Soybean transporter AAT Rhg1 abundance increases along the nematode migration path and impacts vesiculation and ROS

Abstract Rhg1 (Resistance to Heterodera glycines 1) mediates soybean (Glycine max) resistance to soybean cyst nematode (SCN; H. glycines). Rhg1 is a 4-gene, ∼30-kb block that exhibits copy number variation, and the common PI 88788-type rhg1-b haplotype carries 9 to 10 tandem Rhg1 repeats. Glyma.18G022400 (Rhg1-GmAAT), 1 of 3 resistance-conferring genes at the complex Rhg1 locus, encodes the putative amino acid transporter AATRhg1 whose mode of action is largely unknown. We discovered that AATRhg1 protein abundance increases 7- to 15-fold throughout root cells along the migration path of SCN. These root cells develop an increased abundance of vesicles and large vesicle-like bodies (VLB) as well as multivesicular and paramural bodies. AATRhg1 protein is often present in these structures. AATRhg1 abundance remained low in syncytia (plant cells reprogrammed by SCN for feeding), unlike the Rhg1 α-SNAP protein, whose abundance has previously been shown to increase in syncytia. In Nicotiana benthamiana, if soybean AATRhg1 was present, oxidative stress promoted the formation of large VLB, many of which contained AATRhg1. AATRhg1 interacted with the soybean NADPH oxidase GmRBOHG, the ortholog of Arabidopsis thaliana RBOHD previously found to exhibit upregulated expression upon SCN infection. AATRhg1 stimulated reactive oxygen species (ROS) generation when AATRhg1 and GmRBOHG were co-expressed. These findings suggest that AATRhg1 contributes to SCN resistance along the migration path as SCN invades the plant and does so, at least in part, by increasing ROS production. In light of previous findings about α-SNAPRhg1, this study also shows that different Rhg1 resistance proteins function via at least 2 spatially and temporally separate modes of action.


Introduction
Rhg1 (Resistance to H. glycines 1) of soybean (Glycine max) is a complex genetic locus that encodes novel mechanisms of disease resistance (Cook et al. 2012;Bayless et al. 2016;Mitchum 2016). Rhg1 is a central tool used for the control of soybean cyst nematode (SCN, H. glycines), the most economically damaging pathogen of US soybeans (Niblack et al. 2006;Jones et al. 2013;Mitchum 2016;Allen et al. 2017). The mechanisms of Rhg1 activity remain only partially understood.
During pathogenesis, recently hatched J2 SCN migrate toward soybean root exudates and then penetrate soybean roots above the root cap in the zone of elongating cells (Endo 1992). They then migrate intracellularly through root cortical cells, secreting bioactive effectors through their stylet to degrade cell walls and manipulate plant defense responses (Qin et al. 2004;Gheysen and Mitchum 2011;Mitchum et al. 2013;Sato et al. 2019). The protrusible stylet is also used directly to disrupt plant cell walls prior to cell penetration. When it reaches a suitable site adjacent to the vascular cylinder the nematode selects an individual endodermis or endodermis-adjacent cortical cell upon which to feed. Over subsequent days the cell walls are partially dissolved between an increasingly large cluster of dozens of adjacent root cells that also lose their large central vacuoles and become metabolically hyperactive, forming a multinucleate syncytium from multiple cytoplasmically merged cells (Jones 1981;Fenoll et al. 1997;Kyndt et al. 2013). After 3 to 4 wk of deriving nutrients from the host through a syncytium, the life cycle of a fertilized female SCN is completed by forming an egg-filled and durable cyst. Despite a number of previous studies (e.g. Mahalingam and Skorupska 1996;Hermsmeier et al. 1998;Escobar et al. 2011;Kandoth et al. 2011), there is particularly incomplete knowledge regarding how plants respond during the early stages of infection as SCN penetrate and migrate through relatively new root tissues.
The structure of the complex Rhg1 locus has been characterized across a wide array of soybean germplasm (Cook et al. 2012(Cook et al. , 2014Lee et al. 2015;Patil et al. 2019). Increased copy number of a 4-gene block is a hallmark of resistanceconferring Rhg1 haplotypes (Cook et al. 2012). "Peking"-type (rhg1-a) haplotypes typically carry 3 copies while "PI 88788"-type (rhg1-b) haplotypes often carry 9 or 10 copies of the ∼30 kb Rhg1 4-gene block (Cook et al. 2012(Cook et al. , 2014Lee et al. 2015). rhg1-a generally must be combined with an appropriate allele of the unlinked Rhg4 and/ or Rhg2 loci to achieve sufficient SCN resistance (Brucker et al. 2005;Liu et al. 2012;Mitchum 2016;Basnet et al. 2022). Until recent decades most soybean accessions carried single-copy "Williams 82"-type Rhg1 WT loci and were SCN-susceptible (Niblack et al. 2008;Lee et al. 2015;Bayless et al. 2019). In the current US market, approximately 95% of SCN-resistant soybean rely on the PI88788-derived rhg1-b haplotype (Niblack et al. 2008;Tylka and Mullaney 2015;Rincker et al. 2017) Gene silencing and gene overexpression approaches have demonstrated individual contributions to SCN resistance by 3 of the 4 genes within the multicopy Rhg1 segment (Cook et al. 2012;Liu et al. 2017;Butler et al. 2019;Dong and Hudson 2021). The resistance-contributing Rhg1 genes include the subject of the present study, Rhg1-GmAAT (Glyma.18G022400, formerly named Glyma18g02580), which encodes a putative amino acid transporter hereafter referred to as AAT Rhg1 (Cook et al. 2012). Contributions to SCN resistance have also been documented for the adjacent Rhg1 genes Glyma.18G022500 (formerly named Glyma18g02590), encoding a predicted α-SNAP (alpha-soluble NSF [N-ethylmaleimide-sensitive factor] attachment protein), and Glyma.18G022700 (formerly named Glyma18g02610), encoding a protein with a WI12 wound-inducible protein domain (Cook et al. 2012;Liu et al. 2017;Butler et al. 2019;Dong and Hudson 2021). The contributions of AAT Rhg1 and α-SNAP Rhg1 (also known as GmSNAP18) to restriction of SCN progression past the J2 stage in transgenic soybean roots were quantitatively similar, while that of WI12 Rhg1 was possibly greater (Fig. 1B of Cook et al. 2012). Several recent studies on the Rhg1-encoded α-SNAP proteins have revealed their elevated abundance in syncytia and their cytotoxicity, which apparently poisons syncytium cells during the otherwise biotrophic plant-nematode interaction (Cook et al. 2012;Matsye et al. 2012;Cook et al. 2014;Bayless et al. 2016;Lakhssassi et al. 2017;Liu et al. 2017;Bayless et al. 2018Bayless et al. , 2019. Until 2 recent publications, the products of the other 2 Rhg1 genes that contribute to SCN resistance, Glyma.18G022400 (AAT Rhg1 ) and Glyma.18G022700 (WI12 Rhg1 ), had been much less well characterized (Guo et al. 2019;Dong and Hudson 2021). There are no predicted amino acid polymorphisms in the products of those genes between SCN-susceptible Rhg1 WT and the resistance-conferring low-copy rhg1-a and high-copy rhg1-b haplotypes. However, the higher copy numbers of those Rhg1 genes result in constitutive elevation of their transcript abundance in SCN-resistant plants (Cook et al. 2014), and locus copy number has been shown to correlate with the strength of SCN resistance conferred by various rhg1-a and rhg1-b haplotypes (Cook et al. 2014;Lee et al. 2016;Yu et al. 2016;Patil et al. 2019). mRNA encoding AAT Rhg1 was reported to be more abundant upon SCN infection (Kandoth et al. 2011;Matsye et al. 2011). A higher accumulation of resistance-associated α-SNAP Rhg1 in syncytia was reported to play a key role in poisoning the nematode during development. However, the accumulation pattern of AAT Rhg1 was not known.
The amino acids transported by AAT Rhg1 , if any, are not known. Similarly, very little is known about AT3G56200/ AtAVT6C, the Arabidopsis thaliana ortholog of soybean AAT Rhg1 . The family is predicted to export neutral amino acids including aspartate and glutamate, and a recent study of the related protein AtAVT6D reported plasma membrane localization in Nicotiana benthamiana and aspartate transport in Xenopus (Dhatterwal et al. 2022). AAT Rhg1 retains the sequence hallmarks of a bona fide amino acid transporter but, in unpublished work done in the laboratories of our collaborators, AAT Rhg1 has been recalcitrant in yeast and Xenopus oocyte experiments attempting to document the transport of amino acids (A. Reinders, J.M. Ward, B.E. Broeckling, and D.R. Bush, unpublished data). A recent publication showed that overexpression of Rhg1-GmAAT in soybean can increase tolerance of toxic levels of exogenously supplied glutamate, and presented additional indirect evidence showing that AAT Rhg1 impacts glutamate abundance and transport (Guo et al. 2019). Endogenous jasmonic acid (JA) levels and JA pathway genes were also upregulated in Rhg1-GmAAT overexpression soybean lines (Guo et al. 2019). However, evidence is still lacking regarding mechanistic roles of AAT Rhg1 in SCN-soybean interactions.
Reactive oxygen species (ROS) are commonly produced during plant-pathogen interactions and act up-and downstream of various signaling pathways (Apel and Hirt 2004;Camejo et al. 2016;Waszczak et al. 2018). Mechanisms that control ROS production during infections, and the impacts of ROS on infection outcomes, are diverse and continue to be discovered. ROS generated during pattern-triggered immunity (PTI) act as important defense signal transduction molecules (Macho and Zipfel 2014). ROS generated from extracellular and/or intracellular sources can accumulate to more toxic levels during the hypersensitive response associated with effector-triggered immunity (ETI) (Zurbriggen et al. 2010). Plant ROS production can also contribute to disease susceptibility, for example, in soybean interactions with necrotrophic Sclerotinia sclerotiorum (Ranjan et al. 2018). Moderate levels of ROS can help limit the extent of cell death (Torres et al. 2005). This appears to be the case during colonization of Arabidopsis by beet cyst nematodes, for which ROS responses were shown to help limit plant cell death and enhance nematode growth (Siddique et al. 2014;Chopra et al. 2021). In other cases, ROS have been shown to contribute to plant defense against plant parasitic nematodes (Simonetti et al. 2010;Kandoth et al. 2011;Kong et al. 2015;Pant et al. 2015;Teixeira et al. 2016;Habash et al. 2017;Labudda et al. 2018;Lee et al. 2018;Mei et al. 2018;Zhou et al. 2018;Yang et al. 2019aYang et al. , 2019bChen et al. 2020;Hawamda et al. 2020;Labudda et al. 2020).
Plasma membrane-localized NADPH oxidases, encoded by respiratory burst oxidase homolog genes (RBOH), are key enzymes for pathogenesis-associated ROS generation (Averyanov 2009). In soybean, the 17 GmRBOH genes were recently characterized by 2 separate groups (Ranjan et al. 2018;Liu et al. 2019), including their stress induction patterns. In Arabidopsis, a particular RBOH, RBOHD, plays major roles in mediating ROS production during both PTI and ETI (Kadota et al. 2015). As one of many examples, AtRBOHD interacts with the FLS2 immune receptor complex and is phosphorylated by BIK1 to enhance ROS generation that contributes to stomatal closure defense mechanisms against Pseudomonas bacteria (Li et al. 2014). RBOHD orthologs in multiple plant species have been reported to control PTI (e.g. (Simon-Plas et al. 2002;Yoshioka et al. 2003;Trujillo et al. 2006;Kobayashi et al. 2007;Wong et al. 2007;Li et al. 2015;Lee et al. 2020)). RBOHD contributes to defense against root-knot nematodes (Teixeira et al. 2016;Zhou et al. 2018). Soybean GmRBOHG (Glyma.06G162300) is one of the orthologs of Arabidopsis RBOHD (Ranjan et al. 2018;Liu et al. 2019). GmRBOHG is the only RBOH mRNA reported to increase more than 2-fold during SCN infection in both susceptible and resistant soybean lines (Wan et al. 2015;Liu et al. 2019), but other aspects of GmRBOHG behavior during SCN infestations have not been characterized.
The present study investigated soybean AAT Rhg1 . We discovered a unique accumulation of this protein along the SCN root penetration/migration path. Extensive vesiculation occurred in soybean root cortical cells penetrated during SCN migration and AAT Rhg1 was often present on those large vesicle-like bodies (VLB). Overexpression of Rhg1-GmAAT in N. benthamiana also caused AAT Rhg1 association with a subset of the extensively present VLB. Further, we found that AAT Rhg1 interacts with GmRBOHG, a soybean ortholog of Arabidopsis RBOHD. Simultaneous overexpression of AAT Rhg1 and GmRBOHG in N. benthamiana elevated ROS production. The elevated abundance of AAT Rhg1 along the SCN migration path contrasts with the elevated abundance of Rhg1-encoded α-SNAP Rhg1 in syncytia. Hence the α-SNAP Rhg1 and AAT Rhg1 proteins apparently contribute to SCN resistance through temporally, spatially, and biochemically distinct mechanisms.

Soybean AAT Rhg1 protein abundance is elevated along the migration path of SCN
The abundance of native AAT Rhg1 in soybean roots was assessed via standard Western immunoblots using a custom antibody raised against a unique AAT Rhg1 peptide sequence ( Fig. 1). SCN-infested root regions or analogous regions from mock-inoculated controls were harvested at 4 d postinfection (dpi) from non-transgenic Wm82, Forrest, and Fayette cultivars, which, respectively, carry wild-type (WT) Rhg1 (rhg1-c; single-copy/susceptibility-associated), or rhg1-a (3 copies of resistance-associated Rhg1), or rhg1-b (10 copies of resistance-associated Rhg1). Supplemental Fig. S1 confirms antibody recognition of the intended gene product. Figure 1B presents densitometric quantification of the immunoblot band intensities for 4 samples per treatment from 2 independent experiments. In noninfected roots, AAT Rhg1 protein abundance was low and similar in WT and low-copy rhg1-a samples but consistently greater in high-copy rhg1-b samples (Fig. 1, A and B). The detected AAT Rhg1 band migrated with an apparent mass of ∼45 kDa, similar to its predicted mass of 46.9 kDa. A significant increase in AAT Rhg1 protein abundance was observed in SCN-infected samples of high-copy rhg1-b roots (Fig. 1A), about 4.7-fold higher than mock treatment samples of high-copy rhg1-b roots (Fig. 1B). Any AAT Rhg1 abundance increases were more subtle for susceptible/WT and for low-copy rhg1-a roots (Fig. 1, A and B).
Transmission electron microscopy (TEM) immunogold detection experiments were conducted to provide cellular and subcellular resolution regarding AAT Rhg1 protein location and relative abundance. Previously, we used similar methods for Rhg1 locus α-SNAP Rhg1 proteins and discovered more than 10-fold greater accumulation of α-SNAP Rhg1 HC or α-SNAP Rhg1 LC within syncytium cells (the root cells that comprise the differentiated SCN feeding site), relative to the surrounding cells (Bayless et al. 2016;Bayless et al. 2019). Using the anti-AAT Rhg1 antibody to detect native AAT Rhg1 in non-transgenic roots during SCN infection, we observed an entirely different pattern (Fig. 1, C and D). Relative to adjacent cells, AAT Rhg1 protein abundance was elevated in penetrated root cells along the migration path of SCN. , rhg1-a (low-copy), or rhg1-b (high-copy) haplotypes. Detached roots from the 3 varieties were either infected with an Hg Type 0 population of SCN (SCN +) or mock-inoculated with 0.05% sterile agarose water (SCN −), and harvested at 4 dpi. Samples of SCN-infested root regions from 4 roots per treatment were pooled together for the immunoblot. Ponceau staining of the blotted membrane shown as a check for equivalent loading of total protein. B) Densitometry analysis of immunoblot AAT Rhg1 protein levels. N = 4 for each treatment; data obtained from 2 biological replicates. Band intensity was normalized to the intensity for WT mock treatment within each blot; mean ± SE are shown. Bars with the same letter are not significantly different (ANOVA Tukey analysis performed on non-normalized data, P < 0.05). C) Representative electron micrographs taken at 15,000 × magnification showing immunogold-labeled AAT Rhg1 (solid black particles) on vesicle membranes of SCN-penetrated cells in SCN-infested roots of susceptible (upper panels), rhg1-a SCN resistant (low-copy, middle panels), and rhg1-b SCN resistant (high-copy, bottom panels) genotypes, at 3 dpi (middle column) and 7 dpi (right column), but not in normal cells (left column) from the same samples. Arrows indicate immunogold particles in each image. Asterisks highlight vesicle clusters in SCN-penetrated cells. CW, cell wall; M, mitochondrion; Vac, vacuole. All scale bars = 800 nm. D) Number of AAT Rhg1 immunogold particles in SCN-penetrated cells relative to the highest number in a similar 2D area of an adjacent normal cell on the same grid (whose quantity is, therefore, 1.0 for each treatment). At least 30 images, from 3 independent experiments, were used to quantify AAT Rhg1 immunogold particle abundance for each treatment. Values are mean ± SE. Treatments marked with the same letter were not significantly different (ANOVA, P < 0.05).
Representative lower-magnification images showing the SCN body, SCN-penetrated root cells, and adjacent normal root cells are shown in Supplemental Fig. S2 and later figures. Using higher magnification, the experiments of Fig. 1 found 4.4-fold to 12.5-fold more immunogold-labeled AAT Rhg1 in SCN-penetrated cells relative to a similar area in adjacent normal root cells, across 3 independent experiments. Increases were observed in both SCN-susceptible and 2 types of SCN-resistant cultivars (Fig. 1, C and D). At 3 dpi, anti-AAT Rhg1 immunogold particle abundance values relative to adjacent normal cells were lowest for single-copy Rhg1 (susceptible) roots (∼4.4-fold elevation), moderate for lowcopy rhg1-a (resistant) roots (∼7.3-fold elevation), and the highest for high-copy rhg1-b (resistant) soybean roots (∼11.1-fold elevation). However, at 7 dpi, all genotypes accumulated AAT Rhg1 to similar fold-change levels (9.9, 11.0, and 12.5 fold-changes in single-copy, low-copy, and high-copy resistant, respectively). In all cases, the elevated abundance of AAT Rhg1 signal relative to nearby non-penetrated root cells within the same microscopy grid was statistically significant (Fig. 1D).
The anti-AAT Rhg1 immunogold particles were often associated with vesicles and larger VLB, which were strikingly abundant in the areas of nematode penetration ( Fig. 1C; Supplemental Fig. S2; see also additional results below). SCN penetrate root cells and move through the root cortex intracellularly, disrupting the host cells they probe with their stylet and then killing the plant cells they migrate through as they move toward the endodermis. Many of the SCN-adjacent regions with AAT Rhg1 -bearing vesicles and VLB were observed in deceased recently SCN-penetrated root cortex cells lacking a central vacuole, but subsequent work (presented in subsequent Results sections) showed similar induction of VLB as well as multivesicular bodies (MVB) and paramural bodies in the absence of nematodes in live soybean root and N. benthamiana leaf cells expressing AAT Rhg1 from strong constitutive promoters. In the experiments summarized in Fig. 1, C and D and Supplemental Fig. S2, the cellular/subcellular morphologies of cells exhibiting elevated AAT Rhg1 immunogold signal were similar in the SCN-resistant and SCN-susceptible genotypes.
The AAT Rhg1 epitope against which anti-AAT Rhg1 was raised is 15 to 20 amino acids from the N-terminus and is predicted by Deep TMHMM to be oriented internally 7 amino acids from the first transmembrane domain, rather than within 1 of the predicted 11 transmembrane domains (https://dtu.biolib.com/ DeepTMHMM). AAT Rhg1 is predicted to lack a secretion signal peptide (https://services.healthtech.dtu.dk/service.php?SignalP-6.0). Because of the length of primary and secondary antibodies, the immunogold particle can be offset 15 to 30 nm from the location of the bound epitope (Hermann et al. 1996). Hence when AAT Rhg1 is membrane-embedded, gold particles labeling AAT Rhg1 are expected to variously appear on top of or adjacent to that membrane.
Anti-AAT Rhg1 immunogold particles were rarely found in mock treatment samples of all the 3 genotypes tested (Supplemental Fig. S3A). Importantly, in both a susceptible and the 2 types of resistant varieties, anti-AAT Rhg1 immunogold particles were also rare in root syncytium cells (which are readily identifiable by the absence of a large vacuole, abundant presence of organelles, and partially degraded cell walls) (Supplemental Fig. S3B).
In control experiments, no specific immunogold labeling could be found on vesicles or other compartments of SCN-penetrated cells when only the secondary antibody was used (Supplemental Fig. S3C). In further control experiments, competitive binding assays were conducted to confirm the antigen specificity of the anti-AAT Rhg1 antibody in the context of SCN-penetrated soybean roots imaged by the same EM and immunogold labeling method (Supplemental Fig. S4). In those experiments, the N-terminal 44 amino acid peptide that contains the antigen recognized by our custom AAT Rhg1 antibody was purified and preincubated with the AAT Rhg1 antibody at a 1-fold or 10-fold molar excess before use on electron microscopy (EM) sections. Multiple adjacent tissue sections from 1 identical region were examined on separate EM grids. The numbers of AAT Rhg1 immunolabel gold particles within the same penetrated cells were counted for sections probed with AAT Rhg1 antibody pretreated with 1-fold or 10-fold molar excess antigen and compared with the particle numbers for tissue sections probed with AAT Rhg1 antibody not pretreated with peptide antigen. Results showed that both 1-fold and 10-fold molar excess of antigen binding significantly reduced the AAT Rhg1 immunogold signals within SCN-penetrated cells (Supplemental Fig. S4). This indicates the specificity of the anti-AAT Rhg1 immunogold signal for the intended antigen in immunogold-labeled EM soybean root specimens.
We further confirmed this discovery of AAT Rhg1 abundance elevation in SCN-penetrated root cells using a separate method, confocal microscopy with immunofluorescent detection. This method allows broader visualization of AAT Rhg1 distribution within root samples. Roots of non-transgenic SCN-resistant soybean varieties Forrest (rhg1-a) and Fayette (rhg1-b) were inoculated with 200 J2 SCN per root. Four days after inoculation, SCN-infected root regions were chemically fixed. The in situ location and abundance of native AAT Rhg1 protein were monitored by secondary detection of the anti-AAT Rhg1 antibody using an Alexa Fluor 568 dyeconjugated anti-rabbit IgG antibody. Under bright-field illumination, the SCN-penetrated cells could be identified readily due to the visible nematode body or the round hole caused by SCN penetration (Fig. 2). Across multiple samples and in 3 independent experiments, confocal fluorescence microscopy detected elevated anti-AAT Rhg1 antibody signal only in cells that had been penetrated by a nematode (Fig. 2). Signal was detected throughout the entire cell rather than only at the site of penetration. Labeling of individual vesicles was not evident at the resolution provided in these experiments, although infrequent larger aggregations of fluorescence signal were present. Consistent with the findings of Taken together, the above results indicate that SCN-penetrated cells undergo a substantial increase in the abundance of the Rhg1-encoded amino acid transporter-like protein AAT Rhg1 . At the early 3 dpi infection stage the level of AAT Rhg1 accumulation at SCN-penetrated cells positively correlated with Rhg1 copy number, with the 10-copy rhg1-b soybean variety accumulating the most AAT Rhg1 . Cells through which SCN migrates are killed so the observed elevation of AAT Rhg1 protein abundance is apparently activated in advance of that cell death, possibly as nematodes use their stylet to physically probe the cell that they will next invade and release cell wall-degrading effector proteins (e.g. Siddique et al. 2022).

AAT Rhg1 abundance increase not observed upon wounding with needle
To test whether AAT Rhg1 is a wound-inducible protein, roots of Fayette (rhg1-b) were penetrated with a 100 µm diameter microneedle. After 3 d, wounded root regions were isolated and chemically fixed. Then, as in the previous section, TEM immunogold labeling experiments and separate confocal microscopy/immunofluorescent detection experiments were performed using the anti-AAT Rhg1 antibody. The mechanical damage caused by the microneedle did not elicit signal accumulation in or around the site of needle penetration (Supplemental Fig. S5). Although these experiments do not exclude the possibility that the particular forms or patterns of physical damage caused by the nematode can elicit an elevated abundance of AAT Rhg1 signal, the experiments do provide evidence that simple physical penetration of the root cortex is not on its own sufficient to induce elevated abundance of AAT Rhg1 .

Abundance of large vesicles is elevated along the migration path of SCN and AAT Rhg1 protein accumulates on those vesicles
Independent of immunogold label detection, the abovedescribed TEM images revealed a second observation: a strong increase in the abundance of subcellular vesicles in those cells that had been penetrated by SCN ( Fig. 3A; see also Fig. 1C and Supplemental Fig. S2). This increase in vesicle abundance was evident in susceptible roots as well as lowcopy rhg1-a and high-copy rhg1-b soybean roots (Fig. 3A). In addition to numerous vesicles in the ∼50 to 500 nm size range (similar to or larger than common transport or secretory vesicles), some 1 to 2 μm diameter "macrovesicle" VLB were also present. Compared to the adjacent non-penetrated root cortical cells, which retain their large central vacuole, the cells directly surrounding the nematode body showed distinct morphology changes. First, the large central vacuoles were shrunken or otherwise replaced by a nematode body. Second, there were numerous vesicles clustered within the remaining cytoplasm. Third, organelles like mitochondria, ER, or Golgi were rarely observed (Fig. 3A). In all types of soybean roots tested, close inspection of immunogold labeling showed accumulation and co-localization of the AAT Rhg1 protein onto those vesicles formed in penetration cells but not in adjacent normal cells (see, for example, Fig. 1C). The results indicate that SCN-penetrated cells undergo a substantial accumulation of VLB and that a substantial proportion of the AAT Rhg1 protein accumulates on those vesicles.
Additional TEM experiments were performed using epoxy embedding resin, which generally provides better sample integrity than the LR White resin that is superior for immunodetection. With epoxy-embedded samples, the elevated presence of VLB upon the overexpression of AAT Rhg1 was again observed (Fig. 3B and Supplemental Fig. S6). More MVB were also observed, as were paramural bodies (vesicles between the cell membrane and cell wall). MVB structures fused to the cell membrane were observed, apparently releasing the contained vesicles and other content into the apoplast ( Fig. 3B and Supplemental Fig. S6). These MVB and other vesicular structures resemble the cellular structures observed in barley responding to powdery mildew infection (An et al. 2006a(An et al. , 2006b).

Overexpression of AAT Rhg1 in N. benthamiana leads to vesiculation with vesicles containing the AAT Rhg1 protein
Expression of AAT Rhg1 protein increases along the SCN infection path. As one means of investigating impacts of AAT Rhg1 protein accumulation in planta, we expressed soybean AAT Rhg1 in N. benthamiana. N-terminally GFP-tagged AAT Rhg1 or a GFP-only control, driven by a double CaMV 35S promoter, were transiently expressed in N. benthamiana leaves by agroinfiltration. Seventy-two hours after agroinfiltration, the localization of GFP-AAT Rhg1 was analyzed by confocal microscopy. The fluorescence signal for both GFP-AAT Rhg1 and GFP was readily detectable over background. Interestingly, in addition to its distribution throughout the plasma membrane, GFP-AAT Rhg1 was present in the form of multiple primarily cytoplasmic puncta (small spots) as well as large hollow (peripherally fluorescent) or solidly green fluorescent vesicle-like structures (Fig. 4). These large hollow vesicles were of various sizes from ∼0.58 to 5.5 μm diameter, and clearly did not overlap with chloroplasts (Fig. 4A). The structures are analogous in size and shape to the VLB and MVB we observed in soybean roots using EM. The elevated abundance of large VLB and MVB-like structures was also analogous to the elevated abundance observed in soybean cells in the SCN penetration zone. As expected, the GFP-only control localized diffusely throughout the cytoplasm and, unlike GFP-AAT Rhg1 , did not cluster in puncta or small spheres. In these and other N. benthamiana experiments with AAT Rhg1 (see below), bright-field images of the GFP-only control samples showed typical healthy leaf cell cytoplasmic configuration while the cytoplasm of GFP-AAT Rhg1 samples frequently contained a more granular appearance ( Fig. 4A and Fig. 5C).
We conducted additional immunolocalization analyses using untagged AAT Rhg1 . N. benthamiana leaves transiently expressing soybean AAT Rhg1 or empty vector were chemically fixed at 72 hpi. Samples were incubated with the anti-AAT Rhg1 antibody followed by secondary incubation with Alexa Fluor 568 dye for immunofluorescence confocal microscopy. Other samples expressing AAT Rhg1 were incubated with secondary Alexa Fluor 568 dye alone as a control. AAT Rhg1 immunofluorescent signal was detected in cells expressing AAT Rhg1 and not in empty vector controls (Fig. 4B). The signal was specific to the primary anti-AAT Rhg1 antibody as the secondary antibody alone control did not show any signal (Fig. 4B). Interestingly, the specific immunofluorescent signal again accumulated in puncta spots and in vesicle-like structures with an apparent size ranging from 0.94 to 6.6 μm. The free/diffuse cytosolic signal in AAT Rhg1 samples was less prominent, as was also observed for GFP-AAT Rhg1 but not for the GFP controls of Fig. 4A.
To associate the observed GFP-AAT Rhg1 -containing VLB with defined cellular structures, we used confocal laser scanning microscopy to test for co-localization of GFP-AAT Rhg1 and 5 organelle markers in N. benthamiana leaves (Nelson et al. 2007) (Supplemental Fig. S7). GFP-AAT Rhg1 was co-expressed with RFP-tagged markers for Golgi, ER, plasma membrane, plastids, or peroxisomes. Upon coexpression with each of the organelle markers, GFP-AAT Rhg1 displayed the same VLB localization that GFP-AAT Rhg1 alone showed in Fig. 4. Partial co-localization was observed for GFP-AAT Rhg1 and the ER marker, and for GFP-AAT Rhg1 and the plasma membrane marker (Supplemental Fig. S7). GFP-AAT Rhg1 did not show colocalization with the Golgi, plastid, or peroxisome markers (see also Supplemental Fig. S8). Together, the immunofluorescence and green fluorescentprotein tagging showed that, as in soybean root cells along the SCN migration path, AAT Rhg1 overexpressed in N. benthamiana leaves accumulates on profuse vesicles and VLB.

H 2 O 2 accumulates in some cells around the SCN at the early infection stage
ROS signaling can be a mediator of plant defenses against pathogens and ROS production in roots during cyst nematode infections has been studied using multiple approaches (see Introduction). We examined the spatial pattern of ROS accumulation as SCN migrates through roots. ROS A B Figure 4. Soybean AAT Rhg1 localizes to specialized giant vesicles in N. benthamiana. A) Top row: GFP-tagged GmAAT Rhg1 transiently expressed from a CaMV 35S promoter by agroinfiltration into N. benthamiana leaves. Left column: GFP-AAT Rhg1 localized on vesicles of sizes ranging from less than 1 μm (arrowhead) to ∼6 μm (arrow). Chlorophyll signal (second column from left) and bright-field image (third column from left) are from the same imaging layer. Merged images (right column) show that the AAT Rhg1 -containing vesicles were independent of chloroplasts. Lower row: GFP alone expressed similarly as a control. Experiments were replicated on 3 separate dates with similar results. Scale bar = 20 μm. B) Immunofluorescent stain confocal imaging showing that untagged GmAAT Rhg1 also accumulates in puncta of various sizes. N. benthamiana leaves expressing untagged GmAAT Rhg1 were immunostained with anti-AAT Rhg1 antibody and then probed with a secondary antibody conjugated to Alexa Fluor 568 (Top row). Leaf samples expressing the same GmAAT Rhg1 construct immunostained with secondary antibody alone (middle row) or samples expressing empty vector immunostained with both anti-AAT Rhg1 antibody and the secondary antibody (bottom row) served as controls. Images were acquired under the same settings across all 3 rows. Untagged soybean AAT Rhg1 formed puncta with sizes ranging from less than 1 μm (arrowhead) to ∼6 μm (arrow) (top panels). At least 48 fields of view for each treatment were imaged across 3 independent experiments. Scale bars = 20 μm. production was monitored in susceptible and in rhg1-a and rhg1-b SCN-resistant soybean varieties, by tracking bright green fluorescence from the hydrogen peroxide probe 2′,7′-dichlorodihydrofluorescein diacetate (H 2 DCFDA). Roots were examined at 3 d postinfection, a time at which nematode migration is still occurring but some nematodes have initiated feeding and other infections have terminated. We found that H 2 O 2 was induced by SCN in root cells of all the 3 types tested, but not in the mock treatments (Fig. 5). Interestingly, only a portion of root cells around the SCN (indicated by arrows in Fig. 5A) had strong H 2 DCFDA fluorescent signals. Scattered lesions (dead root cells) in the area of nematode invasion are commonly observed a few days after initial root exposure to SCN, but neither the lesions (arrowhead in Fig. 5A) nor the SCN (arrow with stem) showed elevated ROS signals. Reproducibly, there were more cells with positive H 2 DCFDA fluorescent signals in the rhg1-b high-copy variety than the rhg1-a low-copy variety or the WT. Compared with the WT mock treatment, the percent root area with ROS accumulation was the lowest for WT single-copy Rhg1 (susceptible) roots (∼6.8-fold elevation), moderate for low-copy rhg1-a (resistant) roots (∼15.8-fold elevation), and the highest for high-copy rhg1-b (resistant) soybean roots (∼34.6-fold elevation) (Fig. 5B). There were no significant differences between all the mock treatments of the 3 varieties (Fig. 5B).

ROS accumulation enhances endocytosis-associated accumulation of AAT Rhg1 -containing vesicles
Because soybean roots present a challenging experimental system for transient gene expression and for confocal imaging, N. benthamiana leaves were used to initiate the investigation of the potential interaction of AAT Rhg1 with ROS. N. benthamiana leaves transiently expressing GFP-AAT Rhg1 or GFP alone as a control were infiltrated with 20 μm methyl viologen (MV; paraquat). MV is an inhibitor of photosynthetic electron transport chains that induces the elevation of ROS in plant cells (Han et al. 2015). Eight hours after MV treatment, the leaf apoplast was infiltrated with FM4-64 and then imaged 30 min later. Confocal live imaging of FM4-64 dye uptake is a standard technique to monitor vesicle dynamics in endocytic pathways (Bolte et al. 2004). As noted above, expression of GFP-AAT Rhg1 (in the absence of MV) led to the accumulation of green fluorescent puncta as well as green fluorescent vesicle-like structures ( Fig. 5C and Supplemental Fig. S9, GFP column of images). Across independent samples and experiments, expression of GFP-AAT Rhg1 in the presence of MV consistently led to the accumulation of more and larger green fluorescent vesiclelike structures with diameters of approximately 6 to 10 μm (arrowheads, Fig. 5C). Within some large vesicles (MVB) multiple small round vesicles with GFP-AAT Rhg1 fluorescent signals were present, suggesting that the larger vesicles had formed through uptake of multiple small vesicles.
The FM4-64 channel (red fluorescence) in these Fig. 5C experiments also consistently showed abundant accumulation of spherical vesicle/VLB/MVB-like structures in the GFP-AAT Rhg1 + MV samples. FM4-64 labels membranes independent of the presence or absence of GFP-AAT Rhg1 . At the imaged time point 30 to 40 min. after dye application, FM4-64-stained membranes would primarily but not exclusively have recent endocytic origins.
Merging of FM4-64 and GFP images showed that FM4-64-stained membranes colocalized with the fused large vesicles (indicated by GFP fluorescence, arrowheads, Fig. 5C and Supplemental Fig. S9) but not with smaller GFP-AAT Rhg1 -containing vesicles (arrows, Fig. 5C). This suggests that endocytic vesicles fuse with and possibly promote the agglomeration of GFP-AAT Rhg1 into larger clusters. The small GFP-AAT Rhg1 vesicles may have formed prior to the addition of FM4-64 or might not be derived from endocytic events.
To reiterate, we found in N. benthamiana that upon MV-induced oxidative stress, vesicle-associated AAT Rhg1 was more often associated with larger VLB and MVB. These larger vesicles could be stained by a 30-min FM4-64 treatment, indicating that this fusion is associated with an endocytic internalization process. This endocytic vesiculation response to ROS stimuli, which was observed in the presence of elevated AAT Rhg1 expression (Fig. 5), was reminiscent of our

AAT Rhg1 interacts on vesicles with GmRBOHG, an SCN-responsive NADPH oxidase homolog
We hypothesized that AAT Rhg1 physically interacts with 1 or more previously discovered SCN pathogenesis or ROS-associated proteins, and used N. benthamiana agroinfiltration and co-immunoprecipitation (co-IP) to conduct planta tests for interactions. GmRBOHG, the Rhg1-encoded α-SNAP Rhg1 , and WI12 Rhg1 proteins and 8 SCN effectors reported to be expressed during SCN infection (Pogorelko et al. 2020) were tested for interaction with AAT Rhg1 . Only GmRBOHG gave a positive result in the preliminary prescreening, and it was then confirmed upon further testing to be an AAT Rhg1 interactor (Fig. 6). Glyma.06G162300, which encodes GmRBOHG (Ranjan et al. 2018), is a soybean ortholog of the gene encoding defense-associated Arabidopsis RBOHD (Miller et al. 2009;Li et al. 2014;Lee et al. 2020). In previous genomewide expression profiling, Glyma.06G162300 transcript abundance was significantly upregulated after SCN infection in both susceptible and resistant lines, with greater abundance in resistant lines at 3 dpi (Wan et al. 2015;Liu et al. 2019). We co-expressed epitope-tagged GmRBOHG-Myc with GFP-tagged AAT Rhg1 (or GFP-only control) in N. benthamiana leaves. GmRBOHG-Myc co-immunoprecipitated with GFP-AAT Rhg1 but not with GFP alone (Fig. 6A).
To validate these results and to investigate the cellular location of GmRBOHG-AAT Rhg1 interaction, bimolecular fluorescence complementation (BiFC) experiments were carried out. Stringent positive and negative controls are necessary for BiFC experiments (Kudla and Bock 2016); we used the known protein interaction partners NSF and α-SNAP Rhg1 WT of soybean (Bayless et al. 2016) for this purpose. NSF and α-SNAP interact in vitro and in vivo, where they participate in the disassembly of SNARE protein bundles that are associated with vesicle trafficking, including at the plasma membrane (Zhao et al. 2015). Here, AAT Rhg1 with an N-terminal cYFP tag (cYFP-AAT Rhg1 ) was transiently co-expressed in N. benthamiana leaves with either GmRBOHG-nYFP or the negative control NSF-nYFP. cYFP-α-SNAP Rhg1 WT was co-expressed with GmRBOHG-nYFP to serve as another negative control. cYFP-α-SNAP Rhg1 WT and NSF-nYFP were co-transformed within the same leaf as the above samples to serve as a positive expression control for the negative controls (Kudla and Bock 2016). We observed positive interaction signals, indicated by yellow fluorescence, only upon coexpression of cYFP-AAT Rhg1 and GmRBOHG-nYFP, and for the positive control cYFP-α-SNAP Rhg1 WT + NSF-nYFP (Fig. 6B). Interestingly, the signals for interaction between cYFP-AAT Rhg1 and GmRBOHG-nYFP were localized on small vesicle-like puncta within the cytoplasm, which was consistent previous AAT Rhg1 localization results (Figs. 1C and 4). The known vesicle trafficking contributors α-SNAP and NSF also interacted in vesicle-like puncta in our BiFC assay (Fig. 6B). The other control combinations did not give yellow fluorescence under the same confocal detection settings (Fig. 6B).
A substantial portion of the observed AAT Rhg1 localized onto larger fused vesicles under MV-induced ROS stress when overexpressed in N. benthamiana leaves ( Fig. 5C; see also Figs. 1C and 4A for soybean). To test whether the colocalization pattern of AAT Rhg1 /GmRBOHG changed under similar stress, a hemostat wounding method was used after 60 h co-expression of cYFP-AAT Rhg1 /GmRbohG-nYFP or cYFP-α-SNAP/NSF-nYFP control. After wounding, the leaves were left for 30 min in the air before confocal microscopy. After this treatment, the reconstituted YFP signal indicating the interaction of cYFP-AAT Rhg1 and GmRbohG-nYFP was shifted toward larger vesicles (Fig. 6B). The α-SNAP/NSF interaction signal remained on similarly sized vesicles with or without wounding treatment (Fig. 6B). As a point of clarification, the microneedle experiments of Supplemental Fig.  S5 tested if punctures would mimic nematode penetration and cause AAT Rhg1 protein abundance increases, in soybean roots with untagged AAT Rhg1 expressed from its native locus. The present experiments (Fig. 6B) used constitutively expressed AAT Rhg1 in N. benthamiana leaves and tested if its interaction with GmRBOHG changes in the presence of the elevated ROS and other stresses caused by wounding. The experiment indicated that the ROS-generating GmRBOHG protein, previously shown to be transcriptionally upregulated during SCN infection, interacts with AAT Rhg1 in small vesicles in normal conditions and in larger vesicles after wounding.

Simultaneous elevation of AAT Rhg1 and GmRBOHG abundance causes ROS production
Having discovered that AAT Rhg1 and GmRBOHG physically interact in planta, experiments were then carried out to determine if AAT Rhg1 alters ROS generation in concert with GmRBOHG (Fig. 7). GmRBOHG and AAT Rhg1 without epitope tags were co-expressed in N. benthamiana leaves under control of CaMV 35S promoters. GmRBOHG alone, AAT Rhg1 alone, or GFP alone were expressed in the same leaf within the same biological replicate to serve as controls; 72 h after agroinoculation, leaves were detached, stained with nitroblue tetrazolium (NBT) for one-half hour, and then destained. NBT is a standard histochemical stain that detects superoxide (Beauchamp and Fridovich 1971). Numerous NBT-positive spots were detected when coexpressing GmRBOHG and AAT Rhg1 or expressing AAT Rhg1 alone, while within the same leaves few or no NBT-positive spots could be seen in the cells transiently expressing GmRBOHG alone, or GFP (Fig. 7A). Quantification of staining areas confirmed a significant elevation of ROS production when expressing AAT Rhg1 alone. There was even more ROS production when GmRBOHG and AAT Rhg1 were co-expressed (Fig. 7B). GmRBOHG and AAT Rhg1 synergize to promote more ROS production than either protein causes when overexpressed on its own.

Discussion
The Rhg1 locus has for multiple decades been the primary means of control of the most economically damaging pathogen of soybean, SCN, but little is known about how AAT Rhg1 contributes to resistance (Mitchum 2016;Yan and Baidoo 2018). One of the main findings of the present study is that AAT Rhg1 protein levels increase approximately 10-fold along the path of SCN root invasion. The hypothesis that AAT Rhg1 protein abundance differences are a determinant of resistance had previously been proposed (Cook et al. 2012(Cook et al. , 2014, because unlike the α-SNAP Rhg1 proteins, there are no amino acid polymorphisms in AAT Rhg1 between resistant and susceptible varieties. Instead, Glyma.18G022400 mRNA abundance in noninfected tissues had been shown to scale with Rhg1 locus copy number, showing significant elevation in multicopy Rhg1 SCN-resistant genotypes (Cook et al. 2014;Wan et al. 2015). Moreover, whole-genome expression profiling reported that Rhg1-GmAAT mRNA abundance is elevated in SCN-infected root tissues (Kandoth et al. 2011;Matsye et al. 2011). By 7 dpi we observed similar fold-change increases in AAT Rhg1 protein abundance, relative to neighboring cells, in both SCN-susceptible and SCN-resistant genotypes. However, at 3 d after SCN infection the fold-change increase of AAT Rhg1 protein in cells along the SCN migration path, relative to nearby noninfested cells, scaled with Rhg1 locus copy number. This was on top of the baseline (non-inoculated root) elevation of AAT Rhg1 protein abundance in roots carrying the higher copy numbers of the Rhg1 locus. Rhg1 copy number has also been shown to positively correlate with SCN resistance efficacy, especially when isolated from contributions from other loci such as GmRBOHG-nYFP was co-expressed transiently with cYFP-AAT Rhg1 in N. benthamiana leaves under unwounded conditions or wounded conditions (top 2 panels). As a positive control, GmNSF-nYFP and cYFP-α-SNAP Rhg1 WT were co-expressed similarly (middle 2 panels). As a negative control, the same constructs were co-expressed with different pairings (bottom 2 panels). YFP fluorescence indicated by yellow color (left column) was detected from epidermal cells. In cells coexpressing cYFP-AAT Rhg1 and GmRBOHG-nYFP, complemented fluorescence signal was detected in small vesicles (indicated by white arrows in the first panel) in the unwounded condition, or in large vesicles (white arrows in the second panel) in wounded cells. The experiments were repeated on 3 separate dates with similar results. Scale bars = 20 μm.
Rhg4 (Cook et al. 2014;Yu et al. 2016;Patil et al. 2019). The differences in AAT Rhg1 protein abundance and extent of elevation at infection sites are a likely mechanism contributing to the documented differences in resistance efficacy between Rhg1 haplotypes. Equally or more striking, the present study discovered that AAT Rhg1 protein specifically accumulates along the SCN root migration path, relative to its abundance in the root cells a few cells away from the migration path or in any other observed root cells. Most SCN-penetrated root cells would be dead or dying at the time of fixation for microscopy, but they apparently had been stimulated to express elevated levels of AAT Rhg1 protein prior to that event. One reason that this finding is of interest is because α-SNAP Rhg1 , the protein product of the adjacent gene within the same Rhg1 locus, was previously shown to accumulate more than 10-fold specifically within the syncytium cells that, in SCN-susceptible genotypes, serve for a few weeks as the biotrophic interface for cyst nematode feeding (Bayless et al. 2016;Bayless et al. 2019). We observed little or no AAT Rhg1 immunogold signal within syncytia for all 3 of the soybean varieties tested. This indicates that the resistance-contributing genes within the Rhg1 locus not only encode distinctly different proteins, but those proteins also appear likely to act in SCN defense at spatially and temporally separate locations in the infection court. The multi-decade durability of Rhg1-encoded SCN resistance may have been due in part to the lower evolutionary potential of SCN relative to some microbial plant pathogens (McDonald and Linde 2002). However, the present study provides experimental findings supporting the hypothesis that the durability of Rhg1 is enhanced because it is a naturally occurring "resistance stack" encoding more than 1 mode of action.
The mechanism through which AAT Rhg1 activates defenses has remained unknown. As noted in the Introduction, collaborators (A. Reinders, J.M. Ward, B.E. Broeckling, and D.R. Bush, unpublished data) tested AAT Rhg1 for amino acid transport activity in 2 heterologous model systems (Xenopus oocytes and yeast) previously used by those groups to successfully describe plant amino acid transporters. No activity was detected for 29 amino acids and other compounds, possibly because AAT Rhg1 was not functionally expressed. Alternatively, AAT Rhg1 may transport a different substrate than the compounds tested. None of the Rhg1 genes encodes an NB-LRR, RLK, or other protein type that commonly serves the role of pathogen detection and defense activation in plants (Dangl and Jones 2001). Plants sense infection in 1 cell and then can activate defenses in nearby noninfected cells and/or systemic cells using a variety of mediators, including glutamate, ROS, Ca 2+ , salicylic acid, and N-hydroxypipecolic acid, to name just a few examples (Bernsdorff et al. 2016;Hartmann et al. 2018;Toyota et al. 2018;Wang et al. 2019). Local stimuli such as insect herbivory can activate glutamate as a longer-distance wound signal to rapidly initiate defense responses in undamaged parts (Toyota et al. 2018). ROS and electrical signaling, mediated by RBOH proteins and glutamate A B Figure 7. Simultaneous overexpression of GmRBOHG and AAT Rhg1 causes increased superoxide production in N. benthamiana leaves. A) Representative images of N. benthamiana leaf regions stained with NBT 72 h. after agroinfiltration to overexpress (OE) (from left to right) GmRBOHG and AAT Rhg1 protein, AAT Rhg1 alone, GmRBOHG alone, or GFP control protein alone. Images from 3 biological replicates from separate dates are shown. Within each row, the leaf sectors shown are all from the same leaf and were photographed at the same time as part of a single image. Scale bar = 1 cm and applies to all images. B) Quantification of NBT staining intensity of transformed leaf regions described in (A). Total area with NBT staining was measured using ImageJ and divided by total infiltration area, and normalized to the results for GFP alone control within the same replicate. n = 12 plants, mean ± SE are shown, and treatments with the same letter are not significantly different (ANOVA, P < 0.05).
receptor-like proteins, control distal activation of JA signaling during tomato responses to root-knot nematodes (Wang et al. 2019). Those examples may be germane because AAT Rhg1 , while not shown to be a glutamate transporter, was recently reported to increase tolerance to toxic levels of exogenously supplied glutamate, and it impacted glutamate abundance and transport (Guo et al. 2019).
Our testing revealed the physical association of the AAT Rhg1 and GmRBOHG proteins in planta. RBOHG was chosen from the 17 soybean RBOH genes because it is an ortholog of Arabidopsis RBOHD and was the only homolog significantly upregulated in an SCN-resistant variety 3 d after SCN infection (Wan et al. 2015;Ranjan et al. 2018;Liu et al. 2019). We further observed in N. benthamiana that overexpression of GmRBOHG did not elevate ROS, overexpression of AAT Rhg1 was sufficient to raise ROS levels above background, but co-expression of both proteins caused significantly more elevation of ROS. In soybean, we found that AAT Rhg1 abundance increases along the path of nematode invasion and found greater 3 dpi increases in ROS in resistant haplotypes. GmRBOHG mRNA abundance has been shown to increase in SCN-infected tissue (Wan et al. 2015;Liu et al. 2019). The present work hence establishes as a future priority a dissection of cause-effect relationships between AAT Rhg1 /GmRBOHG interaction, ROS elevation, and glutamate elevation in the activation of defense responses during cyst nematode infections.
Another intriguing feature of the present study was the extensive vesicle and VLB (macrovesicle) production observed along the SCN migration path in roots, and the association of AAT Rhg1 with extensive vesicles and VLB both in N. benthamiana and along the SCN infection path in soybean. In split-YFP experiments, GmRBOHG interaction with AAT Rhg1 was primarily observed on these vesicles.
When considering the above and other findings about these AAT Rhg1 -induced/AAT Rhg1 -carrying vesicles, the vesicles do have things in common with the MVB and paramural bodies observed, for example, in barley responding to powdery mildew infection (An et al. 2006a(An et al. , 2006bLi et al. 2018), and with the extracellular vesicles (EVs) recently found to play prominent roles during plant-microbe interactions (Rutter and Innes 2017). First, the vesicular structures we observed were not common in normal cells-their presence was elicited by pathogen infection. Second, they are physically close to the penetration structure, for example, the nematode body (this study) or the haustoria structure (plant-fungal interaction). EVs that accumulate during microbial pathogen invasion have been shown to be defense cargo shuttle vectors, carrying various defense-related proteins, siRNAs, and lipid signals that play roles in plant defense responses (Rybak and Robatzek 2019). In Arabidopsis, defense-related sRNAs could be shuttled into the necrotrophic fungus Botrytis cinerea via EVs (Cai et al. 2018). The AAT Rhg1 -associated vesicles that we observed may have similar roles during SCN infection. However, we have only preliminary evidence (Supplemental Fig. S6) that the vesicles and VLB might be exported across the cell plasma membrane. For most of the present study, they were observed within penetrated soybean root cells, or within N. benthamiana cells overexpressing AAT Rhg1 , or in soybean root apoplastic fluid surrounding the nematode migration path which includes intracellular remnants from recently deceased penetrated root cells. However, in N. benthamiana FM4-64 experiments that monitored uptake of externally labeled plasma membrane, at least some of the AAT Rhg1 -bearing vesicles were associated with endocytic rather than exocytic processes. In N. benthamiana overexpressing AAT Rhg1 , co-localization of AAT Rhg1 with a plasma membrane marker was observed on internally localized macrovesicles as well as the plasma membrane. Hence at least some of the observed AAT Rhg1 -containing vesicles may be endocytic vesicles, reminiscent of those that become more abundant as a transcytosis precursor to papilla formation or when RLKs such as FLS2 have been activated for signaling (Geldner and Robatzek 2008;Beck et al. 2012;Nielsen and Thordal-Christensen 2013;Li et al. 2018).
It bears mention that our AAT Rhg1 protein localization findings, with native protein expressed from the native gene in soybean roots or with native or GFP-tagged protein expressed in N. benthamiana leaves, were internally consistent but differed from the observation of Guo et al. (2019). In limited work, they expressed a GFP-tagged AAT Rhg1 in tobacco and observed plasma membrane localization but also some GFP signal in the nucleus (which resembled the nuclear localization of free GFP in their figure). The resolution of that figure was too low to detect vesiculation, but tobacco may also be less sensitive to the expression of soybean AAT Rhg1 .
Discovery of physical and functional interactions between AAT Rhg1 and RBOHG is perhaps less surprising given the importance of ROS signaling in plant responses to nematode infection. Primarily defensive roles have been established in responses to root-knot nematodes (e.g. Melillo et al. 2006;Zhou et al. 2018;Wang et al. 2019;Chen et al. 2020). The impacts of ROS accumulation during plant responses to cyst nematodes are subtle. ROS production contributes to nematode virulence but also apparently contributes to plant resistance, indicating the probable importance of the level, timing, and location of ROS production (Waetzig et al. 1999;Kandoth et al. 2011;Siddique et al. 2014;Chen et al. 2020;Lakhssassi et al. 2020b;Chopra et al. 2021). Production of ROS scavenging enzymes is also induced, reinforcing the concept that complex homeostatic mechanisms are at play (Kandoth et al. 2011;Chen et al. 2020;Lakhssassi et al. 2020b). Potential engagement of the serine hydroxymethyltransferase Rhg4 gene product GmSHMT08 in ROS homeostasis (as well as in interaction with the rhg1-a α-SNAP Rhg1 LC protein) has also been noted (Lakhssassi et al. 2020a(Lakhssassi et al. , 2020b. Detailed studies of the mechanisms that lead to ROS generation in soybean-SCN interactions are not available. We speculate, as one possibility, that the SCN-induced VLB may provide a longer-lived membrane site for the ROS generation machinery. Within cells damaged by nematode penetration, these types of vesicles could serve as a briefly enduring cellular compartment where those defense responses can continue to function. A similar concept has been presented by Klink and colleagues, who proposed (Pant et al. 2015) that a transiently protected living plant cell could secrete materials in the vicinity of the nematode to disarm it, prior to that plant cell succumbing to its targeted demise. We observed that GmRBOHG physically interacts with AAT Rhg1 within puncta that resemble these VLB. Upon SCN infection, the accumulation of AAT Rhg1 could recruit upregulated GmRBOHG onto those VLB through their interaction. Alternatively, AAT Rhg1 may activate RBOHG and other respiratory burst oxidase homologs at the cell membrane, and then end up on VLB simply as a recycling mechanism. The yellow puncta observed in Fig. 6B and Supplemental Fig. S8 are not colocalized with peroxisomes. We further note that the SCN penetration through the root and secretion of plant cell wall-degrading enzymes is likely to release damage-associated molecular patterns (DAMPs) that activate PTI, including ROS production (Siddique et al. 2022). AAT Rhg1 may play a supplementing/ amplifying role in strengthening DAMP-induced PTI. Furthermore, in the present study, simple root puncture with a microneedle did not elevate AAT Rhg1 protein abundance. Other signals, yet to be discovered, mediate the elevation of AAT Rhg1 protein abundance. Regardless, the elevated coexpression of AAT Rhg1 and GmRBHOG did enhance ROS production, which may be a key early defense against SCN infection that directly weakens the nematode and/or signals to neighboring cells to potentiate defenses.
It remains possible that defense-related functions other than ROS generation are also mediated by the observed AAT Rhg1 -containing vesicles. For example, the Peking-type SCN resistance requires rhg1-a and Rhg4. Rhg4 encodes a serine hydroxymethyltransferase that interconverts serine and glycine, essential for cellular 1-carbon metabolism. The amino acids transported by AAT Rhg1 are not known and may include potential substrates for the Rhg4 hydroxymethyltransferase. These remain as topics for future study.
Taken together, the present study reports distinct tissue and subcellular sites of the elevated abundance of the putative amino acid transporter AAT Rhg1 , along the path of SCN infection. AAT Rhg1 expression is associated with the accumulation of vesicles and VLB, and with activities that elevate ROS production, revealing mechanisms of the successful Rhg1-mediated SCN resistance that might be applied to other plant-nematode interactions.

Materials and methods
The methods utilized are described in detail in the Supplemental Methods S1. The materials and data are available upon request.

Nematode inoculum
Surface-disinfested J2 SCN of HG type 0 were utilized as inoculum.

Plasmid constructs
Transient overexpression of soybean (Glycine max) AAT Rhg1 and GmRBOHG utilized the respective ORF with a double CaMV 35S promoter with TMV omega enhancer and NOS terminator. BiFC vectors and co-IP vectors utilized single CaMV 35S promoters.

Transgenic soybean root and N. benthamiana experiments
Transgenic Williams 82 soybean roots were generated using Agrobacterium rhizogenes ArQua1 strains as described (Melito et al. 2010). Proteins were transiently expressed in N. benthamiana leaves by agroinfiltration as described (Bayless et al. 2016).

Anti-AAT Rhg1 antibody
Affinity-purified anti-AAT Rhg1 polyclonal antibodies were raised using the synthetic peptide "Ac-CSKGTPP(dPEG4) C-amide" and validated by multiple methods as described above. For immunoblots, secondary horseradish peroxidaseconjugated goat anti-rabbit was used for detection.

Transmission electron microscopy
For conventional TEM, soybean root segments in the elongation zone (∼2 mm long) were fixed and stained by standard methods using glutaraldehyde, osmium tetroxide, Epon 812 resin, and uranyl acetate/lead citrate. Representative images were collected from 4 independent root segments for each genotype. Immunodetection TEM experiments were performed similarly to those of Bayless et al. (2019). Soybean (cv. Fayette, Forrest, and Williams 82) root segments previously inoculated with ∼200 J2 SCN (HG 0) per root were hand-sectioned, fixed using glutaraldehyde and paraformaldehyde, embedded in LR White, and sectioned longitudinally with an ultramicrotome. Immunogold labeling used anti-AAT Rhg1 and goat anti-rabbit antibody conjugated to 15-nm gold. Anti-AAT Rhg1 immunogold particles were counted for single 69 μm 2 areas within the sampled cells (e.g. cells penetrated by a nematode) and in the identically sized region that had the highest observable signal in directly adjacent cells with normal root cell morphology.

Immunofluorescent assay
Four days after inoculation SCN-infested root segments were fixed in glutaraldehyde and paraformaldehyde and processed as described in the Supplemental Experimental Procedures. After overnight incubation with anti-AAT Rhg1 roots were treated with secondary antibody Alexa Fluor 568 goat antirabbit IgG H&L (Abcam ab175471) for 2 h and then imaged right away by confocal microscopy.

Confocal microscopy
Confocal imaging was performed using a Zeiss inverted laser scanning confocal microscope (ELYRA LSM 780) with a 20× or 40× water immersion objective. For N. benthamiana experiments samples were imaged ∼72 h. after agroinfiltration and at least 36 images were assessed for each expression treatment across 3 independent experiments. For FM4-64 imaging, 50 μm FM4-64 solution was infiltrated into transformed leaves 0.5 h before microscopy. For BiFC assays reconstituted YFP signal was acquired using 514 nm laser excitation and a 519 to 620 nm range emission filter and a single detector master gain setting across samples.

ROS detection in SCN-infested soybean root
About 400 SCN/root were placed near the root tip of whole 2-wk-old soybean seedlings growing in MS media in PlantCon containers and after 3 d the 2 cm root segments with the greatest SCN infestation were harvested. For mock treatments, similar regions of the root were excised. Root segments were incubated with 50 µM H 2 DCFDA (2′,7′-dichlorodihydrofluorescein diacetate, Invitrogen, D399) for 30 min, washed twice for 10 min. and then imaged (Allan and Fluhr 1997;Shin et al. 2005;Chen et al. 2020). Roots were then washed twice for 10 min and imaged using a 10 × water mount objective on the above confocal microscope (488 nm excitation at 2% laser power, 493 to 598 nm emission). At least 16 confocal fluorescent images from 8 different roots across 2 independent replicates per treatment were used for quantification. Percent area of root cells with ROS signal was calculated using ImageJ software as the number of pixels with H 2 DCFDA fluorescent signal intensity above background, compared to the total imaged root area (with SCN bodies and space outside the root tissue excluded).

MV treatment
MV treatment on N. benthamiana leaves was conducted as described (Han et al. 2015). In brief, 20 μm MV solution was infiltrated into the transformed leaves at 64 h post agroinfiltration and followed by 8 h under the previous light conditions to induce internal ROS generation before the confocal analysis.

Compression wounding
N. benthamiana leaves were compressed gently for 30 s by full release of a pair of reverse-action tweezers for a consistent wounding force.

NBT staining
Nitro blue tetrazolium (NBT) staining was performed as described (Han et al. 2015) by incubating buffer-infiltrated leaves in 0.1% w/v NBT for 30 min. followed by fixation and imaging on a flatbed scanner. ImageJ was used to calculate the percentage of dark blue pixels in the agroinfiltrated area. 32 images taken from 12 independent leaves across 3 independent replicates were used for quantification.

Author contributions
S.H. designed the research, performed research, analyzed data, and wrote the manuscript. J.M.S. designed the research, performed research, and analyzed data. Y.D. performed research, analyzed data, and improved the manuscript. A.F.B. designed the research, analyzed data, and wrote the manuscript.

Supplemental data
The following materials are available in the online version of this article.
Supplemental Figure S1. Confirming the specificity of custom-generated AAT Rhg1 antibodies.
Supplemental Figure S2. Zoomed-out and zoomed-in images of the same SCN-penetrated cells in WT rhg1-c soybean roots, low-copy rhg1-a, and high-copy rhg1-b soybean roots.
Supplemental Figure S3. Little or no immunogold signal in mock-inoculated samples, syncytium cells, or negative controls that omit primary antibody.
Supplemental Figure S4. Confirming by competitive binding control that the AATRhg1 antibody is specific in EM antigen detection.
Supplemental Figure S5. AATRhg1 signal is not generated by microneedle damage.
Supplemental Figure S6. Additional representative TEM micrographs showing large multivesicular compartments common in AATRhg1 overexpression soybean roots (but rarely observed in control roots).
Supplemental Figure S7. GFP-AATRhg1 partially colocalizes with ER and PM markers but not with Golgi, plastid, or peroxisome markers in N. benthamiana cells.
Supplemental Figure S8. Representative confocal micrographs showing that AATRhg1 interaction with GmRBOHG in N. benthamiana does not colocalize with a peroxisome marker.
Supplemental Figure S9. Additional representative confocal micrographs showing that upon cellular ROS stress, AATRhg1-containing vesicles fuse into larger vesicles through an endocytosis pathway in N. benthamiana leaf cells