•Type II secretion (T2S) systems of many plant-pathogenic bacteria often secrete cell wall-degrading enzymes into the plant apoplast.
•Here, we show that the Xps-T2S system from the plant pathogen Xanthomonas campestris pv vesicatoria (Xcv) promotes disease and contributes to the translocation of effector proteins that are delivered into the plant cell by the type III secretion (T3S) system.
•The Xcs-T2S system instead lacks an obvious virulence function. However, individual xcs genes can partially complement mutants in homologous xps genes, indicating that they encode functional components of T2S systems. Enzyme activity assays showed that the Xps system contributes to secretion of proteases and xylanases. We identified the virulence-associated xylanase XynC as a substrate of the Xps system. However, homologs of known T2S substrates from other Xanthomonas spp. are not secreted by the T2S systems from Xcv. Thus, T2S systems from Xanthomonas spp. appear to differ significantly in their substrate specificities.
•Transcript analyses revealed that expression of xps genes in Xcv is activated by HrpG and HrpX, key regulators of the T3S system. By contrast, expression of xynC and extracellular protease and xylanase activities are repressed by HrpG and HrpX, suggesting that components and substrates of the Xps system are differentially regulated.
In our laboratory, we study virulence factors from Xanthomonas campestris pv vesicatoria (Xcv, also designated Xanthomonas euvesicatoria or Xanthomonas axonopodis pv vesicatoria; Jones et al., 2004), the causal agent of bacterial spot disease in pepper and tomato. The type III secretion (T3S) system is essential for colonization of the plant and translocates bacterial effector proteins directly into the plant cell cytosol where they interfere with cellular functions such as defense responses to the benefit of the pathogen (Ghosh, 2004; Gürlebeck et al., 2006; Zhou & Chai, 2008). Translocation of effector proteins into plant cells depends on extracellular Hrp (hypersensitive response and pathogenicity) pili that are associated with the membrane-spanning T3S apparatus and serve as protein transport channels across the plant cell wall (Büttner & Bonas, 2002; Ghosh, 2004). Components of the T3S system are encoded by the chromosomal hrp gene cluster, which is specifically activated when the bacteria enter the plant or are cultivated in certain minimal media (Büttner & Bonas, 2002). Expression of the hrp genes depends on the response regulator HrpG, which controls – in most cases via the transcriptional activator HrpX – the expression of a genome-wide regulon, including effector genes and genes that encode predicted extracellular degradative enzymes (Wengelnik & Bonas, 1996; Wengelnik et al., 1996b; Noël et al., 2001).
The contribution of the type II secretion (T2S) system to virulence of Xcv has not yet been studied. Generally, T2S systems secrete toxins and degradative enzymes that are transported across the inner bacterial membrane by the Sec or TAT (twin-arginine translocation) system and often contribute to virulence of plant pathogenic bacteria (Sandkvist, 2001; Cianciotto, 2005; Jha et al., 2005). Protein transport across the outer bacterial membrane is mediated by the T2S apparatus that consists of 12–15 components, most of which are associated with the inner membrane or form a periplasmic pseudopilus (Sandkvist, 2001). The continuous assembly and disassembly of the pseudopilus presumably pushes T2S substrates through the outer membrane secretin (Sandkvist, 2001; Cianciotto, 2005). Type II-dependent protein transport is probably energized by a cytoplasmic ATPase associated with the secretion apparatus at the inner membrane (Chen et al., 2005; Shiue et al., 2006, 2007).
In this study, we show that the Xps system from Xcv contributes to bacterial virulence and type III effector protein translocation. No obvious role in virulence was observed for the Xcs system, however, xcs genes can partly compensate loss of homologous xps genes. The virulence function of the Xps system presumably results from the secretion of extracellular proteases and xylanases. We identified the virulence-associated xylanase XynC as a substrate of the Xps system and provide experimental evidence that components and substrates of the Xps system are differentially regulated.
Materials and Methods
Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli cells were grown at 37°C in lysogeny broth (LB) or Super medium (Qiagen), Xcv strains were cultivated at 30°C in nutrient–yeast–glycerol (NYG) medium (Daniels et al., 1984) or in minimal medium A (Ausubel et al., 1996), which was supplemented with sucrose (10 mM) and casamino acids (0.3%). Plasmids were introduced into E. coli by electroporation and into Xcv by conjugation, using pRK2013 as helper plasmid in triparental matings (Figurski & Helinski, 1979). Antibiotics were added to the media at the following final concentrations: ampicillin, 100 μg ml−1; gentamycin 15 μg ml−1; kanamycin, 25 μg ml−1; rifampicin, 100 μg ml−1; spectino-mycin, 100 μg ml−1; tetracycline, 10 μg ml−1.
Table 1. Bacterial strains and plasmids used in this study
The near-isogenic pepper lines ‘Early Cal Wonder’ (ECW), ECW-10R and ECW-30R (Minsavage et al., 1990) were grown at 25°C with 60–70% relative humidity and 16 h light. Xcv strains were hand-inoculated with a needleless syringe into the intercellular spaces of pepper leaves at concentrations of 2 × 108 CFU ml−1 in 1 mM MgCl2, unless stated otherwise. Disease symptoms and the hypersensitive response (HR) were scored over a period of 2–10 d post inoculation (dpi). For better visualization of the HR, leaves were bleached in 70% ethanol. In planta growth curves were performed as described (Bonas et al., 1991). All experiments were repeated at least two times.
Enzyme activity assays
For the analysis of extracellular protease and xylanase activity, Xcv strains were adjusted to a density of 109 CFU ml−1 and incubated on NYG agar plates containing 1% skimmed milk (AppliChem) and 0.1% remazol brilliant blue (RBB) xylan (Sigma-Aldrich), respectively (Vroemen et al., 1995). For detection of cellulase and amylase activities, we used plates containing 1% carboxymethyl cellulose (CMC; Sigma-Aldrich) and 1% starch (AppliChem), respectively. The CMC plates were stained with 0.2% Congo red (Sigma-Aldrich) solution and destained with 0.5 M NaCl (Gough et al., 1988). Starch plates were stained with Lugol’s solution (0.4% potassium iodide : 0.2% iodine solution) and destained with 98% ethanol–acetone (1 : 1) (Hu et al., 1995). Bacteria were inoculated into holes that were punched out of the agar. Plates were incubated at 30°C for 1–2 d and bacteria were removed before documentation. All experiments were repeated at least two times.
Generation of deletion constructs
To delete xcsD, xpsD, xcsE, xpsE, XCV4312 and xynC, respectively, flanking regions were amplified by PCR using genomic DNA from Xcv strain 85-10 as template, and the amplicons were introduced into the suicide vector pOK1. Deletion constructs were introduced into the genome of Xcv by crossing-over events as described in Huguet et al. (1998). Details on cloning strategies are available upon request. Primer sequences are listed in the Supporting Infor-mation, Table S1.
Generation of the Golden Gate-compatible expression vectors pBRM and pBRM-P
The Golden Gate system allows efficient one step-cloning of PCR amplicons into a destination vector in a restriction–ligation reaction (Engler et al., 2008). The system is based on type IIs restriction enzymes (e.g. BsaI) that cut DNA outside of the enzyme’s recognition sites. The Golden Gate-compatible expression vectors pBRM and pBRM-P used in this study are derived from pBBR1MCS-5 (see Table 1). To remove the internal BsaI site of pBBR1MCS-5, vector DNA was amplified by PCR with primers pBBR1-for/pBBR2-rev and pBBR3-for/pBBR4-rev and both fragments were religated in a one-step restriction–ligation reaction using BsaI. The resulting vector pBBR1mod1 lacks the internal BsaI site and contains a single EcoRI and a HindIII site that replace the polylinker. To generate pBRM, we amplified the lacZ alpha gene from pUC19 and the triple c-Myc epitope-encoding sequence from pC3003 with primers lacZ-for/lacZ-rev and myc-for/myc-rev, respectively, and cloned the corresponding fragments (digested with BpiI) into the EcoRI/HindIII sites of pBBR1mod, giving pBRM (Fig. S1). To generate pBRM-P, we amplified pBRM with primers pBBR1-for and pBBR5-rev and ligated the PCR product in a one-step restriction–ligation reaction using BsaI and ligase. The resulting vector pBBR1mod2 lacks the lac promoter and contains a single EcoRI and a HindIII site that replace the polylinker. We amplified the lac promoter, the lacZ alpha gene and the c-Myc epitope-encoding sequence from pBRM with primers lacZprom-for and myc-Esp-rev. The corresponding PCR fragment and pBRM were digested with Esp3I and EcoRI/HindIII, respectively, and ligated, thus generating pBRM-P (see Fig. S1).
Generation of expression constructs
For the generation of expression constructs encoding components or putative substrates of T2S systems, corresponding gene fragments were amplified by PCR from Xcv strain 85-10, inserted into pENTR/D-TOPO and recombined into pDGW4M using GATEWAY technology (Invitrogen). Alternatively, PCR amplicons were cloned into pBRM or pBRM-P in a one step restriction–ligation reaction as described in Engler et al. (2008). For fragments with an internal BsaI site, the reaction mix was incubated with ligase after heat inactivation for additional 20 min at 37°C (Engler et al., 2008). All constructs are listed in Table 1.
Protein analysis and secretion experiments
For the analysis of bacterial protein accumulation in planta, bacteria were inoculated at a density of 109 CFU ml−1 into leaves of susceptible ECW pepper plants. Proteins were extracted 6 h post inoculation (hpi) from five leaf discs that were ground in 1 mM MgCl2. For in vitro T2S assays, bacteria were grown overnight in NYG medium, resuspended in fresh NYG medium at a cell density of 1.5 × 108 CFU ml−1 and incubated at 30°C for 2 h. Culture supernatants were separated from bacterial cells by filtration and secreted proteins were precipitated by trichloroacetic acid as described in Rossier et al. (1999). Equal amounts of total bacterial cell extracts and culture supernatants were analysed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting. We used polyclonal antibodies specific for HrpF, AvrBs3, HrcJ and GroEL (Knoop et al., 1991; Rossier et al., 2000; Büttner et al., 2002; Stressgen), and monoclonal anti-c-Myc (Roche) and anti-Strep antibodies (IBA GmbH). Horseradish peroxidase-labeled anti-rabbit and anti-mouse antibodies (Amersham Pharmacia Biotech) were used as secondary antibodies, and antibody reactions were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech).
For transcript analyses via quantitative reverse-transcription PCR (qRT-PCR), bacteria were grown overnight in NYG medium, adjusted to an optical density (OD, 600 nm) of 0.2 and incubated for 4 h at 30°C. RNA was extracted using a Trizol-based protocol and cDNA was synthesized using 4.5 μg RNA in a RevertAid H Minus First Strand cDNA-Synthesis Kit (Fermentas). xcsE, xpsE, xpsF and xopC transcripts were amplified using primers xcsE_RT-for/xcsE_RT-rev, xpsE_RT-for/xpsE_RT-rev, xpsF_RT-for/xpsF_RT-rev and xopC_RT-for/xopC_RT-rev, respectively. The qRT-PCR was performed on an iCycler (Bio-Rad) using a SYBR Green-based PCR reaction mixture (ABsolute QPCR SYBR Green Fluorescein Mix; ABgene Limited) and c. 9 ng of template cDNA. Details on PCR profiles are available upon request. For comparative analysis of transcript levels of different genes, the efficiency of each PCR reaction was determined using a dilution series of template plotted into a standard curve. To ensure amplification specificity, amplicons were subjected to melting curve analysis and analysed on 1.5% agarose gels. Transcript levels were determined as technical triplicates and transcript levels of constitutively expressed 16S rRNA (primers: 16S_RT-for/16S_RT-rev) were used to account for differences in cDNA amounts as described in the ABI user bulletin 2 (Applied Biosystems).
xcs and xps genes from Xcv strain 85-10 are expressed in vitro
Genome sequence analyses of Xcv strain 85-10 revealed the presence of chromosomal xcs and xps gene clusters containing 12 (xcsC to xcsN) and 11 (xpsD to xpsN) genes, respectively (Fig. 1a; Thieme et al., 2005). To investigate expression of xcs and xps genes, we performed RT-PCR analysis with cDNA derived from Xcv strain 85-10 grown in complex NYG medium. Using primers specific for the predicted ATPase-encoding genes xcsE and xpsE, cDNAs corresponding to fragments of both genes were amplified, suggesting that xcsE and xpsE are expressed (Fig. 1b). This was confirmed by qRT-PCR analysis, which revealed that xpsE transcript levels were c. 76-fold higher than xcsE levels (data not shown).
The predicted secretin gene xpsD contributes to bacterial virulence
To investigate a possible contribution of the Xcs and Xps systems to virulence of Xcv strain 85-10, we deleted xcsD and xpsD, respectively, which encode the predicted outer membrane secretins of the two T2S systems. As expected, when bacteria were inoculated into leaves of susceptible ECW pepper plants, strain 85-10 induced water-soaked lesions (Fig. 1c). A similar phenotype was observed for strains 85-10ΔxcsD and 85-10Δxcs, which lacks the complete xcs gene cluster, suggesting that xcs genes do not significantly contribute to virulence (data not shown; Fig. S2). By contrast, strain 85-10ΔxpsD led to reduced disease symptoms (Fig. 1c). This effect was more pronounced when bacteria were inoculated at low population densities (2 × 107 CFU ml−1; compare Fig. 1c,d). The xpsD mutant phenotype was complemented by ectopic expression of xpsD, suggesting that it was specifically caused by the deletion of xpsD (Fig. 1c).
To investigate potential functional redundancies among secretin-encoding genes, we deleted both xpsD and xcsD from the genome of Xcv strain 85-10. XcsD shares 28% sequence identity with XpsD. Notably, the resulting double deletion mutant strain 85-10ΔxpsDΔxcsD caused reduced disease symptoms when compared with strain 85-10ΔxpsD (Fig. 1d). By contrast, the double deletion mutant strain 85-10ΔxpsDΔxcsE caused disease symptoms similar to the xpsD single deletion mutant (Fig. 1d). This suggests that the additional reduction in virulence of strain 85-10ΔxpsDΔxcsD was specifically caused by the lack of xcsD. We therefore conclude that XcsD can at least partially compensate the loss of XpsD. In addition to disease symptoms, we analysed bacterial growth of wild-type and deletion mutant strains in susceptible ECW pepper plants. In planta growth of strain 85-10ΔxcsD 7 dpi was similar to that of the wild-type strain 85-10, whereas 85-10ΔxpsD numbers were c. 100-fold reduced compared with strain 85-10 (Fig. 1e). Additional deletion of xcsD in strain 85-10ΔxpsD led to a further reduction of bacterial counts 7 dpi, which is in good agreement with the observed phenotypes (Fig. 1e). Reduced bacterial growth of single and double deletion mutants in planta was presumably not caused by a general growth deficiency, because bacterial multiplication in minimal medium was not affected (Fig. 1f).
Characterization of Xcv mutants lacking predicted T2S-associated ATPases
To confirm that the virulence phenotypes of secretin mutants were caused by nonfunctional T2S systems and not by alterations in the outer membrane, we deleted xcsE, xpsE and XCV4312, respectively. XCV4312 encodes a predicted ATPase with homology to XcsE and XpsE and is located outside of the xcs and xps gene clusters of Xcv strain 85-10. Deletion of xpsE led to a significant reduction in disease symptoms and in planta bacterial growth (Fig. 2a,b). Reduced disease symptoms but not bacterial growth of the xpsE mutant was more pronounced than for strain 85-10ΔxpsD, which was presumably a result of the functional redundancy of xpsD and xcsD (Fig. 2c,d; see above). By contrast, xcsE and XCV4312 deletion mutants displayed a wild-type phenotype (Fig. 2a; data not shown). Furthermore, strain 85-10ΔxpsEΔxcsE provoked disease symptoms comparable to strain 85-10ΔxpsE (Fig. 2a,c). Notably, however, deletion of xcsE in strain 85-10ΔxpsE led to a further reduction of in planta bacterial growth, suggesting that in this respect XcsE (42% sequence identity with XpsE) can partly complement the xpsE deletion mutant (Fig. 2b). Reduced virulence and in planta growth of single and double deletion mutants was presumably not caused by a general growth deficiency, because bacterial multiplication in minimal medium was not affected (Fig. 2e). The xpsE mutant phenotype was complemented by ectopic expression of xpsE or xpsE-c-myc, which encodes a C-terminally c-Myc epitope-tagged XpsE (Fig. 2f).
Type III-dependent effector protein translocation is reduced in the absence of xpsE
In addition to disease symptoms, we investigated whether Xps and Xcs systems from Xcv contribute to the induction of plant defense responses that are triggered upon recognition of type III effectors (also termed avirulence (Avr) proteins) in resistant plants carrying cognate resistance (R) genes. Avr protein-triggered plant defense is often associated with the HR, a rapid local plant cell death that restricts bacterial multiplication (Dangl & Jones, 2001). For the analysis of the HR induction, strains 85-10, 85-10Δxcs and 85-10ΔxpsE were inoculated into leaves of resistant ECW-10R pepper plants that carry the Bs1 resistance gene and induce the HR upon recognition of the type III effector AvrBs1 that is delivered by strain 85-10 (Ronald & Staskawicz, 1988; Escolar et al., 2001). Deletion of the complete xcs gene cluster in strain 85-10 did not affect the HR (data not shown). By contrast, the HR induced by strain 85-10ΔxpsE was significantly reduced when compared with the wild type (Fig. 3a). We then tested translocation of additional type III effector proteins including AvrBs3, XopC, XopJ and XopF1 using AvrBs3Δ2 as a reporter. AvrBs3Δ2 is a derivative of the type III effector AvrBs3 lacking the secretion and translocation signal. However, AvrBs3Δ2 contains the effector domain and is recognized in resistant ECW-30R pepper plants when fused to a functional T3S and translocation signal (Szurek et al., 2002; Noël et al., 2003). As expected, strain 85-10 expressing AvrBs31-200-, XopC1-200-, XopJ1-155- and XopF11-200-AvrBs3Δ2 fusion proteins induced the HR in ECW-30R leaves (Szurek et al., 2002; Noël et al., 2003; Büttner et al., 2007; Fig. 3a). However, the HR induced by the corresponding xpsE deletion mutant strains was reduced, suggesting that XpsE is required for efficient effector protein translocation (Fig. 3a).
To investigate whether reduced effector protein translocation was caused by reduced T3S, we performed in vitro T3S assays with strains 85-10hrpG* (85*) and 85*ΔxpsE. Both strains contain a constitutively active derivative of the key regulator HrpG and therefore express T3S genes in vitro under noninducing conditions (Wengelnik et al., 1999). For the analysis of in vitro T3S, strains 85* and 85*ΔxpsE were incubated in secretion medium and total cell extracts and culture supernatants were analysed by immunoblotting. Fig. 3b shows that comparable amounts of the T3S translocon protein HrpF and the effector fusion proteins XopJ1-155–AvrBs3Δ2 and XopC1-200–AvrBs3Δ2, which were selected for in vitro T3S assays, were detected in the culture supernatants of both strains. These results suggest that the Xps system is not essential for the activity of the T3S system per se.
Next, we investigated at which time-point after infection the reduction in effector protein translocation is detectable. For this, we blocked bacterial protein synthesis in infected leaf tissue of ECW-10R pepper plants by treatment with spectinomycin. Infiltration of spectinomycin 0 and 4 hpi abolished the HR induction by strains 85-10 and 85-10ΔxpsE (Fig. 3c). However, strain 85-10 induced the HR when spectinomycin was infiltrated 6 hpi, whereas no HR induction was observed for strain 85-10ΔxpsE (Fig. 3c). This suggests that the amounts of effector proteins translocated within 6 hpi by the wild-type, but not by the xpsE deletion mutant are sufficient to elicit the HR. Reduced HR induction 6 hpi was presumably not caused by differences in bacterial counts of strains 85-10 and 85-10ΔxpsE, which were comparable 6 hpi (Fig. 3d).
We also analysed the accumulation of Hrp and effector proteins 6 hpi by immunoblotting. Expression of hrp genes is specifically activated when the bacteria enter the apoplast, presumably by plant-derived molecules (Schulte & Bonas, 1992; Wengelnik et al., 1996a,b). When proteins were extracted 6 hpi from leaves of susceptible ECW pepper plants that were inoculated with Xcv wild-type and xpsE deletion mutants, comparable levels of the predicted lipoprotein HrcJ and the effector fusion XopF1-200-AvrBs3Δ2 were detected in both 85-10 and 85-10ΔxpsE (Fig. 3e,f). A slight reduction in HrpF levels was observed for the xpsE deletion mutant, however, previous experiments indicated that this is not sufficient to cause a reduction in HR induction (C. Lorenz & D. Büttner, unpublished). We conclude from our experiments that xpsE is required for efficient effector protein translocation.
The xpsE mutant phenotype can be complemented in trans
It is conceivable that the reduced virulence of strain 85-10ΔxpsE results from a lack of extracellular bacterial proteins. We therefore wondered whether this deficiency can be complemented in trans. For this, we coinoculated strain 85-10ΔxpsE with strains 85-10ΔhrcN and 85-10ΔhrcNΔxpsE, respectively, both deficient in T3S and therefore nonpathogenic (Lorenz & Büttner, 2009; Fig. 4). Disease symptoms in ECW pepper leaves were partially restored when strain 85-10ΔxpsE was coinoculated with strain 85-10ΔhrcN (Fig. 4). This was not observed when strain 85-10ΔxpsE was coinoculated with strain 85-10ΔhrcNΔxpsE, which is deficient in the Xps system, suggesting that the restoration of the wild-type phenotype was specifically caused by Xps substrates (Fig. 4). Similarly, HR induction of strain 85-10ΔxpsE in resistant ECW-10R pepper plants was enhanced upon coinoculation with strain 85-10ΔhrcN but not with strain 85-10ΔhrcNΔxpsE. We therefore conclude that the efficient translocation of effector proteins by the xpsE deletion mutant can be restored in trans.
The Xps system contributes to extracellular protease and xylanase activities
Next, we investigated the contribution of the Xcs and Xps system to the secretion of degradative enzymes, including proteases, xylanases, cellulases and amylases, by suitable plate assays (see the Materials and Methods section). Strain 85-10 secretes extracellular proteases that degrade milk proteins, which is visible as a cleared halo around the bacteria (Noël et al., 2001; Fig. 5a). The analysis of xcs and xps mutants revealed that halo formation on milk plates by xcs mutants was like wild type (Fig. 5a). By contrast, no halo was detected for strains 85-10ΔxpsE and 85-10ΔxpsD, suggesting that the Xps system is required for extracellular protease activity (Fig. 5b).
To analyse the activity of extracellular xylanases, bacteria were incubated on NYG agar plates containing RBB xylan. Degradation of xylan is visible as a clear zone around the bacteria. Deletion of xps, but not of xcs genes resulted in significantly reduced extracellular xylanase activity 1 dpi (Fig. 5a,b). Notably, when bacteria were incubated for 2 d on RBB xylan plates there was residual extracellular xylanase activity for xps deletion mutants (see Fig. 5b). Because similar results were observed for the triple deletion mutants 85-10ΔxpsEΔxcsEΔ4312 and 85-10ΔxpsDΔxcsDΔhrcC, the residual xylanase activity of xps deletion mutants was presumably not caused by cross-complementation by homologous genes (Fig. 5c). Strain 85-10ΔxpsDΔxcsDΔhrcC lacks both secretins of the T2S systems and HrcC, which is the secretin of the T3S system and shares 25% identity with XpsD. Reduced activity of extracellular proteases and xylanases in xpsD and xpsE deletion mutants was complemented by ectopic expression of xpsD and xpsE-c-myc, respectively (Fig. 5d,e).
Extracellular amylase and cellulase activities are independent of the Xcs and Xps systems
In addition to extracellular protease and xylanase activities, we analysed the activities of extracellular cellulases and amylases on plates containing CMC and starch, respectively. To visualize substrate degradation, CMC plates were stained with Congo red and starch plates with Lugol’s solution. Unexpectedly, extracellular cellulase and amylase activities in xcs and xps deletion mutants were like wild type (Fig. 6a,b). Similar results were obtained with the triple deletion mutants 85-10ΔxpsEΔxcsEΔ4312 and 85-10ΔxpsDΔxcsDΔhrcC (Fig. 6c,d), suggesting that the Xcs and Xps systems are not required for secretion of amylases and cellulases from Xcv. Notably, however, secretion of amylases and cellulases in Xcc and Xoo depends on the T2S system (Hu et al., 1992; Ray et al., 2000; Furutani et al., 2004; Sun et al., 2005).
Analysis of candidate T2S substrates
Next, we analysed secretion of candidate T2S substrates including a putative cellulase (XCV0031), polygalacturonase (XCV0722), xylanase (XCV0965), glucosidase (XCV1823), pectate lyase (XCV2569) and several proteases (XCV0959, XCV0960, XCV2049, XCV2918, XCV3013, XCV3669, XCV4074; Table 2). Candidate proteins were selected based on homologies to known T2S substrates and virulence factors from Xanthomonas spp. and/or their coregulation with the T3S system (Noël et al., 2001; F. Thieme et al., unpublished; see Table 2). All proteins were synthesized in Xcv as C-terminally c-Myc epitope-tagged derivatives. Wild-type and T2S mutant strains carrying the respective expression constructs were incubated in NYG medium, and total cell extracts and culture supernatants were analysed by immunoblotting using a c-Myc epitope-specific antibody.
Table 2. Selected candidate type II secretion (T2S) substrates from Xcv strain 85-10
Candidate T2S substrates
Presence/characteristics of homologous proteins from Xanthomonas spp.4
2HrpG/HrpX-dependent gene expression was analysed by microarray (F. Thieme et al., unpublished) and/or cDNA-amplified fragment length polymorphism (AFLP; Noël et al., 2001). Numbers of corresponding cDNA-AFLP fragments (hgi, hrpG-induced; hgr, hrpG-repressed) are given in brackets. PIP, plant-inducible promoter, consensus TTCGC-N15-TTCGC (Fenselau & Bonas, 1995); na, not analysed.
3Secretion was analysed in Xcv wild-type and T2S mutant strains (see Fig. 7).
Conserved in Xanthomonas spp.; homolog XOO0282 (EglXoB) is required for virulence of Xoo PXO995
Conserved in Xanthomonas spp.; homologs Xc0705 (PghAxc) from Xcc and PglA and Peh-1 from Xac are secreted by the T2S system6,7
Conserved in Xanthomonas spp.
Conserved in Xanthomonas spp.; homolog XOO2356 (CelD) was identified as virulence factor of Xoo8
Conserved in Xac
na, imperfect PIP box (TTCGG-N9-TTCGG) 27 bp upstream of ATG
Conserved in Xanthomonas spp.
Conserved in Xanthomonas spp.
Conserved in Xanthomonas spp.; homolog ClpA from Xoo was identified as virulence factor8
Conserved in Xanthomonas spp.
Conserved in Xanthomonas spp.; homolog CysP2 from Xoo is secreted via the T2S system9
Conserved in Xanthomonas spp.
Conserved in Xanthomonas spp.; homolog XAC3980 was identified as virulence factor from Xac10
Most proteins except the predicted pectate lyase XCV2569-c-Myc and the predicted proteases XCV2918-c-Myc and XCV2049-c-Myc were detected in the culture supernatant of the wild-type strain, indicating that they were secreted (Fig. 7a,b). Unexpectedly, for the predicted proteases XCV0959 and XCV3669 we detected two proteins in the culture supernatants including smaller and/or larger proteins than the proteins detected in total cell extracts. Smaller proteins are presumably cleavage products whereas larger proteins might result from a specific protein modification in the culture supernatant. However, it cannot be excluded that secreted XCV0959-c-Myc and XCV3669-c-Myc are part of protein complexes that are not dissolved by SDS-PAGE. Note that XCV0959-c-Myc was less stable in cell extracts of T2S mutant strains. Unexpectedly, c-Myc epitope-tagged derivatives of XCV0031, XCV0722, XCV0959, XCV0960, XCV1823, XCV3013, XCV3669 and XCV4074 were detected in the culture supernatants of T2S-deficient strains, suggesting that secretion of these proteins was independent of both T2S systems (Fig. 7a,c). This was not caused by cell lysis as, for example, the cytoplasmic T3S system-associated ATPase HrcN that was ectopically expressed as Strep or c-Myc epitope-tagged derivative was only detectable in total cell extracts (Fig. 7a).
We also investigated secretion of two polygalacturonases (PghAxc and PghBxc) from Xcc that were previously identified as T2S substrates (Wang et al., 2008b; Table 3). The corresponding genes (Xc0705 and Xc1849) were amplified from Xcc strain 8004 and expressed as C-terminally c-Myc epitope-tagged derivatives in Xcv strains 85-10 and 85-10ΔxpsEΔxcsEΔ4312. When analysed for secretion, PghAxc-c-Myc and PghBxc-c-Myc were detected in the culture supernatants of both Xcv strains, suggesting that they were secreted but independently of the T2S systems (Fig. 7d). We therefore conclude that T2S substrate recognition differs in Xcv and Xcc.
Table 3. Known type II secretion (T2S) substrates from Xanthomonas spp.
Secretion in Xcv2
Closest homolog in Xcv Protein
1Names of gene products are given in brackets. The accession number of CysP2 from Xoo T7174R is not published. The accession number of the homologous protein from Xoo PXO99A is PXO_04730.
2For secretion assays in Xcv, genes were cloned into vector pBRM and expressed in Xcv wild-type and T2S mutant strains (see Fig. 7). na, not analysed.
3Secretion of homologous proteins from Xcv was analysed in wild-type and T2S mutant strains (see Fig. 7 and Table 2).
The xylanase XynC is secreted by the Xps system and contributes to virulence
In contrast to most candidate substrates tested, secretion of the predicted xylanase XCV0965 (hereafter referred to as XynC) was severely reduced in strain 85-10ΔxpsEΔxcsEΔ4312 when compared with the wild-type strain (Fig. 7a). Secretion assays with single deletion mutants revealed that secretion of XynC-c-Myc was dependent on xps but not on xcs genes (Fig. 7e). To investigate the contribution of the predicted xylanase XynC to extracellular xylanase activity and bacterial virulence, we deleted the corresponding gene from the genome of Xcv strain 85-10. When strains 85-10 and 85-10ΔxynC were grown on NYG agar plates containing RBB xylan, as described earlier, extracellular xylanase activity was severely reduced in strain 85-10ΔxynC (Fig. 8a). We also analysed a potential role of XynC in virulence. When bacteria were inoculated into leaves of susceptible ECW pepper plants, strain 85-10ΔxynC induced reduced disease symptoms compared with the wild-type strain 85-10 (Fig. 8b). Furthermore, deletion of xynC resulted in reduced bacterial counts in planta 4–7 dpi (Fig. 8c). We conclude from these data that XynC is an active xylanase, which is secreted by the Xps system and contributes to bacterial virulence.
Notably, ectopic expression of xynC-c-myc under control of the native or the lac promoter in the wild-type or the xynC deletion mutant led to increased extracellular xylanase activity in vitro, which was already detectable 5–7 hpi, whereas halo formation for the wild-type strain was only observed 24 hpi (Fig. 8d). However, xynC-c-myc expression did not reproducibly restore the virulence of the xynC deletion mutant (data not shown). We speculate that lack of reproducible complementation resulted from the enhanced xylanase activity which might not be favorable for the bacterial interaction with the host (note that xynC is down-regulated by HrpG; see below).
xps genes and T2S substrates are differentially regulated by HrpG and HrpX
Given the potential interplay between T2S and T3S systems, we speculated whether xps genes and xynC are regulated by HrpG and HrpX, which are the key regulators of the T3S system (Wengelnik & Bonas, 1996; Wengelnik et al., 1996b). For this, we performed qRT-PCR analysis of strains 85-10, 85* and 85*ΔhrpX (deleted in the transcriptional activator HrpX) grown in NYG medium using primers specific for xpsE, xpsF and xynC, respectively. The amounts of both xpsE and xpsF transcripts were increased in strain 85* when compared with strains 85-10 and 85*ΔhrpX, suggesting that expression of xpsE and xpsF is enhanced by HrpG and HrpX (Fig. 9a). As control for HrpG/HrpX-dependent gene expression we analysed transcript abundance of the type III effector gene xopC, which is induced by HrpG and HrpX (Noël et al., 2003; Fig. 9a). In contrast to xpsE, xpsF and xopC, however, xynC transcript amounts were significantly reduced in strain 85* compared with strains 85-10 and 85*ΔhrpX, suggesting that expression of xynC is downregulated by HrpG and HrpX (Fig. 9a).
We also analysed potential effects of HrpG and HrpX on extracellular enzyme activities. For this, we performed plate assays as described above with strains 85-10, 85* and 85*ΔhrpX. It was previously shown that extracellular protease activity of Xcv is suppressed by HrpG and HrpX (Noël et al., 2001; Fig. 9b,c). Similarly to protease activity, extracellular xylanase activity was decreased in strain 85* when compared with strains 85-10 and 85*ΔhrpX (Fig. 9b,c). This is in good agreement with the reduced xynC transcript levels in strain 85* (see above). In contrast to protease and xylanase activities, amylase and cellulase activities that are independent of the T2S systems were slightly increased in strain 85* (Fig. 9b,c). Overall, our results suggest that HrpG and HrpX reduce the expression (and hence the extracellular activity) of Xps substrates and thus differentially regulate expression of components and substrates of the Xps system.
T2S systems are important virulence factors of many plant pathogenic bacteria including Erwinia chrysanthemi, Erwinia carotovora, Pseudomonas syringae, Ralstonia solanacearum and Xanthomonas spp. Genome sequence analyses revealed that several Xanthomonas spp., including the tomato and pepper pathogen Xcv, contain two predicted T2S systems, designated Xps and Xcs (Thieme et al., 2005; Lu et al., 2008). In this study, the genetic analysis of both T2S systems from Xcv revealed that xps genes promote disease and bacterial growth in leaves of susceptible pepper plants. By contrast, xcs genes do not significantly contribute to bacterial virulence and secretion of several extracellular enzymes. However, xcs genes partially complement loss of homologous xps genes and therefore presumably encode functional components of a T2S system. To date, a virulence function has not yet been attributed to any Xcs system from Xanthomonas spp. Nevertheless, it is possible that the contribution of Xcs systems to virulence is restricted to a certain stage of the natural infection process that is difficult to monitor under laboratory conditions.
Interestingly, coinoculation experiments revealed that the reduced virulence of the xpsE deletion mutant could phenotypically be rescued by a nonpathogenic Xcv T3S mutant strain, but not by a T3S-T2S double mutant strain. This suggests that the xpsE mutant phenotype resulted from a lack of T2S substrates secreted into the apoplast and not from the absence of nondiffusible factors. It was previously shown that the virulence function of T2S systems from other plant pathogenic bacteria depends on the secretion of extracellular enzymes that presumably degrade components of the plant cell wall, which is a major barrier for invading microbes (Andro et al., 1984; Dow et al., 1987; Ray et al., 2000; Toth et al., 2003; Liu et al., 2005; Sun et al., 2005). Degradation of the plant cell wall might help nutrient acqui-sition and could facilitate the assembly of T3S pili that serve as transport channels for effector proteins across the plant cell wall (Belien et al., 2006; Lagaert et al., 2009). The latter hypothesis is supported by the fact that mutation of the Xps system leads to reduced translocation of type III effectors but does not significantly affect in vitro T3S or the synthesis of T3S system components in planta. Reduced effector protein translocation by xps mutants was not caused by differences in bacterial counts as translocation of effector proteins occurred within 6 hpi when the number of wild-type and xpsE deletion mutant bacteria was similar. It is therefore tempting to speculate that the Xps system promotes effector protein translocation by secretion of degradative enzymes that could facilitate the assembly of extracellular components of the T3S system. However, it remains to be investigated whether additional T2S-dependent factors contribute to the translocation of type III effectors.
To investigate whether the T2S systems from Xcv secrete degradative enzymes, as was reported for T2S systems from other plant pathogens, we performed enzyme activity plate assays and tested individual candidates for secretion. The results revealed that the Xps system is crucial for the activity of extracellular milk protein-degrading proteases. However, the identity of type II-secreted proteases remains to be determined. Genome sequence analysis revealed that Xcv strain 85-10 encodes > 50 predicted extracellular proteases (Thieme et al., 2005). One obvious candidate protease, XCV3669, which is encoded adjacent to the xps gene cluster, is secreted independently of the T2S systems. Similarly to XCV3669, secretion of the predicted proteases XCV0959, XCV0960, XCV3013 and XCV4074 was detected in wild-type and T2S mutant strains and thus depends on alternative secretion systems that are active in complex NYG medium. As halo formation on milk plates was abolished in T2S mutants, we assume that the predicted proteases XCV0959, XCV0960, XCV3669, XCV3013 and XCV4074 do not efficiently degrade milk proteins.
T2S-independent secretion was also observed for the predicted cellulase XCV0031, the endo-polygalacturonase XCV0722 and the glucan 1,4-beta-glucosidase XCV1823. This was unexpected as cellulases, amylases, proteases and polygalacturonases from Xcc, Xac and Xoo including homologs of XCV0722 and XCV3013 were identified as T2S substrates (Ray et al., 2000; Furutani et al., 2004; Chen et al., 2005; Sun et al., 2005; Wang et al., 2008b; Yamazaki et al., 2008; see Tables 2 and 3). Furthermore, we found that two polygalacturonases from Xcc are secreted by Xcv wild-type and T2S mutant strains. Our findings therefore uncovered remarkable differences in the substrate specificity of T2S systems from different pathovars of Xanthomonas spp.
Interestingly, we observed that the Xps system contributes to extracellular xylanase activity. Xylanases are glycoside hydrolases that degrade xylan, which is the most abundant hemicellulose in the cell wall of monocotyledonous plants but also occurs in dicotyledonous plants (Belien et al., 2006). Xylanase activity was not completely abolished in xcs and xps mutants, suggesting that xylanases can also be secreted by an alternative secretion system. We identified the xylanase XynC (XCV0965) as a substrate of the Xps system with a role in virulence. Infection studies revealed that XynC contributes to in planta bacterial multiplication and to disease symptoms. This is in line with the identification of a type II-secreted xylanase as a virulence factor of Xoo (Ray et al., 2000; Rajeshwari et al., 2005; Sun et al., 2005). Deletion of xynC from Xcv led to a reduction of xylanase activity comparable to the effect of xps mutations, suggesting that T2S-dependent extracellular xylanase activity mainly depends on XynC. Notably, the genome of Xcv strain 85-10 encodes at least six predicted xylanases including XynA, XynB2 and XynB3 that are homologous to known and predicted xylanases from other plant pathogenic bacteria (Thieme et al., 2005). It remains to be investigated whether deletion of additional xylanase-encoding genes in Xcv leads to a further reduction of bacterial virulence. It was reported previously that inactivation of single genes for T2S substrates often has a minor effect on bacterial virulence presumably owing to functional redundancies among secreted proteins (Rajeshwari et al., 2005; Jha et al., 2007).
Interestingly, extracellular xylanase activity of Xcv is repressed by HrpG and HrpX, the key regulators of the T3S system. This was confirmed by qRT-PCR analysis of xynC transcript levels. Similarly, extracellular protease activity, which also depends on the Xps system, is repressed by HrpG and HrpX (Noël et al., 2001). We therefore conclude that production of certain T2S substrates from Xcv, including proteases and xylanases, is negatively affected by HrpG and HrpX. Our observations contrast with the finding that synthesis of T2S substrates from Xoo and Xcc is activated by HrpG and HrpX (Furutani et al., 2004; Wang et al., 2008b). Notably, however, HrpG and HrpX from Xcv do not only negatively affect extracellular protease and xylanase activity but also promote expression of xps genes. HrpG/HrpX-dependent activation of xps gene expression was recently also reported for Xac and suggests that T2S and T3S systems are coregulated (Yamazaki et al., 2008). To our knowledge, our findings provide the first experimental evidence for a differential regulation of components and substrates of a T2S system. It remains to be investigated whether HrpG/HrpX-dependent suppression of extracellular enzyme activities promotes the bacterial interaction with the host. It is tempting to speculate that the HrpG/HrpX-dependent reduction of extracellular protease activity during T3S prevents degradation of extracellular components or substrates of the T3S system.
In conclusion, we show that the Xps system contributes to virulence and secretes the xylanase XynC, which is required for disease symptoms and in planta bacterial growth of Xcv. Intriguingly, our data strongly suggest that the virulence function of the Xps system from Xcv might be to facilitate type III-dependent effector protein translocation. Another key finding is the remarkable difference in regulation and identity of T2S substrates in different Xanthomonas spp., which clearly highlights the importance of comparative studies of protein secretion systems in different Xanthomonas pathovars.
We are grateful to S. Marillonnet for helpful suggestions on the Golden Gate cloning strategy, to C. Lorenz for generating pBRM and pBRMhrcN, to J. Hausner for construction of pBRM-P, to C. Kretschmer for sequencing and B. Rosinsky for greenhouse work. Part of this work was supported by grants from the Deutsche Forschungsge-meinschaft, the Sonderforschungsbereich SFB 648 ‘Molekulare Mechanismen der Informationsverarbeitung in Pflanzen’ to DB and UB, and the Federal Ministry of Education and Research (BMBF GenoMik+) to UB.