Leucine zipper motif in RRS1 is crucial for the regulation of Arabidopsis dual resistance protein complex RPS4/RRS1

Arabidopsis thaliana leucine-rich repeat-containing (NLR) proteins RPS4 and RRS1, known as dual resistance proteins, confer resistance to multiple pathogen isolates, such as the bacterial pathogens Pseudomonas syringae and Ralstonia solanacearum and the fungal pathogen Colletotrichum higginsianum. RPS4 is a typical Toll/interleukin 1 Receptor (TIR)-type NLR, whereas RRS1 is an atypical TIR-NLR that contains a leucine zipper (LZ) motif and a C-terminal WRKY domain. RPS4 and RRS1 are localised near each other in a head-to-head orientation. In this study, direct mutagenesis of the C-terminal LZ motif in RRS1 caused an autoimmune response and stunting in the mutant. Co-immunoprecipitation analysis indicated that full-length RPS4 and RRS1 are physically associated with one another. Furthermore, virus-induced gene silencing experiments showed that hypersensitive-like cell death triggered by RPS4/LZ motif-mutated RRS1 depends on EDS1. In conclusion, we suggest that the RRS1-LZ motif is crucial for the regulation of the RPS4/RRS1 complex.


Results
Transient expression of dual R proteins. To investigate the expression of dual R proteins, RPS4 and RRS1, in Nicotiana benthamiana leaves using Agrobacterium-mediated transient expression assays, 3 × FLAG and 4 × Myc tags were fused into the N-terminus of RPS4 and RRS1, respectively. These tagged constructs were driven by the cauliflower mosaic virus 35S promoter, while the omega leader sequence, which is known to act as a translational enhancer in plants, served as the 5′ UTR 26,27 (Fig. 1a).
Immunoblot assays of transiently expressed or co-expressed FLAG-RPS4 and Myc-RRS1 in N. benthamiana showed that co-expressed full-length RPS4 and RRS1 gave a strong signal on immunoblot compared with transiently expressed RRS1 or RPS4 that were both weakly detected (Fig. 1b). In addition, transiently co-expressed FLAG-RPS4 and Myc-RRS1 were mainly localised in N. benthamiana microsomal fractions and weakly localised in nucleus fractions (Fig. 1c). Several previous studies also showed that RPS4 was localised in the microsomal fraction 25,28 and that when transiently expressed in N. benthamiana, RPS4 and SNC1 formed a common protein complex in cytoplasmic microsomal compartments 28 .
Co-immunoprecipitation (Co-IP) analysis on full-length RPS4 and RRS1 proteins isolated 44 h after Agrobacterium infiltration showed that they interacted with one another in the microsomal fraction (Fig. 2). The interaction between the RPS4 TIR domain and full-length RRS1 was not detected in the microsomal fraction with Co-IP, although both were detected in this fraction with immunoblotting. Our results suggested that other domains were also required for RPS4/RRS1 interaction, even though the TIR domains of RPS4 and RRS1 formed a heterodimer 21 .
To test whether tag-fused RPS4 and RRS1 confer resistance against C. higginsianum, we complementarily introduced N-or C-terminal HA-tagged RRS1 and N-terminal Myc-tagged RRS1 into the rrs1-1 A. thaliana Ws-2 accession mutant line. Resistance to C. higginsianum was fully restored in the transgenic plants with N-terminal tagged RRS1, but was not restored in the transgenic plants with C-terminal tagged RRS1 (Fig. S1). These results suggested that a proper C-terminus structure in RRS1 is important for C. higginsianum resistance in Arabidopsis. Similarly, we complementarily introduced N-terminal FLAG-tagged RPS4 and C-terminal YFP-tagged RPS4 into the rps4-21 A. thaliana Ws-2 accession mutant line. Resistance to C. higginsianum was fully restored in the transgenic plants with N-terminal tagged RPS4, but was not restored in the transgenic plants with C-terminal tagged RPS4 (Fig. S1). Thus, a proper C-terminus structure in RPS4 is likely also important for C. higginsianum resistance in Arabidopsis. In addition, we showed via immunoblotting that N-terminal tagged 4 × Myc-RRS1 and 3 × FLAG-RPS4 were detected in transgenic Arabidopsis, but the transgenic plants with C-terminal tagged RRS1-3 × HA, RPS4-YFP and N-terminal tagged 3 × HA-RRS1 were not detected (Fig. S1). The amount of these proteins may be undetectable with immunoblotting under our conditions. It is important that the N-terminal tagged RRS1 and RPS4 used are functional.
The LZ motif is completely conserved in these 19 accessions, which contain both RRS1-R and RRS1-S. Saucet et al. 31 had reported that RRS1B and RPS4B, similar to RRS1 and RPS4, confer recognition of AvrRps4 but not PopP2. In RRS1B from Ws-2, we also found two LZ motifs, LKGSLSSLPNVLRLLHWENYPL and LRVRYAGLQEIYKALFLYIAGL.
The LZ is a protein-protein interaction domain consisting of an α -helical conformation with a leucine residue at every seventh position, which often facilitates dimerisation 32 . To investigate the role of the LZ motif in RRS1 (from Ws-2), we generated mutations in three leucine residues within the zipper (Fig. 3b) and named the mutated protein RRS1Δ lz. We observed that the rrs1-1/RRS1Δlz mutants grew abnormally and constitutively expressed the inducible defence gene, PR1. Compared with expression in wild-type Ws-2, PR1 gene expression was three hundred thousand-(rrs1-1/RRS1Δlz#1) and one hundred fifty thousand-fold (rrs1-1/RRS1Δlz#2) higher. Therefore, introducing RRS1Δlz induced autoimmune response (Figs 3c and 4). Interestingly, the rrs1-1/rps4-21/ RRS1Δlz mutants, in which RPS4 was absent, grew normally and did not constitutively express PR1, suggesting that RPS4 was required for autoimmune response in the rrs1-1/RRS1Δlz mutant (Figs 3 and 4).

Transient expression of dual R proteins induces an HR-like cell death in N. benthamiana.
Transient expression of both full-length RPS4 and RRS1Δlz induced HR-like cell death (Fig. 5a). The HR symptoms appeared 3 d after the injection of full-length RPS4 and RRS1Δlz into N. benthamiana. However, transient expression of full-length RRS1, RRS1Δ lz, or RPS4/RRS1 did not cause HR-like cell death (Fig. 5a). In addition, we used immunoblotting to verify that HR absence was not due to absence of the R proteins. Co-expressed RRS1/RPS4 and RRS1Δlz/RPS4 yielded strong and weak signals on immunoblots, respectively (Fig. 5b). Moreover, RRS1Δ lz and RRS1 were weakly detected in the total protein extraction from RRS1Δlz-and RRS1-injected N. benthamiana (Fig. 5b). These results confirm that the lack of HR was not due to the absence of R proteins.
Immunoprecipitation analysis of RPS4 and RRS1Δlz. Tissue samples of N. benthamiana were harvested 44 h after Agrobacterium infiltration. The nucleus and microsomal fractions were extracted and immunoassayed with anti-FLAG and anti-Myc antibodies. Under these conditions, FLAG-RPS4 and Myc-RRS1 were detected as strong signals on immunoblotting in the microsomal fraction and as weak signals in the nucleus fraction (Fig. 1c). In contrast, RPS4 and RRS1Δ lz isolated 44 h after Agrobacterium infiltration were weakly detected in the microsomal fraction but not in the nucleus fraction, although both mRNAs were detected in the leaves ( Fig. 1c  and S2). Microsomal (Micro) and nuclear (Nucl) fractions are 10-and 60-fold more concentrated, respectively, compared with the soluble (Sol) fraction. The degree of fraction enrichment was determined using antibodies against marker proteins (cytoplasmic soluble; anti-GAPDH, microsomes; anti BiP, and nucleus; anti-Histone H3).
Scientific RepoRts | 6:18702 | DOI: 10.1038/srep18702 RPS4/RRS1Δlz-triggered cell death depends on EDS1. To investigate whether RPS4/ RRS1Δ lz-triggered cell death depends on EDS1, virus-induced gene silencing (VIGS) was used 33 . An N. benthamiana seedling was silenced for NbEDS1 by inoculation with Agrobacterium-carrying tobacco rattle virus TRV:EDS1. The N. benthamiana plants that were inoculated with Agrobacterium harbouring TRV:GFP were used as control. Following the first inoculation, RPS4 and RRS1Δlz were transiently expressed in the upper leaves of silenced N. benthamiana plants. The HR-like cell death phenotype was monitored at 7 d post-inoculation (dpi). The results indicated that co-expression of RPS4 and RRS1Δlz conferred an HR-like cell death in the TRV:00 silenced control plants (Fig. 6a). Silencing of NbEDS1 in N. benthamiana completely abolished the HR phenotype (Fig. 6a). This result indicated that VIGS of NbEDS1 impaired HR because of RPS4 and RRS1Δlz co-expression, and that NbEDS1 was required for RPS4/RRS1Δlz-triggered HR in N. benthamiana. Using VIGS, we found that cell death resulting from the co-expression of RPS4 and RRS1Δlz was due to the signalling component EDS1.
In addition, we used immunoblotting of the total protein extraction to analyse RRS1 and RRS1Δ lz accumulation in both the presence and absence of RPS4 and EDS1 (Fig. 6b). With RPS4/RRS1 co-expressed N. benthamiana, EDS1 absence resulted in a decrease of RRS1 and RPS4 accumulation, compared with EDS1 presence. However, RRS1Δ lz was weakly detected in both the presence and absence of EDS1. Finally, with RRS1Δlz expressed N. benthamiana, RRS1Δ lz was weakly detected in the absence of RPS4.
VIGS is known to reduce the level of target mRNA 34,35 . To investigate whether TRV:EDS1 silencing causes a reduction in NbEDS1 mRNA, we performed mRNA expression analysis. NbEDS1 mRNA drastically decreased in TRV:EDS1-silenced plants after the second inoculation (Fig. S3). However, NbEDS1 mRNA was expressed in the control plants infected with TRV:GFP after the second inoculation (Fig. S3).
Characterisation of RPS4/RRS1Δlz-triggered cell death. Nuclear localisation of RPS4, containing nuclear localisation sequence (NLS), is necessary for AvrRps4-triggered cell death 19,36 . We investigated whether HR was abolished when RRS1Δ lz was co-expressed with RPS4Δ nls. We found that RRS1Δ lz/RPS4Δ nls did not induce HR (Fig. 7). Therefore, nuclear localisation of RPS4 is necessary for RRS1Δ lz-mediated cell death.
Co-expression of RRS1WT was reported to prevent RRS1-SLH1/RPS4-dependent constitutive HR, indicating that RRS1-SLH1-dependent auto-activation is recessive 36 . We also investigated whether RRS1Δ lz/RPS4-dependent constitutive HR was prevented by co-expression of RRS1WT. Our results revealed that HR was weakened, but not abolished, when triggered by RRS1Δ lz/RPS4/RRS1WT as opposed to RRS1Δ lz/RPS4. These data suggest that RRS1WT did not completely interfere with RRS1Δ lz/RPS4 triggered HR (Fig. 7). Therefore, auto-active alleles of RRS1Δlz are semi-dominant.

Discussion
Dual R proteins, RPS4 and RRS1, function as a complex and mediate defence response 16,17 . Zhang and Gassmann 37 previously reported that 35S:FLAG-genomicRPS4-Ler construct produced low levels of full-length protein in N. benthamiana. In this study, we showed that full-length N-terminally FLAG-tagged RPS4 was only slightly detected in the microsomal fraction eluted from N. benthamiana leaves in the absence of RRS1. In contrast, co-expressed full-length RPS4 and RRS1 gave strong signals on immunoblots, indicating that RPS4/RRS1 is stable in planta.   The asterisks indicate significant differences compared with wild-type Ws-2 (Dunnett's method, P < 0.05) 50 . The nucleotide sequence of the gene-specific primer is listed in Table S1. Using co-immunoprecipitation assays, we found that full-length RPS4 and RRS1 strongly interact with one another in planta. Williams et al. 2014 also showed that RPS4 interacts with the RRS1 TIR domain 21 . However, co-expressed full-length RPS4 and RRS1Δ lz were weakly detected in the microsomal fraction and the total protein extraction eluted from N. benthamiana leaves. Additionally, we verified with quantitative real-time polymerase chain reaction (qRT-PCR) that RPS4 and RRS1 mRNAs accumulated at similar levels in N. benthamiana leaves (Fig. S2).  Previous studies showed that the N-terminal NBS and LZ motif are critical for the function of NBS-LRR type R protein, RPS2 38 , and that WRKY transcription factors play an important role in plant defence 39 . WRKY 18, WRKY 40, and WRKY60 have potential LZ motifs at the N-terminus that are involved in the physical interaction of these WRKY proteins 40 . WRKYs with no LZ motifs are unable to interact with themselves and with each other. The C-terminal portion of RRS1 also possesses two conspicuous structural motifs, the LZ and the WRKY domain, that play a key role in protein-protein interaction. In this study, direct mutagenesis revealed that the C-terminal LZ motif in RRS1 is important for RPS4/RRS1 immunity. Noutoshi et al. also showed that a 3-bp insertion mutation of the WRKY domain in RRS1 (named RRS1-SLH1) causes autoimmunity 41 .
In this study, transient co-expression of RPS4/RRS1Δlz in N. tabacum and N. benthamiana induced strong HR-like cell death. However, we did not observe HR-like cell death in N. benthamiana induced by the transient expression of full-length RRS1, RRS1Δlz, or RPS4/RRS1. In N. tabacum, when transient RPS4 expression in the absence of AvrRps4 causes HR, it is called the R gene overdose effect 19,42 . We also observed stunting and mimic lesions in RPS4/RRS1Δlz-transgenic Arabidopsis plants, but not in rps4/RRS1Δlz transgenic plants without RPS4. We assume that RRS1 regulates the activation of RPS4 in the absence of Avr proteins, and Avr proteins either direct or indirect modify RRS1 to activate RPS4. The structural change of RRS1 by Avr proteins is necessary for the activation of downstream defence responses. Therefore, the coexistence of RRS1 and RPS4 is essential for pathogen recognition and defence responses. On the other hand, RRS1B and RPS4B also confer recognition of AvrRps4 but not PopP2 31 . In our preliminary experiments, we found that RRS1B was not required for resistance to C. higginsianum.
Interestingly, we found that Arabidopsis mutants complemented by C-terminally tagged RPS4-or RRS1-transgenes did not confer resistance to C. higginsianum, whereas those complemented by N-terminally tagged RPS4-or RRS1-transgenes did. These results imply that the C-termini of both RPS4 and RRS1 play a key role in the activation of RPS4/RRS1-mediated defence responses. We previously reported that RRS1 alleles from three C. higginsianum-susceptible accessions, Bur-0, Col-0, and Cvi-0, contain a premature stop codon 16 . In these accessions, the C-terminal region of RRS1 that follows the WRKY domain is relatively short compared with several resistant accessions (Ws-2 etc.). Recent reports 22,23 suggest that C-terminal extension is necessary for resistance signalling, because the C-terminal region is required for PopP2 but not for AvrRps4. Thus, the C-terminal region of RRS1 likely plays a key role in the activation of RPS4/RRS1-mediated defence responses.
EDS1 encodes a lipase-like protein that functions in R protein-mediated and basal plant disease resistance 43 . In addition, EDS1 and NbEDS1 are required for RPS4-mediated disease resistance in Arabidopsis and tobacco, respectively 42,44 . In present study, we found that RPS4/RRS1Δ lz-triggered HR is completely abolished in NbEDS1-silenced N. benthamiana plants. Our results suggested that RPS4/RRS1Δlz-triggered HR also requires NbEDS1 as a signalling component in N. benthamiana plants.
In conclusion, our data indicated that RPS4/RRS1 is required for active defence responses, and some structural domains that constitute these R proteins contribute to the interaction of RPS4 with RRS1. Construction of the R-gene plasmid. All DNA fragments containing RRS1 and/or RPS4 used in this study derived from genome of the A. thaliana Ws-2 accession. Plasmids used in this study were constructed by Gateway ® technology following manufacturer protocol (Life Technologies, USA). All clones were verified with DNA sequencing. The pCR8GW-RR-Ws plasmid was cloned using a 10.9-kbp genomic fragment containing RPS4 and RRS1 as previously described 17 . The genomic fragment of RRS1 (coding region) was PCR-amplified using pCR8GW-RR-Ws as template and cloned into pCR8GW-TOPO (named pCR8GW-gRRS1). To generate the destination vector pGWB18Ω (35S:Ω :4xMyc), the HindIII and XbaI regions in pGWB18 were replaced with 35S enhancers and the omega leader sequence cassette in pBE2113 45 . To create the destination vector pGWB18-RRS1p, the 35S promoter region in pGWB18 was replaced with the RRS1 promoter region (1.8 kbp upstream of the start codon) using the In-Fusion ® HD Cloning Kit (Takara Bio Inc., Japan). To generate the 35S:Ω :4 × Myc-gRRS1 and RRS1p:4 × Myc-gRRS1 constructs, LR reactions were performed to recombine the entry clone, containing genomic RRS1, into the Gateway ® -compatible destination vectors: pGWB18Ω for Agrobacterium-mediated transient expression and pGWB18-RRS1p for stable Arabidopsis transformation via LR reaction. Site-directed mutagenesis of genomic RRS1 with or without the promoter region 16 was performed by a custom cloning service (Takara Bio Inc., Japan) to generate RRS1Δlz carrying Leu to Ala mutations (L1089A, L1096A, and L1103A) at the LZ motif. Subsequently, LR cloning was used to generate binary constructs in pGWB18Ω for Agrobacterium-mediated transient expression and in pGWB1 for stable Arabidopsis transformation. The 6.3-kbp genomic RPS4 fragment, including approximately 2.1-kbp upstream and 109-bp downstream regions, was cloned into pCR8GW-TOPO (named pCR8GW-RPS4p:gRPS4). To generate the RPS4p:3 × FLAG-gRPS4 construct, synthetic 3 × FLAG and a spacer sequence, MDYKDHDGDYKDHDIDYKDDDDKGGGS, was inserted just before the RPS4 start codon by a custom cloning service (Life Technologies, USA). To generate the pCR8GW-35S:Ω: 3 × FLAG-gRPS4, the RPS4 promoter region in pCR8GW-RPS4p:gRPS4 was replaced by 35S enhancers and the omega leader sequence cassette in pBE2113, using the In-Fusion ® HD Cloning Kit (Takara Bio Inc., Japan).
Arabidopsis transformation. Arabidopsis transformation was carried out according to the floral inoculating method using Agrobacterium tumefaciens strain GV3101 (pMP90) 46 .

C. higginsianum inoculation and quantification of C. higginsianum actin mRNA. C. higginsianum
Saccardo isolates (MAFF305635) were obtained from the Ministry of Agriculture, Forestry and Fisheries (MAFF) Genebank, Japan. Arabidopsis plants were inoculated as described previously 47 and harvested at 5 dpi for qRT-PCR analysis. The quantification of C. higginsianum was performed as described previously 47 . N. benthamiana and N. tabacum. N. benthamiana and N Protein fractionation and immunoblotting. Microsomal and soluble fractions were prepared as described previously 28 . Extraction buffer (50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.5, 250 mM sucrose, 15 mM ethylenediaminetetraacetic acid, 5% glycerol, 0.5% polyvinylpyrrolidone, 3 mM dithiothreitol, 2× Roche protease-inhibitor cocktail and 1× Roche phosphatase-inhibitor cocktail) (Roche, Switzerland) was added to the plant materials, which was ground with mortar and pestle, then with a Potter-Elvehjem tissue grinder. The extracts were filtered through two layers of miracloth pre-wetted with extraction buffer and centrifuged at 2,000 × g for 10 min at 4 °C. The supernatant consisting of the cytoplasmic fraction was further subjected to ultracentrifugation at 100,000 × g to separate the soluble and microsomal (pellet) fractions. The pellet was resuspended in extraction buffer containing 0.5% Igepal CA-630 (Sigma-Aldrich, USA).

Transient expression assay in
Nuclear extracts were prepared with a semi-pure preparation method using the CelLytic TM PN Isolation/ Extraction Kit (Sigma-Aldrich, USA) and following manufacturer protocol, with one modification: 1× Roche phosphatase-inhibitor cocktail was added into NIBA buffer.
Total proteins were extracted according to previously described methods 23 . Total protein was separated on a 4-15% sodium dodecyl sulfate polyacrylamide gel (BioRad, USA) and transferred onto a polyvinylidene difluoride membrane. Immunoblots were performed with monoclonal anti-FLAG (Sigma-Aldrich, USA) and anti-c-Myc (Roche, Switzerland) antibodies, as well as a secondary HRP-conjugated anti-mouse antibody (Promega, USA), then visualised with chemiluminescence (ECL; Bio-Rad, USA). The degree of enrichment in cellular fractionation was determined by immunoblot analyses with anti-GAPDH (Genscript, USA), anti-BiP (Agrisera, Sweden), and anti-histone H3 (Abcam, USA) antibodies. Secondary HRP-conjugated anti-mouse, anti-goat, and anti-rabbit antibodies (Promega, USA) were used for detecting chemiluminescent immunoblots.
The co-immunoprecipitation assays were performed using Anti-FLAG M2 affinity and anti-c-Myc affinity gels (Sigma, USA) following manufacturer protocol. The protein concentrations of fractions were determined with Bradford assays 48 using bovine serum albumin as a standard. Protein samples (1 mg) of microsomal fractions Scientific RepoRts | 6:18702 | DOI: 10.1038/srep18702 were incubated with Anti-FLAG M2 for 2 h or anti-c-Myc for 1 h affinity gels (Sigma-Aldrich, USA) at 4 °C. The immunoprecipitates were analysed with immunoblotting using monoclonal anti-FLAG (Sigma-Aldrich, USA) and anti-c-Myc (Roche, Switzerland) antibodies. A special secondary antibody (Mouse TrueBlot anti-Mouse Ig HRP; Rockland, USA) that only detects full-length immunoglobulin was used for immunoblotting to avoid any nonspecific bands.

VIGS.
Using a tobacco rattle virus vector, VIGS of the N. benthamiana homologue EDS1 was performed as described previously 49 . Nicotiana benthamiana plants (14-day-old) were infiltrated with Agrobacteria carrying TRV:EDS1 to silence NbEDS1. After 3 to 4 weeks, the upper leaves of these plants were used for an Agrobacterium-mediated transient assay.