Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
•The pathogenicity of the Gram-negative plant-pathogenic bacterium Xanthomonas campestris pv. vesicatoria (Xcv) is dependent on type III effectors (T3Es) that are injected into plant cells by a type III secretion system and interfere with cellular processes to the benefit of the pathogen.
•In this study, we analyzed eight T3Es from Xcv strain 85-10, six of which were newly identified effectors. Genetic studies and protoplast expression assays revealed that XopB and XopS contribute to disease symptoms and bacterial growth, and suppress pathogen-associated molecular pattern (PAMP)-triggered plant defense gene expression.
•In addition, XopB inhibits cell death reactions induced by different T3Es, thus suppressing defense responses related to both PAMP-triggered immunity (PTI) and effector-triggered immunity (ETI).
•XopB localizes to the Golgi apparatus and cytoplasm of the plant cell and interferes with eukaryotic vesicle trafficking. Interestingly, a XopB point mutant derivative was defective in the suppression of ETI-related responses, but still interfered with vesicle trafficking and was only slightly affected with regard to the suppression of defense gene induction. This suggests that XopB-mediated suppression of PTI and ETI is dependent on different mechanisms that can be functionally separated.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Plants defend themselves against microbial invaders by basal defense responses including the production of reactive oxygen species, activation of mitogen-activated protein kinase (MAPK) cascades, expression of pathogenesis-related (PR) genes and callose deposition into the plant cell wall (Nürnberger et al., 2004). Usually, these defense reactions are activated on recognition of pathogen-associated molecular patterns (PAMPs), such as flagellin, lipopolysaccharides and elongation factor EF-Tu, by specific receptors in the plant plasma membrane (Nürnberger et al., 2004; Jones & Dangl, 2006). However, specialized bacterial pathogens have evolved sophisticated strategies to avoid or manipulate plant defense responses and to proliferate in the plant’s apoplast. Essential for the pathogenicity of Gram-negative bacteria is often a type III secretion (T3S) system consisting of a complex membrane-spanning injection apparatus that is associated with an extracellular pilus and a channel-like translocon in the plant plasma membrane (Büttner & He, 2009). T3S systems translocate type III effectors (T3Es) directly into the host cell cytosol where they interfere with plant cell processes to the benefit of the pathogen, often leading to the suppression of PAMP-triggered immunity (PTI) (Jones & Dangl, 2006). However, individual effectors can also act as avirulence (Avr) proteins that are recognized in plants carrying a corresponding resistance (R) gene. Recognition leads to the elicitation of host defense reactions that often culminate in the hypersensitive response (HR), a rapid, localized programmed cell death reaction that restricts pathogen ingress (Klement & Goodman, 1967; Greenberg & Yao, 2004). To circumvent effector-triggered immunity (ETI), bacterial pathogens have evolved T3Es that interfere with the induction of R gene-mediated defense responses (Jones & Dangl, 2006).
Most phytopathogenic bacteria translocate 20–30 different T3Es into the plant cell (Büttner & Bonas, 2010; Hann et al., 2010). Notably, the deletion of individual effector genes often does not lead to reduced virulence, presumably because of functional redundancies among T3Es (Büttner & Bonas, 2010; Hann et al., 2010). Although the molecular functions of most T3Es inside the plant cell are still unknown, a number of T3Es from phytopathogenic bacteria have been shown to interfere with signaling cascades, proteasome-dependent protein degradation and the transcription machinery (Kay & Bonas, 2009; Block & Alfano, 2011).
Our laboratory studies the T3S system and T3Es from Xanthomonas campestris pv. vesicatoria (Xcv, also termed X. euvesicatoria (Jones et al., 2004) and X. axonopodis pv. vesicatoria (Vauterin et al., 2000)), the causal agent of bacterial spot disease on pepper and tomato. The T3S system of Xcv is encoded by the 23-kb chromosomal hrp (hypersensitive response and pathogenicity) gene cluster, which is essential for bacterial growth and disease symptoms on susceptible plants, as well as for HR induction in resistant host and nonhost plants (Bonas et al., 1991). The expression of hrp genes is induced in planta by the OmpR-family regulator HrpG, which controls the expression of a genome-wide regulon (Noël et al., 2001) including hrpX, which encodes an AraC-type transcriptional activator (Wengelnik & Bonas, 1996; Wengelnik et al., 1996a). HrpX binds to cis-regulatory PIP (plant-inducible promoter) boxes in the promoter regions of hrp and other genes that contribute to virulence (Koebnik et al., 2006).
In addition to a functional T3S apparatus, efficient translocation of effectors by the Xcv T3S system requires the T3S chaperone HpaB, which has a broad substrate specificity (Büttner et al., 2004, 2006; Szczesny et al., 2010a). T3S chaperones specifically bind T3S substrates and promote their secretion and/or stability (Parsot et al., 2003; Wilharm et al., 2007). Interestingly, T3Es in Xcv differ in their HpaB dependence and are therefore grouped into two classes. While class A effectors depend on HpaB for translocation, class B effectors are translocated even in the absence of HpaB, albeit in reduced amounts. It is conceivable that HpaB imposes a hierarchy on T3E translocation and that class A effectors are preferentially translocated during a certain stage of the infection process (Büttner et al., 2006).
On the basis of experimental and bioinformatic analyses, individual Xanthomonas strains express 23–37 different T3Es (Büttner & Bonas, 2010). The largest effector class, although not present in all strains, is the AvrBs3/PthA family of transcription activator-like (TAL) effectors, which mimic eukaryotic transcription factors and induce the transcription of plant genes to support bacterial growth and dispersal (Marois et al., 2002; Yang et al., 2006; Kay et al., 2007; Boch et al., 2009; Antony et al., 2010). Intriguingly, they can also activate R gene promoters leading to the induction of the HR (Gu et al., 2005; Römer et al., 2007). Plant transcript levels are also modulated by XopD (Xop, Xanthomonas outer protein) from Xcv, which negatively regulates the expression of defense- and senescence-related genes via ethylene response factor amphiphilic repression motifs (Kim et al., 2008). In addition, XopD acts as a cysteine protease (Hotson et al., 2003). All biochemical activities of XopD contribute to its virulence function, that is, a delay of plant chlorosis and necrosis and the promotion of bacterial multiplication in tomato (Kim et al., 2008). Cysteine protease activity has also been demonstrated for effectors of the YopJ/AvrRxv family from plant- and animal-pathogenic bacteria (Orth et al., 2000; Hotson & Mudgett, 2004; Ma et al., 2006; Sweet et al., 2007; Szczesny et al., 2010a). Their exact mode of action, however, is a controversial issue because an acetyltransferase activity has also been described (Mukherjee et al., 2006). Members of the YopJ/AvrRxv family, for example XopJ and AvrBsT from Xcv, are involved in the suppression of plant defense by the inhibition of cell wall-associated defense responses and ETI, respectively (Bartetzko et al., 2009; Szczesny et al., 2010a). A role in the suppression of plant immunity has also been proposed for other Xcv T3Es, including XopX and XopN (Metz et al., 2005; Kim et al., 2009). The latter presumably suppresses PTI by targeting an atypical receptor-like kinase in tomato involved in immune signaling (Kim et al., 2009).
Escherichia coli cells were cultivated in LB (lysogeny broth) medium at 37°C. Agrobacterium tumefaciens was grown at 30°C in YEB (yeast extract broth) medium and Xcv at 30°C in NYG (nutrient yeast glycerol; Daniels et al., 1984), hrp gene-inducing medium (XVM2; Wengelnik et al., 1996b) or secretion medium (minimal medium A; Ausubel et al., 1996) supplemented with 10 mM sucrose and 0.3% casamino acids. Plasmids were introduced into E. coli and A. tumefaciens by electroporation, and into Xcv by conjugation, using pRK2013 as helper plasmid in triparental matings (Figurski & Helinski, 1979). Bacterial strains and plasmids are listed in Supporting Information Table S1.
Plant material and inoculations
The near-isogenic pepper (Capsicum annuum) cultivars ECW, ECW-10R, ECW-20R and ECW-30R (Minsavage et al., 1990), Nicotiana benthamiana and N. tabacum plants were grown at 25°C with 65% relative humidity and 16 h light. Xcv strains were hand-inoculated with a needleless syringe into the apoplast of leaves at 108 colony-forming units (cfu) ml−1 in 10 mM MgCl2. For in planta growth curves, bacteria were inoculated at 104 cfu ml−1, and bacterial growth was determined as described by Bonas et al. (1991). For in planta transient expression studies, A. tumefaciens strain GV3101 was grown overnight in YEB medium, resuspended in inoculation medium (10 mM MgCl2, 5 mM MES, pH 5.3, 150 μM acetosyringone) and inoculated into leaves at 4 × 108 cfu ml−1 for localization studies and 2 × 108 cfu ml−1 for secGFP secretion assays. For the analysis of cell death suppression, we used final bacterial densities of 4 × 108 cfu ml−1 (Agrobacterium delivering avrBs1, avrBs2, avrBsT, avrRxv, xopG and xopJ) and 6 × 108 cfu ml−1 (Agrobacterium delivering xopS, xopB and mutant derivatives), respectively. For fractionation studies, Agrobacterium strains were inoculated at 6 × 108 cfu ml−1. For co-expression, Agrobacterium strains were mixed before inoculation.
RNA extraction from Xanthomonas, cDNA synthesis and semiquantitative reverse transcription polymerase chain reaction (RT-PCR) experiments were performed as described by Noël et al. (2001) and Thieme et al. (2007). For oligonucleotide sequences, see Table S2. Experiments were performed at least three times for each gene with two independent cDNA preparations each.
Xanthomonas protein secretion experiments were performed as described by Rossier et al. (1999) and Büttner et al. (2002). Equal amounts of total bacterial cell extracts and culture supernatants were analyzed by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting following standard protocols. To check for bacterial lysis, blots were routinely reacted with an antibody specific for the intracellular protein HrcJ (data not shown; Rossier et al., 2000). For Agrobacterium-mediated expression studies, two 0.64-cm2 leaf disks per strain were frozen in liquid nitrogen, ground and resuspended in 100 μl of 8 M urea and 50 μl of 5 × Laemmli buffer, and boiled for 10 min. Proteins were separated by SDS-PAGE and analyzed by immunoblotting. We used specific polyclonal antibodies for the detection of AvrBs3 (Knoop et al., 1991), GFP (green fluorescent protein; Invitrogen GmbH, Karlsruhe, Germany), RFP (red fluorescent protein; Antibodies-online GmbH, Aachen, Germany), cFBPase (Agrisera AB, Vaennaes, Sweden) and XopB (H. Berndt & U. Bonas, unpublished), and a monoclonal c-Myc-specific antibody (Roche Diagnostics, Mannheim, Germany). Horseradish peroxidase-labeled α-rabbit and α-mouse antibodies (Amersham Pharmacia Biotech, Piscataway, NJ, USA) were used as secondary antibodies. Antibody reactions were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech).
Golden Gate vectors
The suicide vector pOGG2 was derived from the suicide plasmid pOK1 (Huguet et al., 1998) and contains a lacZ gene for blue–white selection. The binary vector pGGA1 contains the backbone of pBGWFS7 (Karimi et al., 2005), a chloramphenicol resistance-ccdB selection cassette, and allows the expression of genes 3′ translationally fused to GFP under the control of the 35S promoter. To allow cloning of DNA fragments by BsaI/T4-ligase cut-ligation (Engler et al., 2008), additional BsaI restriction sites were removed during vector construction. Cloning details are available on request.
GATEWAY and Golden Gate expression constructs
To generate binary expression constructs, the coding sequences of avrRxv, xopB, xopBA313V, xopBK455R, xopG, xopI, xopK, xopM, xopR, xopS and xopV were amplified by PCR, cloned into pENTR/D-TOPO (Invitrogen) and recombined into pGWB5, pGWB17, pK7FWG2 and pK7WGF2 (Karimi et al., 2005; Nakagawa et al., 2007) using Gateway® technology (Invitrogen). Oligonucleotides are listed in Table S2.
For expression in Xcv, xopS was amplified from strain 85-10 and cloned into the Golden Gate-compatible expression vector pBRM (Szczesny et al., 2010b). xopB and xopBA313V PCR amplicons were cloned downstream of plac into the Golden Gate-compatible expression vector pLAND, which allows the insertion of the cloned fragment into the genome by homologous recombination (C. Lorenz & D. Büttner, unpublished).
To generate avrBs3Δ2 fusions, the promoters and 5′ coding sequences of xopG, xopI, xopK, xopM, xopR, xopS and xopV were amplified by PCR from genomic DNA of Xcv 85-10, cloned into pENTR/D-TOPO and recombined into pL6GW356 (Noël et al., 2003). In the case of xopM, the complete coding region was amplified. For xopB, the 5′ coding sequence without promoter was amplified and cloned downstream of plac into pBR356, which allows a 3′ fusion of the gene to avrBs3Δ2 encoding an N-terminally truncated AvrBs3 derivative with a C-terminal FLAG epitope (C. Lorenz & D. Büttner, unpublished).
secGFP, which contains a basic chitinase signal sequence at the N-terminus of GFP (Haseloff et al., 1997), was generated by cloning annealed oligonucleotides into pGGA1. xopJ was recombined into pGWB17 (Nakagawa et al., 2007) using an available pENTR/D-TOPO construct (Thieme et al., 2007). xopJC235A was derived from xopJ by splicing by overlap extension (SOE)-PCR, cloned into pENTR/D-TOPO and recombined into pGWB17 (Nakagawa et al., 2007). AtSYP121_Sp2 (Tyrrell et al., 2007) was amplified from Arabidopsis thaliana (Col-0) cDNA, cloned into pENTR/D-TOPO and recombined into pGWB17 (Nakagawa et al., 2007). Oligonucleotides are listed in Table S2.
Effector deletion strains
To generate deletions of xopG, xopI, xopM, xopR, xopS and xopV, 0.6–1-kb fragments upstream and downstream of the respective gene were amplified from genomic DNA of Xcv 85-10 by PCR using oligonucleotides harboring appropriate restriction sites (Table S2). For the deletion of xopK, corresponding fragments were synthesized by Eurofins MWG Operon (Ebersberg, Germany). Fragments were cloned into the suicide vectors pK18mobsac (Schäfer et al., 1994) (xopG, xopI, xopM), pOK1 (Huguet et al., 1998) (xopS) or pOGG2 (xopK, xopR, xopV). The resulting constructs were conjugated into Xcv strain 85-10, and mutants were selected by PCR. The xopB deletion mutant was available (Noël et al., 2001).
Mesophyll protoplast transient expression assay
Transient expression experiments with A. thaliana (Col-0)-derived protoplasts were performed according to Ranf et al. (2011). Protoplast samples were co-transformed with the NHL10 promoter-luciferase construct (Boudsocq et al., 2010; Ranf et al., 2011), pUBQ10-GUS (Norris et al., 1993) and either p35S-effector gene constructs (xopB, xopBA313V, xopS) or p35S-H2B (At5g59910) as a control (10 μg total DNA per 100 μl protoplasts; ratio 1 : 1 : 1).
Electrolyte leakage measurements
Triplicates of five leaf disks each (0.64 cm2) were harvested at 1 and 2 d post-inoculation (dpi), respectively. Measurements were carried out as described by Szczesny et al. (2010a).
Lower epidermal cells of N. benthamiana were inspected with a confocal laser scanning microscope LSM 510 and LSM Image Browser software (Carl Zeiss GmbH, Göttingen, Germany) according to the manufacturer’s protocol. mCherry Golgi (G-rk, CD3-967) was used for co-localization experiments (Nelson et al., 2007). To visualize plant cell nuclei, leaves were infiltrated with 0.1% (w/v) 4’,6-diamidino-2-phenylindole (DAPI) solution 1 h before inspection. GFP was excited with an argon laser at 488 nm, mCherry with an HeNe laser at 543 nm and DAPI with a krypton (UV) laser at 364 nm. The emission filter wavelengths were 505–530 nm for GFP, 560–615 nm for mCherry and 385–470 nm for DAPI.
Transmission electron microscopy was performed as described by Thieme et al. (2007) using an EM Libra 120 (Carl Zeiss GmbH).
Gene regulation of xopB, xopG and six new T3E gene candidates in Xcv strain 85-10
For the analysis of effector proteins from Xcv strain 85-10, we chose XopB, previously shown to be type III secreted into the medium (Noël et al., 2001), and XopG, a recently identified T3E (Potnis et al., 2011) with homology to the HopH1 family from Pseudomonas syringae (Thieme et al., 2005). In addition, we identified six new candidate effectors in Xcv strain 85-10 as a result of homology to known effectors (XopK, XopR and XopV), predicted eukaryotic motifs (XopI), indication of gene acquisition by horizontal gene transfer because of significantly lower G + C content (xopS) and the presence of a PIP box in the respective promoters (xopI, xopM, xopR and xopS) (Table 1). Because the presence of a PIP box suggests the co-regulation of a gene with the T3S system, we performed RT-PCR analyses of xopG and the candidate effector genes in the Xcv wild-type strain 85-10, its derivative 85*, which expresses a constitutively active HrpG point mutant resulting in constitutive expression of the T3S system (Wengelnik et al., 1999), and the hrpX deletion mutant 85*ΔhrpX (Noël et al., 2001). When the bacteria were cultivated in complex NYG medium, mRNA of xopG was detectable at similar levels in strains 85-10, 85* and 85*ΔhrpX, suggesting constitutive expression (Fig. 1a). The transcripts of xopB, xopI, xopK, xopM, xopR, xopS and xopV were amplified from strain 85*, suggesting co-expression with T3S genes. As the amounts of amplified transcripts were clearly reduced with RNA preparations from strains 85-10 and 85*ΔhrpX, transcription of the candidate genes is presumably controlled by both HrpG and HrpX. The HrpX-dependent induction of xopR has been described previously (Koebnik et al., 2006).
Table 1. Characteristics of effector genes from Xanthomonas campestris pv. vesicatoria (Xcv) strain 85-10 analyzed in this study
Gene (gene no.)
G + C (%)a
aG + C content of the DNA within the coding region (G + C content of the Xcv 85-10 chromosome: 64.75% (Thieme et al., 2005)).
bPutative function of gene product, eukaryotic motifs and homology to known type III effectors.
cHomologs were determined using BLAST algorithms. −, absence; + presence of a (partial) homolog. X., Xanthomonas spp.; P., Pseudomonas spp.; R., Ralstonia solanacearum; A., Acidovorax spp.; o., other organisms.
dPresence of a PIP and -10 box (TTCGB-N15-TTCGB-N30–32-YANNNT; B represents C, G, or T; Y represents C or T) in the respective promoter (Koebnik et al., 2006). +, presence; −, absence of distinct motifs.
eHrpG- and HrpX-dependent co-regulation with the T3S system (+, co-regulation; −, constitutive expression).
To investigate whether the effector candidates are indeed type III dependently secreted and translocated into the plant cell, we generated translational fusions with the reporter protein AvrBs3Δ2, a derivative of the TAL effector AvrBs3 which lacks a T3S and translocation signal (Szurek et al., 2002; Noël et al., 2003). Fusion of a functional T3S signal to AvrBs3Δ2 enables its translocation and thus the induction of the HR in pepper cultivar ECW-30R plants that harbor the corresponding resistance gene Bs3 (Noël et al., 2003; Thieme et al., 2007). The native promoters and 5′ coding regions of candidate genes (xopG, xopI, xopK, xopR, xopS and xopV) or the complete coding region (xopM) were fused to avrBs3Δ2. In the case of xopB, the 5′ coding region without promoter was used and the fusion construct was expressed from the lac promoter. As controls, we used an empty vector and avrBs3Δ2 alone (Szurek et al., 2002). All plasmids were conjugated into Xcv strain 85* and the T3S mutant 85*ΔhrcV, which lacks an essential inner membrane component of the T3S system (Rossier et al., 2000). When the bacteria were incubated in T3S medium, XopB1–177-, XopG1–100-, XopI1–140-, XopK1–74-, XopM1–520-, XopR1–152-, XopS1–157- and XopV1–148-AvrBs3Δ2 were detected in the culture supernatant of strain 85*, but not of 85*ΔhrcV, by an AvrBs3-specific antibody (Fig. 1b). These results demonstrate that the effector candidates contain functional T3S signals in their N-terminal regions.
To test for type III-dependent translocation, Xcv strains 85* and 85*ΔhrcV expressing avrBs3Δ2 or the corresponding effector fusions, as described above, were inoculated into leaves of AvrBs3-responsive pepper plants (ECW-30R) and the near-isogenic susceptible pepper line ECW, which lacks the Bs3 resistance gene. Derivatives of strain 85* expressing XopB1–177-, XopG1–100-, XopI1–140-, XopK1–74-, XopM1–520-, XopR1–152-, XopS1–157- and XopV1–148-AvrBs3Δ2 induced the HR in ECW-30R (Fig. 1c), but not in ECW (data not shown). As expected, no HR induction was observed in plants infected with derivatives of strain 85*ΔhrcV (Fig. 1c). Taken together, these findings confirm the type III-dependent secretion and translocation of XopB, XopG, XopI, XopK, XopM, XopR, XopS and XopV, and thus their nature as T3Es. In case of XopG, our data confirm a recent publication which showed type III-dependent translocation of the protein using AvrBs2 as reporter (Potnis et al., 2011).
HpaB-dependent translocation of the T3Es
It has been shown previously that the translocation of some T3Es from Xcv is dependent on the general T3S chaperone HpaB (Büttner et al., 2004, 2006). To address this question for the Xcv effectors analyzed here, we introduced the T3E-AvrBs3Δ2 fusion constructs into strain 85*ΔhpaB and inoculated the bacteria into leaves of resistant ECW-30R pepper plants. As shown in Fig. 1(c), XopB1–177-, XopG1–100-, XopI1–140-, XopK1–74-, XopM1–520- and XopV1–148-AvrBs3Δ2 induced the HR even in the absence of HpaB, albeit more or less reduced compared with derivatives of strain 85*. By contrast, XopR1–152- and XopS1–157-AvrBs3Δ2 failed to induce the HR when analyzed in strain 85*ΔhpaB, although both proteins were expressed (Supporting Information, Fig. S1). Thus, according to the published definition (Büttner et al., 2006), XopR and XopS belong to class A, which includes effectors that are not detectably translocated in the absence of HpaB, whereas XopB, XopG, XopI, XopK, XopM and XopV, which are still translocated by the 85*ΔhpaB strain, belong to class B (Table 1).
XopB and XopS contribute to the virulence of Xcv strain 85-10
To study the contribution of the T3Es to bacterial virulence, all effector genes were individually deleted in Xcv strain 85-10, and the mutants were inoculated into leaves of susceptible ECW pepper plants. In addition, induction of the HR in pepper ECW-10R was analyzed, which is based on the recognition of the T3E AvrBs1 by the Bs1 resistance gene (Cook & Stall, 1963; Ronald & Staskawicz, 1988; Escolar et al., 2001). Bacterial strains carrying deletions of xopG, xopI, xopK, xopM, xopR and xopV showed no difference in the induction of disease symptoms and the HR compared with wild-type strain 85-10 (data not shown). By contrast, deletion of xopB or xopS led to significantly reduced disease symptoms, whereas the HR induction was not impaired (Fig. 2a,b and data not shown). The mutant phenotypes of 85-10ΔxopB and 85-10ΔxopS were complemented by ectopic expression of the respective effector gene, suggesting that reduced virulence was not caused by polar effects of the deletions on downstream genes (Fig. 2a,b). Although the growth of both individual effector mutants in ECW plants did not differ significantly from that of the wild-type strain (Fig. S2), multiplication of an 85-10ΔxopBΔxopS double mutant was reduced significantly, suggesting that XopB and XopS fulfill redundant functions (Fig. 2c).
XopB and XopS suppress defense gene expression
To test whether the positive effect of XopB and XopS on disease symptoms and bacterial growth can be explained by the suppression of the plant PTI, we analyzed the influence of the effectors on basal and PAMP-induced defense-related gene expression. Therefore, we performed Arabidopsis leaf protoplast assays, a well-established system for PAMP signaling studies (Boudsocq et al., 2010; Ranf et al., 2011). We tested the influence of XopB, a XopB mutant derivative (XopBA313V, see the following section) and XopS on the activity of the A. thaliana NHL10 (NDR1/HIN1-LIKE 10) (Zipfel et al., 2004) promoter fused to the firefly luciferase gene (LUC) after application of different elicitor-active epitopes of bacterial PAMPs. The luciferase reporter assays showed that the expression of xopB and xopS decreased the pNHL10 basal activity significantly, that is, in the absence of an elicitor (Fig. 3a). In addition, both effectors completely inhibited the activation of pNHL10 by flg22, a bacterial flagellin epitope (Felix et al., 1999), or elf18, a fragment of bacterial EF-Tu (Kunze et al., 2004) (Fig. 3b,c). XopBA313V was only affected slightly in its ability to suppress the elf18-dependent pNHL10 induction (Fig. 3c). The flg22-mediated induction of pNHL10 depends, at least in part, on MAPKs (Boudsocq et al., 2010). Therefore, the activation of the MAPKs MPK3, MPK4, MPK6 and MPK11, which are involved in plant immune signaling (Tena et al., 2011; Bethke et al., 2012), might be affected by XopB and XopS. However, immunoblot analysis using an antibody that specifically detects activated kinases revealed no differences in MAPK activity between protoplasts expressing the respective effector genes and protoplasts expressing CFP (cyan fluorescent protein) as negative control (Fig. S3; Methods S1). The T3E AvrPto from P. syringae served as a positive control (He et al., 2006). Taken together, XopB and XopS suppressed both the basal and PAMP-induced activity of the NHL10 promoter, but they probably act downstream or independent of MAPK activation (Fig. S2).
XopB, XopG, XopM and XopS trigger cell death in different Solanaceae
To identify additional virulence phenotypes, as well as defense reactions, mediated by the analyzed T3Es, we inoculated leaves of pepper ECW, N. benthamiana and N. tabacum, the latter two being nonhost plants of Xcv 85-10, with Agrobacterium strains mediating the in planta expression of the eight effector genes fused to GFP. XopB triggered a cell death reaction in N. benthamiana at 5–6 dpi, but not in N. tabacum (Fig. 4a). These data are in accordance with recent findings (Salomon et al., 2011). We also tested the transient expression of two xopB derivatives with point mutations, accidentally introduced during PCR amplification, for their cell death-inducing activity. Although XopBK455R was still active, XopBA313V did not elicit cell death in N. benthamiana (Fig. 4b). Immunoblot analysis of protein extracts from infected plant material, using a GFP-specific antibody, revealed that XopBA313V protein levels were reduced slightly compared with the wild-type protein. We therefore inoculated a dilution series of Agrobacterium strains. The Agrobacterium strain mediating xopB expression triggered cell death even at low density, corresponding to low XopB protein amounts in the plant tissue, whereas Agrobacterium-mediated synthesis of XopBA313V did not induce any visible cell death reactions (Fig. S4). This suggests that functional loss rather than a reduced protein level is responsible for the lack of cell death induction by XopBA313V.
We also observed a XopG-triggered HR-like cell death in pepper ECW and N. tabacum at 2–3 dpi. Furthermore, XopM elicited a cell death reaction in N. benthamiana at 3–5 dpi, and XopS caused a weak necrosis (compared with the XopG-triggered reaction) in pepper ECW at 3–4 dpi (Fig. 4a). No distinct plant reactions were observed with the other effectors, although they were expressed, as confirmed by immunoblot analyses of protein extracts from N. benthamiana leaves using a GFP-specific antibody (data not shown).
XopB suppresses cell death reactions triggered by XopG and other T3Es
As described above, XopG induces the HR in pepper ECW when transiently expressed in planta, whereas Xcv 85-10, which naturally expresses XopG, does not. The latter might be caused by other Xanthomonas T3Es with cell death suppressing (CDS) activity. Therefore, we tested whether XopB or XopS, which contribute to bacterial virulence (see earlier in the Results section), suppress the XopG-elicited cell death reaction. Co-expression experiments using Agrobacterium-mediated gene delivery revealed that the XopG-dependent HR in pepper and N. tabacum was strongly reduced or fully abolished in the presence of XopB, but not XopS (Fig. 5a). To validate this finding, we used electrolyte leakage assays, a quantitative measure of early plant cell death (Stall et al., 1974), and found that XopB completely suppressed the XopG HR in N. tabacum at early time points (Fig. 5b). Interestingly, no HR suppression was observed with XopBA313V, whereas XopBK455R showed wild-type XopB activity (Fig. 5a,b), although the in planta expression of both effectors could be detected (Fig. 5c). Hence, the CDS activity in N. tabacum and pepper (Fig. 5a,b) and necrosis induction in N. benthamiana (Fig. 4b) seem to be functionally linked.
To explore whether the CDS activity of XopB is restricted to XopG-mediated cell death, we transiently co-expressed xopB with avrBs1, avrBs2, avrBsT, avrRxv and xopJ, and tested for cell death induction in corresponding resistant plants. XopB suppressed the cell death reactions elicited by AvrBsT, AvrRxv and XopJ in N. benthamiana (Fig. 6a). The AvrBs1- and AvrBs2-dependent HRs in pepper ECW-10R and ECW-20R, respectively, were not affected (data not shown). Similar to the effect on XopG-triggered cell death, XopBA313V exhibited no suppression activity, whereas XopBK455R did. Expression of the effector genes was confirmed by immunoblot (Fig. 6b).
To exclude the possibility that the XopB CDS activity is based on unintended cellular changes induced by Agrobacterium-mediated overexpression, we analyzed the effect on the AvrBsT-dependent HR in pepper using Xanthomonas infection. Xcv-mediated additional ectopic expression of xopB, but not xopBA313V, suppressed the HR caused by Xcv strain 75-3, which naturally expresses avrBsT, in pepper ECW (Fig. 7a,b). This was correlated with increased bacterial growth (Fig. 7c). If avrBsT was deleted, strain 75-3 caused no HR and grew significantly better, as expected. In this case, xopB overexpression provided no additional growth advantage. This indicates that the positive effect of XopB on bacterial growth is based on the specific suppression of AvrBsT-induced defense responses, and therefore represents a biologically relevant activity of the effector.
XopB localizes to Golgi vesicles and the cytoplasm
Analysis of the subcellular localization of T3Es might provide some clues about their site of action inside the plant cell. To investigate the localization of XopB, we transiently expressed a xopB::GFP fusion in N. benthamiana using Agrobacterium-mediated gene delivery. Subcellular localization was determined by confocal laser scanning microscopy at 24 h post-inoculation (hpi). GFP alone was clearly detectable in the cytoplasm and nuclei (Fig. 8a). By contrast, the fluorescence of XopB::GFP was confined to vesicle-like structures and the cytoplasm, and was not detectable in the nucleus (Fig. 8a). We assumed that the vesicle-like structures might be part of the Golgi system. To test this hypothesis, we transiently co-expressed XopB::GFP with Golgi-mCherry, a fluorescence marker for the Golgi apparatus (Nelson et al., 2007). Both proteins co-localized (Fig. 8a), suggesting that XopB indeed associates with Golgi vesicles. On the ultrastructural level, electron microscopy following immunolabeling showed that, in contrast to free GFP, XopB::GFP predominantly localized to vesicle structures and was strongly under-represented in vesicle-free areas (Fig. 8b).
Furthermore, immunoblot of the subcellular fractionation of N. benthamiana extracts confirmed that XopB::c-Myc, similar to Golgi-mCherry, is predominantly associated with the plant membrane fraction (Fig. S5; Methods S1). In addition, intact XopBA313V and XopBK455R were detectable (Fig. S6) and showed a similar localization pattern to the wild-type protein in microscopic as well as fractionation studies (Figs 8a and S5). This suggests that the functional loss of XopBA313V is not caused by mislocalization.
XopB interferes with plant cell protein secretion
The Golgi apparatus provides the cellular basis for intracellular vesicle trafficking, for example protein transport to the plasma membrane and secretion into the apoplast, which plays an important role in plant immunity and has been shown to be targeted by several T3Es (Bartetzko et al., 2009; Frey & Robatzek, 2009). To investigate whether XopB interferes with protein secretion of the plant cell, we used secGFP, a GFP variant that is secreted to the apoplast (Batoko et al., 2000). secGFP only accumulates in a fluorescent form when its transport to a post-Golgi compartment is prevented, for example, by the secretion inhibitor Brefeldin A (BFA). To analyze the influence of XopB on secGFP secretion, both proteins were transiently co-expressed in N. benthamiana mediated by Agrobacterium. As positive controls, we used BFA and AtSYP121_Sp2, a cytosolic fragment of a plasma membrane-bound regulator of vesicle trafficking, which has a dominant negative effect on membrane trafficking (Tyrrell et al., 2007), and, in addition, XopJ, a T3E from Xcv which suppresses plant cell secretion (Bartetzko et al., 2009). XopJC235A, a XopJ mutant that does not block secretion served as a negative control (Bartetzko et al., 2009). Only weak fluorescence was detectable inside the epidermal cells if secGFP was expressed either alone or together with the negative control (Fig. 9a). By contrast, co-expression of secGFP and XopB resulted in a strong accumulation of GFP fluorescence, forming an intracellular reticulate pattern, comparable with the fluorescence pattern obtained after co-expression of secGFP with AtSYP121_Sp2 and XopJ or the application of BFA (Fig. 9a). XopBA313V also resulted in the accumulation of intracellular secGFP fluorescence; however, less distinct networks and punctate structures were observed (Fig. 9a). By contrast, XopBK455R caused a fluorescence pattern indistinguishable from that induced by wild-type XopB. Immunoblot analysis showed that differences in secGFP fluorescence were not a result of different protein levels (Fig. 9b).
In this study, we analyzed eight Xcv effector proteins, six of which were newly identified, so that there are now 23 experimentally verified T3Es in Xcv strain 85-10. A major finding is that XopB is a virulence factor that suppresses plant PTI as well as ETI. The T3Es were classified on the basis of whether or not their translocation into plant cells requires the general chaperone HpaB. XopR and XopS belong to Xcv translocation class A, comprising T3Es whose translocation into plant cells is completely dependent on HpaB, whereas XopB, XopG, XopI, XopK, XopM and XopV were assigned to class B, because they are still translocated in the absence of HpaB (Büttner et al., 2006). Both new class A effectors lack homology to known proteins or motifs, so that their molecular function remains elusive. By contrast, the class B effectors comprise the putative enzyme XopG, a member of the HopH family (Lindeberg et al., 2005) of putative zinc metalloproteases. Other effectors possess interesting features, for example XopI contains an F-box motif typical for eukaryotic proteins playing a role in the ubiquitin-26S proteasome system (UPS). The UPS controls protein stability in eukaryotes (Willems et al., 2004) and appears to be a favorable target for many T3Es, for example members of the GALA family, which strongly contribute to the virulence of R. solanacearum (Angot et al., 2006), and the E3 ubiquitin ligase AvrPtoB from P. syringae (Abramovitch et al., 2006; Janjusevic et al., 2006).
As HpaB is essential for pathogenicity, it was speculated that class A effectors play a key role in the establishment of a pathogenic interaction of Xanthomonas with the host (Büttner et al., 2004, 2006). Indeed, XopS is involved in the severity of disease symptoms, the promotion of bacterial growth and the suppression of PTI, whereas the role of XopR in the virulence of Xcv strain 85-10 is probably more subtle. Interestingly, deletion of the xopR homolog in X. oryzae pv. oryzae resulted in reduced virulence on host rice plants (Akimoto-Tomiyama et al., 2012). The fact that individual deletion mutants revealed no major contribution of XopR and most other effectors to the virulence of Xcv 85-10 is not unexpected and is probably caused by functional redundancies. Surprisingly, the class B effector XopB clearly contributed to the disease symptoms and bacterial growth of 85-10 on susceptible plants and the suppression of PTI (Figs 2 and 3). This finding differs from our previous study, where no difference in disease symptoms was observed between the wild-type strain and xopB mutant (Noël et al., 2001), and may be caused by different environmental conditions (glasshouse, growth chamber). In addition to its effect on PTI, XopB suppressed the ETI-related HR induced by AvrBsT, as well as cell death reactions triggered by XopG, XopJ and AvrRxv (Figs 5 and 6). Although XopB localizes to the Golgi system and the cytoplasm, the inducers of cell death reactions suppressed by XopB localize to different cellular compartments: XopG to the nucleus (Fig. S7), XopJ to the plasma membrane (Thieme et al., 2007), AvrRxv to the cytoplasm (Bonshtien et al., 2005) and AvrBsT to the nucleus and cytoplasm (Szczesny et al., 2010a). As the co-expression of T3Es with XopB did not change their subcellular localization (Fig. S8), XopB probably does not interfere with effector recognition, but rather with downstream signaling.
In contrast with AvrPtoB, it is unlikely that XopB acts on top of signaling cascades because it does not inhibit the flg22-triggered activation of MAPKs (Fig. S3) and the CDS activity of XopB is not dependent on membrane-bound receptor kinases because XopG, AvrRxv and AvrBsT are probably not recognized at the plasma membrane (see earlier in the Discussion section). XopB may therefore target a later step of the convergent cellular pathways following T3E recognition. Our studies point to XopB-dependent inhibition of intracellular vesicle trafficking as a possible mode of action to suppress plant immunity. Vesicle trafficking plays an important role in plant defense, for example for the correct localization of PAMP receptors in the plasma membrane. During PTI, genes encoding receptor kinases are induced, including the PAMP receptors themselves (Zipfel et al., 2006; Miya et al., 2007). This results in an increase in receptors and an amplification of the PAMP response (Zipfel et al., 2006). In addition, vesicle transport is involved in the export of antimicrobial molecules, for example PR proteins, phytoalexins and cell wall-bound compounds, and the localization of plasma membrane ABC transporters, which release small antimicrobial molecules to the cell surface (Kwon et al., 2008). Intriguingly, our studies suggest that the inhibition of vesicle transport might explain the XopB effect on PTI, but is insufficient for ETI suppression. XopBA313V completely loses CDS activity, although it still inhibits secGFP secretion and is only slightly affected in PTI suppression (Fig. 3). This suggests that the suppression of PTI and ETI is mediated by separate activities of XopB, and that ETI suppression involves as yet unknown mechanisms.
There are two other T3Es from phytopathogenic bacteria which suppress immunity and interfere with plant protein secretion: (1) HopM1 from P. syringae, which is targeted to Arabidopsis endomembranes, suppresses PTI and contributes to disease symptoms and bacterial growth in planta (DebRoy et al., 2004; Nomura et al., 2006). HopM1 mediates the UPS-dependent degradation of a key component of the plant vesicle trafficking system (Nomura et al., 2006). (2) XopJ from Xcv has been proposed to localize, at least in part, to the Golgi apparatus (Bartetzko et al., 2009). However, the respective analyses were performed in N. benthamiana, where XopJ induces necrosis at 3–4 dpi (Thieme et al., 2007), raising the possibility that the occasional localization in punctate structures might be caused by morphological changes induced by ongoing cell death. Nevertheless, XopJ inhibits secGFP secretion and suppresses PTI (Bartetzko et al., 2009). In contrast with XopB, however, HopM1 and XopJ do not appear to affect ETI.
The next challenge is the identification of plant targets, especially of XopB. The putative XopB target appears to be conserved in different plant families as the effector has a virulence activity in pepper, the natural host plant of Xcv, and also in the nonhost Brassicaceae A. thaliana, as demonstrated by our protoplast assays. That XopB homologs are found in a wide range of plant-pathogenic bacteria (Table 1), infecting various host plants, supports this hypothesis. Interestingly, a recent study has shown that XopB inhibits yeast growth (Salomon et al., 2011). The application of caffeine, which induces cell wall stress, strongly increases the negative effect of XopB on yeast (Salomon et al., 2011). In the light of our results and the fact that vesicle transport is important for yeast cell wall assembly, for example, during budding (Smits et al., 2001), we believe that XopB targets a conserved component of eukaryotic vesicle trafficking.
We thank A. Urban, B. Rosinsky, C. Kretschmer, M. Jordan, S. Jahn and N. Bauer for excellent technical assistance. We are grateful to C. Lorenz and H. Berndt for providing unpublished material. This work was funded by grants from the Deutsche Forschungsgemeinschaft to U.B., D.B., J.L. and D.S. (SFB 648 ‘Molekulare Mechanismen der Informationsverarbeitung in Pflanzen’) and from the Bundesministerium für Bildung und Forschung to J.L. and D.S. (‘tools, targets & therapeutics – ProNet-T3’).