TRANSPARENT TESTA GLABRA (TTG) proteins that contain the WD40 protein interaction domain are implicated in many signalling pathways in plants. The salicylic acid (SA) signalling pathway regulates the resistance of plants to pathogens through defence responses involving pathogenesis-related (PR) gene transcription, activated by the NPR1 (nonexpresser of PR genes 1) protein, which contains WD40-binding domains. We report that tobacco (Nicotiana tabacum) NtTTG2 suppresses the resistance to viral and bacterial pathogens by repressing the nuclear localisation of NPR1 and SA/NPR1-regulated defence in plants. Prevention of NtTTG2 protein production by silencing of the NtTTG2 gene resulted in the enhancement of resistance and PR gene expression, but NtTTG2 overexpression or NtTTG2 protein overproduction caused the opposite effects. Concurrent NtTTG2 and NPR1 gene silencing or NtTTG2 silencing in the absence of SA accumulation compensated for the compromised defence as a result of the NPR1 single-gene silencing or the absence of SA. However, NtTTG2 did not interact with NPR1 but was able to modulate the subcellular localisation of the NPR1 protein. In the absence of NtTTG2 production NPR1 was found predominantly in the nucleus and the PR genes were expressed. By contrast, when NtTTG2 accumulated in transgenic plants, a large proportion of NPR1 was retained in the cytoplasm and the PR genes were not expressed. These results suggest that NtTTG2 represses SA/NPR1-regulated defence by sequestering NPR1 from the nucleus and the transcriptional activation of the defence-response genes.
The WD40 domain is a protein interaction domain that is present as a series of repeats in a variety of proteins (Xu and Min, 2011). The name WD40 was derived from the conserved tryptophan (W) and aspartic (D) dipeptide and the length of approximately 40 amino acid residues in a single repeat (Neer et al., 1994). The definition of this motif was further refined as a 44- to 60-residue sequence that typically contains a glycine (G) and histidine (H) dipeptide 11-24 residues from its N-terminus and the WD dipeptide at the C-terminus (Xu and Min, 2011). WD40 repeats form β-propeller structures that act as a platform for stable or reversible associations with binding partners to fulfil a range of functions, including adaptor/regulatory modules in multiple processes of cellular signal transduction (Biedermann and Hellmann, 2011; Stirnimann et al., 2010).
In the plant proteome, the WD40 domain is found in all TRANSPARENT TESTA GLABRA (TTG) proteins (Bouyer et al., 2008; Pang et al., 2009; Wang et al., 2010) except for Arabidopsis thaliana AtTTG2, which is a WRKY transcription factor (Johnson et al., 2002). TTG proteins were characterised as regulators of trichome and seed development in plants (Bouyer et al., 2008; Morohashi et al., 2007; Petroni and Tonelli, 2011; Szymanski et al., 2000; de Vetten et al., 1997). As the WD40 domain has multiple modes for recruiting different substrates, WD40-containing proteins can serve as interchangeable substrate receptors and interact with diverse types of proteins (Xu and Min, 2011). Due to this characteristic, WD40-containing TTG proteins are potential regulators of both plant development and defence pathways (Bouyer et al., 2008; Wang et al., 2009).
Dual roles in both development and defence have been elucidated for several TTG proteins in plants. In Arabidopsis, for instance, AtTTG1 interacts with the bHLH transcription factor GL3 while recruiting MYB GL1, thereby forming the WD40-bHLH-MYB complex that acts to regulate trichome development (Morohashi et al., 2007; Szymanski et al., 2000; Zhao et al., 2008). Our previous study showed that WD40 enabled tobacco (Nicotiana tabacum) NtTTG1 to interact with ParA1 (Wang et al., 2009), an elicitin protein that is produced by an oomycete pathogen and induces hypersensitive cell death (HCD) in the plant (Kamoun et al., 1993; Liang et al., 2004). The NtTTG1-ParA1 interaction was essential for the induction of HCD in leaf trichomes (Wang et al., 2009). An NtTTG1-like gene was also cloned from tobacco by amplification with NtTTG1 primers using the polymerase chain reaction (PCR) (Wang et al., 2010). Based on high sequence identity with WD40-containing proteins identified in other plant species, NtTTG1-like has been renamed NtTTG2 (GenBank accession number FJ795022). NtTTG2 differs from NtTTG1 in its organ-specific expression. While ParA1 induces NtTTG1 expression and interacts with the NtTTG1 protein only in the trichome (Wang et al., 2009), NtTTG2 is expressed in the leaves and other organs of the tobacco plant (Wang et al., 2010). Because many phytopathogens invade plant leaves rather than trichomes, NtTTG2 is more likely related to pathogen defence than NtTTG1. Although the NtTTG2 and NtTTG1 proteins share four WD40 repeats (Wang et al., 2009; Wang et al., 2010), it is unclear whether NtTTG2 performs similarly to NtTTG1 by interacting with other proteins and regulating HCD.
Plant HCD can be induced by such biotic elicitors as elicitin proteins (Kamoun et al., 1993; Wang et al., 2009) and certain phytopathogens (Dangl et al., 1996; Durrant and Dong, 2004). In response to the inducers, salicylic acid (SA) synthesis is elevated to induce expression of pathogenesis-related (PR) genes (Ryals et al., 1996) through activation by the NPR1 (nonexpresser of PR genes 1) protein (Cao et al., 1994; Cao et al., 1997; Delaney et al., 1995; Dong, 2004; Ryals et al., 1997). In the absence of infection or at steady-state levels of SA, NPR1 is predominantly oligomeric and sequestered in the cytoplasm. Upon pathogen infection or an elevation of the SA levels, NPR1 is reduced to a monomeric state, and a large amount of monomeric NPR1 translocates to the nucleus (Mou et al., 2003). Within the nucleus, NPR1 interacts with the TGA-bZIP transcription factors to activate PR genes (Després et al., 2000; Fonseca et al., 2010; Zhou et al., 2000). The E3 ligase-dependent proteasome mediates the degradation of NPR1 following PR gene activation such that fresh NPR1 is used in every round of transcriptional regulation (Mukhtar et al., 2009; Spoel et al., 2009).
Gaps exist in the NPR1-regulated defence (Mukhtar et al., 2009). For the transcriptional activity, NPR1 must be recruited to the target gene promoter, although the protein that brings NPR1 to the PR gene promoter has not been identified. For NPR1 to be degraded in the nuclear proteasome pathway, an unknown adaptor protein is required to position NPR1 for the E3 ligase. The positioning of interacting partners to create a binding site for E3 is a common theme for the function of the WD40 domain (Biedermann and Hellmann, 2011; Orlicky et al., 2003; Stogios et al., 2005; Xu and Min, 2011). NPR1 contains a BTB/POZ (broad-complex, tramtrac, bric-à-brac/poxvirus, zinc finger) domain, an ankyrin domain and an IκB domain (Cao et al., 1997; Rochon et al., 2006; Ryals et al., 1997), and BTB and IκB are both binding partners of WD40 (Ryals et al., 1997; Stogios et al., 2005). Therefore, certain WD40-domain-containing proteins may play a role in the transcriptional activity or proteasome-mediated degradation of NPR1. Additionally, NtTTG2 may affect the nuclear translocation of NPR1 because NtTTG2 is predicted to localise in the cytoplasm with a high probability (WoLF PSORT at http://wolfpsort.org/). If this prediction is correct, then NtTTG2 may retain NPR1 in the cytoplasm, thereby preventing NPR1 from nuclear localisation and PR gene activation. According to this hypothesis, NtTTG2 is a repressor of the NPR1-regulated defence and offers an alternative mechanism to proteasome-mediated NPR1 degradation (Spoel et al., 2009) to avoid a constitutive defence response in the absence of an infection. Examining of this hypothesis is a major purpose of this study.
NtTTG2 expression is repressed by SA and tobacco pathogens
To obtain information on the defensive role of the NtTTG2 gene, we evaluated its expression in tobacco variety NC89 following treatment with SA or inoculation with viral and bacterial pathogens. As shown in Fig. 1A, NtTTG2 expression declined in the SA-treated plants compared to the control plants that were treated with a buffer solution. The SA treatment caused a 4-fold decrease in the level of NtTTG2 transcription (Fig. 1B; ANOVA test, P<0.01). Inoculation with pathogens also repressed NtTTG2 expression (Fig. 1A). The viral pathogens cucumber mosaic virus (CMV) and tobacco mosaic virus (TMV) were more repressive than the bacterial pathogen Pectobacterium carotovora subsp. carotovora (Pcc; Fig. 1A,B). In plants inoculated with CMV, TMV and Pcc, the relative amounts of the NtTTG2 transcript were 18%, 13% and 35%, respectively, compared with the control plants (Fig. 1B).
NtTTG2 represses the PR gene expression
To test whether NtTTG2 affects the SA/NPR1 signalling pathway, we analysed the effect of NtTTG2 silencing and overexpression on the expression of genes involved in the pathway. A hairpin-based gene silencing (HBGS) unit, which was generated with a plant binary vector (Fig. 2A), was introduced into the NC89 genome (supplementary material Fig. S1). NtTTG2-silenced (TTG2RNAi) plants exhibited substantial downregulation of NtTTG2 transcription compared with the transgenic control (wild-type/RFP, WT/RFP) plant, which was generated by the transformation of NC89 with a vector containing an insert of the RFP gene encoding red fluorescent protein (RFP; Fig. 2B,C). Among all of the tested TTG2RNAi lines, TTG2RNAi4 induced the most effective silencing of NtTTG2 RNA (Fig. 2B,C) and reduction of the NtTTG2 protein (Fig. 2D,E). NtTTG2 overexpression was achieved by expressing the NtTTG2::RFP fusion gene in NC89 plants under the control of the cauliflower mosaic virus 35S promoter (p35S) (Fig. 2F). In the tested p35S::TTG2::RFP lines, p35S::TTG2::RFP1 produced the highest level of NtTTG2 RNA (Fig. 2G,H) or NtTTG2 protein (Fig. 2I,J).
We analysed the effect of NtTTG2 alterations on the transcription of the PR-1a and PR-2a genes, both of which respond to SA and require NPR1 for their transcription (Cao et al., 1997). Both genes were highly expressed in the TTG2RNAi lines but were repressed in the p35S::TTG2::RFP lines compared to the WT/RFP plants following treatment with SA (Fig. 3). These differences were consistent with the production of the NtTTG2 protein (Fig. 2D,E,I,J) as the PR-1a and PR-2a transcripts were highest in TTG2RNAi4 and lowest in p35S::TTG2::RFP1 (Fig. 3), which suggests a repressive effect of NtTTG2 on the SA-induced PR gene expression.
NtTTG2 is a repressor of disease resistance
To test whether NtTTG2 plays a role in the resistance of tobacco plants to pathogens, we compared the responses of TTG2RNAi4, p35S::TTG2::RFP1 and WT/RFP to leaf inoculation with CMV, TMV and Pcc compared to a buffer used in a mock inoculation. In NC89, both viruses, which were inoculated in lower leaves, caused chlorosis and a deformation of newly created (systemic) leaves, while Pcc caused necrosis of the inoculated leaves (Sun et al., 2010a; Wu et al., 2010). The severity of these symptoms varied with the plant genotype, as surveyed 20 days post-inoculation (dpi) with the viruses and 5 dpi with Pcc (Fig. 4A). It was consistently observed that the symptoms were less severe in TTG2RNAi4 but more severe in p35S::TTG2::RFP1 compared to WT/RFP, which suggested that TTG2RNAi4 was more resistant, while p35S::TTG2::RFP1 was more susceptible than WT/RFP to infection by the three pathogens. Indeed, TMV and CMV multiplications, which were indicated by the expression of their CP genes that encode coat proteins (Fig. 4B), and Pcc multiplication, as indicated by the bacterial population in leaf tissues (Fig. 4C), were decreased in TTG2RNAi4 but increased in p35S::TTG2::RFP1. Clearly, NtTTG2 silencing enabled the plant to repress pathogen growth and alleviated the severity of the symptoms, while NtTTG2 overexpression had the opposite effect, which suggested that NtTTG2 is a repressor of disease resistance.
The repressive effect of NtTTG2 on disease resistance was confirmed in Xanthi (NN), which is an N. tabacum variety that incurs systemic chlorosis and deformation when infected by CMV and local necrosis when infected by TMV or Pcc. With protocols similar to those used for NC89, we generated p35S-controlled NtTTG2-overexpressing transgenic Xanthi (p35S::TTG2/X) and control transgenic Xanthi (TC-Xanthi) plants. With the virus-induced gene silencing (VIGS) technique (Wang et al., 2009; Sun et al., 2010b), we generated TTG2-silencing Xanthi (TTG2RNAi/X) plants. Based on the inoculation tests, the repression of resistance was a consistent characteristic of the p35S::TTG2/X lines tested, while p35S::TTG2/X1 was the most effective repressor (supplementary material Table S1). Compared with TC-Xanthi, p35S::TTG2/X1 supported CMV (Fig. 5A), TMV (Fig. 5B) and Pcc (Fig. 5C) multiplication and displayed more severe symptoms (Fig. 5D), which suggested impaired resistance as a result of NtTTG2 overexpression. Conversely, the resistance due to TTG2RNAi/X impeded the pathogen growth (Fig. 5A-C) and alleviated symptom severities (Fig. 5D) compared to TC-Xanthi. Enhanced resistance was found to be a consistent characteristic of the TTG2RNAi/X lines tested, while TTG2RNAi/X1 showed the highest level of resistance (supplementary material Table S1). The performances of these Xanthi genotypes confirmed the role of TTG2 in repressing plant resistance.
NtTTG2 suppresses resistance by repressing SA/NPR1-regulated defence
We sought to elucidate whether the repressive effect of NtTTG2 on resistance was associated with a repression of SA/NPR1-regulated defence. We first examined SA production in the WT/RFP, TTG2RNAi4 and p35S::TTG2::RFP1 plants. As shown in Fig. 6A, the SA content was markedly elevated in the pathogen-inoculated plants compared to the basal level detected in the mock-inoculated plants. However, the SA elevation due to pathogen exposure was equivalent (approximately eightfold) in TTG2RNAi4, p35S::TTG2::RFP1 and WT/RFP. Clearly, NtTTG2 did not affect SA production, and thus SA production is not likely to be a reason for the opposite effects of NtTTG2 silencing and overexpression on resistance.
We then determined the PR gene expression in WT/RFP, TTG2RNAi4 and p35S::TTG2::RFP1 plants (Fig. 6B). In WT/RFP, the PR-1a and PR-2a expression was induced by the pathogen inoculation compared to the mock infection. Substantial amounts of both gene transcripts accumulated in TTG2RNAi4, but there was little accumulation in p35S::TTG2::RFP1 compared to WT/RFP. This result suggests that the repressive effect of TTG2 on resistance was associated with a repression of the SA/NPR1-regulated defence.
We next tested the effect of NtTTG2 silencing on resistance in the absence of SA. The VIGS protocol (Wang et al., 2009; Sun et al., 2010b) was used to manipulate NtTTG2 silencing in NahG, a transgenic Xanthi (NN) plant that cannot accumulate SA and is therefore highly susceptible to pathogens (Gaffney et al., 1993). NtTTG2-silenced NahG (TTG2RNAi/NahG) plants were compared with NahG and its parent Xanthi in terms of their responses to CMV, TMV and Pcc. After inoculation with CMV, NahG incurred acute chlorosis and a greater deformation than Xanthi, which confirmed a resistance impairment due to SA deficiency (Fig. 5E). If NtTTG2 represses resistance by repressing SA-mediated defence, NtTTG2 silencing should increase the resistance level in NahG. As anticipated, TTG2RNAi/NahG behaved similarly to Xanthi, but it was more resistant than NahG to pathogens (Fig. 5E-H). The CMV-induced leaf chlorosis and deformation were more aggressive in NahG than in TTG2RNAi/NahG, which incurred similar degrees of symptom severities as Xanthi (Fig. 5E). In contrast to Xanthi and TTG2RNAi/NahG, NahG supported pathogen growth effectively, which was indicated by the expression of viral CP genes (Fig. 5F,G) and the bacterial population that was recovered from leaves (Fig. 5H). Therefore, NtTTG2 silencing compensates for the compromised resistance by inducing an SA deficiency.
We further tested the interaction between NtTTG2 and NPR1 in the plant resistance. We used VIGS to manipulate NPR1 silencing in TC-Xanthi and TTG2RNAi/X1 and generated NPR1-silenced (NPR1RNAi) and concurrent NtTTG2-NPR1-silenced (TTG2-NPR1RNAi) plants. Inoculation experiments indicated that NPR1RNAi1 and TTG2-NPR1RNAi1 should be investigated further (supplementary material Table S1). TTG2-NPR1RNAi1 was more resistant than NPR1RNAi1, but more susceptible than Xanthi or TTG2RNAi/X1 based on pathogen growth (Fig. 5A-C) and symptom severity (Fig. 5D). Thus, NtTTG2 silencing compensated for compromised resistance through the effect of NPR1 silencing. Together, the data described above suggest the possibility that NtTTG2 suppresses the resistance to pathogens through repression of the SA/NPR1-regulated defence pathway in plants.
NtTTG2 is a typical WD40-domain protein, but it does not interact with NPR1
To reveal the molecular mechanism by which NtTTG2 represses the SA/NPR1-regulated defence pathway, we first analysed the canonical WD40 domain and WD40 repeats present in NtTTG2. Based on modelling by the PROSITE ExPASy programme, the GH-WD motif contains conserved but replaceable GH and WD dipeptide residues that differentially contributed to the role of GH-WD40 in molecular interactions (Fig. 7A). However, NtTTG2 lacked the upstream GH dipeptide residues and contained just four repeats of the WD dipeptide residues (Fig. 7B) resembling the traditionally defined canonical WD40 domain, especially in regard to the functional properties of the inner residues (Fig. 7C). Based on this modelling, NtTTG2 is a typical WD40-domain protein.
Next, we conducted yeast two-hybrid (Y2H) assays to test whether NtTTG2 might interact with NPR1 because NPR1 contains a BTB domain and an IκB domain, which are both binding partners of the WD40 domain (Ryals et al., 1997; Stogios et al., 2005). However, the interactions were not detected in replicate Y2H assays (Fig. 7D). To determine the possibility that NtTTG2 might interact with NPR1 in the plant cell, we performed a bimolecular fluorescence complementation (BiFC) protocol using yellow fluorescent protein (YFP) as a probe. The NtTTG2 fused to the N-terminal region of YFP and NPR1, which was ligated to the C-terminal region, were subjected to transient expression tests. Subsequent BiFC monitoring did not detect an NtTTG2-NPR1 interaction (Fig. 7E). It is apparent that NtTTG2 does not physically interact with NPR1.
NtTTG2 sequesters NPR1 in the cytoplasm
In the absence of a direct interaction (Fig. 7D,E), NtTTG2 still affects the subcellular localisation of NPR1 (Fig. 8). Fluorescence imaging indicated that the subcellular localisation of NRP1 varied with NtTTG2 alterations in the plant cell (Fig. 8A). The WoLF PSORT programme predicted 65%, 45%, 60% and 20% possibilities that NtTTG2 would localise to the cytoplasm only, the nucleus only, both compartments, and the chloroplast. However, the NtTTG2-RFP fusion protein was found to localise in both the cytoplasm and nucleus of p35S::TTG2::RFP1 cells (Fig. 8A, left). In the plant cell, NtTTG2-RFP fluorescence was conspicuous in comparison with the RFP in the transgenic WT/RFP control plant and the absence of fluorescence in the NtTTG2-silenced plant, TTG2RNAi4.
In WT/RFP, TTG2RNAi4 and p35S::TTG2::RFP1 plants, NPR1 performed differently based on a fusion to the green fluorescent protein (GFP). The NPR1-GFP fusion protein was found mainly in the cytoplasm and was detected sporadically in the nucleus of WT/RFP (Fig. 8A, right), which was the control plant that produced NtTTG2 at a basal level (Fig. 2D,E,I,J). In TTG2RNAi4, in which the production of NtTTG2 was almost eliminated (Fig. 2D,E), the NPR1-GFP protein was detected in the nucleus in addition to the cytoplasm (Fig. 8A, right). However, NPR1-GFP was mainly found in the cytoplasm of p35S::TTG2::RFP1 (Fig. 8A, right), which was the NtTTG2-overexpressing line that produced approximately sevenfold more NtTTG2 than WT/RFP (Fig. 2J). Based on the fluorescence intensities, the relative amount of NPR1-GFP localised to the cytoplasm or nucleus was negatively correlated with the amount of NtTTG2-RFP in the plant cell (Fig. 8B,C). In TTG2RNAi4, there was little cytoplasmic or nuclear NtTTG2-RFP, while a large proportion of NRP1-GFP was detected in the nucleus. In contrast, a predominant proportion of NRP1-GFP was localised to the cytoplasm of p35S::TTG2::RFP1 in which abundant amounts of TTG2-RFP accumulated in both the cytoplasm and nucleus.
The fluorescence imaging results were confirmed by immunoblot analyses of protein samples from cytoplasmic and nuclear fractions (Fig. 8D). Hybridisation with the NtTTG2 antibody detected abundant amounts of the protein in both the cytoplasmic and nuclear fractions of WT/RFP and p35S::TTG2::RFP1 cells, but little protein was measured in TTG2RNAi4. Based on hybridisation with the GFP antibody, the NPR1-GFP fusion protein was more highly expressed in the cytoplasmic fraction than in the nucleus of WT/RFP cells, but it was conversely localised in TTG2RNAi4 cells. Noticeably, a large quantity of NPR1-GFP was detected in the cytoplasm with a small amount in the nucleus of p35S::TTG2::RFP1 cells. These analyses, together with fluorescence imaging, suggest that excess NtTTG2 sequestered NPR1 in the cytoplasm, but the inhibited NtTTG2 production allowed NPR1 to localise to the nucleus.
The NtTTG2-modulated nuclear localisation of NPR1 affects PR gene transcription
To quantify the effect of NtTTG2 alterations on the subcellular localisation of NPR1, we counted the number of tobacco cells that showed NPR1-GFP fluorescence only in the cytoplasm, only in the nucleus or in both compartments. Nearly 90% of WT/RFP cells and more than 95% of p35S::TTG2::RFP1 cells showed fluorescence only in the cytoplasm, whereas more than 90% of TTG2RNAi4 cells showed fluorescence only in the nucleus (Fig. 9A). This confirms that NtTTG2 overexpression sequestered NPR1 in the cytoplasm, but the lack of NtTTG2 allowed NPR1 to localise to the nucleus. Because nuclear localisation is a prerequisite for the role of NPR1 in gene expression regulation (Kinkema et al., 2000), it is logical to hypothesise that the NtTTG2-modulated subcellular localisation of NPR1 affects the expression of the PR gene. This hypothesis was validated by parallel tests of NtTTG2 expression and the transient expression of NPR1 and PR in WT/RFP, TTG2RNAi4 and p35S::TTG2::RFP1 cells (Fig. 9B). Little NtTTG2 was expressed in TTG2RNAi4, while the gene transcript was 6.5-fold higher in p35S::TTG2::RFP1 than in WT/RFP cells. Intriguingly, NPR1 did not change substantially among the three plant genotypes, which implied that the effect of NtTTG2 (NtTTG2) was posttranscriptional. However, the expression of PR-1a and PR-2a was induced in TTG2RNAi4 cells but was repressed in p35S::TTG2::RFP1 compared to WT/RFP cells. This analysis suggests that the NtTTG2-modulated nuclear localisation of NPR1 affects PR expression, which indicates the activation of the SA/NPR1-regulated defence pathway (An and Mou, 2011).
In plants, inactivation of signal transduction for pathogen resistance is critical to avoiding the fitness consequences that are associated with a constitutive defence response in the absence of a pathogen infection (Mukhtar et al., 2009). In this study, we have characterised NtTTG2 as a repressor of the SA/NPR1-regulated defence pathway in tobacco. We found that NtTTG2 suppresses the resistance of tobacco to viral and bacterial pathogens through repression of the SA/NPR1-regulated defence (Figs 1-Fig. 2,Fig. 3,Fig. 4,Fig. 5,6). Evidence was found not only for SA-repressed NtTTG2 expression (Fig. 1) but also for NtTTG2-mediated obstruction of the SA signal transduction for PR gene expression (Fig. 3), which indicates an activation of the NPR1-regulated SA signalling pathway in plants following a pathogen infection (An and Mou, 2011; Cao et al., 1997; Mukhtar et al., 2009). Evidence was further found for the inhibitive effect of NtTTG2 on the pathogen-induced expression of PR genes (Fig. 6) and the role of NtTTG2 silencing in counteracting the resistance compromise due to either the absence of SA or NPR1 silencing (Fig. 5).
We demonstrate that NtTTG2 executes a repressive effect on the defence by modulating the subcellular localisation of NPR1. NtTTG2 localised in both the cytoplasm and nucleus, while changes in the NtTTG2 production status impacted the nuclear localisation of NPR1 (Fig. 8). The mechanism was thought to consist of the role of the WD40 domain in molecular interactions (Bouyer et al., 2008; Walker et al., 1999; Johnson et al., 2002; Xu and Min, 2011, Todd et al., 2010; de Vetten et al., 1997; Wang et al., 2009). NtTTG2 is characteristic of a typical WD40-domain protein, but it does not directly interact with NPR1 (Fig. 7). Interestingly, NtTTG2 was able to affect the subcellular localisation of NPR1 and the subsequent PR gene expression (Figs 8, 9). When NtTTG2 production was compromised, NPR1 localised to the nucleus, and the PR gene expression was induced. On the contrary, NtTTG2 overexpression retained NPR1 in the cytoplasm, and the PR genes were not expressed. Therefore, the NtTTG2-modulated nuclear localisation of NPR1 had a subsequent effect on the gene expression downstream of the SA signalling pathway, which regulates the resistance to pathogens (Delaney 1997; Dong, 2004; Cao et al., 1997; Ryals et al., 1996).
The NtTTG2-modulated nuclear localisation of NPR1 may reflect the real effect of NtTTG2 on NPR1 under the normal condition when SA remains at a basal level (Fig. 6). In this case, the basal level of NtTTG2 applied an inhibitive effect on NPR1, which was also present at a basal level (Fig. 8). Likewise, elevated NtTTG2 was inhibitory towards NPR1 elevation, as in the case of concurrent NtTTG2 and NPR1 overexpression, such that the nuclear localisation of NPR1 was inhibited (Fig. 8). However, we do not have evidence to support a physiological connection between NtTTG2-modulated subcellular localisation of NPR1 and defence responses of the plant. Although our genetic analysis (Fig. 5) suggests an antagonism between NtTTG2 and the SA/NPR1-regulated defence, the present study does not exclude an alternative model that the effect of NtTTG2 is independent of SA and NPR1. Therefore, it is important to characterise the regulatory mechanism that underlies the effect of NtTTG2 on NPR1.
In regard to how NtTTG2 is connected with NPR1, possible linkages exist in associating NtTTG2, as a WD40-domain-containing protein (Wang et al., 2010), with the proteasome-mediated NPR1 turnover (Cao et al., 1997; Spoel et al., 2009). NtTTG1 and NtTTG2 share a common WD40 domain that is characterised by four repeats of a tryptophan-aspartic dipeptide and a length of 38–45 amino acid residues in a single repeat (Wang et al., 2010). In contrast to the binding of NtTTG1 to the defence-eliciting protein ParA1 (Wang et al., 2009), NtTTG2 was unable to interact with NPR1 (Fig. 7), suggesting the requirement for an additional mediator that may interact with NtTTG2 and NPR1.
Hypothetically, NtTTG2 may function together with potential mediators for the proteasome-mediated NPR1 turnover that is required for SA signalling in the defence response to pathogens (Cao et al., 1997; Spoel et al., 2009). This hypothesis is related to several questions. One is whether the speculated mediator assists NtTTG2 in the sequestration of NPR1 in the cytoplasm or if it antagonises NtTTG2 in regulating the nuclear localisation of NPR1 and promoting the E3 ligase-dependent proteasome-mediated degradation required for maintenance of the defence responses. Additionally, it is that this could be realised via the WD40 domain present in NtTTG2 (Fig. 7), which is widely implicated in the interactions between F-box proteins and E3 ligase protein complexes (Xu and Min, 2011; Biedermann and Hellmann, 2011). NPR1 contains BTB and IκB domains, which are both binding partners of WD40 (Stogios et al., 2005; Ryals et al., 1997). Moreover, an unknown adaptor protein has been proposed to function in positioning NPR1 for the E3 ligase (Mukhtar et al., 2009), while positioning by interacting partners to create or stabilise a binding site for E3 is a common function of the WD40 domain (Biedermann and Hellmann, 2011; Stogios et al., 2005; Orlicky et al., 2003). In addition, the interactions between the E3 ligases and NPR1 occur at an elevated SA level (Mukhtar et al., 2009; Spoel et al., 2009). Thus, although it is unable to bind NPR1, NtTTG2 is still likely to play a role in NPR1 turnover by participating in E3 ligase-mediated proteasomal degradation when SA is at a basal level (Spoel et al., 2009). Future studies to validate this prediction will reveal the molecular mechanism that underlines the NtTTG2-modulated nuclear localisation of NPR1 as a means of coordinating plant development and defence responses.
Materials and Methods
Plant treatment and inoculation
Plants were grown in potting soil in a greenhouse at 22-26°C for 40 days. An aqueous solution of 2 mM SA amended with 0.02% (v/v) of the surfactant Silwet-77 was applied by spraying plant tops with a low-pressure atomiser. Phosphate buffer (0.2 mM, pH 7.4) containing Silwet-77 (0.02%) was applied similarly in control. The fourth youngest leaves were inoculated with CMV, TMV (Dong and Beer, 2000, Peng et al., 2003) and Pcc (Sun et al., 2010b), or the buffer in control. Treated and inoculated plants were used in tests of gene expression and plant resistance.
Gene expression analysis
RNA was isolated from the sixth youngest leaves. RT-PCT and real-time RT-PCR were conducted using specific primers (supplementary material Table S2) and previously described protocols (Chen et al., 2008; Liu et al., 2010). The stably expressed EF1α or ACTIN2 gene (Wu et al., 2010; Peng et al., 2004; Liu et al., 2011; Dong et al., 2004; Dong et al., 2005) was used as a reference. Gene expression levels were quantified as described (Livak and Schmittgen, 2001). Northern blots of RNA samples were hybridised to specific probes labelled with digoxigenin (Peng et al., 2004).
Gene silencing and overexpression
Genes were cloned by RT-PCR using leaf RNA samples. HBGS was constructed by using the pBSSK-in vector (Stratagene), which contains a 200-bp intron from the Phaseolus vulgaris nitrite reductase gene (Kieleczawa, 2005), and sense and antisense strands of a 352-bp NtTTG2 fragment. VIGS was constructed with the virus Y35 DNA1 vector and a 565-bp NtTTG2 fragment or a 536-bp NPR1 fragment (Sun et al., 2010b; Tao and Zhou, 2004). The HBGS or VIGS unit was cloned into the binary vector pCAMBIA1301. For gene overexpressing construction, NtTTG2 and NPR1 fused to RFP and GFP, respectively, were cloned into pCAMBIA1301. Confirmed constructs were transferred into cells of the Agrobacterium tumefaciens strain EHA105, followed by plant transformation and characterisation of transgenic plants as described (Peng et al., 2004; Sun et al., 2010b; Wang et al., 2009). T3 homozygous progenies of NtTTG2-modified and T2 homozygous progenies of other transgenic plants were used in this study.
The fifth to seventh youngest leaves were used to prepare total protein (Liu et al., 2006), cytoplasmic protein (Zhang et al., 2011) and nuclear protein (Kinkema et al., 2000) samples. Western blots were hybridised with GFP antibody from Novagen or NtTTG2 antibody produced by immunising New Zealand white rabbits (Oryctolagus cuniculus) as previously described (Sun et al., 2010b). PEPC and histone H3 used as cytoplasmic and nuclear markers (Garcίa et al., 2010) were hybridised with PEPC antibody (Rockland) and histone H3 antibody (Abcam), respectively. Hybridised blots were probed by horseradish peroxidase-conjugated secondary antibody (Beyotime).
Modelling of WD40 domain was established with the PROSITE ExpASy program (http://prosite.expasy.org/). WD40 repeats in NtTTG2 were positioned by the SMART program (http://smart.embl-heidelberg.de/), and contributions of core residues to the domain function were determined by referring to NtTTG2 WD40 modelling.
Roots of 20-day-old WT/RFP, TTG2RNAi4 and p35S::TTG2::RFP1 plants grown on Murashige and Skoog agar medium (23-26°C) were immersed with a suspension of EHA105 cells transformed with pCAMBIA1301::NPR1::GFP or the empty vector. Sixty hours later, root samples were observed under the ZEISS LSM710 confocal microscope. Red and green fluorescence was captured between 591-630 and 493-540 nm using argon laser excitation at 561 and 488 nm, respectively (Wang et al., 2009).
Y2H and BiFC assays
Previously described protocols were used (Wang et al., 2009).
We thank Dr Jian Hua and Dr Hailing Jin for suggestions on the presentation.
This work was supported by the Key Basic Scientific Study and Development Plan 973 [grant number 2012CB114003 to H.D.] and the Natural Science Foundation [grant number 31272027 to H.D.] of China; and Jiangsu Provincial Priority Academic Program Development of Higher Education Institutions.