A Truncated NLR Protein, TIR-NBS2, Is Required for Activated Defense Responses in the Mutant

Summary

Secretory pathways play an important role in the plant immune response by delivering antimicrobial compounds and metabolites to the site of infection. The evolutionarily conserved exocyst complex is involved in exocytosis, the final step in the secretory pathway. We showed that loss of the function of EXO70B1, a subunit of exocyst complex, results in activated defense responses, and enhanced resistance to a range of pathogens. We found that EXO70B1 associates with the SNARE complex protein SNAP33, which is involved in focal secretion of defense-related proteins. Enhanced disease resistance and cell death in the exo70B1 mutant are dependent on TIR-NBS2 (TN2), an atypical intracellular immune receptor-like protein that lacks leucine-rich repeats. TN2 physically associates with EXO70B1, and TN2 transcripts accumulate at much higher levels in the exo70B1 mutant. These data are consistent with a model where activation of a receptor pathway containing TIR-NBS2 is responsible for activated defense responses and cell death in exo70B1. Our data further suggest that this, and possibly other, exocyst components could be targets of effectors that are guarded by immune receptors.

Introduction

Powdery mildew fungi are obligate biotrophic pathogens that cause widespread disease in many plant species, including economically important crops such as barley [1], wheat [2] and tomato [3] as well as model plants such as Arabidopsis [4]. Powdery mildew fungal spores germinate on the leaf surface, and then extend hyphae, which form appressoria at the attempted penetration sites. These appressoria penetrate the epidermal cell wall. The fungus then invaginates the plant plasma membrane and develops a haustorium, a feeding structure [5]. Evolutionarily adapted powdery mildew fungal pathogens can complete the infection cycle by producing conidia, asexual spores, on the host leaf surfaces. For instance, the adapted powdery mildew pathogen for Arabidopsis, Golovinomyces cichoracearum, produces abundant conidia on Col-0 leaves after infection. Several types of barley and Arabidopsis mutants show altered responses to adapted powdery mildew pathogens; these include mlo (mildew locus O) [1,6], pmr (powdery mildew resistant) [6–10] and edr (enhanced disease resistance) [11–14]. These mutants generally are more resistant to adapted powdery mildew pathogens, indicating that MLO, PMR, and EDR have negative roles in powdery mildew resistance.

Adapted powdery mildew fungi can infect the host plant; by contrast, non-adapted powdery mildew pathogens usually cannot penetrate plant epidermal cell walls or form haustoria. Genetic analyses demonstrated that the full penetration resistance of Arabidopsis to non-adapted powdery mildew requires the functions of PEN1, PEN2 and PEN3 [15–17]. PEN1 encodes syntaxin SYP121, which participates in vesicle fusion events by forming ternary soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) complexes with the synaptosome-associated membrane protein 33 (SNAP33) and vesicle-associated membrane protein721 (VAMP721) and VAMP722 [15,18].

Before fusion occurs, SNARE-mediated vesicles must dock at the plasma membrane; this requires the exocyst, which mediates the early steps of exocytosis [19]. The exocyst complex, originally defined in yeast [20], consists of 8 proteins: Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, EXO70, and EXO84. Among these eight proteins, Sec3 and EXO70 are targeted to the plasma membrane by binding to phosphatidyl (4,5) biphosphate [21,22].

Homologs of all eight exocyst proteins have been found in plants [23–25]. Plant exocyst proteins function in regulation of polarity and morphogenesis [26]. In Arabidopsis, Sec6, Sec8, and EXO70A1 localize to the tips of growing pollen tubes and function in polar secretion required for growth and pollen incompatibility [25,27]. Sec3 and EXO70A1 function in polarized secretion in elongating root hairs [28] and in the interaction between the stigma and compatible pollen [29]. Sec6, Sec8, Sec15b, EXO70A1, and EXO84b also function in secretory processes during cytokinesis, and Sec8 and EXO70A1 are required for the localized deposition of seed coat pectin [30,31]. In addition, EXO70A1 also affects auxin efflux carrier recycling, polar auxin transport [32] and tracheary element development [33].

EXO70 clade members also function in plant immunity [26]. For instance, Arabidopsis EXO70B2 and EXO70H1 are transcriptionally induced by Pseudomonas syringae pv. maculicola and the fungal pathogen Bgh and are required for full plant defense responses. Further, EXO70B2 associates with SNAP33, and contributes to plant immunity [34]. EXO70B2 is a target of the plant U-box-type ubiquitin ligase 22 (PUB22) and is required for PAMP-triggered responses [34,35]. EXO70B1 also affects autophagy-related membrane traffic to the vacuole, and the exo70B1 mutant displays ectopic hypersensitive reaction [36].

Plant disease resistance (R) proteins play critical roles in plant immunity [37]. The R proteins directly or indirectly recognizes pathogen effectors, and leads to effector triggered immunity. Most of the R proteins are intracellular, and belongs to nucleotide binding (NB) domain and leucine-rich repeat (LRR)-containing (NLR) immune receptors, which can be divided to two subclasses, based on the N-terminal domains. The first class of NLR immune receptors contains a Toll-Interleukin-1 Receptor (TIR) domain and the second class contains a coiled-coil (CC) domain [37]. Activation of NLR leads to hypersensitive responses which is characterized by localized cell death at site of infection [37]. Besides those full length NLR immune receptors, there are a number of atypical NLRs that lack LRR domain in plant genome [38]. Although those truncated NLR proteins are implicated in plant immunity, the function of those proteins are not well understood [39].

Here, we report that the exo70B1 mutant displays enhanced resistance to powdery mildew. During the course of this work, it was shown that exo70B1 displayed spontaneous cell death, which was delayed by npr1 mutation suggesting that at least part of the mutant phenotype is due to constitutive SA signaling [36]. Consistent with this, exo70B1 exhibits up-regulation of PR1, and accumulates higher levels of SA in the absence of pathogen attack [36]. Here, we show that EXO70B1 associates with the SNARE complex protein SNAP33. In a suppressor screen, we found that the exo70B1-associated phenotypes require the TIR-NBS gene TN2. These data indicate that mutation of EXO70B1 leads to activation of a defense pathway that requires TN2, a truncated form of the classical TIR-NBS-LRR intracellular immune system receptors.

Results

The exo70B1-3 mutant displays enhanced disease resistance to Golovinomyces cichoracearum

To study the molecular interactions between Arabidopsis and powdery mildew fungus, we screened for mutants that displayed enhanced disease resistance and an exaggerated cell death response following infection with G. cichoracearum UCSC1 [14]. We designated one of the recessive mutants identified in this screen exo70B1-3, based on our subsequent characterization. Under standard short day conditions, exo70B1-3 grew normally up to 4 weeks of age. After 4 weeks, exo70B1-3 plants were slightly smaller than the wild type, and displayed HR-like cell death after 5 weeks (Fig. 1A and S1 Fig.). After inoculation with G. cichoracearum at high inoculum density, the leaf surface of the wild type was covered with abundant fungal spores at 7 dpi, but exo70B1-3 exhibited extensive necrotic lesions and no visible powder of fungal spores on leaves (Fig. 1A). Trypan blue staining revealed abundant fungal hyphae and conidiophores on the leaves of the wild type, but showed very few fungal hyphae and large patches of dead cells on the leaves of exo70B1-3 (Fig. 1B). To further assess the powdery mildew resistance of exo70B1-3, we quantified fungal growth in plants at 5 dpi. Consistent with the trypan blue staining, exo70B1-3 supported significantly fewer conidiophores than the wild type (Fig. 1C), indicating that fungal reproduction was inhibited in the exo70B1-3 mutant.

Plant responses to pathogen infection include: callose deposition, hydrogen peroxide (H₂O₂) accumulation, and MAPK activation [40]. We used aniline blue to visualize callose deposition in the infected leaves of exo70B1-3 at 2 dpi. As shown in Fig. 1D, very little callose was deposited at the infection sites of pathogen in the wild type, but a large amount of callose was produced at the sites of fungal infection in exo70B1-3 at 2 dpi, when cell death did not occur. In plants infected with pathogens, cell death triggered by specific recognition is usually associated with H₂O₂ accumulation [41,42]. We examined H₂O₂ accumulation in wild type and exo70B1-3 by staining the infected leaves with 3,3’-diamino benzidine hydrochloride (DAB) and trypan blue and found that exo70B1-3 leaves accumulated more H₂O₂ upon fungal attack than wild type (Fig. 1E).

We also examined the activation of MPKs upon infection with G. cichoracearum, by immunoblotting with anti-pTEpY antibody, which detects phosphorylated MPK3 and MPK6 [43]. As shown in Fig. 1F, MPK3 and MPK6 were activated in both wild type and exo70B1-3 at 3 dpi; however, activation of MPK3, MPK6 in exo70B1-3 was much stronger than in wild type. Protein levels of MPK3 and MPK6 were similar in infected and uninfected wild type and mutant plants (Fig. 1F).

To examine whether exo70B1-3 affects the expression of immune output genes, we measured the transcript levels of the pathogenesis-related gene PR1, a marker for salicylic acid signaling, at various time points from 0 to 5 dpi with G. cichoracearum, using quantitative real time RT-PCR. Prior to inoculation, PR1 transcript levels were very low in both genotypes, but the basal expression levels of PR1 were higher in exo70B1-3 than in wild type. Upon infection with G. cichoracearum, PR1 transcript accumulated at much higher levels in exo70B1-3 than in wild type (Fig. 2A). We also examined expression of other immune output genes, including PR2 (Fig. 2B), PR5 (Fig. 2C), ALD1, FMO1, PAD4 and SID2 (Fig. 2D). In general, the relative transcript levels of those genes were low at day 0, but were much higher in exo70B1-3 than in wild type at day 0 and 3 dpi. However, for PAD4, SID2 and NPR1, we observed significant differences in transcript levels between exo70B1-3 and wild type only at day 3, not at day 0 (Fig. 2D). Taken together, these data indicated that exo70B1-3 displayed enhanced disease resistance, and upon infection, defense responses were activated at higher levels in exo70B1-3 than in wild type.

PAD4, EDS5 and NPR1 contribute to exo70B1-3 mediated powdery mildew resistance, but only PAD4 is required for ectopic cell death

Salicylic acid (SA) plays critical roles in the defense response to pathogens and activation of SA signaling pathways often leads to disease resistance and cell death in response to pathogen attack [44]. To assess the role of SA signaling in the disease resistance observed in exo70B1-3, we crossed exo70B1-3 with pad4, sid2 and eds5 mutants, which express lower levels of pathogen-induced SA [45–47] or npr1, which has defects in SA signaling [48]. Double mutants were then infected with G. cichoracearum and phenotypes were scored at 7 dpi. Only pad4 completely suppressed both powdery mildew resistance and mildew-induced cell death in exo70B1-3 (Fig. 3A-C). The eds5 and npr1 mutations suppressed powdery mildew resistance, but did not fully suppress cell death in exo70B1-3. The sid2 mutation affected neither resistance nor cell death in exo70B1-3 (Fig. 3A-C). Similarly, spontaneous cell death in exo70B1-3 was only suppressed by pad4, but not by sid2, npr1 or eds5 (S2 Fig.). Taken together, these observations indicate that PAD4, EDS5 and NPR1 contribute to exo70B1-3 mediated powdery mildew resistance. Those observations also suggest that the cell death and ectopic disease resistance in exo70B1-3 can be uncoupled.

To further assess the role of SA in exo70B1-3 mediated resistance, we measured SA levels in exo70B1-3. Free SA levels were significantly higher in exo70B1-3 plants than in wild type in the absence of pathogen or at 3 dpi with G. cichoracearum (Fig. 3D, 3E). The SA levels in exo70B1-3 were reduced to wild type levels (or even lower) in combination with pad4, eds5 and sid2. However, the SA levels in exo70B1-3 npr1 were much higher than in exo70B1-3, which was consistent with previous findings that npr1 accumulates higher levels of SA, and that npr1 mutation often further increases SA levels in mutants that over-accumulate SA [49–51]. The observation that exo70B1-3 sid2 mutants accumulated low levels of SA, but still displayed enhanced disease resistance and mildew-induced cell death, indicates that the disease resistance and cell death phenotypes in response to G. cichoracearum in exo70B1-3, similar to the lsd1-mediated immune responses, do not result from high accumulation of SA [52].

We examined the role of ethylene and JA in exo70B1-3 mediated defenses by examining the effect of mutations in EIN2 and COI1, which block all known ethylene and JA responses, respectively [53,54]. The exo70B1-3 coi1 and exo70B1-3 ein2 double mutants did not show any alteration in disease resistance or cell death in response to G. cichoracearum infection compared to exo70B1-3 (S3 Fig.), indicating that the ethylene and JA pathways do not affect cell death and disease resistance observed in exo70B1-3.

We also crossed exo70B1-3 with two well-characterized powdery mildew resistant mutants, edr1 and pmr4, and then infected the double mutants with G. cichoracearum. As shown in S4 Fig., the edr1 and pmr4 mutations enhanced exo70B1-3 resistance phenotypes, indicating that the disease resistance in exo70B1-3 may differ from the previously characterized powdery mildew resistance mediated by edr1 and pmr4.

exo70B1-3 displayed enhanced disease resistance to the oomycete pathogen H. arabidopsidis Noco2 and the bacterial pathogen Pseudomonas syringae pv. tomato strain DC3000

To examine whether exo70B1-3 causes resistance to other pathogens, we challenged exo70B1-3 with the virulent oomycete pathogen Hyaloperonospora arabidopsidis (H. a.) Noco2 and the bacterial pathogen Pseudomonas syringae pv. tomato (Pto) DC3000. Two-week old seedlings were infected with H.a. Noco2, and we found that exo70B1-3 supported significantly fewer spores than the wild type at 7 dpi (Fig. 4A). Similarly, exo70B1-3 were more resistant to the virulent strain Pto DC3000 and the avirulent bacterial strain Pto DC3000 avrRpt2 (Fig. 4B, 4C).

EXO70B1 is an exocyst subunit

We identified the exo70B1-3 mutation by map-based cloning (S5A Fig.). To confirm that the EXO70B1 we defined is the gene responsible for the exo70B1-3 phenotype, we examined the phenotypes of two additional EXO70B1 T-DNA insertion lines N328818 (exo70B1-1) and N717829 (exo70B1-2) (S5A-S5B Fig.), and noted that these two mutants also displayed enhanced disease resistance and lesions upon inoculation with G. cichoracearum, a phenotype similar to that of exo70B1-3 (S5C Fig.). As an additional confirmation, we introduced an EXO70B1 genomic clone into exo70B1-3 plants, and found that the EXO70B1 genomic clone complemented the exo70B1-3 phenotypes (S5C Fig.). We also examined PR1 expression in the three exo70B1 alleles, the exo70B1-3 gEXO70B1 transgenic line and wild type upon mildew infection. We found that accumulation of the PR1 transcripts in the exo70B1-3 transgenic lines was similar to that in the wild type, which was much lower compared to the three exo70B1 alleles (S5D Fig.). Taken together, these data demonstrated that loss of EXO70B1 is responsible for the enhanced resistance to powdery mildew in exo70B1-3 mutants.

The 23 annotated Arabidopsis EXO70 genes fall into nine clusters, and EXO70B1 falls into the EXO70B cluster, which is very conserved in land plants [55]. Among the 23 EXO70 family members in Arabidopsis, EXO70B2 shares the highest sequence similarity with EXO70B1. However, exo70B2-1 was as susceptible as wild type in response to G. cichoracearum, supporting abundant fungal growth on the leaf surface and developing no visible lesions at 7 dpi. The exo70B1-3 exo70B2-1 double mutant was similar to exo70B1-3 (S6A-S6C Fig.).

As EXO70B1 is predicted to be a subunit of the exocyst complex, we performed yeast two-hybrid assays to examine whether EXO70B1 interacts with other exocyst subunits. Although EXO70B1 seems to be weakly auto-activating in our experimental conditions, the yeast two-hybrid assays showed that EXO70B1 interacts with SEC3a, SEC5a, SEC15b and EXO84bN (S7 Fig.). However, Kulich et al. previously showed that EXO70B1 interacted with only SEC5a and EXO84bN, but not SEC3a. [36]. The discrepancy may be because the interaction between EXO70B1 and SEC3a is relatively weak.

To further examine EXO70B1 expression, we expressed a GUS reporter gene in wild type under the control of the EXO70B1 promoter. We obtained a number of EXO70B1::GUS transformants and analyzed 20 transgenic lines. We observed GUS activity in seedlings, leaves, calyx, stigma and siliques, demonstrating that EXO70B1 is ubiquitously expressed (S8 Fig.).

EXO70B1 interacts with SNAP33

The ternary SNARE complex PEN1-SNAP33-VAMP721/VAMP722 affects powdery mildew resistance [56]. To examine whether EXO70B1 functions in the PEN1-SNAP33-VAMP721/VAMP722 pathway, we first used yeast two-hybrid assays to test whether EXO70B1 interacts with the PEN1, SNAP33 and VAMP721. As shown in Fig. 5A, EXO70B1 interacted with SNAP33, but not PEN1 or PEN1 without its transmembrane domain. To confirm the interaction between EXO70B1 and SNAP33, we used bimolecular fluorescence complementation (BiFC) in Nicotiana benthamiana [57] to examine whether EXO70B1 and SNAP33 can interact. EXO70B1 and SNAP33 were fused to the N-terminal and C-terminal fragments of YFP, respectively. In N. benthamiana leaves co-transformed with the EXO70B1-YFP^N and SNAP33-YFP^C or PEN1-YFP^N and SNAP33-YFP^C constructs, we observed YFP fluorescence (Fig. 5B and S9 Fig.). However, we observed no YFP fluorescence in leaves co-transformed with EXO70B1-YFP^N and VAMP721-YFP^C or PEN1-YFP^C, suggesting that EXO70B1 interacts with SNAP33 but not PEN1 in N. benthamiana. These results were consistent with the observations from the yeast two-hybrid assays.

To confirm the interaction of EXO70B1 and SNAP33, we also performed co-immunoprecipitation (Co-IP) in N. benthamiana. We co-expressed EXO70B1-GFP with SNAP33-Myc in N. benthamiana leaves and expressed EXO70B1-GFP or SNAP33-Myc alone as negative controls. SNAP33 protein was immunoprecipitated by anti-Myc antibody, and anti-GFP antibody detected EXO70B1-GFP in the precipitate only from the leaves that expressed both EXO70B1-GFP and SNAP33-Myc, not from the negative controls (Fig. 5C), indicating that EXO70B1 and SNAP33 associate in N. benthamiana. We also examined whether EXO70B1 interacts with PEN1 by Co-IP assays; however, as shown in S10 Fig., EXO70B1 did not co-immunoprecipitate with PEN1, indicating that EXO70B1 and PEN1 did not associate, consistent with the results from the yeast two-hybrid and BiFC assays.

To further confirm the association of EXO70B1 and SNAP33, we performed Co-IP assays using stably transformed transgenic Arabidopsis plants expressing both EXO70B1-GFP and SNAP33-HA under their respective native promoters. The EXO70B1-GFP construct complemented the exo70B1-3 mutant phenotypes (S11 Fig.), indicating that the EXO70B1-GFP protein is functional. Total protein was extracted from the transgenic plants and SNAP33-HA protein was immunoprecipitated by anti-HA antibody. We then examined whether EXO70B1-GFP was in the precipitate by immunoblot analysis with the GFP antibody. EXO70B1-GFP protein was only detected in the transgenic plants that expressed both EXO70B1-GFP and SNAP33-HA, not in the negative control (Fig. 5D), indicating that EXO70B1-GFP associates with SNAP33-HA in Arabidopsis.

EXO70B1 interacts with SNAP33, a component of the PEN1-containing ternary SNARE complex involved in penetration resistance. To examine whether the exo70B1 mutant phenotype is altered in the absence of PEN1, we made pen1-1 exo70B1-3 double mutants, and then infected wild type, exo70B1-3, pen1-1, and pen1-1 exo70B1-3 with adapted and non-adapted powdery mildew pathogens. Both exo70B1-3 and pen1-1 displayed enhanced resistance to G. cichoracearum, and pen1-1 exo70B1-3 was very similar to pen1-1 or exo70B1-3 (Fig. 6A-C). However, exo70B1-3 mutants were similar to wild type when infected with non-adapted wheat powdery mildew pathogen Blumeria graminis f. sp. tritici. But pen1-1 showed a significantly higher frequency of epidermal single cell death, and exo70B1-3 pen1-1 was very similar to pen1-1 (Fig. 6D), suggesting that the loss of EXO70B1 may not affect PEN1-mediated resistance to a non-adapted pathogen, which is consistent with the observations that EXO70B1 does not interact with PEN1.

Enhanced disease resistance in exo70B1 requires the TIR-NBS gene TN2

We performed a genetic screen to identify suppressors of exo70B1. We mutagenized exo70B1-3 seeds with EMS, and screened the M2 plants for mutants that displayed wild-type like phenotypes in response to G. cichoracearum. We obtained more than 20 mutants, which fell into 5 complementation groups, including nine alleles of At1g17615, which we defined by standard map-based cloning (S12 Fig.), seven alleles of PAD4 (consistent with our double mutant analysis, above) and four mutant alleles of a gene encoding a calcium dependent protein kinase (which will be reported elsewhere). The mutation in At1g17615 fully suppresses exo70B1-associated phenotypes, including spontaneous cell death (S13A Fig.), powdery mildew resistance and cell death (Fig. 7A-C), increased accumulation of H₂O₂ and SA (S13B-S13C Fig.) and enhanced resistance to Pto DC3000 (Fig. 7D), indicating that enhanced disease resistance in exo70B1 requires At1g17615. In addition, the up-regulation of PR1 expression in exo70B1-3 during G. cichoracearum infection was also suppressed by At1g17615 mutation (Fig. 7E). However, activation of MAPK in exo70B1-3 was not suppressed by At1g17615 mutation (S13D Fig.), indicating that activation of MAPK in exo70B1-3 is not dependent on At1g17615. Sequence analysis revealed that At1g17615 encodes a TIR-NBS protein, previously designated as TIR-NBS2 (TN2). TN2, and a further 20 genes in Arabidopsis ecotype Col-0, lack leucine-rich repeats [38]. Among the nine tn2 alleles obtained in the screen, three carry missense mutations in the TIR domain, four in the NBS domain, and two carry premature stop mutations (Fig. 7F). TN2 does not express a conserved canonical P-loop, which is likely involved in nucleotide binding, but rather encodes GRS instead of the canonical GKT in the P-loop motif (S14 Fig.). However, the mutant allele tn-6 (G222R) is in the traditionally defined P-loop (S14 Fig.), and G222 of TN2 is highly conserved among GRS containing NB-TIRs (S14 Fig.), suggesting that the GRS motif of TN2 and its relatives might retain some P-loop function. TN2 transcripts accumulated at much higher levels in exo70B1 compared to wild type (Fig. 7G); and the accumulation of TN2 transcripts was further increased upon G. cichoracearum infection (Fig. 7H). The tn2 mutations suppress exo70B1-associated resistance and cell death phenotypes, indicating that enhanced disease resistance and cell death in exo70B1 requires TN2. The pad4 mutation suppressed the increased accumulation of TN2 transcripts in exo70B1-3 (S15 Fig.), consistent with the observations that pad4 suppressed exo70B1-3 associated phenotypes (Fig. 3) and suggesting that the up-regulation of TN2 is important for the exo70B1-3 mutant phenotype. In addition, TN2 transcripts over-accumulated in older exo70B1-3 plants (S16 Fig.), which is consistent with the observations that exo70B1-3 displays spontaneous cell death after 5 weeks.

To further assess the role of TN2 in plant immunity, we examined the tn2 single mutant for resistance to fungal and bacterial pathogens. The tn2 single mutant displayed wild type like responses to virulent fungal pathogen G. cichoracearum (S17A-S17C Fig.) and bacterial pathogen Pto DC3000 (S17D Fig.), indicating that TN2 does not contribute to basal defense to G. cichoracearum or Pto DC3000. To further investigate the phenotypes of exo70B1-3, tn2 and exo70B1-3 tn2 mutants in response to Pto DC3000, we also inoculated those plants by spray infection. As shown in S18 Fig., exo70B1-3 displayed enhanced resistance to Pto DC3000, and the resistance was fully suppressed by the tn2-1 mutation, which was consistent with our findings usinginfiltration assays. In addition, TN2 was not involved in powdery mildew resistance and mildew-induced cell death in edr2, a well characterized powdery mildew resistant mutant, which accumulates higher levels of SA [58] (S19 Fig.).

To further investigate the relationship between TN2 and EXO70B1, we examined whether TN2 interacts with EXO70B1 by yeast two-hybrid assays. As shown in Fig. 8A, EXO70B1 interacts with TN2 and the TIR domain of TN2. We extended this finding by using a BIFC assay in N. benthamiana to examine whether EXO70B1 and TN2 associates. TN2 was fused to the N-terminal fragments of YFP (YFP^N), and was co-transformed to N. benthamiana with EXO70B1-YFP^C, SNAP33-YFP^C, VAMP721-YFP^C or PEN1-YFP^C, we observed YFP fluorescence only in leaves co-transformed with the EXO70B1-YFP^N and TN2-YFP^C, but not in leaves co-transformed with SNAP33-YFP^C and TN2-YFP^N, and VAMP721-YFP^C and TN2-YFP^N, or PEN1-YFP^C and TN2-YFP^N, indicating that TN2 can interact with EXO70B1 in N. benthamiana (Fig. 8B and S20 Fig.), To further confirm the interaction between EXO70B1 and TN2, we performed Co-IP assays by co-expressing EXO70B1-FLAG and TN2-Myc in N. benthamiana leaves. And EXO70B1-FLAG alone was expressed as a negative control. TN2 was immunoprecipitated using anti-Myc antibody, and in the precipitate, EXO70B1 was detected with anti-FLAG antibody only from the leaves that co-expressed both EXO70B1-FLAG and TN2-Myc, not from the negative control leaves (Fig. 8C). Taken together, those observations indicate that EXO70B1 and TN2 interact. These data are consistent with the hypothesis that TN2 or TN2 associated nucleotide binding domain and leucine-rich repeat-containing receptor is a guard directly monitoring the integrity of EXO70B1.

Discussion

EXO70B1 belongs to the EXO70 gene family, and encodes a subunit of the exocyst complex, which plays a key role in anchoring and tethering vesicles to specific sites of the plasma membrane [19]. In yeast and mammals, exocyst complex-mediated vesicle tethering is followed by SNARE protein-mediated fusion to the targeted membrane, an important step in vesicle trafficking that leads to site-directed secretion. We showed that the loss-of-function exo70B1-3 mutant displayed spontaneous HR-like cell death, enhanced disease resistance to G. cichoracearum, and mildew-induced cell death upon infection. Recent work showed that the exo70B1 mutant displayed HR-like lesions, and accumulated higher levels of SA and PR1 transcripts [36]. In addition, exo70B1 exhibited a variety of additional defects, including vacuole anthocyanin accumulation, nitrogen starvation and reduced internalized autophagosomes inside the vacuole, suggesting a role of EXO70B1 in autophagic membrane transport [36]. It will be interesting in the future to address whether these phenotypes also require TN2. Recently, Stegmann et al. reported that the exo70B1 mutant was more resistant to H. a. Emco5, and that exo70B1 exhibited increased cell death in response to the non-host oomycete pathogen Phytophtora infestans [59], consistent with our findings. However, Stegmann et al. also showed that exo70B1 displayed enhanced susceptibility to Pto DC3000 [59], in contrast to our results that exo70B1 was more resistant to this bacterial pathogen. The differences between our study and Stegmann et al. [59] could be caused by different growth conditions used in these two studies, such as humidity and light intensity.

The HR-like and enhanced disease resistance phenotypes indicate that exo70B1 behaves like lesion mimic mutant. The exo70B1-3 associated phenotypes require TN2, a TIR-NBS protein. The fact that we recovered loss of function tn2 missense changes in a conserved glycine residue in the catalytic domain of the P-loop suggest that activation of TN2 causes the exo70B1 loss of function phenotypes. Recessive mutations that give rise to enhanced disease resistance and lesion mimic phenotypes are increasingly associated with ectopic activation of NLR immune receptors, as demonstrated for Arabidopsis acd11 (accelerated cell death 11) and lsd1 (lesion simulating disease 1) [52,60,61]. Both lsd1 and acd11 result in constitutively activated immune responses in the absence of pathogens; these phenotypes are suppressed by mutations in the TIR-NLR LAZ5 or the CC-NLRs, ADR1-L1 and ADR1-L2, respectively [52,60,61]. Similar to exo70B1, the lsd1 and acd11-associated autoimmune phenotypes correlate with altered levels of SA [52,60,61].

We showed that the truncated NLR protein TN2 associates with EXO70B1, suggesting a direct functional link between TN2 and EXO70B1. Considering the important role of exocytosis in plant immunity, it would not be surprising if some pathogens evolved effectors that target the exocytosis machinery. It is worth noting that the EXO70 gene family in plants is hugely expanded, with 23 copies in Arabidopsis, compared to only one in yeast and animals [55]. Many exo70 subunits have adopted alternative functions [26], which may contribute to the expansion. One additional reason for this expansion might be that pathogen effectors target EXO70 proteins to manipulate plant immunity. The evolution of new EXO70 family members could allow the plant to evade pathogen effectors and retain antimicrobial secretory function. An alternative evolutionary possibility, and the one most consistent with our data, is that EXO70B1 has retained its function, but evolved associations with TN2, which can be activated by an as yet undiscovered effector that has evolved to manipulate EXO70B1 function. Considering that EXO70B1 contributes to autophagy, another possibility is that the plant may keep the TN2 protein at a low level via degradation of TN2 in an EXO70B1-autophagy dependent pathway. In this scenario, TN2 is activated, and also up-regulated, when EXO70B1 is not functional. Many studies have shown that autophagy plays important role in plant immunity and pathogen triggered cell death. For instance, the atg5 (autophagy-related 5) mutant displays unrestricted chlorotic cell death after infiltration of Pto DC3000 avrRpm1 [62]. Similarly, the atg2 mutant shows mildew-induced cell death and enhanced powdery mildew resistance [42]. In addition, Lenz et al. show that atg5, atg10 and atg18 mutants display enhanced resistance to Pto DC3000 [63]. Hofius et al. show that the atg7 and atg9 mutants display enhanced susceptibility to Pto DC3000 and the Emwa isolate of H. arabidopsidis, and cell death triggered by RPS4 and RPP1 was suppressed in those two mutants [64]. Although autophagy is involved in plant immunity, it has not been shown that any NLR receptor protein is specifically degraded by autophagy. It would be very interesting to examine whether the exo70B1-related autophagy defects are dependent on TN2, and whether accumulation of TN2 protein is regulated by EXO70B1-associated autophagy.

The TN2 is one of 21 TIR-NBS proteins in the Arabidopsis ecotype Col-0 genome [65]. Although the specific roles of TN proteins are not known, some evidence demonstrates that they can contribute to immune system function. For instance, overexpression of some TN genes induces cell death in tobacco, and results in stunted growth and high free SA in Arabidopsis [39,65]. Li et al. [39,65] showed that cell death in the copine protein double mutant bon1 (bonzai1) bon3 was partially suppressed by loss-of-function of some NLRs and TN2. However, the suppression was very mild since none of the plants survived to bolt. In addition, TN2 transcripts did not over-accumulate in bon1 bon3. Collectively, these data suggesting that cell death in bon1 bon3 is different from exo70B1.

Some TN proteins can interact with effector proteins from pathogens [39], suggesting that the TN proteins are likely involved in plant immunity. We cannot exclude the possibility that EXO70B1 and TN2 have other cellular functions in addition to biotic interactions. Many TN genes occur in complex genomic clusters with TIR-NBS-LRR genes. For instance, TN2 is in a cluster of 3 TIR-containing genes, including one TIR-NBS-LRR (At1g17600), and another TN gene (At1g17610, TN1, also called CHILLING SENSITIVE 1 or CHS1). Mutations in TN1/CHS1 caused auto-immunity at low temperature. As expected of a TIR-containing protein, this phenotype required EDS1 and PAD4 [66,67]. However, as no function has been assigned to the TIR-NBS-LRR gene (At1g17600) in this cluster, it will be interesting to examine whether At1g17600 is required for At1g17610 (TN1) and At1g17615 (TN2) function.

In conclusion, we found that Arabidopsis exocyst subunit EXO70B1 associates with the SNARE complex component SNAP33. Loss-of-function of EXO70B1 leads to a syndrome of defense related phenotypes, which require the TIR-NBS2 protein. We speculate that exocyst components could be targets of effectors that are, at least in this case, guarded by TN2. Alternatively, TN2 could function as an adaptor in combination with an undefined classical NLR receptor. However, we note that we did not identify any other NLR receptor in our fairly deep mutant screen (recall that we isolated many alleles of TN2 and PAD4). This would suggest that redundant classical NLR might use TN2 as an adaptor is such a model. Irrespective of the ultimate mechanistic model, we found that EXO70B1 protein interacts with TN2.

Although the role of EXO70B1 in plant immunity is not clear, work by Stegmann et al. showed that the exo70B1 mutant is compromised in flg22-triggered responses suggesting that EXO70B1 contributes to PTI [35], which makes it a potential target for pathogen effector. Important questions remain to be addressed, including: the nature of the cargos transported by EXO70B1-mediated trafficking, and how those cargos affect plant immunity, as well as how mutation of EXO70B1 activates TN2, a protein that lacks canonical regulatory domains associated with full length NLR receptors, and whether TN2 is involved in the autophagic pathway mediated by EXO70B1. Answering these questions will shed new light on plant immunity.

Materials and Methods

Plant materials and growth conditions

The Arabidopsis thaliana exo70B1-3 mutant was identified from a T-DNA insertion population [14]. T-DNA insertion lines exo70B1-1 (N328818), exo70B1-2 (N717829), exo70B2-1 (SALK_091877C) were from the Arabidopsis Biological Resource Center or Nottingham Arabidopsis Stock Center. The mutant alleles used for the double mutant analyses were as described: pad4-1 [47], npr1-1 [16], eds5-1 [45], sid2-2 [46], ein2-1 [68], coi1-1 [69], edr1 [70], edr2 [12], pmr4-1 [9], pen1-1 [15]. All of the mutants used were in the Columbia-0 (Col-0) background, except for coi1-1 (in Col-6 background). Double mutants were created by standard genetic crosses and confirmed by PCR.

Arabidopsis plants were grown in a growth room at 22–24°C, at a light intensity of 8000 lux, and a 9-h-light /15-h-dark photoperiod for phenotype analysis or 16-h-light / 8-h-dark photoperiod for seed setting, as described previously [71]. Nicotiana benthamiana plants were grown in the same short-day conditions as Arabidopsis.

Fungal infection

Adapted powdery mildew strain Golovinomyces cichoracearum UCSC1 [4] was maintained on susceptible pad4-1 mutant plants. Four-week-old plants were inoculated with powdery mildew using a settling tower at low or high densities [14] [42]. Low spore density was used to quantify the number of conidiophores per colony and high density was used to evaluate hyphae growth and lesion formation. The plants were moved to a growth chamber immediately after infection with G. cichoracearum to facilitate growth of the fungi. Hyphae and dead plant cells in the infected leaves were visualized by trypan blue staining [70]. H₂O₂ accumulation was detected at 2 dpi by 3,3’-diamino benzidine hydrochloride (DAB) staining [72]. Callose deposition was detected by aniline blue staining [73]. The samples were observed and photographed with an Olympus BX60 microscope. Non-adapted wheat powdery mildew Blumeria graminis f. sp. tritici E09 was grown on susceptible Chinese wheat variety Jing411 [74]. Assessment of epidermal cell death in response to graminis f. sp. tritici E09 was performed as described [16] with minor modifications. Briefly, four-week-old plants were inoculated with B. graminis f. sp. tritici using a settling tower, and the infected leaves was stained with trypan blue, and epidermal single cell death in the infected siytes was visulized with an Olympus BX60 microscope.

Pseudomonas syringae infection was performed by leaf infiltration (OD₆₀₀ = 5 ×10⁻⁴ for growth curve analysis) as described [75], unless otherwise indicated. For spray infection, 5-week-old plants were sprayed with bacterial suspensions (OD₆₀₀ = 0.2) in 10 mM MgCl₂ containing 0.025% Silwet L-77, and the infected plants were placed under plastic domes for 24 hr.

Inoculation of Hyaloperonospora arabidopsidis (H. a.) Noco2 (5 × 10⁴ spores/mL) was performed according to [76].

Suppressor screen

The exo70B1-3 seeds were incubated with 0.3% EMS for 16 h. Around 60,000 M2 plants derived from 5,000 M1 plants were scored by infection with G. cichoracearum. Plants showing wild type-like phenotypes were selected for further characterization.

Map-based cloning of EXO70B1 and TN2

The exo70B1-3 mutant was crossed to ecotype Landsberg erecta (Ler). F2 plants displaying exo70B1-3 mutant phenotypes were selected for mapping. Using rough mapping markers [77] and our newly developed SSLP, CAPS or dCAPS markers, we narrowed down the mutation to an 83-kb region spanning BAC MCK7 (GenBank accession AB019228) and MQJ2 (GenBank accession AB025632) on chromosome 5. We subsequently designed PCR primers to amplify the coding sequences of these genes. The suppressor gene TN2 was also identified by standard map-based cloning.

For complementation analysis, a 4.3-kb genomic sequence including 1531 bp upstream of the start codon, and 835 bp downstream of the stop codon of EXO70B1 was amplified and cloned into binary vector pCAMBIA1300. The construct was then introduced into Agrobacterium tumefaciens strain GV3101, and transformed into exo70B1-3 plants by the floral dip method [78]. Transformants were selected on 1/2 MS medium containing 80 μg/mL hygromycin. T1 transgenic plants were used for phenotyping.

Quantitative real-time RT-PCR

Extraction of total RNA, synthesis of first-strand cDNA and analysis of gene expression were performed as described previously [79].

GUS activity assay

The promoter region of EXO70B1, including 1531 bp upstream of the start codon, was cloned into the pBI121 vector. Homozygous transgenic plants in the T3 generation were used for GUS activity analysis according to [80].

SA quantification

SA extraction and HPLC assays were conducted as described previously [49,81].

Yeast two-hybrid assays

The MATCHMAKER GAL4 Two-Hybrid System 3 (Clontech) was used to examine the interactions between EXO70B1 and other exocyst subunits, as well as PEN1 and SNAP33. EXO70B1-pGBKT7 was co-transformed into Saccharomyces cerevisiae strain AH109 with SEC3a, SEC5a, SEC6, SEC8, SEC10, SEC15b, EXO84bN (the N terminal domain of EXO84b), PEN1 and SNAP33 fused with pGADT7. The constructs containing SEC3a, SEC5a, or SEC15b fused with pGADT7 were kindly provided by Dr. Viktor Zárský [25]. The constructs containing SEC6, SEC8, SEC10, EXO84bN, PEN1 and SNAP33 fused with pGADT7 were made in this study using primers listed in S1 Table. A single colony grown on SD/-Leu/-Trp plate was incubated in liquid media, the overnight culture was diluted to OD₆₀₀ = 0.5 with sterile water, and a 10 μL drop of dilution was plated on SD/-Ade/-His/-Leu/-Trp plates to test the interaction.

The interaction of EXO70B1 and TN2 was assayed using Gateway-modified versions of the traditional pEG202 and pJG4-5 vectors [82]. pDONR207 Gateway Entry clones containing ORFs (S1 Table for cloning primers) were recombined into pEG202 and pJG4-5 using LR clonase II (Invitrogen). The resulting vectors were sequence verified and then transformed into EGY48 (pEG202) and RFY206 containing the LacZ reporter pSH18-34 (pJG4-5). The resulting strains were mated and diploids selected on synthetic media (SD-HUW). Diploids were spotted onto SD-HUW +X-gal (80 μg/mL). Protein accumulation of fusion proteins in diploids was verified with standard Western blotting using anti-LexA (rabbit pAB, AbCam) and anti-HA (rat mAB, Roche) antibodies.

BiFC analysis

To create the constructs used in BiFC analysis, the coding sequences of EXO70B1 and TN2 were first cloned into pSY736 vector in frame with YFP^N. The derived YFP^N-fused-sequences were then amplified and inserted into the Gateway entry vector pDONR207 and subsequently the destination vector pMDC32. Similarly, to create YFP^C fused to EXO70B1, SNAP33, VAMP721 and PEN1, the coding sequence of each gene was cloned into pSY735 vector in frame with YFP^C. The inserts were subcloned into pDONR207 and pMDC32 vectors. Agrobacterium GV3101 carrying different YFP^N and YFP^C pairs were infiltrated into four-week-old Nicotiana benthamiana [83]. YFP fluorescence was observed by confocal microscopy (Carl Zeiss LSM710) after 48hr.

Protein extraction and immunoprecipitation in N. benthamiana and Arabidopsis

To create the constructs used in coimmunoprecipitation analysis in N. benthamiana, EXO70B1 was cloned into the pGWB11 vector with C-terminal FLAG fusion using gateway system (Invitrogen). Similarly, PEN1 was cloned into pEarleyGate 201 vector with 35S promoter and N-terminal HA fusion, while SNAP33 and TN2 were cloned into pEarleyGate 203 vector with 35S promoter and N-terminal Myc fusion. Transient expression, total protein extractions and immunoblots were performed as described previously [58]. Briefly, about 20 μg protein was resuspended in sample buffer and separated by SDS-PAGE. The protein was detected by immunoblot with primary antibody anti-GFP (Roche 1:1000) or anti-HA (Sigma 1:1000), and then with anti-mouse (rabbit) HRP-linked secondary antibody, followed by detection with the chemiluminescent HRP substrate kit (Millipore).

For immunoprecipitation, proteins prepared from NB1 buffer [84] were incubated with primary antibody (3 μL /1 mL), at 4 °C for 3 hours with gentle shaking, and 20 μL protein G agarose beads (Millipore) were added and kept shaking for another 3 hours. The beads were washed 4 times with PBS buffer, and resuspended in PBS buffer. The samples were separated by SDS-PAGE and analyzed by immunoblotting. To produce an EXO70B1-GFP fusion, a 3.4-kb genomic sequence, containing 1531bp promoter and 1872 bp coding sequence of EXO70B1 without the stop codon, was cloned into the binary vector pEGAD [85] between the StuI and AgeI cloning sites to remove the 35S promoter from the vector. The derived EXO70B1pro:EXO70B1-GFP construct was introduced into the A. tumefaciens strain GV3101 and transformed into exoB1-3 plants using the floral dip method [78]. To test the interaction between SNAP33 and EXO70B1 in Arabidopsis, we amplified SNAP33 genomic sequence (without stop codon) and fused with an HA tag, and cloned it into pMDC163 vector using gateway system (Invitrogen). And exo70B1-3 plants expressing EXO70B1pro:EXO70B1-GFP were transformed with derived SNAP33pro:SNAP33-HA construct, and the T1 plants were used for coimmunoprecipitation immunoblot assays.

To detect MAP kinase activation, 4-week-old plants were inoculated with G. cichoracearum for 3 days, and the infected leaves were excised and quickly frozen in liquid nitrogen. Total protein was isolated and subjected to immunoblot with rabbit anti-pTEpY (Cell Signaling), anti-MPK3 or anti-MPK6 (Sigma) as the primary antibody, following the manufacturer’s instructions.

Oligonucleotide Sequences

The primers used in this work are listed in S1 Table.

Supporting Information

Attachment 1

Attachment 2

Attachment 3

Attachment 4

Attachment 5

Attachment 6

Attachment 7

Attachment 8

Attachment 9

Attachment 10

Attachment 11

Attachment 12

Attachment 13

Attachment 14

Attachment 15

Attachment 16

Attachment 17

Attachment 18

Attachment 19

Attachment 20

Attachment 21