly in the other rhg1-ds PIs identified in this study.
Unlike domesticated soybean, G. soja accessions are not known to contain the resistance-associated allele of Rhg4, which is necessary for the efficacy of SCN resistance conferred by rhg1-a in G. max (Brucker et al., 2005;Meksem et al., 2001;Wu et al., 2016;Yu et al., 2016). Our findings are consistent in these observations. The accessions we examined did not have a resistance-type Rhg4 or other attributes of Peking-type SCN resistance: presence of RAC or the nonfunctional chromosome 11 α-SNAP allele (although in PI 507613 there does seem to be a novel α-SNAP Ch11 ). Previous GWAS studies with G. soja have either not detected significant contributions from the Rhg1 locus when testing with a Race 1 (HG type 2.5.7) SCN population (Zhang et al., 2017a) or detected only a minor contribution (R 2 = 13.6%) from a locus 5-12 Mb (>8 genes) away from Rhg1 when testing with a separate HG type 2.5.7 SCN population (Zhang et al., 2016). Those GWAS results may have been impacted by a very low rate of occurrence of multicopy Rhg1 loci in G. soja. However, they are also consistent with poor expression of the SCN resistance phenotype that might be predicted if a haplotype such as rhg1-ds is not coupled with complementary Rhg4/RAC/chr11 α-SNAP genotypes. This makes it all the more interesting to understand the other genetic features that contribute to any observed SCN resistance in G. soja (possibly including loci identified by Kabelka Zhang et al. [2016Zhang et al. [ , 2017aZhang et al. [ , 2017b, and other studies). Presumably, novel cellular mechanisms that could enhance the durability of SCN resistance are encoded at those loci.
The evolutionary history of the economically important rhg1-b and rhg1-a soybean loci is a point of interest, and not only because of the unique structure of these loci (copy number variation of up to 10 copies of a four gene chromosomal block, in which three of the four tightly linked genes contribute to SCN resistance). Investigations of Rhg1 evolution may also reveal new Rhg1 haplotypes or suggest approaches to additional functional engineering of Rhg1 components. The previous cloning and characterization of both rhg1-a and rhg1-b demonstrated that the bridge junction (the junction between two Rhg1 blocks) is conserved (Cook et al., 2012). This suggested that Rhg1 duplication preceded the split between rhg1-a and rhg1-b. Further studies have revealed an array of copy numbers for rhg1-a and rhg1-b haplotypes in various soybean accessions (Cook et al., 2012(Cook et al., , 2014Lee et al., 2015bLee et al., , 2016Patil et al., 2019;Yu et al., 2016). Wholegenome resequencing had suggested the existence of a rhg1a-like Rhg1 within G. soja, providing evidence for an early origin of Rhg1 outside of G. max (Lee et al., 2015b). The data we present provides more complete evidence of an early origin of multicopy Rhg1, and revealed a novel haplotype and novel α-SNAP Rhg1 . Further, this study identified rhg1ds as a candidate progenitor of rhg1-a and rhg1-b and α-SNAP Rhg1 Gsm as a candidate progenitor of α-SNAP Rhg1 LC and α-SNAP Rhg1 HC. The multiyear undertaking has been initiated to generate transgenic soybean lines in which isogenic presence or absence of the allele encoding α-SNAP Rhg1 Gsm can be associated with resistance to different HG types of SCN.
Vesicle trafficking is essential to eukaryotic cellular homeostasis. The aberrant α-SNAP encoded by resistanceassociated Rhg1 haplotypes has been implicated in conferring resistance to SCN (Bayless et al., 2016;Cook et al., 2012Cook et al., , 2014Lakhssassi et al., 2017Lakhssassi et al., , 2020Liu et al., 2017). Interestingly, the only SNPs causing nonsynonymous codon differences in the Rhg1 proteins from resistant and susceptible cultivars occur within the gene body of the α-SNAP, particularly at the C-terminus-the region that stimulates NSF activity for SNARE disassembly (Barnard et al., 1997;Bayless et al., 2016;Cook et al., 2014;Marz et al., 2003;Zhao et al., 2015). Here, we report an additional resistanceassociated α-SNAP with a C-terminus distinct from that of either known resistant types or the standard wild-type α-SNAP. Variation in nematode functions that interface with host plant α-SNAPs might be associated with SCN virulence. Recent work elucidating effectors of SCN suggests that SNARE-like proteins-proteins that normally interact with α-SNAP and NSF to form the 20S complex-constitute a large and variable effector family in SCN (Gardner et al., 2018). Further, there is some molecular evidence for direct phys-ical interaction between SNARE-like protein effectors and α-SNAP with virulence outcomes (Bekal et al., 2015). Moreover, recent evidence suggests a potential physical interaction between the Rhg4-encoded serine hydroxymethyltransferase and the α-SNAP Rhg1 , as well as α-SNAP Rhg1 and syntaxin proteins, both with implications for resistance (Dong et al., 2020;Lakhssassi et al., 2020). In light of this, varying the soybean α-SNAP repertoire (for example by use of rhg1-ds and α-SNAP Rhg1 Gsm) may hinder the capacity of nematodes to overcome Rhg1-mediated resistance.
The present study also discovered a novel α-SNAP on chromosome 11 in PI 507613. From previous work, it is apparent that SCN resistance involves not just presence of a α-SNAP Rhg1 but also modification of the relative amount of wild-type α-SNAP protein in syncytia or in whole plants (Bayless et al., 2016(Bayless et al., , 2018(Bayless et al., , 2019. It is possible that the novel α-SNAP Ch11 of PI 507613 is a bona fide resistance-associated QTL. Indeed, under a model where HG type 0 or HG type 2.5.7 nematode populations produce effectors that interact with α-SNAP WT toward facilitating infection or reproduction, modifying the α-SNAP WT could change virulence outcomes as a result of reduced interaction. Future work should be directed toward understanding whether the novel α-SNAP Ch11 cosegregates with SCN resistance. It may be possible to engineer soybean with novel α-SNAPs by genetic modification and gene-editing methods, or by classical breeding to bring in different α-SNAPs such as the G. soja rhg1-ds product. Through work such as that of Marz et al. (2003), a library of mammalian α-SNAP mutations and known consequences is already available, and this type of structure-function knowledge (see also Zhao et al., 2015) is ready to be translated to soybean α-SNAP. Vesicle trafficking has been implicated in many plant-pathogen interactions beyond SCN (Hoefle & Hückelhoven, 2008;Inada & Ueda, 2014). Because of the seeming centrality of vesicle trafficking in plant defense, α-SNAP protein sequences and expression patterns may also represent appealing targets to modify in order to confer resistance to other biotrophic pathogens.

D A T A AVA I L A B I L I T Y S T A T E M E N T
The data supporting the conclusions of this article are included within the article and the supplementary materials.

A C K N O W L E D G M E N T S
We thank Adam Bayless for early contributions to the study and Alison Colgrove for assistance with nematode experiments. Funding for this study was provided by United Soybean Board awards to A.B. and B.D., by the Department of Plant Pathology and the Agricultural Experiment Station Hatch program at University of Wisconsin-Madison, and by National Science Foundation Predoctoral Fellowship and NIH Molecular Biosciences Training Grant awards to D.G.

C O N F L I C T O F I N T E R E S T
The authors declare that they have no competing interests.