Plants protect themselves from the harmful effects of pathogens by resistance and tolerance. Disease resistance, which eliminates pathogens, can be modulated by bacterial type III effectors. Little is known about whether disease tolerance, which sustains host fitness with a given pathogen burden, is regulated by effectors.
Here, we examined the effects of the Xanthomonas effector protein XopDXcc8004 on plant disease defenses by constructing knockout and complemented Xanthomonas strains, and performing inoculation studies in radish (Raphanus sativus L. var. radiculus XiaoJinZhong) and Arabidopsis plants.
XopDXcc8004 suppresses disease symptoms without changing bacterial titers in infected leaves. In Arabidopsis, XopDXcc8004 delays the hormone gibberellin (GA)-mediated degradation of RGA (repressor of ga1-3), one of five DELLA proteins that repress GA signaling and promote plant tolerance under biotic and abiotic stresses. The ERF-associated amphiphilic repression (EAR) motif-containing region of XopDXcc8004 interacts with the DELLA domain of RGA and might interfere with the GA-induced binding of GID1, a GA receptor, to RGA.
The EAR motif was found to be present in a number of plant transcriptional regulators. Thus, our data suggest that bacterial pathogens might have evolved effectors, which probably mimic host components, to initiate disease tolerance and enhance their survival.
Plants are constantly exposed to a battery of potential pathogens, including viruses, fungi, bacteria, parasites and insects. Unlike animals, which move to escape environmental challenges, plants are sessile organisms and tend to protect themselves from potential pathogens through either resistance mechanisms, which prevent or limit pathogen infection and growth, or tolerance, which alleviates the host fitness costs from pathogens without limiting infection (Best et al., 2008; Medzhitov et al., 2012). During plant–pathogen interactions, plants have evolved specific mechanisms to mount resistance towards most pathogens. These processes are referred to as pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) and effector-triggered immunity (ETI). In turn, pathogens have evolved so-called type III effectors that are injected into the host cell to suppress PTI and ETI (Chisholm et al., 2006; Jones & Dangl, 2006). Interestingly, several studies on the phytopathogenic bacterium Xanthomonas have suggested that bacterial effectors are delivered by the type III secretion system into plant cells to regulate disease tolerance. For example, the Xanthomonas campestris pv vesicatoria (Xcv) strain 85-10 effector XopDXcv85-10 has been shown to delay disease symptoms and promote bacterial growth in susceptible tomato leaves (Kim et al., 2008). XopJXcv85-10, another effector protein from Xcv85-10, has been described to delay the onset of cell death in infected tissues of susceptible pepper leaves (Üstün et al., 2013). XopDXccB100, an effector protein from X. campestris pv campestris strain B100 (XccB100), leads to leaf chlorosis rather than the hypersensitive response (HR) symptom in Arabidopsis plants (Canonne et al., 2011). However, very little is known about the mechanisms of delayed disease symptoms that are regulated by effectors.
Among c. 35 putative effectors from different Xanthomonas species, the XopD protein family plays crucial roles in the suppression of plant defenses and the promotion of bacterial infections (Ryan et al., 2011; Canonne et al., 2012). Xcc strain 8004 (Xcc8004), the causal agent of black rot disease, infects Brassica crops and Arabidopsis plants and produces XopDXcc8004, a XopD family member. In Arabidopsis cells, XopDXcc8004 is localized in the cytoplasm and the nucleus (Canonne et al., 2011; Kim et al., 2011). Compared with the subnuclear foci-localized effectors XopDXcv85-10 and XopDXccB100 (Canonne et al., 2011), XopDXcc8004 contains three N-terminal ERF-associated amphiphilic repression (EAR) motifs and a C-terminal small ubiquitin-like modifier (SUMO) domain, but it lacks the N-terminal DNA-binding helix-loop-helix (HLH) domain. The EAR motif has previously been described in plant transcriptional regulators where is responsible for the repression of other transcriptional activators (Kazan, 2006), whereas the SUMO domain is similar to the yeast ubiquitin-like protease 1, which removes SUMO from SUMO-conjugated proteins (Hotson et al., 2003). Considering the distribution of XopD effector proteins in different plant pathogenic bacteria and the differences in the domain structure between individual members of this effector family (Kim et al., 2011), it can be hypothesized that a certain functional specialization of XopD proteins might have occurred during evolution. In order to shed further light on the molecular mechanism of XopD proteins, we conducted a functional characterization of XopDXcc8004.
In this study, we found that XopDXcc8004 suppressed disease symptoms and did not affect bacterial growth in Xcc-infected radish (Raphanus sativus L. var. radiculus XiaoJinZhong) leaves and Pseudomonas syringae-infected Arabidopsis leaves. Furthermore, we investigated the potential host targets of XopDXcc8004 and identified interactions of XopDXcc8004 with DELLA proteins. Previous reports have shown that DELLA proteins are localized in the nucleus of plant cell and are plant growth repressors (Sun & Gubler, 2004). The plant hormone gibberellin (GA) stimulates the binding of the GA receptor GID1s (GID1a, GID1b and GID1c) to DELLA proteins; this binding in turn promotes DELLA degradation via the 26S ubiquitin-proteasome system (Feng et al., 2008). In Arabidopsis, DELLA proteins contain five members: RGA, GAI, RGL1, RGL2 and RGL3, all of which share an N-terminal DELLA regulatory domain (containing the DELLA and VHYNP motifs) and a C-terminal GRAS domain (for GAI, RGA and Scarecrow) with homology to the GRAS family of putative transcription factors (Sun & Gubler, 2004). Importantly, DELLA proteins promote plant tolerance under adverse biotic and abiotic stresses (Achard et al., 2008). Thus, interactions of XopDXcc8004 with DELLA proteins provide a possible explanation for related disease tolerance mechanisms.
Materials and Methods
Pathogens, plants and growth conditions
The bacterial strains used in this study included Xcc8004, P. syringae pv tomato (Pst) DC3000 and Agrobacterium tumefaciens strain GV3101, all of which were cultivated at 28°C. The Arabidopsis thaliana (L.) Heynh plants used for transgene experiments included the Columbia (Col)-0 ecotype and the Landsberg erecta ecotype background RGAp::GFP-RGA (Silverstone et al., 2001), a transgenic line that expresses the green fluorescent protein (GFP)-RGA fusion protein under the control of the native RGA promoter. Arabidopsis and Nicotiana benthamiana plants were grown in a growth room at 23°C at 70% relative humidity with a 10-h photoperiod. Radish plants were grown in a glasshouse at 22–28°C at 80% relative humidity.
Constructs, generation of knockout and complemented strains, and transgenic plants
The primers used in this study are listed in Supporting Information Table S1. To knock out the xopD gene from Xcc8004, the 500-base pair sequences at both ends of the xopD coding region were amplified and cloned into the suicide plasmid pK18mob (Schäfer et al., 1994). The resulting construct was transferred into Xcc8004 by triparental conjugation (Figurski & Helinski, 1979). The dual crossover mutants were confirmed by polymerase chain reaction (PCR) and the xopD knockout strain was designated as Xcc8004ΔxopD. The entire xopD coding regions of Xcc8004 and XccB100, fused a C-terminal 6× histidine (His) tag coding sequence (Dan et al., 2009), were amplified and cloned into the pHM1 vector under the control of the constitutive lacZ promoter (Aparna et al., 2009). The resulting constructs were transferred into Xcc8004ΔxopD by triparental conjugation to generate two complemented strains, Xcc8004ΔxopD(XopDXcc8004) and Xcc8004ΔxopD(XopDXccB100).
For the yeast two-hybrid screening, XopDXcc8004 and its derivatives were amplified and cloned into the pGBKT7 vector (Clontech, Mountain View, CA, USA) to generate XopDXcc8004-pGBKT7 (BD), XopDXcc80041-210-BD and XopDXcc8004211-442-BD. DELLA genes were amplified and cloned into the pGADT7 vector (Clontech) to generate RGA-pGADT7 (AD), GAI-AD, RGL1-AD, RGL2-AD, RGL3-AD and RGAΔDELLA (a RGA derivative that lacks the DELLA and VHYNP motifs)-AD. The pGADT7 vector contains a HA affinity tag and the pGBKT7 vector contains a c-myc affinity tag.
For the protoplast transformation assays, GID1a, RGA and XopDXcc8004 and its derivatives were amplified and cloned into the pUC19-35S-FLAG-RBS vector (Li et al., 2005) and the pUC19-35S-HA-RBS vector (Li et al., 2005) to generate fusion proteins GID1a-FLAG, RGA-FLAG, XopDXcc8004-FLAG, XopDXcc80041-210-FLAG, XopDXcc8004211-442-FLAG, RGA-HA, XopDXcc8004-HA, XopDXcc80041-210-HA and XopDXcc8004211-442-HA.
For the bimolecular fluorescence complementation (BiFC) assays, XopDXcc8004 was amplified and cloned into the pSPYNE vector to generate XopDXcc8004-nYFP. The pSPYNE vector expresses proteins of interest fused to the N-terminal fragment of the yellow fluorescent protein (YFP) (Walter et al., 2004). RGA was cloned into the pSPYCE vector, which expresses proteins of interest fused to the C-terminal fragment of the YFP (Walter et al., 2004), to generate RGA-cYFP. The pSPYNE vector contains a c-myc epitope tag and the pSPYCE vector contains a HA epitope tag.
In order to generate XopDXcc8004 transgenic plants, the XopDXcc8004-FLAG sequence was amplified from the pUC19-35S-XopDXcc8004-FLAG-RBS plasmid and cloned into the pER8 vector, that is, remaining under the control of an estradiol-inducible promoter (Zuo et al., 2000; Li et al., 2005). The resulting construct was transformed into Col-0 and RGAp::GFP-RGA plants by a A. tumefaciens-mediated transformation system (Zhang et al., 2006). Two independent XopDXcc8004 transgenic lines from Col-0 and one transgenic line from RGAp::GFP-RGA were generated. The transgenic plants were sprayed with 50 mM estradiol to induce the expression of XopDXcc8004.
In order to analyze the effects of XopDXcc8004 on the ubiquitination of RGA, GFP and XopDXcc8004 sequences were cloned into the pUC19-35S-FLAG-RBS vector (Li et al., 2005) to substitute the FLAG tag and to generate XopDXcc8004-GFP. The RGA-FLAG, RGAK33A,K65A–FLAG (K, lysine; A, alanine) and XopDXcc8004-GFP sequences were amplified from the pUC19-35S-RGA-FLAG-RBS and pUC19-35S-XopDXcc8004-GFP-RBS plasmids and cloned into the pCAMBIA1300 vector (Cambia, Brisbane, QLD, Australia), which is used for the transient expression of A. tumefaciens-mediated proteins in N. benthamiana leaves.
Bacterial growth of Xcc was assessed by clipping- or piercing-inoculation with a 105 colony-forming units (CFU) ml−1 inoculum of Xcc8004 or Xcc8004ΔxopD on 4-wk-old radish or Arabidopsis leaves. Bacterial growth of PstDC3000 was assessed by infiltrating-inoculation with a 104 CFU ml−1 inoculum of PstDC3000 on 4-wk-old Arabidopsis leaves. Eight leaf discs (0.74 cm2 per disc, two leaf discs from each plant) per treatment from each time point were ground in 10 mM MgCl2 and diluted. Dilutions of bacteria Xcc and PstDC3000 were spotted onto NYG (nutrient-yeast-glycerol) agar medium (5 g peptone, 3 g yeast extract, 15 g agar and 20 g glycerol l−1) containing 100 μg ml−1 rifampicin and KB (King's B) agar medium (29 g Bacto™ Proteose peptone, 1.5 g K2HPO4, 0.74 g MgSO4, 15 g Bacto™ agar and 8 g glycerol l−1) containing 100 μg ml−1 rifampicin, respectively. Four biological replicates (i.e. four plants) were used, and the experiment was repeated at least three times.
For the disease symptom assays, 4-wk-old radish and Arabidopsis leaves were inoculated with a 105 CFU ml−1 inoculum of Xcc8004, Xcc8004ΔxopD, Xcc8004ΔxopD(XopDXcc8004), Xcc8004ΔxopD(XopDXccB100) or Xcc8004(+GA3) (spraying 100 μM GA3 on radish leaves). Photographs of Xcc-infected radish leaves were taken at 6 d postinoculation (dpi). Photographs of Xcc-infected Arabidopsis leaves were taken at 3 dpi. Four-week-old Arabidopsis plants were sprayed with 50 mM estradiol for 24 h and/or 50 μM GA3 for 2 h before bacterial inoculation with a 104 CFU ml−1 inoculum of PstDC3000 accompanied with (or without) a 105 or 106 CFU ml−1 inoculum of Xcc8004, Xcc8004ΔxopD or Xcc8004ΔxopD(XopDXcc8004). Photographs of leaves were taken at 2 dpi. These experiments were repeated at least three times.
Yeast two-hybrid screening
The Matchmaker™ GAL4 two-hybrid system 3 (Clontech) was used for screening in this study. The full-length XopDXcc8004 coding sequence was cloned in the pGBKT7 vector to generate bait to screen the Arabidopsis cDNA library of mature leaves and roots (CD4-10; The Arabidopsis Biological Resource Center (ABRC)) (Wang et al., 2011). The yeast strain AH109 was cotransformed with the bait and cDNA library. Approximately 106 primary yeast clones were screened. Potential yeast transformants containing cDNA clones that interacted with XopDXcc8004 were selected and sequenced. Interactions of XopDXcc8004, XopDXcc80041-210 and XopDXcc8004221-442 with RGA, RGAΔDELLA, GAI, RGL1, RGL2 and RGL3 in yeast cells were analyzed on the SD (selective minimal synthetic dropout) media (0.67% yeast nitrogen base (without amino acids but with ammonium sulfate), 2% glucose, 2% agar and 1× Amino Acid Dropout Mix), including SD/-Trp-Leu (lacking tryptophan and leucine), SD/-Trp-Leu/X-α-Gal (supplemented with 4 mg ml−1 X-α-Gal), SD/-Trp-Leu-His (lacking tryptophan, leucine and histidine) and SD/-Trp-Leu-His-Ade (lacking tryptophan, leucine, histidine and adenine). Protein expression in the yeast cells was detected by immunoblot assays using anti-HA (Qiagen) or anti-myc (Sigma) antibodies.
Confocal laser microscopy
Leica TCS SP8 Confocal Microscope (Leica Microsystems, Wetzlar, Germany) with ×40 and ×63 oil objectives was used. To detect the GFP fluorescence, the excitation wavelength was 488 nm and a band-path filter of 510–525 nm was used for emission. To detect YFP fluorescence, the excitation wavelength was 514 nm and a band-path filter of 520–550 nm was used for emission. RGAp::GFP-RGA and its XopDXcc8004 transgenic seedlings were grown for 7 d on MS (Murashige-Skoog) medium (Duchefa, Amsterdam, Netherlands) supplemented with 1% sucrose and 1% agar under a 16 h photoperiod at 23°C. The seedlings were transferred to MS liquid medium containing 1 mM estradiol and incubated for 24 h. Before fluorescence observation, the Arabidopsis plants were incubated for 2 h in MS liquid medium containing 10 μM GA3. The Arabidopsis root tips were mounted on microscope slides and GFP fluorescence was detected by confocal microscopy. Arabidopsis protoplasts transfected for 12–18 h were mounted on microscope slides and YFP fluorescence was detected by confocal microscopy.
Arabidopsis protoplasts were cotransfected with the cYFP-fused full-length RGA and the nYFP-fused full-length XopDXcc8004 according to a previously described protocol (Yoo et al., 2007). The complementation of YFP was detected by confocal microscopy.
Arabidopsis RGAp::GFP-RGA and its XopDXcc8004 transgenic seedlings were cultivated for 10 d in a plant growth room on MS medium plates containing 1 mM estradiol or 10 μM GA3 plus 1 mM estradiol at 23°C in 70% relative humidity with a 10 h photoperiod. The fresh-weight ratio of RGAp::GFP-RGA vs transgenic seedlings with different treatments were measured for the growth-inhibition assay.
Transient protein expression in N. benthamiana
Agrobacterium tumefaciens strain GV3101-containing constructs were grown overnight at 28°C in LB (Luria Bertani) agar medium (10 g tryptone, 5 g yeast extract, 15 g Agar and 10 g NaCl l−1) containing 100 μg ml−1 rifampicin and 35 μg ml−1 kanamycin. Bacteria were incubated in induction medium (10 mM MES, pH 5.6, 10 mM MgCl2 and 150 M acetosyringone; Acros Organics, Geel, Belgium) for 2 h, and then bacteria were resuspended to a final concentration of OD600 = 0.5 using induction medium before inoculation. Bacterial suspensions were infiltrated into young but fully expanded N. benthamiana leaves. After infiltration, plants were immediately covered with plastic bags and placed at 23°C for 48 h before protein extraction.
Xcc8004ΔxopD(XopDXcc8004) and Xcc8004ΔxopD(XopDXccB100) strains expressing His-tagged XopD were cultivated overnight at 28°C in MOKA rich medium (Blanvillain et al., 2007) or in secretion medium (Rossier et al., 1999). Secretion experiments were performed as described previously (Rossier et al., 1999). Total protein extracts and trichloroacetic acid-precipitated filtered supernatants were used for the immunoblot assay using anti-His antibody (Sigma) and anti-GroEL antibody (Sigma). Yeast total proteins were extracted according to the protocols provided by Clontech and were then used for the immunoblot assay using anti-HA and anti-myc antibodies. Arabidopsis protoplasts transfected with the indicated constructs were incubated for 12 h (with/without 1 μM GA3 treatment for 2 h) and N. benthamiana leaves transfected with A. tumefaciens strain GV3101 were incubated for 48 h. Total protein was extracted using protein extraction buffer (50 mM HEPES-KOH (pH 7.5), 150 mM KCl, 1 mM EDTA, 0.2% Triton-X100, 1 mM DTT and complete protease inhibitors (Roche)). To investigate the expression of XopDXcc8004 in transgenic plants, total proteins of Arabidopsis leaves sprayed with 50 mM estradiol were extracted and were used for anti-FLAG (Sigma) immunoblot. In the BiFC assay, total proteins of Arabidopsis protoplasts cultivated overnight were extracted and used for anti-HA and anti-myc immunoblot. For the anti-FLAG co-immunoprecipitation (co-IP) assay, 10-d-old RGAp::GFP-RGA and its XopDXcc8004 transgenic seedlings were sprayed with 50 mM estradiol for 12 h. Total proteins of Arabidopsis seedlings and N. benthamiana leaves was extracted and were incubated with an agarose-conjugated anti-FLAG antibody (Sigma-Aldrich) for 4 h and washed 6 times with protein extraction buffer. The bound protein was eluted with 0.5 mg ml−1 3× FLAG peptide. Total proteins and the eluted proteins were detected using anti-FLAG, anti-HA, anti-GFP (Roche) or anti-ubiquitin (Abcam, Cambridge, UK) antibody.
Leaves of Col-0 and the XopDXcc8004 transgenic plants were cut into 1-mm strips and incubated in 200 μl water in a 96-well plate for 12 h before the addition of 1 μM bacterial flagellar peptide flg22 in 200 μl reaction buffer supplemented with 20 mM luminol and 1 μg horseradish peroxidase (Sigma). Luminescence was recorded with a Luminometer (Promega).
Total RNA was extracted with an RNeasy Plant Mini kit (Qiagen). RNA samples were treated with DNase Turbo DNA-free (Promega), and 5 g of treated RNA was used for reverse transcription with SuperScript III reverse transcriptase (Invitrogen). Quantitative RT-PCR was performed using an ABI 7500 Fast RT-PCR instrument and SYBR Premix Ex Taq kit (TaKaRa, Otsu, Shiga, Japan). Gene transcripts were standardized using ACTIN2 (Li et al., 2010) as an internal control for RGA and GAPC as an internal control for GA20ox2, GA3ox1, SCL3 and bHLH137 (Zentella et al., 2007). The primers used in this assay are listed in Table S1.
XopDXcc8004 delays disease symptoms and does not change bacterial growth
XopD-like proteins share the conserved EAR motif and SUMO domain with the N-terminal variant region (Fig. 1a). To assess the role of XopDXcc8004 in bacterial virulence, we generated the xopD deletion mutant strain Xcc8004ΔxopD from wild-type Xcc8004. Susceptible radish plants were inoculated by clipping with a 105 CFU ml−1 inoculum of Xcc8004 or Xcc8004ΔxopD. The result showed no significant difference in bacterial titers in radish leaves infected by Xcc8004 and Xcc8004ΔxopD (Supporting Information Fig. S1). However, Xcc8004ΔxopD caused earlier necrosis symptoms than Xcc8004 at 6 dpi (Fig. 1c), which indicated that XopDXcc8004 may suppress disease symptom development. To further verify the role of XopDXcc8004 in regulating disease symptoms, we complemented Xcc8004ΔxopD using the entire xopDXcc8004 encoding region, which is controlled by the lacZ promoter, to generate the complemented strain Xcc8004ΔxopD(XopDXcc8004). Immunoblot assays for total proteins and supernatant of Xcc were conducted using anti-His and anti-GroEL antibodies and showed that XopDXcc8004 proteins can be stably synthesized from the complemented strains (Fig. 1b). Bacterial inoculation assays in radish leaves showed that Xcc8004ΔxopD(XopDXcc8004) recovered disease symptoms to the level caused by Xcc8004 (Fig. 1c), which confirmed the importance of XopDXcc8004 in repressing disease symptoms. Meanwhile, we also complemented Xcc8004ΔxopD using the xopDXccB100 encoding region to generate the complemented strain Xcc8004ΔxopD(XopDXccB100) (Fig. 1b). Bacterial inoculation assays showed that no different disease symptoms were present in Xcc8004ΔxopD(XopDXccB100)- and Xcc8004ΔxopD-infected radish leaves (Fig. 1c). This finding indicated that XopDXccB100 might differ from XopDXcc8004 in regulating Xcc8004-infected radish defenses.
We next performed a bacterial inoculation assay by infecting the model Arabidopsis plants with Xcc8004 and Xcc8004ΔxopD and found no difference in bacterial titers and disease symptoms between Xcc8004 and Xcc8004ΔxopD at 3 dpi (Fig. S2). These results suggested that XopDXcc8004 might not be a dominant determinant in Xcc virulence. To investigate whether XopDXcc8004 affects the severity of the disease symptoms induced by the virulent bacterium PstDC3000, PstDC3000 and Xcc were coinoculated on Arabidopsis leaves. The result showed that disease symptoms in Arabidopsis leaves, which are inoculated with a 104 CFU ml−1 inoculum of PstDC3000, were not altered when a 105 CFU ml−1 inoculum of Xcc8004, Xcc8004ΔxopD or Xcc8004ΔxopD(XopDXcc8004) were coinoculated, whereas a 106 CFU ml−1 inoculum of Xcc8004 or Xcc8004ΔxopD(XopDXcc8004) delayed disease symptoms (Fig. 1d). This finding suggested that a certain amount of XopDXcc8004 proteins produced by Xcc might be important in detecting the contribution of XopDXcc8004 to delaying disease symptoms in PstDC3000-infected Arabidopsis leaves.
In order to characterize XopDXcc8004 further in planta, we generated four independent XopDXcc8004 transgenic lines (two lines were used in this study) from the Col-0 ecotype, all of which expressed the FLAG-tagged XopDXcc8004 protein under the control of the estradiol-inducible promoter. Col-0 and transgenic lines were inoculated with a 104 CFU ml−1 inoculum of PstDC3000. We detected estradiol-induced XopDXcc8004 expression in transgenic plants using anti-FLAG antibody and found that estradiol-induced XopDXcc8004 expression did not alter PstDC3000 multiplication in transgenic lines (Figs 1e, S3). However, much fewer chlorosis symptoms were present in XopDXcc8004 transgenic lines compared with wild-type plants at 2 dpi (Figs 1f, S4), indicating a role of XopDXcc8004 in triggering disease tolerance.
XopDXcc8004 interacts with DELLA proteins in the nucleus of plant cells
In order to elucidate the mechanisms that underlie this disease tolerance process, XopDXcc8004-associated proteins were identified from an Arabidopsis cDNA library of mature leaves and roots (obtained from The Arabidopsis Information Resource) by yeast two-hybrid screening using the full-length XopDXcc8004 as bait. Approximately 106 colonies were screened and 30 positive clones were sequenced. Three truncated RGA genes were deduced to encode truncated RGA protein that contained the DELLA and VHYNP motifs (Fig S5). Yeast two-hybrid assays showed that XopDXcc8004 interacted with the full-length DELLA proteins RGA, GAI, RGL1, RGL2 and RGL3 (Figs 2a,b, S6, S7). The truncated protein XopDXcc80041-210 (amino acids 1-210 EAR motif-contained region), but not XopDXcc8004211-442 (amino acids 221-442 SUMO domain-contained region), interacted with RGA (Figs 1a, 2a, S6), indicating that the EAR region is required to interact with DELLA proteins.
In order to confirm interactions between XopDXcc8004 and RGA in vivo, RGAp::GFP-RGA and its XopDXcc8004 transgenic plants were used for anti-FLAG co-IP and anti-GFP immunoblot assays. After 24-h estradiol-induced XopDXcc8004 expression in 10-d-old seedlings, FLAG-tagged XopDXcc8004 proteins were precipitated from total tissue proteins using anti-FLAG beads, and the precipitates were detected using anti-FLAG and anti-GFP antibodies. We found that XopDXcc8004 interacted with RGA in planta (Fig. 2d). BiFC assays were used to further examine colocalization and interaction of XopDXcc8004-RGA in plant cells. The N-terminal half of YFP (nYFP) was fused to the C terminus of full-length XopDXcc8004, and the C-terminal half of YFP (cYFP) was fused to the C terminus of full-length RGA. We found that Arabidopsis protoplasts cotransfected with XopDXcc8004-nYFP and RGA-cYFP showed YFP fluorescence. However, those cotransfected with XopDXcc8004-nYFP and the empty cYFP plasmid or with RGA-cYFP and the empty nYFP plasmid did not display fluorescence complementation (Fig. 2e), suggesting that XopDXcc8004 targets RGA in the plant cell nucleus.
XopDXcc8004 delays the GA-mediated degradation of DELLA proteins
Unexpectedly, we found that estradiol-induced XopDXcc8004 expression caused dwarfism in transgenic seedlings when the in vivo interaction between XopDXcc8004 and RGA was analyzed. Considering that DELLA proteins negatively regulate plant growth and development and GA induces the degradation of DELLA proteins (Sun & Gubler, 2004; Feng et al., 2008), we investigated whether the dwarf phenotype was related to GA. The Arabidopsis seeds were cultivated for 10 d on MS medium containing 1 mM estradiol and/or 10 μM GA3. We found that the dwarfism seedlings upon estradiol-induced XopDXcc8004 expression were partially rescued by GA3 at either configuration or fresh-weight ratio (Fig. 3a,b), which indicated that the possible stabilization of DELLA proteins might be required for the XopDXcc8004-induced growth inhibition, although diverse internal or external factors may restrain plant growth (Achard et al., 2008). To investigate whether XopDXcc8004 affected the GA-mediated degradation of DELLA proteins, we generated one RGAp::GFP-RGA transgenic line of Arabidopsis, in which GFP-RGA fusion protein is driven by native promoter of RGA. The GFP fluorescence of root tips of 7-d-old XopDXcc8004 transgenic seedlings was observed. We found that a 2-h GA3-treatment accelerated the disappearance of the GFP-RGA fused protein in RGAp::GFP-RGA plants, while 24-h estradiol-induced XopDXcc8004 expression slowed the GA3-mediated degradation of GFP-RGA (Fig. 3c). In addition, no difference in GFP-RGA protein amounts was evident between RGAp::GFP-RGA and XopDXcc8004 transgenic seedlings in the absence of estradiol treatment (Fig. 3d), suggesting that estradiol-induced XopDXcc8004 can delay GA-mediated DELLA degradation.
Exogenous GA reduces disease symptoms promoted by XopDXcc8004
Considering that DELLA proteins are involved in plant tolerance to biotic and abiotic stresses (Achard et al., 2008), we reasoned that disease tolerance triggered by XopDXcc8004 might depend on DELLA stabilization. As exogenous GA treatment eliminates the majority of DELLA proteins, which therefore mimics the della mutant phenotype (Feng et al., 2008), we analyzed the effects of GA treatment on disease symptoms in Xcc8004-infected radish leaves. Radish leaves that were treated by spraying GA3 and incubated for 6 h, and then inoculated with a 105 CFU ml−1 inoculum of Xcc8004. These leaves exhibited severe necrosis symptoms at 6 dpi, in a similar manner to those caused by Xcc8004ΔxopD (Figs 1c, S4). Although XopDXcc8004 suppressed leaf disease symptoms and did not change bacterial growth in PstDC3000-infected Arabidopsis plants (Figs 1f, S3, S4), GA3 treatment reduced these phenotypes in XopDXcc8004 transgenic lines (Figs 1f, S4, S8), suggesting that DELLA stabilization might be important for XopDXcc8004-triggered disease tolerance.
XopDXcc8004 affects the GA-induced binding of GID1 to DELLA proteins
Given that DELLA proteins are degraded via the 26S ubiquitin-proteasome system (Feng et al., 2008), we examined whether XopDXcc8004 can remove ubiquitination from RGA. To identify whether RGA ubiquitination occurs independently from the binding of other ubiquitinated proteins in N. benthamiana leaves, we generated mutant RGAK33A,K65A by changing the lysine (K), the deduced ubiquitination site (Peng et al., 2003), to alanine (A) at residues 33 and 65. Using an A. tumefaciens-mediated transient expression system, we expressed FLAG-tagged RGA, FLAG-tagged RGAK33A,K65A and GFP-tagged XopDXcc8004 in 4-wk-old N. benthamiana leaves. After protein transient expression for 48 h, the anti-FLAG co-IP and anti-ubiquitin immunoblot assays were performed. We found that RGAK33A,K65A could not be ubiquitinated (Fig. 4a), which suggested that the detected ubiquitination of RGA was mainly RGA self-ubiquitination. Meanwhile, XopDXcc8004 did not reduce RGA ubiquitination levels when XopDXcc8004 and RGA were coexpressed in N. benthamiana leaves (Fig. 4a). We then generated FLAG-tagged XopDXcc8004 and its truncated proteins XopDXcc80041-210 and XopDXcc8004211-442 (Fig. 1a), and examined the effect of these proteins on DELLA protein amounts in Arabidopsis protoplasts. The result showed that XopDXcc8004 and XopDXcc80041-210, but not XopDXcc8004211-442, had the ability to delay the GA3-mediated degradation of RGA protein (Fig. 4b). The EAR region of XopDXcc8004 appears to be important in delaying GA-mediated DELLA degradation.
As DELLA and VHYNP motifs are important for the GA-induced binding of GID1s to DELLA proteins and for subsequent DELLA proteolysis (Sun, 2010), we hypothesized that XopDXcc8004 might interfere with the binding of GID1 to DELLA proteins. HA-tagged XopDXcc8004, XopDXcc80041-210 and XopDXcc8004211-442 were each cotransformed with FLAG-tagged GID1a into Arabidopsis protoplasts. In the absence of GA3 treatment, XopDXcc8004, XopDXcc80041-210 and XopDXcc8004211-442 did not change GID1a protein amounts (Fig. 4c). However, XopDXcc8004 and XopDXcc80041-210, but not XopDXcc8004211-442, interfered with the GA3-induced binding of GID1a to RGA when cotransforming FLAG-tagged GID1a and HA-tagged RGA with HA-tagged XopDXcc8004, XopDXcc80041-210 or XopDXcc8004211-442 into Arabidopsis protoplasts (Fig. 4d). In addition, as there is no difference in the amounts of RGA-HA in protoplasts cotransforming FLAG-tagged GID1a and HA-tagged XopDXcc8004, XopDXcc80041-210 or XopDXcc8004211-442 (Fig. 4d), we speculated that the presence of GID1a might interfere with the RGA degradation delayed by XopDXcc8004. Meanwhile, XopDXcc8004, XopDXcc80041-210 and XopDXcc8004211-442 did not interact with GID1a and did not change GID1a protein levels (Fig. 4d). Furthermore, XopDXcc8004 and XopDXcc80041-210 interacted with full-length RGA but not with the DELLA domain truncated protein RGAΔDELLA in yeast cells (Figs 2c, S9). These results suggest that the EAR region that delays GA-mediated DELLA degradation might interfere with the binding of GID1 to DELLA proteins.
XopDXcc8004 does not appear to alter transcription of RGA and GA-responsive genes
We next determined whether XopDXcc8004 affects RGA protein levels by altering RGA gene expression. Estradiol-induced XopDXcc8004 expression was detected and confirmed by immunoblot assays using anti-FLAG antibody (Fig. 5a). After estradiol treatment for 24 h, we extracted total RNA from Arabidopsis plants and performed quantitative RT-PCR. We found that there was no difference in RGA transcripts between wild-type and transgenic plants (Fig. 5b), which suggested that XopDXcc8004 may not affect RGA gene transcripts. Despite a slight upregulation of RGA (At2g01570) gene transcript at 24 h after estradiol treatment (Fig. 5b), RGA protein levels were not changed in XopDXcc8004 transgenic plants (Fig. S4). To analyze whether XopDXcc8004 affects GA signaling by regulating GA-responsive gene transcripts, we investigated the transcriptional profiles of GA20ox2 (At5g51810), GA3ox1 (At1g15550), SCL3 (At1g50420) and bHLH137 (At5g50915), which are all genes that are downregulated by GA and upregulated by DELLA proteins (Zentella et al., 2007). We found that GA3 treatment reduced GA20ox2, GA3ox1, SCL3 and bHLH137 gene transcripts, and both GA3 and estradiol treatment barely altered the levels of these transcripts in either wild-type or transgenic plants. In addition, the estradiol-induced XopDXcc8004 expression increased GA20ox2, GA3ox1, SCL3 and bHLH137 gene transcripts in transgenic lines, whereas additional GA3 treatment recovered gene transcript levels to those of wild-type plants (Fig. 5c). Therefore, XopDXcc8004 does not appear to alter transcription of GA-responsive genes.
XopDXcc8004 represses PAMP-induced ROS production
In order to further investigate the possible role of XopDXcc8004 in regulating plant defenses, we tested a molecular event of the PTI signaling pathway, the production of bacterial flagellar peptide flg22-triggered reactive oxygen species (ROS) (Segonzac & Zipfel, 2011), in XopDXcc8004 transgenic plants. We found that estradiol-induced XopDXcc8004 expression repressed H2O2 production in transgenic lines, whereas plants that were pretreated with sprayed GA3 suppressed levels of H2O2 production (Fig. S10). It appears that XopDXcc8004 is a potential suppressor of PTI and DELLA proteins might be important for the suppression of H2O2 production caused by XopDXcc8004.
Defense against pathogens by plant hosts can be divided into resistance and tolerance (Best et al., 2008). Disease resistance mechanisms have been well illustrated in the study of plant–pathogen interactions (Chisholm et al., 2006; Jones & Dangl, 2006). Despite the fact that disease tolerance as a defense strategy has been the focus of extensive research in the field of plant pathology for decades (Restif & Koella, 2003), very little is known about the molecular basis of plant disease tolerance in pathogen infections. Our investigation of the phytopathogenic bacterium Xanthomonas effector protein XopDXcc8004 in regulating plant defenses uncovered a possible disease tolerance mechanism that is manipulated by bacterial pathogens. In this study, we found that XopDXcc8004 has the ability to delay the development of disease symptoms in bacteria-infected plant leaves, whereas exogenous GA treatment makes these disease symptoms more severe. Furthermore, XopDXcc8004 interacts with DELLA proteins in the nucleus of the plant cell and delays, but not blocks, the GA-mediated degradation of DELLA proteins. These data suggest that XopDXcc8004 might promote plant disease tolerance by targeting and partially stabilizing DELLA proteins.
The Xanthomonas effector-mediated suppression of disease symptoms has been previously reported. In those studies, XopDXcv85-10 and XopJXcv85-10 were shown to delay disease symptoms in Xcv-infected tomato/pepper plants (Kim et al., 2008; Ustun et al., 2013), and XopDXccB100 was shown to delay disease symptoms in Xcc-infected Arabidopsis plants (Canonne et al., 2011). In addition, the P. syringae effector protein, HopN1, was also shown to suppress the development of disease symptoms in tomato leaves (López-Solanilla et al., 2004). Considering the characterization of XopDXcc8004 in delaying disease symptoms, we hypothesized that effector-triggered disease tolerance might be an important strategy for some phytopathogenic bacteria to regulate plant defenses in order to gain survival advantages. Although XopD-like proteins share the conserved EAR motif and SUMO domain (Kim et al., 2011), the N-terminal variant region that contains a DNA-binding HLH domain is absent in XopDXcc8004 compared with XopDXcv85-10 and XopDXccB100. Recent reports have shown that the HLH domain is required for XopDXcv85-10 and XopDXccB100 to target MYB30 and to repress MYB30-mediated resistance (Canonne et al., 2011). Upon the overexpression of MYB30 in Arabidopsis leaves, XopDXcc8004 cannot recover the mutant strain XccB100ΔxopD-induced HR to levels of XccB100-induced chlorosis symptoms (Canonne et al., 2011). Similarly, our results showed that XopDXccB100 failed to recover mutant strain Xcc8004ΔxopD-induced necrosis to levels of Xcc8004-induced chlorosis symptoms in radish leaves. Additionally, XopDXcv85-10 and XopDXccB100 are localized in subnuclear structures as nuclear foci or nuclear bodies in plant cells (Canonne et al., 2011), whereas XopDXcc8004 is localized in the nucleus. Under the assumption that all XopD variants are properly translocated by the respective bacterial strain into the host cell, these data imply that the mechanisms of disease tolerance regulated by XopDXcv85-10 and XopDXccB100 might be different from XopDXcc8004. However, it remains unclear whether XopDXcv85-10 and XopDXccB100 can interact with DELLA proteins and delay GA-mediated DELLA degradation. Nevertheless, previous studies have shown that XopDXcv85-10-mediated inhibition of disease symptoms is related to the suppression of salicylic acid production (Kim et al., 2008) and XopDXcv85-10 directly targets the tomato ethylene responsive transcription factor, SlERF4, to reduce ethylene production and, in turn, limit disease symptom development (Kim et al., 2013). Considering that DELLA proteins are signaling repressors of the hormone GA (Sun & Gubler, 2004), it is possible that components of plant hormone signaling pathways are important targets for XopD effectors to regulate disease tolerance.
DELLA proteins have previously been shown to promote plant tolerance responses to adverse biotic and abiotic stresses (Achard et al., 2008). DELLA-stabilization mutant plants increase the growth of PstDC3000 and the quadruple-DELLA mutant plant (deficient in GAI, RGA, RGL1 and RGL2) reduces bacterial titers (Navarro et al., 2008). In particular, the quadruple-DELLA leaves infected with PstDC3000 show microHR, while no microHR symptoms are observed on wild-type infected leaves (Navarro et al., 2008), demonstrating a crucial role of DELLA proteins as positive regulators in regulating plant disease tolerance. Our results showed the following: XopDXcc8004 delays disease symptoms; XopDXcc8004 targets DELLA proteins; exogenous GA enhances disease symptoms delayed by XopDXcc8004; and XopDXcc8004 delays the GA-mediated degradation of DELLA proteins. Taken together, these results suggest that XopDXcc8004 might promote disease tolerance by positively regulating host disease tolerance positive regulators.
Interestingly, plant proteins that negatively regulate disease tolerance have also been reported. For example, the Arabidopsis ethylene-insensitive ein2 mutant represses disease symptoms and promotes P. syringae growth (Bent et al., 1992). The leucine zipper transcription factor TGA2 mutant, which can no longer bind DNA, enhances rice disease tolerance to X. oryzae (Fitzgerald et al., 2005). However, whether these disease-tolerant negative regulators are targeted and regulated by bacterial effectors to trigger disease tolerance remains unknown.
The finding that XopDXcc8004 delays GA-induced DELLA degradation is important to help us understand the possible mechanisms that underlie XopDXcc8004-triggered disease tolerance. A current model of the GA-induced DELLA degradation is that GA is recognized by the GID1 receptor. GID1-GA then binds DELLA proteins to form the GA-GID1-DELLA complex, which triggers the ubiquitination of DELLA proteins and their degradation by the 26S proteasome (Sun, 2010). It has been shown that the Yersinia virulence factor, YopJ, a cysteine protease, cleaves a reversible post-translational modification in the form of ubiquitin or a ubiquitin-like protein (Orth, 2002). Interestingly, the XopD-like protein XopDXcv85-10 acts as a cysteine protease with plant-specific SUMO substrate specificity (Hotson et al., 2003) and XopDXcc8004 contains a putative cysteine protease SUMO domain. Despite YopJ belonging to the cysteine peptidase (C55) family (see MEROPS Protease Database, http://merops.sanger.ac.uk/) while XopD belongs to the C48 peptidase family (Kim et al., 2011), we still hypothesize that XopDXcc8004 might have the ability to deubiquitinate DELLA proteins. However, A. tumefaciens-mediated transient protein expression assays showed that XopDXcc8004 cannot deubiquitinate the ubiquitination of RGA. Therefore, we paid close attention to the EAR motif-containing region. The EAR motif is one of the most predominant forms of transcriptional repression motif in plant transcriptional regulators and is emerging as one of the principal mechanisms of gene regulation (Kagale & Rozwadowski, 2011). We speculated that XopDXcc8004, as a nuclear-localized and EAR-contained protein, might regulate DELLA gene transcripts. GA-responsive gene transcripts, which are feedback regulated by DELLA proteins (Zentella et al., 2007), might also be regulated by XopDXcc8004. However, no obvious change of RGA gene transcript was observed in estradiol-induced XopDXcc8004 transgenic plants. Although transcription of GA-responsive genes was upregulated by estradiol-induced XopDXcc8004, transcription of these genes can be downregulated by application of exogenous GA. These data indicate that XopDXcc8004 delays GA-induced DELLA degradation, which does not appear to alter transcription of GA-responsive genes.
Indeed, growing amounts of evidence support a role for the EAR motif in protein–protein interactions. For example, the EAR-containing proteins ERF3 and ERF4 interact with the transcriptional corepressor SAP18, which, in turn, binds to histone deacetylase HDA19 to remodel chromatin (Ohta et al., 2001). In addition, the stress-responsive mitogen-activated protein kinases, MPK3 and MPK6, were shown to interact with the EAR motif of a poplar zinc finger protein and mediate its degradation (Hamel et al., 2011). Our data showed that the N-terminal EAR motif region of XopDXcc8004 interacts with the DELLA domain of DELLA protein RGA, in which the DELLA and VHYNP motifs are essential, and interferes with the GA-induced binding of GID1a to RGA. These results suggest that XopDXcc8004 delays GA-induced DELLA degradation mainly by affecting the levels of DELLA proteins. As a result, the DELLA and VHYNP motifs are required for the binding of GID1 to DELLA and the induction of DELLA degradation (Dill et al., 2001; Liu et al., 2010). We propose that the EAR motif region might delay GA-induced degradation of DELLA proteins by interfering with the binding of GID1 to the DELLA domain of DELLA proteins. These data explain a possible mechanism for XopDXcc8004-triggered disease tolerance, by which Xcc8004 grows normally and host plant fitness is not affected.
The bacterial component that delays the GA-induced degradation of DELLA proteins has been revealed by investigating the eliciting bacterial flagellin peptide flg22 (Navarro et al., 2008). However, flg22 as a bacterial PAMP induces plant PTI resistance when it is recognized by the surface receptor FLS2 (Zipfel et al., 2004). Considering that XopDXcc8004 delays GA-induced DELLA degradation and promotes disease tolerance, DELLA proteins may be important plant defense-associated targets whose functions are mediated by bacterial components. Additionally, a previous report has shown that DELLA proteins regulate plant immune responses by modulating the balance of jasmonic acid (JA) and salicylic acid (SA) signaling (Navarro et al., 2008). Although we do not know whether XopDXcc8004 regulates JA- or SA-mediated plant defenses, XopDXcc8004 represses the flg22-triggered ROS production and exogenous GA reduces the suppression of ROS production restrained by XopDXcc8004, indicating a suppressor role for XopDXcc8004 in PTI. These data suggest that XopDXcc8004 might regulate disease resistance in addition to disease tolerance, which is consistent with empirical studies regarding resistance and tolerance defense mechanisms that are likely to be present within a single population (Best et al., 2008).
Nevertheless, this study still has several limitations. For example, the estradiol treatment might lead to major XopDXcc8004 overexpression artifacts. Coinfiltration of PstDC3000 with Xcc8004 may avoid the overexpression artifacts, whereas Xcc8004 infects plants via the vascular rather than the mesophyll tissues (Feng et al., 2012). Furthermore, Arabidopsis plants treated by exogenous GA cannot substitute, the DELLA mutants (deficient in RGA, GAI, RGA, RGL1 and RGL2) and GA might affect other signaling pathways related to defense responses, due to hormone cross-talk in plant disease and defense (Robert-Seilaniantz et al., 2011). Future work that aims at addressing the disease tolerance mechanisms regulated by bacterial effectors in model plants should be based on effectors that are secreted by the type III secretion system.
In conclusion, we have shown that XopDXcc8004 acts as a disease tolerance-promoting factor, resulting in the suppression of plant disease symptom development, which might function by targeting DELLA proteins. We hypothesize that this plant disease tolerance mechanism might be utilized by bacterial pathogens to sustain their maximum niches.
We thank Dr Jian-min Zhou for comments and Dr Xiangdong Fu for providing the RGAp::GFP-RGA line. This work is supported by the Ministry of Science and Technology of China (grant no. 2012CB722206) and the Department of Science and Technology of Hainan Province (grant no. ZDZX2013023).