SGT1 (Suppressor of G2 allele of SKP1) is required to maintain plant disease Resistance (R) proteins with Nucleotide-Binding (NB) and Leucine-Rich Repeat (LRR) domains in an inactive but signaling-competent state. SGT1 is an integral component of a multi-protein network that includes RACK1, Rac1, RAR1, Rboh, HSP90 and HSP70, and in rice the Mitogen-Activated Protein Kinase (MAPK), OsMAPK6. Tobacco (Nicotiana tabacum) N protein, which belongs to the Toll-Interleukin Receptor (TIR)-NB-LRR class of R proteins, confers resistance to Tobacco Mosaic Virus (TMV).
Following transient expression in planta, we analyzed the functional relationship between SGT1, SIPK – a tobacco MAPK6 ortholog – and N, using mass spectrometry, confocal microscopy and pathogen assays.
Here, we show that tobacco SGT1 undergoes specific phosphorylation in a canonical MAPK target-motif by SIPK. Mutation of this motif to mimic SIPK phosphorylation leads to an increased proportion of cells displaying SGT1 nuclear accumulation and impairs N-mediated resistance to TMV, as does phospho-null substitution at the same residue. Forced nuclear localization of SGT1 causes N to be confined to nuclei.
Our data suggest that one mode of regulating nucleocytoplasmic partitioning of R proteins is by maintaining appropriate levels of SGT1 phosphorylation catalyzed by plant MAPK.
Infection of resistant cultivars of tobacco plants with TMV triggers a multi-component defense that culminates in tissue collapse at pathogen entry sites. This programmed cell death-like process, known as the Hypersensitive Response (HR), is thought to restrict systemic spread of the virus (van Loon, 1983). Resistance to TMV is mediated by the tobacco N protein of the TIR-NB-LRR class of plant disease resistance proteins (Whitham et al., 1994). Recognition of a helicase (p50) domain of the viral 126-kDa replicase by N (Mestre & Baulcombe, 2006; Ueda et al., 2006; Burch-Smith et al., 2007) initiates downstream signaling pathways that include rapid activation of two tobacco MAP kinases, Salicylic acid-Induced Protein Kinase (SIPK) and Wounding-Induced Protein Kinase (WIPK), that share a common upstream kinase, NtMEK2 (Jin et al., 2003). A body of evidence suggests that SIPK and WIPK are involved in induction of defense-related genes and HR cell death (Zhang et al., 1998, 2000; Lee et al., 2001).
N-mediated resistance requires SGT1 (Liu et al., 2002b), a conserved HSP90 co-chaperone that was originally identified in Saccharomyces cerevisiae as an essential component of cell cycle progression (Kitagawa et al., 1999). In yeast, SGT1 functions as an adaptor in assembly of a number of multi-protein complexes, including kinetochore CBF3 and ubiquitin ligase SCF (Catlett & Kaplan, 2006). In plants, SGT1, HSP90 (Heat Shock Protein 90kD) and RAR1 (Required for MlaResistance 1) form a dynamic molecular chaperone complex whose role is to maintain NB-LRR receptors in an inactive but signaling-competent form (Liu et al., 2003; Botër et al., 2007; Zhang et al., 2010; Kadota & Shirasu, 2012). Current evidence suggests that this chaperone complex both stabilizes NB-LRR proteins and mediates their subsequent degradation (Holt et al., 2005; Leister et al., 2005; Sangster & Queitsch, 2005; Azevedo et al., 2006; Kadota & Shirasu, 2012), thus maintaining a finely tuned balance between signal transmission and attenuation. In mammals, structurally similar NOD-LRR sensors of microbial infection also require SGT1 and HSP90 for proper function (Mayor et al., 2007; da Silva Correia et al., 2007).
Few protein substrates have been unequivocally identified for plant MAPKs, although large-scale protein screens have demonstrated that many substrates are likely to exist in plants (Feilner et al., 2005; Popescu et al., 2009) and that some of these are involved in controlling disease (Andreasson et al., 2005; Menke et al., 2005; Bethke et al., 2009; Popescu et al., 2009). Also, cell death elicited in Nicotiana benthamiana by transient overexpression of AtMKK7 and AtMKK9, two defense-related MAPK kinases from Arabidopsis thaliana, depends on SGT1 (Popescu et al., 2009). Although previous studies have demonstrated an essential role for SIPK in N-gene mediated TMV resistance (Jin et al., 2003), it has remained unclear how SIPK and its MAPK6 orthologs execute this function. Here we provide evidence that N. benthamiana SGT1 undergoes specific phosphorylation by SIPK, thus establishing that a protein known to play a central role in the plant defense is also the direct target of a MAPK cascade. We further show that SIPK phosphorylation of SGT1 is required for functional N-mediated resistance. We propose that phosphorylation of SGT1 fine-tunes the nucleocytoplasmic distribution of the N-receptor, a process that is necessary for an effective plant resistance response to TMV infection.
Materials and Methods
Nicotiana tabacum L. cv Xanthi-nc plants of all genotypes, wild-type nn and transgenic NN N. benthamiana plants (BN3) were grown in soil under controlled environmental conditions (21°C, 16 h light, 8 h dark), as described previously (Talarczyk et al., 2002).
Recombinant protein preparation
The full-length cDNA of NbSGT1 (AY899199) was cloned into the pET-28a vector. The full-length cDNA of SIPK was cloned into the pGEX-6P-1 vector. Full-length cDNAs of NtMEK2DD and NtMEK2KR were cloned into the pGEX-6P-2 vector. Escherichia coli (BL21) cells were induced with 0.25 mM IPTG at 18°C for 4 h. His-tagged recombinant protein was purified using Ni-NTA resin (Qiagen). Glutathione S-transferase (GST)-tagged recombinant proteins were purified using Glutathione-Agarose resin (Sigma-Aldrich).
Proteins were extracted in buffer containing 100 mM Tris HCl pH 8.0, 1 mM EDTA, 150 mM NaCl, 7.2 mM β-mercaptoethanol, 0.5 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF) and 0.03 μM PMSF (phenylmethylsulfonyl fluoride). Protein content was measured by the Bradford method using a commercially available reagent (Bio-Rad, www.bio-rad.com). Extracts were fractionated by 12.5% SDS–PAGE and subjected to immunoblot analysis using specific primary antibodies: monoclonal mouse anti-c-myc (Sigma-Aldrich), polyclonal rabbit anti-SGS or anti-TMV (Bioreba, Reinach BL1, Switzerland) and appropriate alkaline phosphatase-conjugated anti-mouse or anti-rabbit antibodies from Sigma-Aldrich. Immunoblots were developed using the NBT/BCIP colorimetric detection kit from Roche Applied Science.
In vitro phosphorylation assay
The purified recombinant proteins (NbSGT1, SIPK, NtMEKDD or NtMEK2KR; c. 2 μg each) were incubated with 50 μM ATP supplemented with 1.5 μCi [γ32P] ATP in the reaction buffer (20 mM Tris HCl pH 7.5, 40 mM MgCl2, 10 mM EGTA) at 30°C for 30 min. The reaction was terminated by adding 3× Laemmli sample buffer. In the nonradioactive reaction followed by mass spectrometry analysis, the concentration of ATP was increased to 200 μM and the incubation time extended to 2 h.
Strep-tag II affinity purification
The full-length cDNA of NbSGT1 was cloned into the pROK2 vector with a (6xHis)-(Ala-Ala)-(Strep-tag II) coding sequence at the 3′-end. The full-length cDNA of SIPK was cloned into the pROK2 vector. The pROK2 constructs and pTA7002 carrying cDNA of NtMEK2DD or NtMEK2KR were electroporated into Agrobacterium tumefaciens (GV3101) cells, which were infiltrated into 4-wk-old N. benthamiana leaves as described earlier (Romeis et al., 2001). Expression of the NtMEK2 variant cDNAs was induced by 30 μM dexamethasone 40–48 h later (Yang et al., 2001). Treated leaves were collected 5 h after dexamethasone infiltration. Ground leaf material (c. 3 g) was thawed in 10 ml Ex-strep buffer (100 mM Tris HCl pH 8.0; 5 mM EGTA; 5 mM EDTA; 150 mM NaCl; 10 mM DTT; 0.5 mM AEBSF; 5 μg ml−1 antipain; 5 μg ml−1 leupeptin; 50 mM NaF; 1% (v/v) Phosphatase Inhibitor Cocktail 1 (Sigma-Aldrich); 0.5% Triton X-100; and 100 μg ml−1 avidin) as described previously (Witte et al., 2004). The slurry was centrifuged for 10 min at 4°C (15 000 g), the supernatant filtered through Miracloth and 0.5 ml StrepTactin Sepharose (IBA GmbH, Göttingen, Germany) was added. Binding was performed by incubation of this suspension on a rotator for 1 h at 4°C. The slurry was transferred into a Poly-Prep column (Bio-Rad) and the flow-through was discarded. The resin was washed twice with 10 ml W-buffer (100 mM Tris HCl pH 8.0; 150 mM NaCl; 1 mM EDTA). Four times 250 μl E-buffer (100 mM Tris HCl pH 8.0; 150 mM NaCl; 1 mM EDTA; 2.5 mM desthiobiotin) was added and eluates were collected. The samples were concentrated on a Microcon YM-10 (Millipore) for 30 min at 4°C (13 000 g) to a volume of c. 20 μl and resolved by SDS-PAGE.
Mass spectrometry analysis
Gel bands containing the proteins of interest were subjected to a standard proteomic procedure as described in Supporting Information Methods S1. Briefly, reduced and alkylated proteins were subjected to trypsin digestion. The resulting peptides were eluted from the gel and phosphopeptide enrichment was carried out on SwellGel Gallium-Chelated Discs (Phosphopeptide Isolation Kit, Thermo Fisher Scientific Rockford, IL, USA). LC-MS analyses of the peptide mixtures were carried out on LTQ FTICR spectrometer (Thermo). The MASCOT program was used for database searches, and MS/MS spectra of phosphorylated peptides were also curated manually.
Heterologous transient complementation assay
For VIGS of NbSGT1, the TRV-based vector system was adopted (Liu et al., 2002b). An NbSGT1 cDNA fragment was PCR amplified using primers that add specific restriction sites (bold): TCTAGACGAGACATTTACAAAGATGCCG and GGATCCAGGGCTTCCTTCGACCTTCTTT and ligated into the pTRV2 vector. For the GFP construct, the first 130 bp of the GFP were PCR-amplified using primers that add specific restriction sites (bold): GGATCCATGGTAGATCTGACTAGTAAAGG and CTCGAGGTATGTTGCATCACCTTCACC and cloned into the pTRV2 vector. To generate phospho-null (AP) and phospho-mimic (DP) variants of AtSGT1b by PCR, primers were designed that incorporated mutations T346A and T346D, respectively. Next, wild-type and both phospho-variants of AtSGT1b cDNA were amplified by PCR with an upstream primer AGATCTGCTATGGAACAAAAGCTTATTTCTGAAGAAGACTTGCTCGAGATGGCCAAGGAATTAGCAGAGAAAGCT adding restriction sites and c-myc encoding sequence (bold) to the 5′ end, and a downstream primer GGTACCCGGGTCGACTCGAGTCAATACTCCCACTTCTTGAGCTCCAT adding restriction sites (bold) to the 3′ end of the product. All three inserts were cloned into the pROK2 vector. For silencing of NbSGT1, Agrobacterium cultures (OD600 = 0.5) containing pTRV1 or pTRV2 derivative plasmids were mixed at a 1 : 1 ratio and infiltrated into the lower leaves of four-leaf-stage N. benthamiana plants using a 1-ml needleless syringe. For transient expression of wild-type and phospho-variants of AtSGT1b, the systemic leaves of the plants silenced for NbSGT1 were infiltrated with Agrobacterium cultures containing c-myc-AtSGT1b constructs. Twenty-four hours later, the same leaves were dusted with carborundum and rubbed with TMV strain U1 suspension (1 μg ml−1). To detect the presence of TMV, total proteins extracted from inoculated leaves 4 d after virus infection were analyzed by Western blots. Each experiment was repeated four times, and each experiment included at least three independent biological replicates.
Subcellular localization of AtSGT1b
The pCambia1302 vector was used to express AtSGT1b protein variants N-terminally tagged with GFP. cDNAs encoding wild-type, AP and DP variants of AtSGT1b (with T346A and T346D substitutions, respectively) were amplified by PCR with an upstream primer adding a NheI restriction site to the 5′ end, and a downstream primer adding a PmlI restriction site to the 3′ end of the product. As controls, wild-type AtSGT1b cDNA was PCR-amplified with the upstream primers adding a NheI restriction site and ‘NLSSV40’ (PKKKRKV), ‘NLSSLCV’ (SYVKTVPNRTRTYIK), ‘NES’ (NELALKLAGLDINK) or ‘nes’ (NELALKAAGADANK) encoding sequences to the 5′ ends of the products, and the downstream primer as above. NcoI/NheI GFP and NheI/PmlI AtSGT1b inserts were cloned into the NcoI/PmlI pCambia1302 vector by a triple-ligation strategy.
In order to determine the subcellular localization of GFP-AtSGT1b fusion proteins, leaves from 4-wk-old N. benthamiana plants were used for transient gene expression in epidermal cells. Plasmid DNA (2 μg) was adsorbed onto tungsten M17 particles (diameter 1.1 μm, 350 μg). Micro-bombardment was performed at a pressure of 1100 psi using a biolistic particle delivery system (model PDS-1000/He, Bio-Rad), and tissues were analyzed 24 h after bombardment.
Hypersensitive response assay
In order to obtain N-NXS constructs, the Citrine sequence was amplified with an upstream primer that added a SacI restriction site and a sequence encoding the last three amino acids (ING) of the N to the 5' ends of the products, and downstream primers that added ‘NLSSLCV’ ‘NLSSV40’ or ‘nlsSV40’ encoding sequences with a STOP-codon and a SacI restriction site to the 3′ ends of the products. The SacI inserts were then cloned into the SacI restriction site of gN (Dinesh-Kumar et al., 2000).
In order to generate p50-U1-Cerulean-NLS and p50-U1-Cerulean-nls, the p50-U1-Cerulean construct (Burch-Smith et al., 2007) was PCR-amplified with primers including ‘NLSSV40’ or ‘nlsSV40’ and the inserts were cloned into the pROK2 vector. The N and p50 derivative constructs were electroporated into A. tumefaciens (GV3101) cells, which were infiltrated into 4-wk-old N. benthamiana leaves, at OD600 = 1.7 and 1.0, respectively. For co-infiltrations, Agrobacterium strains carrying N and p50 derivative constructs were mixed at a 1:1 ratio. To enhance the expression levels, the infiltration suspensions were mixed at 10:1 ratio with an Agrobacterium suspension (OD600 = 2–3) expressing p19 (Witte et al., 2004). Photographs of leaves were taken 9 or 10 d after infiltration.
The pGWB 445 vector (Nakagawa et al., 2007a,b) was used to express AtSGT1b protein variants N-terminally tagged with CFP. cDNAs encoding wild-type, AP, DP and variants containing ‘NLSSV40’ (PKKKRKV), ‘nlsSV40’ (PKTKRKV), ‘NES’ and ‘nes’ coding sequences were amplified by PCR to produce blunt-end products for TOPO Cloning (pENTR/D-TOPO vector, Invitrogen). The resulting entry clones were LR recombined with the Gateway pGWB 445 destination vector. cDNAs encoding TIR (amino acids from 1 to 179) or NB domains (amino acids from 180 to 525) were amplified by PCR with primers containing attB sequences. As a template for subcloning of the TIR domain, a cDNA fragment consisting of exon 1 and exon 2 of the N gene from Nicotiana glutinosa was used. attB-flanked PCR products were BP-recombined with the pDONR201 donor vector (Invitrogen). Subsequently generated entry clones were LR-recombined with the Gateway pGWB 441 destination vector to express TIR or NB domain polypeptides C-terminally tagged with YFP. To obtain a cDNA encoding the N LRR domain (amino acids from 589 to 1144), a fragment of exon 3 of the N gene from N. glutinosa was PCR-amplified with an upstream primer adding an EcoRI site to the 5′ end of the product and a downstream primer adding the sequence of this domain from exon 4 (up to BglII site) to the 3′ end of the product. Next, the sequence of exon 4 was PCR-amplified with the upper primer starting from the beginning of this sequence and the lower primer adding the sequence of exon 5 and a XhoI site to the 3′ end of the product. EcoRI/BglII exon 3 and BglII/XhoI exon 4+5 inserts were cloned into EcoRI/XhoI pGEX-6P-1 vector by a triple-ligation strategy. The coding sequence of the LRR domain was PCR-amplified to produce a blunt-end product for TOPO Cloning (pENTR™/D-TOPO vector). The resulting entry clone was LR-recombined with the Gateway pGWB 442 destination vector to express the LRR domain N-terminally tagged with YFP. The mixture of CFP-AtSGT1b variants (1.5 μg) with either one of three YFP-N domain constructs (YFP-TIR, YFP-NB or YFP-LRR; 1.5 μg) and p19 (0.5 μg) plasmid DNAs was used for microbombardment and tissue samples were analyzed 24–48 h afterwards.
For co-localization of AtSGT1b with N, a combination of CFP-AtSGT1b variants (1.5 μg) with N-Citrine (1.5 μg) and p19 (0.5 μg) plasmid DNAs was used for microbombardment and tissue samples were analyzed after 24–48 h.
Confocal laser scanning fluorescence microscopy
Transient intracellular fluorescence was observed by confocal laser scanning microscopy using a Nikon TE2000E EZ-C1 inverted confocal microscope equipped with 60× oil immersion objective lens (numerical aperture = 1.4). YFP, Citrine and GFP were excited with the 488 nm line from an argon ion laser, and images of CFP and Cerulean were obtained using 408 nm diode laser excitation. The fluorescence signals were detected using the 515/30 and 450/35 emission filters for the first and latter fluorophore groups, respectively. Scanning was performed in sequential mode to prevent bleed-through. Images were collected from a single optical section and processed using the EZ C1 program (Nikon Instruments B.V. Europe, Amstelveen, the Netherlands). Optimal imaging parameters were set up for each experiment and were equal for each image dataset. Quantification of fluorescence intensities in the nuclear and cytoplasmic regions was performed using ImageJ software (Abramoff et al., 2004).
SGT1 is phosphorylated by SIPK in vitro and in planta
Genetic analyses have linked SGT1 with diverse biological processes in plants, and the SGT1 protein was found to form part of a multi-protein functional network that includes a rice MAPK, OsMAPK6 (Thao et al., 2007), whose tobacco ortholog is SIPK (NtSIPK). Amino acid sequence analysis of eukaryotic SGT1 sequences (Fig. S1) revealed a canonical MAPK phosphorylation motif (S/T-P) in the C-terminal SGS (SGT1-Specific) domain of NbSGT1 and most known plant SGT1 proteins, as well as in the Saccharomyces cerevisiae Sgt1p. In addition, NbSGT1 contains a predicted MAPK docking site of medium stringency (scansite algorithm, http://scansite.mit.edu).
These observations prompted us to determine whether SGT1 can be phosphorylated by SIPK. When recombinant SIPK was co-incubated with either constitutively active NtMEK2DD or the catalytically inactive form, NtMEK2KR (Jin et al., 2003), SIPK activated by NtMEK2DD was able to phosphorylate the control substrate (myelin basic protein, Fig. 1a, lane 4) and was also able to phosphorylate recombinant NbSGT1 (Fig. 1a, lane 5). Two phosphopeptides derived from the C-terminal part of SIPK-treated NbSGT1 were detected by mass spectrometry (LC-MS-MS/MS), and Ser358 within the KVEGSPPDGMELK-peptide that includes the predicted MAPK phosphorylation site (Fig. 1b) was unambiguously identified as phosphorylated (Fig. S2).
A transient expression assay was used to determine whether SGT1 could also be phosphorylated in planta. Agrobacterium strains carrying either 35S::NbSGT1-Strep II – a dexamethasone (Dex)-inducible form of NtMEK2DD – or 35S::SIPK constructs were co-infiltrated into Nicotiana benthamiana leaves and after 2 d incubation to ensure expression of the constitutive promoter constructs, NtMEK2DD expression was induced by treatment of the infiltrated leaves with Dex (Fig. 2a). Protein was extracted from leaf samples collected 5 h after induction and SGT1 was affinity-purified via its Strep tag II. Only one phosphopeptide, that encompassing Ser358, was detected by mass spectrometry under these conditions (Fig. 2b). As a control, we co-expressed kinase inactive NtMEK2KR with SGT1. In this case, SIPK was not included because its transient expression results in slightly increased SIPK activity and associated activation of defense responses (Zhang & Liu, 2001). Under these conditions, SGT1 phosphorylation was not detected (Fig. 2c). We therefore concluded that SGT1 can also be phosphorylated by SIPK in vivo in N. benthamiana cells. However, we cannot fully exclude the possibility of SGT1 being phosphorylated by an endogenous N. benthamiana kinase.
Functional relevance of SGT1 phosphorylation for N-mediated resistance
SGT1 is a positive regulator of plant defense responses and plant cell death (Austin et al., 2002; Azevedo et al., 2002; Wang et al., 2010). Inappropriate activation of an immune response is detrimental for plant growth, which suggests that SGT1 function should also require tight regulation. Because MAPK6 orthologs, including SIPK, are activated during immune responses (Pitzschke et al., 2009), we asked whether modification of the SGT1 MAPK phosphorylation motif would affect the ability of the plant to mount an effective resistance response. To test this, we used a heterologous transient complementation assay in N. benthamiana (Azevedo et al., 2006). Expression of the native NbSGT1 was first suppressed by means of Virus Induced Gene Silencing (VIGS), and this genotype was then complemented with ectopic expression of modified forms of heterologous AtSGT1b. Ectopic expression of the heterologous gene in the SGT1-silenced background thus allows the phenotypic consequences of specific SGT1 amino acid substitutions to be assessed in planta.
We used Tobacco Rattle Virus (TRV)-based VIGS to silence endogenous SGT1 in N. benthamiana BN3 (Liu et al., 2002a), a genotype that carries the tobacco N gene (N. benthamiana NN) which confers resistance to the U1 strain of TMV. To minimize the risk of simultaneously suppressing expression of both the endogenous NbSGT1 and the heterologous AtSGT1b, we designed a VIGS construct that targets a short region (nt 949-1076) of NbSGT1 where it possesses low homology to AtSGT1b. Because expression of empty VIGS vectors carrying virus components (TRV1 and TRV2) can sometimes lead to severe disease symptoms (Hartl et al., 2008), we used constructs containing fragments of GreenFluorescentProtein (GFP) sequence (TRV:GFP) as a negative control. The efficiency of the NbSGT1 silencing was monitored by immunodetection. Nine days after TRV application, NbSGT1 protein was detected in extracts from nonsilenced plants and from plants infiltrated with TRV:GFP, but not in plants silenced with the SGT1-targeting construct (Fig. 3a). Binary vectors encoding either wild-type AtSGT1b or its variants in which the phosphorylation motif serine residue was mutated to either AP or DP were next introduced via agroinfiltration into systemic leaves of these SGT1-silenced plants. All AtSGT1b variants were expressed as N-terminally c-myc-tagged fusion proteins under control of the 35S promoter. Twenty-four hours after infiltration, the leaves were inoculated with TMV. Due to reported differences in virulence of the TMV U1 and TMV:GFP strains (Peart et al., 2002), the wild-type TMV U1 strain was used. Because N. benthamiana NN occasionally forms necrotic lesions in response to TMV (Peart et al., 2002; Jin et al., 2003), resistance was monitored by directly measuring the accumulation of viral Coat Protein (TMV-CP) in the inoculated tissue, rather than scoring the dimensions of necrotic lesions.
As shown in Fig. 3(b), 4 d post-infection (dpi) little or no signal corresponding to TMV-CP was detected in extracts from nonsilenced plants or plants infiltrated with TRV:GFP. Viral particles also could not be detected in SGT1-silenced plants that were complemented with a wild-type copy of AtSGT1b, indicative of a strong resistance response. In contrast to this, complementation of SGT1-silenced plants with either the phospho-null or the phospho-mimic AtSGT1b variants resulted in levels of TMV-CP accumulation in the inoculated plants that were approximately equal to those seen in the empty vector controls (Fig. 3b). Moreover, plants complemented with mutated AtSGT1b proteins or empty vector control, in contrast to plants expressing wild-type AtSGT1b, developed macroscopic disease symptoms manifested as chlorotic lesions and tissue desiccation (Fig. 3c). We concluded that neither the phospho-mimic nor the phospho-null variant of AtSGT1b is functional in N-mediated resistance. This pattern of increased susceptibility is not a consequence of different levels of expression of the AtSGT1b phospho-variants (Fig. 3b). Notably, we found that the phospho-mimic and phospho-null alleles of AtSGT1b were still able to complement cell division defects of yeast sgt1-3 and sgt1-5 mutant strains (Fig. S3), which argues against general loss of function of these modified proteins. Because both the phospho-mimic and phospho-null AtSGT1b variants failed to complement NbSGT1 function, we reasoned that either the simultaneous presence of both phospho-forms of SGT1 (null and mimic) or dynamic changes in the phosphorylation state of SGT1might be an important feature of N-mediated resistance to TMV.
An SGT1-phospho-mimic variant shows enhanced nuclear localization
Phosphorylation can affect various properties of proteins, including subcellular localization, ability to form complexes, stability and binding of divalent cations (Cohen, 2000). In Arabidopsis, wild-type AtSGT1b accumulates in both the cytosolic and nuclear compartments, whereas a truncated form, AtSGT1beta3, which lacks the MAPK phosphorylation site and is nonfunctional, was found only in the nucleus-depleted fraction of leaf tissue extracts (Noël et al., 2007). We therefore tested whether phosphorylation would affect AtSGT1b subcellular distribution. Constructs encoding GFP fused N-terminally to wild-type AtSGT1b or to its phospho-variants were transiently expressed in N. benthamiana epidermal cells via particle bombardment. Wild type GFP-AtSGT1b localized to the nucleus in c. 25% of all transformed cells (Fig. 4). Upon bombardment with constructs encoding AtSGT1b fused to a strong Nuclear Localization Signal derived from SV40 large T antigen (NLSSV40), or a Nuclear Export Signal (NES), c. 100% and c. 1% of the transformed cells, respectively, showed nuclear fluorescence (Kalderon et al., 1984; Wen et al., 1995). Control-like subcellular distributions could be restored upon mutagenesis of the NES (Fig. 4). Thus, changes in protein localization could be readily assessed in our bombardment assay. Similar to the wild-type control protein, the phospho-null (AtSGT1bAP) variant localized to the nucleus in c. 27% of transformed cells (Fig. 4). By contrast, the phospho-mimic variant of AtSGT1b (AtSGT1bDP) was observed in the nucleus of > 50% of transformed cells. These data suggest that phosphorylation facilitates movement of SGT1 into the nucleus. Interestingly, a similar proportion of cells with nuclear localization of SGT1 was observed for AtSGT1bDP and AtSGT1b fused to the NLS from Squash Leaf Curl Virus (NLSSLCV) movement protein, BR1 (Sanderfoot et al., 1996). Apparently, this nuclear targeting signal differs significantly from NLSSV40 in its capacity to relocate proteins.
The nuclear pool of N protein is sufficient for HR development
Controlled nucleocytoplasmic partitioning of the potato NB-LRR receptor Rx was found to be necessary for full resistance to Potato Virus X (Slootweg et al., 2010; Tameling et al., 2010), and impaired Rx nuclear localization in SGT1-silenced plants depended on the presence of the Rx-LRR domain (Slootweg et al., 2010). As the phosphorylation state of SGT1 alters its nucleocytoplasmic distribution (Fig. 4), we hypothesized it might also lead to improper partitioning of R proteins that it associates with. If correct, this might explain the observed lack of complementation of SGT1 phospho-variants in N-mediated resistance.
In order to test this, we first evaluated the relevance of N's subcellular distribution for development of HR in response to TMV p50 elicitor. N protein C-terminally fused to Citrine was observed to accumulate in both the nucleus and cytoplasm (Fig. 5b). Exclusion of N from the nucleus by adding a NES sequence at the end of the N and the Citrine fusion protein was previously shown to abolish HR induced in tobacco by p50 (Burch-Smith et al., 2007). To gain a more complete picture, we inserted NLS sequences in our N-Citrine construct in a similar fashion, and assessed the ability of the resulting nuclear N pool to initiate HR. The N-Citrine-NLSSV40 protein was efficiently relocated to the nucleus (Fig. 5b), whereas N fused to NLSSLCV remained nucleocytoplasmic. Expression of a control construct carrying a nonfunctional nlsSV40 restored wild-type N-Citrine localization. When co-expressed with p50, both wild-type N-Citrine and its variants carrying mutated nes- and nls-signals initiated HR in N. benthamiana tissues, as anticipated (Fig. 5a). Surprisingly, both N-NLS-Citrine versions (NLSSV40 and NLSSLCV) were able to initiate HR (Fig. 5a) equally as well as the N-Citrine fusion. By contrast, no HR occurred when the N-Citrine-NES version was co-expressed (Fig. 5a) with p50, consistent with earlier reports (Burch-Smith et al., 2007). As previously described (Burch-Smith et al., 2007), and as shown in Fig. S4, p50 variants with either nuclear, cytoplasmic or nucleocytoplasmic distribution were all readily recognized by N in two experimental systems, that is, upon co-expression with N (Fig. S4a) or in transgenic N. benthamiana NN plants (Fig. S4b). These findings support the model that the nuclear pool of N is indispensable for tissue necrotization, and that upon appropriate stimulation, N protein is translocated to the nucleus. However, it does not seem critical in which of the two cellular compartments the p50 recognition event takes place.
The distribution of SGT1 determines N receptor localization
The above results collectively suggest that establishment of resistance to TMV requires translocation of N into the nucleus and is affected by the phosphorylation state of SGT1. We asked therefore whether SGT1 could control nucleocytoplasmic partitioning of N. To this end, we co-expressed N-Citrine together with CFP-AtSGT1b fused to the nuclear targeting signals in N. benthamiana plants that had been silenced for endogenous SGT1. Under these conditions, AtSGT1b with the strong NLSSV40 sequence was detected exclusively in the nucleus in some cells (Fig. 6a panel A), whereas in others it distributed between the nucleus and cytoplasm (Fig. 6a panel C). In the cells of both classes, relative fluorescence intensities in nuclei were very high at 99% and 80%, respectively (Fig. 6b). NLSSV40-AtSGT1b was never found exclusively in the cytoplasm of any of the cells examined. The distribution of the N protein strongly resembled the pattern of NLSSV40-AtSGT1b distribution based on both visual inspection of the images (Fig. 6a panels B, D) and quantification of their relative fluorescence intensities (Fig. 6b). However, in cells displaying only a small nuclear pool of AtSGT1b, due to its fusion with NES (Fig. 6a panel F, Fig. 6b), > 50% of the N protein was still present in the nucleus. This suggests that SGT1 may assist nuclear import of N protein.
We next compared the distribution patterns of N-derived domains and SGT1. Individual domains of N fused to YFP were co-expressed in N. benthamiana with AtSGT1b localization variants N-terminally fused to CFP. We detected no changes in relative nuclear fluorescence intensities for N-TIR and N-NB domains when these were co-expressed with NLS-AtSGT1b or NES-AtSGT1b variants (Fig. 7a). By contrast, the localization parameters were strictly correlated for N-LRR and SGT1 variants (Fig. 7a). Most tellingly, upon co-expression of YFP-N-LRR with CFP-AtSGT1b variants, we found that N-LRR only localized to the nucleus in those cells where AtSGT1b variants also localized to the nucleus (Figs 7b, S5). Upon forced nuclear localization or exclusion of AtSGT1b, the co-expressed N-LRR localized again exclusively to the AtSGT1b-containing compartment, and the same spatial relationship was observed upon co-expression with the nls/nes AtSGT1b constructs (Fig. 7b). Again, this pattern was not observed for TIR and NB domains (Fig. 7).
Collectively, these results suggest that the N-LRR domain mediates SGT1-dependent trafficking of N receptor into the nucleus. Despite this apparent functional relationship, however, we were unable to detect a direct interaction of N-LRR polypeptide with AtSGT1b in yeast two-hybrid assays (Fig. S6). A similar result was reported in studies of the MLA6 receptor (Bieri et al., 2004), whereas physical interaction of SGT1 with the LRR domains of two non-TIR class NB-LRR receptors (MLA1 and Bs2) has been observed (Bieri et al., 2004; Leister et al., 2005).
Upon perception of a pathogen by NB-LRR receptors, a signal is conveyed to the nucleus where transcriptional reprogramming takes place. There is an increasing body of evidence that several activated plant NB-LRR receptors partition dynamically between the cytoplasm and the nucleus (Shen et al., 2007; Wirthmueller et al., 2007; Caplan et al., 2008a; Slootweg et al., 2010; Tameling et al., 2010; Bhattacharjee et al., 2011; Heidrich et al., 2011). Because some of these receptors have been shown to interact with transcription factors (Heidrich et al., 2012; Chang et al., 2013; Padmanabhan et al., 2013), it has been proposed that they regulate expression of defense-related genes.
The functions of NB-LRR receptors are tightly controlled at multiple levels by an HSP90-SGT1-RAR1 chaperone complex(Kadota & Shirasu, 2012). Here, we provide evidence that, besides the known role of the chaperone complex in regulating the activity and stability of NB-LRR recognition complex, SGT1 might also help to maintain appropriate nucleocytoplasmic partitioning of the NB-LRR receptors. We propose that SGT1 dynamically shuttles to and from the nucleus, and that under normal conditions the nuclear export rate is higher than import rate, which results in SGT1 residing primarily in the cytoplasm (Fig. 4) (Noël et al., 2007). This process appears to be finely tuned by phosphorylation of SGT1, which can be catalyzed by an MAPK6 ortholog (SIPK), a modification that results in a higher proportion of cells displaying predominantly nuclear localization of SGT1 (Fig. 4). We found that the presence of N protein in the nucleus is correlated with forced SGT1 nuclear localization (Fig. 6), and thus we suggest that one role of SGT1 phosphorylation is to increase the nuclear pool of the NB-LRR receptors.
While previous studies have pointed to roles of both SGT1 and MAPK6 in the ability of plants to mount effective disease resistance responses, the functional relationship between these two components has remained unclear. Our observation of specific phosphorylation of the SGS domain in SGT1 by SIPK reveals a new level of SGT1 regulation. Importantly, in terms of the biological relevance of this modification, we were able to demonstrate that SGS phosphorylation influences N resistance to TMV infection. AtSGT1b constructs modified in the MAPK site are not able to restore full TMV resistance in transgenic N. benthamiana plants in which endogenous NbSGT1 expression has been silenced (Fig. 3). Interestingly, both phospho-null and phospho-mimic modifications compromised resistance. This outcome resembles a situation in which both gain-of-function and loss-of-function of the transcription factor ERF104, one of the AtMPK6 substrates from Arabidopsis, resulted in enhanced susceptibility to Botrytis cinerea and Pseudomonas syringae pv. phaseolicola (Bethke et al., 2009). The increased susceptibility to TMV observed for both phospho-mimic and phospho-null modifications does not necessarily imply that both SGT1 phospho-variants play the same role in defense response or that both are nonfunctional. On the contrary, the fact that these two variants differ in their impact on N protein nucleocytoplasmic partitioning might indicate that two sub-pools of SGT1, phosphorylated and nonphosphorylated, are required in the cell in order to establish resistance. Because either of the mutated SGT1 proteins is able to restore viability in sgt1 mutant yeast strains (Fig. S3), the overall protein architecture does not appear to be unduly disturbed by the introduced amino acid substitutions. It was previously shown that a frame-shift mutation that eliminated the target phosphorylation motif (TP) in the distal part of AtSGT1b had a dominant negative effect on resistance to PVX in a heterologous complementation assay in N. benthamiana (Botër et al., 2007). Also, the resistance phenotype of the sgt1beta3 Arabidopsis mutant, which produces a truncated protein lacking the C-terminal 36 aa containing the SIPK phosphorylation site, is indistinguishable from that of the sgt1b-3 null mutant (Gray et al., 2003; Noël et al.,2007).
These observations, taken together with our results, strongly suggest that reversible phosphorylation of SGT1 plays an important role in plant NB-LRR resistance to pathogen attack. The finding that the phospho-mimic substitution of SGT1 results in a marked increase in its partitioning into the nucleus (Fig. 4) suggests that phosphorylation at this residue dynamically regulates the cytoplasmic and nuclear SGT1 pools, and that the plant's ability to mount an effective resistance response is conditioned by an appropriate balance between the phosphorylated and nonphosphorylated forms.
The N protein function during the tobacco TMV defense response can be separated temporally and spatially into two distinct phases: pathogen recognition, which is thought to occur in the cytoplasm but requires the chloropastic protein NRIP1 (N Receptor-Interacting Protein 1); and subsequent defense-oriented physiological reprogramming, which is orcheastrated mainly in the nucleus (Burch-Smith et al., 2007; Caplan et al., 2008b). This model is consistent with the observed distribution of the receptor and its ligand, because the N protein is found in the host cytoplasm and nucleus, whereas the TMV 126-kDa replicase protein (and its p50 fragment that is sensed by N) is associated with the viral replication complex at the ER (Padmanabhan et al., 2006). Unexpectedly, however, our results show that restriction of N to the nucleus does not compromise HR development (Fig. 5). Re-direction of p50 into the nucleus also did not change its elicitation activity, and a typical HR was consistently observed when p50-NLSSV40 was co-expressed with N in N. benthamiana (Fig. S4a). Similarly, p50-NLSSV40 infiltration into transgenic N. benthamiana BN3 plants (expressing N) triggered HR (Fig. S4b). These data suggest that the nuclear pool of N is itself competent for association with ligand (Figs 5, S4), and that events occurring subsequent to p50 recognition, such as conformational changes and oligomerization of the receptor (Mestre & Baulcombe, 2006), can also proceed outside the cytoplasmic milieu. A recent report showed that recognition of the Coat Protein of Potato Virus X (CP-PVX) mediated by the potato Coiled-Coil (CC)-NB-LRR type Rx protein occurs mainly in the cytoplasm (Tameling et al., 2010). At this point, we cannot exclude the possibility that establishment of full resistance to TMV also requires a small cytoplasmic pool of N because nuclear-targeted proteins have to travel through the cytoplasm. Indeed, bifurcating cytoplasmic cell death and nuclear pathogen resistance pathways were recently reported for the Arabidopsis RPS4 and barley MLA10 NB-LRR receptors (Heidrich et al., 2011; Bai et al., 2012).
The finding that nuclear localization of N is critical for its function in TMV resistance (Burch-Smith et al., 2007) and Fig. 5 raises the question of how trafficking of the receptor complexes into the nucleus is controlled. It was previously hypothesized that transport of the human NB-LRR type NOD1 receptor into the nucleus could be facilitated by SGT1 (da Silva Correia et al., 2007). Consistent with this idea, and with the evolutionarily conserved nature of SGT1 (Fig. S1), nuclear localization of the potato Rx receptor protein was dramatically reduced in plants silenced for SGT1 expression (Slootweg et al., 2010). Nucleocytoplasmic partitioning of an Rx variant lacking the LRR domain was not affected by SGT1 silencing, as predicted by a model in which SGT1 interacts with the LRR domain of NB-LRR proteins (Bieri et al., 2004; Leister et al., 2005). Reduced nuclear localization of N protein lacking the LRR domain has also been observed (Burch-Smith et al., 2007).
Our data support the notion that SGT1 plays an important role in maintaining an appropriate subcellular distribution of the N receptor. When the N-derived LRR domain was transiently expressed either with wild-type AtSGT1b or its ectopic forms, both proteins co-localized (Fig. 7). Interestingly, when AtSGT1b was co-expressed with the full-length N protein, AtSGT1b was able to relocate N in only one direction, towards the nucleus. This distinction might indicate that control of the normal direction of signal flow upon TMV recognition involves factors in addition to SGT1. Alternatively, there might be intramolecular constraints within the N receptor that normally hinder nuclear import, but which are relieved upon ligand recognition. Further experiments are necessary to establish which scenario operates within the N signaling cascade.
The co-localization pattern of N-LRR and AtSGT1b was similar for all the inspected cells, independently of whether they expressed AtSGT1b phospho-variants or AtSGT1b with forced cytoplasmic or nuclear localization, which suggests that presence or absence of phosphorylation probably does not change the association of AtSGT1b with N (Fig. 7 and data not shown). However, because we were able to show that phospho-mimic substitution, in particular, results in a more pronounced redistribution of SGT1 between the nucleus and the cytoplasm, we propose a model in which phosphorylation of SGT1 is required for establishing a pattern of distribution of SGT1, and consequently N, between the nucleus and cytoplasm, and that this distribution pattern enables the host cell to mount an effective TMV resistance response (Fig. 8). SGT1 seems to contribute to two steps of N shuttling. First, by maintaining an appropriate level of cytoplasmic receptors (Holt et al., 2005; Li et al., 2010), SGT1 might affect nuclear flow rate of N, as has been shown for other cargoes and their chaperones (Timney et al., 2006). Second, it might control the translocation process itself (Fig. 6). However, the observation that the N protein still shows nucleocytoplasmic partitioning in SGT1-silenced plants (Fig. 6) suggests that, in contrast to Rx (Slootweg et al., 2010), the critical role for SGT1 resides not in control of N-shuttling per se, but rather in its ability to shift the equilibrium of the receptor distribution towards the nucleus upon effector recognition.
In summary, our results provide new insights into the mechanism(s) by which SGT1 modulates plant responses to pathogen attack, and reveal exciting new avenues for exploring the underlying regulatory network.
We thank: S. Zhang for MEK2 constructs; N-H. Chua for the pTA7002 vector; B. Baker for BN3 seeds; S. P. Dinesh-Kumar for the TRV VIGS system and gN-Citrine and p50-Cerulean constructs; M. Taube for N cDNA subclone; K. Kitagawa for yeast strains; I. Fijałkowska for plasmids used as a positive control in yeast two-hybrid assays; A. Anielska-Mazur for assistance in confocal analysis; and members of the G. Muszynska and G. Dobrowolska labs for help and stimulating discussions. This work was supported by grants from Ministry of Science and Higher Education Republic of Poland to M.K. (No N302 015 31/1618 and NN301 163235) and J.H. (N N301 318039), and from NSERC Canada to B.E. The fellowship for M.Z. was funded by the Foundation for Polish Science, project MPD/2009-3/2. J.S. and J.E.P. were supported by a Deutsche Forschungsgemeinschaft ‘SFB 635′ grant.