Disease resistance through impairment of α-SNAP-NSF interaction and vesicular trafficking by soybean Rhg1

Abstract

α-SNAP [soluble NSF (N-ethylmaleimide-sensitive factor) attachment protein] and NSF proteins are conserved across eukaryotes and sustain cellular vesicle trafficking by mediating disassembly and reuse of SNARE protein complexes, which facilitate fusion of vesicles to target membranes. However, certain haplotypes of the Rhg1 (resistance to Heterodera glycines 1) locus of soybean possess multiple repeat copies of an α-SNAP gene (Glyma.18G022500) that encodes atypical amino acids at a highly conserved functional site. These Rhg1 loci mediate resistance to soybean cyst nematode (SCN; H. glycines), the most economically damaging pathogen of soybeans worldwide. Rhg1 is widely used in agriculture, but the mechanisms of Rhg1 disease resistance have remained unclear. In the present study, we found that the resistance-type Rhg1 α-SNAP is defective in interaction with NSF. Elevated in planta expression of resistance-type Rhg1 α-SNAPs depleted the abundance of SNARE-recycling 20S complexes, disrupted vesicle trafficking, induced elevated abundance of NSF, and caused cytotoxicity. Soybean, due to ancient genome duplication events, carries other loci that encode canonical (wild-type) α-SNAPs. Expression of these α-SNAPs counteracted the cytotoxicity of resistance-type Rhg1 α-SNAPs. For successful growth and reproduction, SCN dramatically reprograms a set of plant root cells and must sustain this sedentary feeding site for 2-4 weeks. Immunoblots and electron microscopy immunolocalization revealed that resistance-type α-SNAPs specifically hyperaccumulate relative to wild-type α-SNAPs at the nematode feeding site, promoting the demise of this biotrophic interface. The paradigm of disease resistance through a dysfunctional variant of an essential gene may be applicable to other plant-pathogen interactions.

Keywords: Rhg1; plant disease resistance; soybean cyst nematode; α-SNAP.

Fig. 1.

Rhg1 resistance-type α-SNAPs are deficient in NSF interactions and destabilize 20S complexes. (A) Alignment of Rhg1 single-copy (wild-type, SCN-susceptible), low-copy (SCN-resistant), and high-copy (SCN-resistant) α-SNAPs (9), showing resistance-type amino acid polymorphisms, and an alternate splice form of the low-copy α-SNAP. Asterisks indicate identical amino acid residues; colons indicate similar residues. (B) Silver-stained SDS/PAGE of recombinant soybean NSF_Ch13 bound in vitro by recombinant wild-type, low-copy (LC), or high-copy (HC) Rhg1 α-SNAP proteins. BSA, bovine serum albumin. (C) Densitometric quantification of NSF_Ch13 bound by Rhg1 α-SNAPs as in B; data are from three independent NSF_Ch13 experiments. Error bars show SEM. (D) Immunoblot of coimmunoprecipitation of endogenous WT α-SNAP and α-SNAP_Rhg1HC upon anti-HA immunoprecipitation (IP) of soybean NSF_Ch07-HA in transgenic Fayette roots. α-SNAP detection was with custom antibodies (Fig. S2 A–C). Input: total protein samples before immunoprecipitation. EV, empty vector. All panels were exposed for 20 s, except for an 8-min exposure for final panel, labeled with an asterisk. (E) Immunoblot of density gradient fractions to detect the presence of NSF in 20S complexes. Total solubilized membrane proteins were loaded from N. benthamiana leaves expressing either α-SNAP_Rhg1LC, α-SNAP_Rhg1WT, or empty vector, and anti-NSF antibody was used to detect endogenous N. benthamiana NSF after SDS/PAGE immunoblotting of the resulting fractions. (F) Quantification of NSF present in 20S complexes. Densitometric data are from four independent experiments, calculated as the combined density of NSF signal in ∼20S-migrating fractions (fractions 11 to 13) over the total NSF density (fractions 3 to 13). Error bars show SEM.

Fig. S1.

α-SNAP extreme C terminus is highly conserved among eukaryotes and critical for NSF binding. (A) Logo showing conservation of the final 10 C-terminal α-SNAP residues from model organisms across diverse phyla, similar to ref. . α-SNAP C-terminal consensus was generated from the following species: Chlamydomonas reinhardtii, Saccharomyces cerevisiae, Physcomitrella patens, Arabidopsis thaliana, Medicago truncatula, Nicotiana tabaccum, Caenorhabditis elegans, Drosophila melanogaster, Danio rerio, Xenopus laevis, Gallus gallus, Rattus norvegicus, and Homo sapiens. The conservation logo was generated using WebLogo (55). (B) Silver-stained SDS/PAGE of recombinant soybean NSF_Ch07 bound in vitro by recombinant wild-type (WT), low-copy (LC), or high-copy (HC) Rhg1 α-SNAP proteins. (C) Densitometric quantification of NSF_Ch07 bound by Rhg1 α-SNAPs as in Fig. 1C; data are from four independent NSF_Ch07 experiments. Error bars show SEM. (D) Agarose gel showing RT-PCR product generated due to the presence of both full-length transcript and the alternate splice product. RT-PCR was performed on low-copy Rhg1 line Forrest cDNA with a primer directly upstream of the splice site and at a sequence unique to the low-copy Rhg1 α-SNAP C terminus. Alternate splicing represents roughly 20% of total low-copy α-SNAP transcripts. (E) Silver-stained SDS/PAGE of recombinant soybean NSF_Ch07 bound to recombinant Rhg1 α-SNAPs, including the alternately spliced low-copy α-SNAP protein (LC_Splice) or a 10-residue C-terminal truncation of α-SNAP_Rhg1WT [WT(-10)]. (F) As in B, but with a Rhg1-encoded α-SNAP binding assay with recombinant Chinese hamster ovary NSF (NSF_CHO). (G) Densitometric analysis of in vitro NSF_CHO binding from four independent experiments. Error bars show SEM. (H) NSF coimmunoprecipitation upon anti-GFP immunoprecipitation (IP) of GFP-α-SNAP_Rhg1WT or GFP-α-SNAP_Rhg1HC coexpressed with soybean NSF_Ch07-HA in N. benthamiana leaves. Input: total protein samples before immunoprecipitation.

Fig. S2.

Confirming the specificity of custom-generated α-SNAP and NSF antibodies. (A) Immunoblot test of anti–α-SNAP WT on root lysates from Fayette (Fay.) or Williams 82 (Wm82), recombinant WT α-SNAP truncated at the final 10 C-terminal residues and thereby lacking the epitope region [Rec. WT(-10)], or recombinant α-SNAP_Rhg1LC protein (Rec. LC). Note: α-SNAP WT antibody was raised to the highly conserved α-SNAP C terminus and is thus cross-reactive with most WT α-SNAPs. (B) Immunoblot test of anti–α-SNAP_Rhg1LC (low-copy) on root lysates from Fayette (endogenous high-copy Rhg1), Forrest (endogenous low-copy Rhg1), or transgenic Williams 82 (single-copy Rhg1) roots expressing α-SNAP_Rhg1LC or an empty vector control (EV), or purified recombinant α-SNAP_Rhg1LC or recombinant α-SNAP_Rhg1HC protein. (C) Similar to B, but an immunoblot test of anti–α-SNAP_Rhg1HC (high-copy). Note: α-SNAP_Rhg1HC antibody is cross-reactive with α-SNAP_Rhg1LC but not with WT α-SNAPs. (D) Immunoblot test of anti-NSF on recombinant NSF_Ch07 or NSF_Ch13 or root lysates from Fayette or Williams 82. As expected, the anti-soybean NSF antibody is also cross-reactive with the N. benthamiana NSF protein (e.g., Fig. 2).

Fig. S3.

Density gradient fractionation of protein standards of known sedimentation, performed in the same run as one of the fractionations that detected the presence of NSF in 20S complexes (e.g., Fig. 1G). Sedimentation was performed similar to refs. and . Protein standards were detected by SDS/PAGE and Coomassie blue stain. Protein standards used were thryoglobulin (19.4S), ∼250-kDa dimer; catalase (11.3S), ∼60-kDa tetramer; yeast alcohol dehydrogenase (7.6S), ∼37-kDa tetramer; and BSA (4.5S), ∼65 kDa.

Fig. 2.

Rhg1 resistance-type α-SNAP expression disrupts secretory trafficking, triggers NSF hyperaccumulation, and eventually causes cell death in N. benthamiana. (A) N. benthamiana leaf expressing Rhg1 α-SNAPs with no epitope tag, or an empty vector control, 6 d after agroinfiltration. HC, α-SNAP_Rhg1HC; LC, α-SNAP_Rhg1LC; LC_Splice, α-SNAP_Rhg1LC_Splice; WT, α-SNAP_Rhg1WT. (B) Immunoblot of endogenous N. benthamiana NSF abundance in leaves expressing the indicated α-SNAP_Rhg1 constructs from A or empty vector control. The same samples were probed with anti–α-SNAP_Rhg1 antibodies raised against peptides from the indicated source. Leaf tissue was harvested 3 d after agroinfiltration; Ponceau S staining shows similar loading of total protein. (C) Confocal images of N. benthamiana epidermal cells coexpressing sec-GFP and Rhg1-encoded α-SNAPs denoted as in A, or empty vector control. The sec-GFP assay detects GFP signal if there is failed secretion (retention in ER–Golgi). Images are for 3 d after agroinfiltration. (Scale bars, 20 µm.) (D) Quantification of sec-GFP fluorescence with the respective Rhg1-encoded α-SNAPs as shown in C using ImageJ; n = 25 for each construct. Error bars show SEM. (E and G) N. benthamiana leaves 5 d after agroinfiltration to express the indicated Rhg1 α-SNAPs with no epitope tag, Rhg1 α-SNAPs mutagenized to carry different residues at the penultimate amino acid (no epitope tag), or an empty vector control. (F and H) Endogenous N. benthamiana NSF abundance at 3 d as in B, upon expression of the indicated α-SNAP_Rhg1 constructs from E or G, respectively, or empty vector control.

Fig. S4.

α-SNAP protein encoded by alternate splicing of the low-copy α-SNAP transcript does not appreciably accumulate in soybean roots or N. benthamiana leaves. (A) Anti–α-SNAP_Rhg1LC immunoblot of three separate samples of agroinfiltrated N. benthamiana leaves expressing α-SNAP WT, α-SNAP_Rhg1LC, α-SNAP_Rhg1LC_Splice, or empty vector. Ponceau S staining shows relative protein levels. Immunoblot labels: EV, empty vector; LC, α-SNAP_Rhg1LC; LC_Splice, α-SNAP_Rhg1LC_Splice; WT, α-SNAP_Rhg1WT. (B) Anti–α-SNAP_Rhg1LC immunoblot of soybean Forrest root lysates, transgenic root lysates from Williams 82 expressing α-SNAP_Rhg1LC, empty vector, or α-SNAP_Rhg1LC_Splice, or purified recombinant α-SNAP_Rhg1LC_Splice to confirm anti–α-SNAP_Rhg1LC recognition of the α-SNAP_Rhg1LC_Splice protein. Note that a low-abundance band is present in Wm82 transgenic roots agroinfiltrated with α-SNAP_Rhg1LC_Splice construct but not empty vector.

Fig. S5.

Rhg1 resistance-type α-SNAP cytotoxicity is dosage-dependent and occurs independent of the other Rhg1 locus-encoded genes. (A) N. benthamiana leaf agroinfiltrated with native genomic Rhg1 three-gene blocks (3G Native Rhg1) containing Glyma.18G022400 or Glyma.18G022700 and the Glyma.18G022500 alleles encoding the respective single-copy, low-copy, or high-copy Rhg1 α-SNAPs. Overexpressed α-SNAP_Rhg1LC (OX LC) and an empty vector were agroinfiltrated as controls. Cytotoxic symptoms in N. benthamiana still occur from expression of α-SNAPs driven by native soybean Rhg1 promoters, albeit at a decreased rate and severity compared with expression from a strong ubiquitin promoter. All constructs were infiltrated at OD₆₀₀ 0.60. An image is shown for 9 d after agroinfiltration. LC, α-SNAP_Rhg1LC expressed from the soybean ubiquitin promoter. (B) N. benthamiana leaf agroinfiltrated with serial twofold dilutions of α-SNAP_Rhg1LC or an empty vector control. Leaf shown 6 d after agroinfiltration. (C) N. benthamiana leaf agroinfiltrated with a 1:3 vs. a 3:1 mixture of α-SNAP_Rhg1LC and α-SNAP_Rhg1WT shows further decreases in cytotoxic progression compared with α-SNAP_Rhg1LC alone. Leaf shown ∼8 d after infiltration.

Fig. S6.

Expression of an NSF lacking ATPase activity phenocopies α-SNAP Rhg1 expression and is cytotoxic to N. benthamiana. N. benthamiana leaf expressing soybean NSF_Ch07-HA, the ATPase-null NSF_Ch07-HA (E₃₃₂Q), α-SNAP_Rhg1LC, α-SNAP_Rhg1HC, α-SNAP_Rhg1WT, or empty vector control at 7 d after agroinfiltration. HC, α-SNAP_Rhg1HC. NSF and α-SNAP expression was from the soybean ubiquitin promoter. NSF-HA_Ch07E₃₃₂Q but not WT NSF_Ch07-HA expression causes cell death similar to α-SNAP_Rhg1LC or HC.

Fig. S7.

Expression of α-SNAP_Rhg1WT(-10) raises NSF levels in N. benthamiana leaves, but paraquat treatment of N. benthamiana leaves, or transgenic expression of Rhg1 resistance-type α-SNAP in soybean hairy roots, does not detectably raise abundance of NSF. (A) Immunoblot of N. benthamiana leaf lysates 24 h after infiltrating with 50 µM paraquat (methyl viologen) or 3 d after agroinfiltration delivery of the indicated α-SNAPs. (B) Anti-NSF immunoblots on transgenic Williams 82 root lysates expressing the indicated α-SNAPs. (C) Anti-NSF immunoblots on transgenic Fayette root lysates expressing the respective α-SNAPs. Ponceau S staining shows relative protein levels. WT(-10), α-SNAP_Rhg1WT(-10).

Fig. S8.

Resistance-type α-SNAP expression appears to disrupt localization of the trans-Golgi network/early endosome marker Syp61-mCherry in N. benthamiana. Confocal images of N. benthamiana mesophyll cells coexpressing Syp61-mCherry and α-SNAP_Rhg1WT, α-SNAP_Rhg1LC, or empty vector. Images are at 3 d after agroinfiltration; n = 20 for each construct. (Scale bars, 20 µm.)

Fig. S9.

Penultimate leucine substitutions of α-SNAP_Rhg1WT are not macroscopically cytotoxic, but removing the final 10 C-terminal residues is strongly cytotoxic. (A) N. benthamiana leaf expressing α-SNAP_Rhg1HC-I₂₈₉L or -I₂₈₉A or α-SNAP_Rhg1WT-L₂₈₈I or -L₂₈₈A shows that substitutions at the penultimate amino acid position influence α-SNAP_Rhg1HC cytotoxicity but do not confer macroscopic cytotoxicity to α-SNAP_Rhg1WT. Image shown at ∼6 d post agroinfiltration. Respective penultimate residue substitutions are as indicated. (B) N. benthamiana leaf expressing α-SNAP_Rhg1WT truncated at the C terminus causes cell death similar to Rhg1 resistance-type α-SNAPs. Agroinfiltrated constructs were LC, α-SNAP_Rhg1LC; LC(-10), α-SNAP_Rhg1LC(-10); LC_Splice(-10), α-SNAP_Rhg1LC_Splice(-10); WT, α-SNAP_Rhg1WT; WT(-10), α-SNAP_Rhg1WT(-10); and empty vector.

Fig. 3.

Coexpression of wild-type soybean α-SNAPs with Rhg1 resistance-type α-SNAPs alleviates cell-death symptoms and secretion defects; a penultimate leucine is required. (A) N. benthamiana leaves 6 d after agroinfiltration with a 3:1 Agrobacterium culture mixture (three parts α-SNAP_Rhg1LC to one part wild-type soybean α-SNAP or empty vector control). The soybean wild-type α-SNAPs are WT (α-SNAP_Rhg1WT), Glyma.18G022500; Ch₀₂, Glyma.02G260400; Ch₀₉, Glyma.09G279400; Ch₁₁, Glyma.11G234500. (B) Confocal imaging of sec-GFP assays as in Fig. 2C, but including leaves treated with a 3:1 Agrobacterium culture mixture as in A. (Scale bars, 20 μm.) (C) Immunoblot of leaf samples taken 3 d after agroinfiltration as in A. LC:WT constructs were infiltrated at a 3:1 ratio. (D) Similar to A, with a 3:1 culture mixture of α-SNAP_Rhg1LC to either α-SNAP_Rhg1WT, α-SNAP_Rhg1WT with A or I penultimate residue substitutions, or empty vector control.

Fig. 4.

α-SNAP_Rhg1HC hyperaccumulates at SCN infection sites in high-copy Rhg1 soybean accession Fayette and depletes 20S complexes. (A) Immunoblot of tissue samples from SCN-infested root regions harvested 4 d after SCN infection. Blots were probed with the indicated antibodies; quantitative comparisons are valid within rows but not within columns. (B) Densitometric ratio of α-SNAP_Rhg1HC to WT α-SNAPs calculated from the band intensities in A. Error bars show SEM. (C) Brightness-adjusted electron micrograph showing immunogold-labeled α-SNAP_Rhg1HC in syncytial cells (Syn.) and adjacent cells (Adj.) 4 d after SCN infection of high-copy Rhg1 soybean accession Fayette. Arrows highlight 8 of the ∼400 immunogold particles in this image. CW, cell wall; M, mitochondrion; Vac, vacuole. (D) Average α-SNAP_Rhg1HC immunogold particle counts in syncytial vs. adjacent cells from 30 images across three independent experiments. See Fig. S10 for raw immunogold particle counts, additional images, and antibody specificity controls. Error bars show SEM. (E) Anti-NSF immunoblot of density gradient fractions to detect 20S complexes from SCN-infested Fayette root regions harvested 4 d after SCN infection. (F) Densitometric analysis of NSF from 20S-migrating fractions (fractions 11 to 13) over total NSF (fractions 3 to 13) in SCN-infested root regions. Data from three independent experiments were normalized to 20S NSF abundance from noninfected root regions within each experiment. Paired t test, *P = 0.022 for similarity to mock. Error bars show SEM.

Fig. S10.

Quantification of Rhg1 α-SNAPs in developing syncytia and confirmation of α-SNAP_Rhg1HC specificity when used in immunogold labeling of electron microscopy sections of SCN-infested roots. (A) Immunoblot of Williams 82 tissue samples from SCN-infested root regions harvested 4 d after SCN infection. Blots were probed with the indicated antibodies. (B) Number of α-SNAP_Rhg1HC immunogold particles detected in syncytial cells vs. adjacent cells in SCN-infested Fayette roots. Data from three independent experiments are shown. (C) Contrasted electron micrograph of the syncytium and adjacent cell of Fayette root infested with SCNs, after immunogold label detection using anti–α-SNAP_Rhg1HC primary antibody [similar to Fig. 4C (noncontrasted)]. Adj., adjacent cell; CW, cell wall; ER, endoplasmic reticulum; M, mitochondrion; Syn., syncytial cell; Vac, vacuole. Arrows highlight four of many gold particle-labeled α-SNAP_Rhg1HC regions. (D) Contrasted electron micrograph of mock-inoculated Fayette root after immunogold label detection using anti–α-SNAP_Rhg1HC primary antibody. (E) Electron micrograph of a syncytium site of Fayette root infested with SCNs, where the primary anti–α-SNAP_Rhg1HC antibody was competitively bound with a 10-fold molar excess of antigen (recombinant α-SNAP_Rhg1HC protein) before immunolabeling of the microscopy section. After the initial competitive binding, anti–α-SNAP_Rhg1HC primary antibody was incubated with fixed cross-sections of SCN-infested Fayette roots and probed with secondary goat anti-rabbit antibody conjugated to 15-nm gold particles. Multiple cross-sections were examined using competitively bound α-SNAP_Rhg1HC primary antibody and little to no gold particle labeling was observed, indicating high antigen specificity. (F) Immunogold labeling using only secondary goat anti-rabbit antibody on SCN-infested roots. No previous incubations with α-SNAP_Rhg1HC antibody were performed. Little to no gold particle labeling is present, indicating α-SNAP_Rhg1HC labeling in SCN-infected roots is highly specific. (Scale bars in E and F, 1 μm.)