HpaB from Xanthomonas campestris pv. vesicatoria acts as an exit control protein in type III-dependent protein secretion

Authors


Summary

The hrp (hypersensitive response and pathogenicity) gene cluster of the plant pathogenic bacterium Xanthomonas campestris pv. vesicatoria encodes a type III secretion (TTS) system, which injects bacterial effector proteins into the plant cell. Here, we characterized hpaB (hpa, hrp-associated), which encodes a pathogenicity factor with typical features of a TTS chaperone. We show that HpaB is important for the efficient secretion of at least five effector proteins but is dispensable for the secretion of non-effectors such as XopA and the TTS translocon protein HrpF. GST pull-down assays revealed that HpaB interacts with two unrelated effector proteins, AvrBs1 and AvrBs3, but not with XopA. The HpaB-binding site is located within the first 50 amino acids of AvrBs3. This region also contains the targeting signal for HpaB-dependent secretion, which is missing in HrpF and XopA. Intriguingly, the N-termini of HrpF and XopA target the AvrBs3Δ2 reporter for translocation in a ΔhpaB mutant but not in the wild-type strain. This indicates that HpaB plays an essential role in the exit control of the TTS system. Our data suggest that HpaB promotes the secretion of a large set of effector proteins and prevents the delivery of non-effectors into the plant cell.

Introduction

Gram-negative plant pathogenic bacteria have evolved a variety of strategies to establish a favourable environment in their host plants. The successful infection and colonization of the host tissue often depends on a type III secretion (TTS) system, a specialized protein transport machinery that spans both bacterial membranes and is associated with an extracellular pilus-like structure (Hueck, 1998). The TTS system mediates the secretion of bacterial proteins into the extracellular milieu as well as the translocation of so-called effector proteins directly into the host cell cytosol (He, 1998; Cornelis and Van Gijsegem, 2000). Type III-secreted proteins that are not translocated into the host cell (e.g. extracellular components of the secretion system) will be referred to as non-effectors hereafter.

Bacterial effector proteins presumably interfere with host metabolic pathways to the benefit of the pathogen (Cornelis, 2002; Van’t Slot and Knogge, 2002). However, a number of effectors from plant pathogenic bacteria elicit defence responses in corresponding resistant plants and thus restrict bacterial growth. They were therefore designated avirulence (Avr) proteins (White et al., 2000). Avr protein-triggered defence reactions are mediated by matching plant resistance (R) proteins (Flor, 1971; Dangl and Jones, 2001) that often initiate a hypersensitive response (HR), a rapid, localized and programmed cell death at the infection site (Klement, 1982). In the absence of a corresponding R protein, i.e. in susceptible hosts, several Avr proteins were shown to promote pathogen fitness suggesting that these effectors have a role in both virulence and avirulence (Kjemtrup et al., 2000; Van’t Slot and Knogge, 2002). The mechanisms by which effectors are recognized by the TTS system and targeted for secretion are not well understood. Type III-dependent secretion is presumably mediated by a signal within the N-terminal 15–20 amino acids of the protein (e.g. Lloyd et al., 2001; 2002; Guttman et al., 2002; Petnicki-Ocwieja et al., 2002; Schechter et al., 2004). Furthermore, a signal in the 5′ region of the mRNA is discussed (Anderson and Schneewind, 1997; Anderson et al., 1999; Ramamurthi and Schneewind, 2003). In addition to the secretion signal, effector protein delivery into the host cell often requires a translocation signal that is located within the first 50–100 amino acids (Sory et al., 1995; Schesser et al., 1996; Mudgett et al., 2000; Schechter et al., 2004). Efficient type III-dependent export of several effector proteins from plant and animal pathogens depends on the presence of a corresponding TTS chaperone (Parsot et al., 2003). TTS chaperones are generally small, leucine-rich and acidic proteins that bind to and often stabilize secreted proteins (Feldman and Cornelis, 2003; Parsot et al., 2003). While TTS chaperones were initially identified in animal pathogenic bacteria, their importance for the secretion of effector proteins from plant pathogens was recently also demonstrated in Erwinia amylovora and Pseudomonas syringae (Gaudriault et al. 2002;van Dijk et al., 2002; Badel et al., 2003; Shan et al., 2004; Wehling et al., 2004).

Our laboratory studies the plant pathogen Xanthomonas campestris pathovar (pv.) vesicatoria, which is the causal agent of bacterial spot disease in pepper and tomato. The TTS system of X. campestris pv. vesicatoria is encoded by the hrp (hypersensitive response and pathogenicity) gene cluster, which is essential for the bacterial ability to cause disease in susceptible plants as well as to induce the HR in resistant plants. The hrp gene cluster is located in a 23 kb chromosomal region and is organized in six operons, designated hrpA to hrpF (Bonas et al., 1991; Fenselau et al., 1992; Fenselau and Bonas, 1995; Wengelnik et al., 1996a; Huguet and Bonas, 1997; Huguet et al., 1998; Rossier et al., 2000). hrp gene expression is activated in planta by the products of two regulatory genes, hrpG and hrpX (Schulte and Bonas, 1992). HrpG is a member of the OmpR family of two-component response regulators and controls – in most cases via the AraC-type transcriptional activator HrpX – a large regulon including genes for type III-secreted proteins (Xops, Xanthomonas outer proteins; Wengelnik and Bonas, 1996; Wengelnik et al., 1996a,b; Astua-Monge et al., 2000; Noël et al., 2001; 2002; 2003). The promoter regions of several but not all HrpX-regulated genes contain a PIP (plant-inducible promoter, consensus TTCGC-N15-TTCGC) box motif, which presumably serves as a control element (Büttner and Bonas, 2002a).

Nine hrp genes from X. campestris pv. vesicatoria– designated hrc for hrp conserved – are conserved in different plant pathogenic bacteria and probably encode core components of the TTS system that are also present in animal pathogens (Bogdanove et al., 1996; Lindgren, 1997; Hueck, 1998). The role of most non-conserved hrp genes is not well understood, and only in some cases a possible function of the corresponding gene products has been uncovered. One example is HrpF, which is secreted and required for the translocation of effector proteins (Casper-Lindley et al., 2002; Szurek et al., 2002; Hotson et al., 2003; Noël et al., 2003; Roden et al., 2004). Because of its pore-inducing ability, HrpF was proposed to be part of the predicted TTS translocon, which presumably inserts as a channel-like protein complex into the eukaryotic plasma membrane and mediates protein translocation (Büttner and Bonas, 2002b; Büttner et al., 2002). The hrp gene cluster of X. campestris pv. vesicatoria also contains hpa (hrp-associated) genes, which contribute to the plant–pathogen interaction (Huguet et al., 1998). Here, we studied hpaB, which is located in the hrpE region. hpaB is essential for bacterial growth in planta and contributes to the type III-dependent export of at least five effector proteins. HpaB interacts with the effectors AvrBs3 and AvrBs1 in vitro. Domain swapping experiments revealed that the targeting signal for HpaB-dependent secretion is located within the first 50 amino acids of AvrBs3. By contrast, the non-effector proteins HrpF and XopA do not contain a signal for HpaB-dependent export. However, in vivo translocation assays showed that the N-termini of HrpF and XopA mediate translocation of a reporter protein in the absence of hpaB. Thus, in the wild-type pathogen HpaB does not only promote effector protein secretion but might also prevent the delivery of non-effectors into the plant cell.

Results

hpaB is required for bacterial pathogenicity

Sequence and codon preference analyses of the hrpE region from X. campestris pv. vesicatoria revealed the presence of an open reading frame, designated hpaB, which is located downstream of hrpE1. hpaB encodes a predicted protein of 162 amino acids with an isoelectric point of 4.3 and a leucine content of 13.6%. Homologous proteins are encoded in the hrp gene clusters of other xanthomonads as well as of the plant pathogens Ralstonia solanacearum and Burkholderia pseudomallei(Table 1).

Table 1. Characteristics of HpaB.
ProteinCharacteristicsHomologous proteins (organism; accession number)aIdentity/Similarity
  • a

    . Homologous proteins were found using the blastp program (http://www.ncbi.nlm.nih.gov/blast/) in non-redundant databases based on the matrix BLOSUM 62.

  • aa, amino acids, refers to the predicted size of the encoded protein; pI, isoelectric point.

HpaB162 aa; pI = 4.3; 13.6% leucineHpaB (X. campestris pv. vesicatoria; AAD21328)100%/100%
HpaB (Xanthomonas oryzae pv. oryzae; AAS48655) 95%/98%
HpaB (Xanthomonas axonopodis pv. citri; AAM35287) 93%/98%
HpaB (Xanthomonas axonopodis pv. glycines; AAP34355) 93%/98%
HpaB (X. campestris pv. campestris; AAM40518) 85%/91%
Putative type III secretory pathway protein  (Burkholderia pseudomallei; AAT44860) 52%/68%
Conserved hypothetical protein (Ralstonia  solanacearum; CAD18004) 50%/65%

In X. campestris pv. vesicatoria, the analysis of Tn3-gus insertion mutants in strain 85-10 revealed that disruption of hpaB leads to a loss of both disease symptom formation in susceptible plants and the HR induction in resistant plants (Bonas et al., 1991). To exclude that the phenotypes were attributed to a polar effect of the transposon on genes located downstream of hpaB, we created a hpaB deletion mutant in X. campestris pv. vesicatoria strains 85-10 and 85* respectively. Strain 85-10 expresses the avrBs1 gene and induces the HR in pepper line Early Cal Wonder (ECW)-10R, which contains the corresponding R gene Bs1 (Minsavage et al., 1990). Strain 85* is a derivative of strain 85-10 that expresses a constitutively active version of the regulator HrpG, HrpG*, which allows expression of the hrp regulon in vitro (Rossier et al., 1999; Wengelnik et al., 1999). Infection studies showed that strain 85-10ΔhpaB was severely reduced in growth and disease symptom formation in the susceptible pepper line ECW and did not induce the HR in ECW-10R plants (Fig. 1A and B). However, strain 85*ΔhpaB was able to induce a partial and delayed HR in ECW-10R plants indicating that constitutive hrp gene expression can partially rescue the ΔhpaB mutant phenotype. Yet, no disease symptoms appeared when this strain was infiltrated into susceptible  plants  (Fig. 1A).  The  phenotypes  of  strains 85-10ΔhpaB and 85*ΔhpaB could be complemented by plasmid pLhrpE, which carries the hrpE region downstream of a lac promoter. By contrast, expression of hpaB alone under the control of a triple lacUV promoter only partially complemented the phenotype of strain 85*ΔhpaB but not of strain 85-10ΔhpaB (Fig. 1A, plasmid pDhpaB). We speculate that overexpression of hpaB from plasmid pDhpaB exerts a negative effect on pathogenicity (see below) and therefore does not complement the phenotype of ΔhpaB mutants. Consistent with this hypothesis, strain 85*(pDhpaB) induced a reduced HR in resistant pepper plants when compared to strain 85* carrying the empty vector (infiltration at 4 × 106 cfu ml−1). By contrast, no reduction in the HR was observed upon infiltration of strain 85*(pLhrpE) (data not shown).

Figure 1.

hpaB encodes an important pathogenicity factor.
A. Phenotypes of X. campestris pv. vesicatoria strains 85-10 and 85* and the corresponding hpaB deletion mutants in susceptible (ECW) and resistant (ECW-10R) pepper plants. Bacteria were infiltrated into leaves at 4 × 108 cfu ml−1. Reactions were scored over a period of 3 days. Plasmid pDhpaB expresses hpaB under the control of a triple lacUV promoter. Plasmid pLhrpE carries the hrpE region downstream of the lac promoter. HR, hypersensitive reaction; hr, delayed and reduced HR; (hr), partial, delayed and reduced HR; WS, water-soaked lesions; ws, delayed and reduced water-soaked lesions.
B. Bacterial growth of strains 85-10, 85-10ΔhpaB and the hrp mutant 85-10ΔhrpG in the susceptible pepper line ECW. Bacteria were infiltrated into leaves at 104 cfu ml−1. Growth was monitored over a period of 8 days. Values are the mean of three samples from three different plants taken at each time point. The result of one representative experiment is shown.

HpaB is not secreted by the TTS system

To study the expression of hpaB we performed reverse transcription polymerase chain reaction (RT-PCR) analyses of X. campestris pv. vesicatoria strains 85-10, 85* and 85*ΔhrpX, which is deleted in the regulatory gene hrpX. The hpaB transcript was amplified in strain 85* but not in strains 85-10 and 85*ΔhrpX(Fig. 2A), indicating that expression of hpaB is controlled by HrpG and HrpX. Sequence analysis did not reveal the presence of a PIP box motif in the region upstream of hpaB. For protein analysis, the hpaB gene product was tagged with a c-myc epitope at the C terminus and expressed in X. campestris pv. vesicatoria strain 85*. When bacteria were incubated in secretion medium, a protein of the expected size was detected by Western blot analysis in total cell extract, but not the culture supernatant (Fig. 2B). The blot was reincubated with an antibody against the type III-secreted protein HrpF. As HrpF was detected in the culture supernatant, we conclude that type III secretion had occurred but could not be monitored for HpaB-c-myc. Interestingly, strain 85* expressing hpaB-c-myc secretes less HrpF when compared to strain 85* carrying the empty vector. Similar results were obtained with strain 85* expressing hpaB without epitope tag (Fig. 2B), suggesting that overexpression of hpaB interferes with the efficiency of in vitro type III secretion. To investigate whether overexpression of hpaB also affects the amounts of secreted effector proteins, we introduced hpaB expression constructs into strain 82*. This strain is a hrpG* derivative of strain 82-8 and contains the effector protein AvrBs3, which can be detected by a corresponding polyclonal antibody (Knoop et al., 1991). Figure 2B shows that AvrBs3 secretion was reduced in hpaB-overexpressing strains when compared to strain 82* carrying the empty vector.

Figure 2.

HpaB is coexpressed with the TTS system but presumably not secreted.
A. RT-PCR analysis of hpaB in X. campestris pv. vesicatoria strains 85-10, 85* and 85*ΔhrpX grown in NYG medium. RT-PCR amplicons were separated on a 1% agarose gel and stained by ethidium bromide. 16S ribosomal DNA (rDNA) was amplified as a constitutive control.
B. In vitro secretion analysis of X. campestris pv. vesicatoria expressing HpaB. In the upper two panels, HrpF and HpaB from strain 85* carrying pDSK602 (vector) or expressing HpaB-c-myc or HpaB (no c-myc tag) from plasmids pDMhpaB and pDhpaB, respectively, were analysed. The lower panel shows the analysis of AvrBs3, an endogenous protein in strain 82* carrying pDSK602 (vector) or hpaB expression constructs as above. Equal amounts of total cell extracts (TE) and culture supernatants (SN), respectively, were analysed by SDS-PAGE and immunoblotting, using the anti-c-myc antibody and antibodies directed against HrpF and AvrBs3. The smaller proteins detected by the anti-HrpF antibody are HrpF degradation products (Büttner et al., 2002).

The data presented above indicate that overexpression of hpaB exerts a negative effect on in vitro type III secretion. We can therefore not exclude that HpaB itself interferes with its own secretion. To further test whether HpaB is secreted by the TTS system, we performed a second assay using a fusion between the N terminus of HpaB and AvrBs3Δ2. AvrBs3Δ2 lacks amino acids 2–152 and is no longer delivered by the TTS system but still induces the HR when expressed in Bs3 pepper plants (Szurek et al., 2002). AvrBs3Δ2 was therefore established as a reporter for type III-dependent protein secretion and translocation (Noël et al., 2003). Here, we fused the first 65 codons of hpaB to avrBs3Δ2 and expressed the resulting fusion construct in X. campestris pv. vesicatoria strain 85*. When bacteria were incubated in secretion medium, HpaB1-65-AvrBs3Δ2 was not detected in the culture supernatant (data not shown). Furthermore, bacteria expressing HpaB1-65-AvrBs3Δ2 did not induce the HR in AvrBs3-responsive ECW-30R pepper plants (Minsavage et al., 1990), indicating that the fusion protein was not translocated into plant cells. Taken together, these data support our hypothesis that HpaB is not secreted by the TTS system.

HpaB specifically contributes to the secretion of different effector proteins

To better understand the role of hpaB in bacterial pathogenicity, we investigated the effect of the hpaB deletion on in vitro type III secretion. Strains 82* and 82*ΔhpaB were incubated in secretion medium and total cell extracts and culture supernatants were analysed by immunoblotting using antibodies directed against HrpF and AvrBs3. While secretion of HrpF was identical, AvrBs3 was only detected in the culture supernatant of the wild-type but not the ΔhpaB mutant strain (Fig. 3A). However, the amounts of AvrBs3 in total cell extracts of both strains were comparable. The lack of detectable AvrBs3 secretion in strain 82*ΔhpaB could be complemented by plasmid pLhrpE (Fig. 3B), which also complemented the virulence defect (see above, Fig. 1A). Next, we analysed secretion of additional effector proteins in ΔhpaB mutants. For this, AvrBs1, AvrBsT, both c-myc-tagged, and XopC1-200- and XopJ1-155-AvrBs3Δ2 reporter fusions were expressed from corresponding expression constructs in wild-type and ΔhpaB mutant strains. XopC and XopJ are recently identified effector proteins that contain a secretion and translocation signal within the first 200 and 155 amino acids respectively (Noël et al., 2003). While AvrBs1, AvrBsT, XopC1-200-AvrBs3Δ2 and XopJ1-155-AvrBs3Δ2 were present in bacterial total cell extracts of the ΔhpaB mutant, they were not or hardly visible in the culture supernatant (Fig. 3A).

Figure 3.

hpaB specifically contributes to the secretion of effector proteins.
A. The in vitro secretion of effector and non-effector proteins was analysed in wild-type (wt; 85* and 82*) and the corresponding hpaB deletion mutant (ΔhpaB) strains. The following strains were used: 82* and 82*ΔhpaB for secretion of AvrBs3 and HrpF, 85*(pDSM100) and 85*ΔhpaB(pDSM100) for AvrBs1-c-myc secretion, 82*(pDSM440) and 82*ΔhpaB(pDSM440) for AvrBsT-c-myc secretion, 85*(pL6xopC356) and 85*ΔhpaB(pL6xopC356) for XopC1-200-AvrBs3Δ2 secretion, 85*(pL6xopJ356) and 85*ΔhpaB(pL6xopJ356) for XopJ1-155-AvrBs3Δ2 secretion and 85* and 85*ΔhpaB for XopA secretion. Bacterial total cell extracts (TE) and culture supernatants (SN) were analysed by SDS-PAGE and immunoblotting, using a monoclonal anti-c-myc antibody and polyclonal antisera directed against AvrBs3, HrpF and XopA respectively. Blots were also reacted with the antiserum against the intracellular protein HrcN to ensure that no bacterial lysis had occurred. The result of one representative experiment is shown.
B. Complementation of the AvrBs3 secretion deficiency of strain 82*ΔhpaB by plasmid pLhrpE. Strains 82*, 82*ΔhpaB and 82*ΔhpaB(pLhrpE) were incubated in secretion medium and bacterial total extracts (TE) and culture supernatants (SN) were analysed as in (A), using polyclonal antisera directed against AvrBs3 and HrcN.

As deletion of hpaB did not affect the protein amounts of components of the TTS system such as HrcN, HrpF (Fig. 3A) and HrcC (data not shown), the strong reduction of type III secretion in ΔhpaB mutant strains cannot be attributed to reduced expression of the TTS system itself. We therefore speculate that HpaB specifically promotes the secretion of various effector proteins but that it is dispensable for the secretion of non-effectors such as HrpF. To confirm this hypothesis, we also analysed the secretion of XopA. XopA was suggested to be involved in the translocation process because it is essential for full virulence and avirulence but dispensable for type III secretion in vitro (Noël et al., 2002). Figure 3A shows that XopA is present in similar amounts in the culture supernatants of both strains 85* and 85*ΔhpaB. This indicates that HpaB is not required for the secretion of XopA.

The signal for HpaB-dependent secretion is located in the N terminus of AvrBs3

To investigate whether effector proteins are targeted for HpaB-dependent secretion by their N-terminal secretion signal, we performed domain swapping experiments. For this, we generated a fusion protein between the first 50 amino acids of AvrBs3 and the HrpF derivative HrpFΔN, which lacks the first 152 amino acids and is no longer secreted by the TTS system (Büttner et al., 2002). To avoid detection of the native HrpF protein by the polyclonal anti-HrpF antibody, the fusion protein AvrBs31-50-HrpFΔN was expressed in hrpF deletion mutants. Figure 4A shows that AvrBs31-50-HrpFΔN was detected in the culture supernatant of strain 85*ΔhrpF but not of strain 85*ΔhpaBΔhrpF. This observation indicates that the targeting signal for HpaB-dependent secretion is located within the first 50 amino acids of AvrBs3. We also performed the reciprocal experiment, using AvrBs3Δ2 as a C-terminal fusion partner for the N terminus of HrpF. Because a fusion between the first 50 amino acids of HrpF and AvrBs3Δ2 was unstable, we fused the first 200 codons of hrpF to avrBs3Δ2 and expressed the resulting fusion protein in X. campestris pv. vesicatoria strains 85* and 85*ΔhpaB. HrpF1-200-AvrBs3Δ2 was detected in culture supernatants of both the wild-type and the ΔhpaB mutant strain (Fig. 4B) indicating that the first 200 amino acids of HrpF target AvrBs3Δ2 for secretion even in the absence of HpaB. Secretion of the fusion protein in ΔhpaB mutants was clearly dependent on the TTS system, as HrpF1-200-AvrBs3Δ2 was not detected in the culture supernatant of strain 85*ΔhrcVΔhpaB, which does not express a functional TTS system (data not shown). We also fused the first 51 and 100 codons, respectively, of xopA to avrBs3Δ2 and expressed the resulting fusions under the control of the xopA promoter in both strains 85* and 85*ΔhpaB. Figure 4C shows that XopA1-51-AvrBs3Δ2 and XopA1-100-AvrBs3Δ2 were secreted by both the wild-type and the ΔhpaB mutant strain. Interestingly, secretion of XopA1-51-AvrBs3Δ2 was significantly increased in the absence of hpaB. By contrast, similar amounts of XopA1-100-AvrBs3Δ2 were detected in the culture supernatants of both strains. At this point, we can only speculate that the increase in secretion of XopA1-51-AvrBs3Δ2 in ΔhpaB mutants is attributed to an intrinsic property of the fusion protein. In summary, our data show that the N-termini of HrpF and XopA target AvrBs3Δ2 for secretion in a ΔhpaB mutant. Thus, HrpF and XopA contain a secretion signal that is HpaB-independent.

Figure 4.

The targeting signal for HpaB-dependent secretion is located in the N terminus of AvrBs3.
A. In vitro secretion analysis of strains 85*ΔhrpF and 85*ΔhpaBΔhrpF expressing AvrBs31−50-HrpFΔN. Total cell extracts (TE) and culture supernatants (SN) were analysed by immunoblotting, using antibodies directed against HrpF and HrcN.
B. In vitro secretion analysis of strains 85* and 85*ΔhpaB expressing HrpF1-200-AvrBs3Δ2. Total cell extracts (TE) and culture supernatants (SN) were analysed as described in (A), using antibodies directed against AvrBs3 and HrcN.
C. In vitro secretion analysis of strains 85* and 85*ΔhpaB expressing XopA1-51-AvrBs3Δ2 and XopA1-100-AvrBs3Δ2, respectively, under the control of the xopA promoter. Total cell extracts (TE) and culture supernatants (SN) were analysed as described in (A), using antibodies directed against AvrBs3 and HrcN.

HpaB has a negative effect on the delivery of non-effectors into the plant cell

Previously, inoculation of X. campestris pv. vesicatoria strain 85* expressing HrpF1-200-AvrBs3Δ2 and HrpF1-387-AvrBs3Δ2, respectively, into pepper ECW-30R plants did not induce the HR (Büttner et al., 2002; Fig. 5). We concluded that the N terminus of HrpF does not contain a functional translocation signal. Similar results were obtained with strain 85* expressing XopA1-51-AvrBs3Δ2 or XopA1-100-AvrBs3Δ2 (Fig. 5), suggesting that also XopA belongs to the class of non-effectors that are secreted into the extracellular space but are not translocated into the plant cell.

Figure 5.

The N termini of HrpF and XopA contain functional translocation signals. X. campestris pv. vesicatoria strains 85* and 85-10 (wild-type) and corresponding ΔhpaB mutant (ΔhpaB) strains expressing AvrBs3, AvrBs3Δ2 or a fusion to the AvrBs3Δ2 reporter were inoculated into leaves of pepper ECW-30R plants. Proteins were expressed under the control of the lac (plac) or the xopA (pxopA) promoter as indicated. Proteins that were delivered by hrpG* strains are marked with an asterisk. Three days after inoculation, leaves were bleached in ethanol for a better visualization of the HR. The dashed lines represent the infiltrated areas.

Surprisingly, however, HrpF- and XopA-fusions to AvrBs3Δ2 induced the HR in ECW-30R plants when delivered by the ΔhpaB mutant strain (Fig. 5, right side). The HR induction was clearly dependent on the R gene Bs3 because the fusion proteins were not recognized in ECW plants that lack Bs3. Furthermore, the HR elicitation was dependent on the TTS system because the double mutant 85*ΔhrcVΔhpaB expressing XopA1-51-AvrBs3Δ2 and HrpF1-200-AvrBs3Δ2, respectively, did not induce the HR in ECW-30R plants (data not shown).

We also studied the effect of different promoters as well as the use of hrpG wild-type and hrpG* strains on reporter protein translocation. In the case of XopA1-51-AvrBs3Δ2, translocation of the protein by the ΔhpaB mutant strain occurred irrespective of whether expression was driven from the lac or the xopA promoter (Fig. 5). However, a slight reduction of the HR was observed when HrpF- and XopA-AvrBs3Δ2 fusion proteins were delivered by strain 85-10ΔhpaB compared with strain 85*ΔhpaB (shown for HrpF1-387-AvrBs3Δ2 in Fig. 5). These results indicate that hrpG*, as expected, promotes protein translocation. Taken together, our data suggest that the N-termini of the non-effector proteins HrpF and XopA contain type III export signals that target the AvrBs3Δ2 reporter for translocation in the absence of HpaB.

HpaB specifically binds to AvrBs3 and AvrBs1

To investigate whether HpaB is able to interact with proteins that are secreted by the TTS system, we performed glutathione S-transferase (GST) pull-down assays. For this, GST and GST-fusion proteins of AvrBs3, AvrBs1 and XopA, respectively, were expressed in Escherichia coli, immobilized on glutathione sepharose and incubated with an E. coli lysate containing HpaB-c-myc. We did not test the interaction of HpaB with HrpF, because HrpF is a sticky protein (D. Büttner and U. Bonas, unpubl. data). HpaB-c-myc specifically bound to GST-AvrBs3 and GST-AvrBs1, but not to GST alone or to GST-XopA (Fig. 6A). This shows that HpaB interacts with different effector proteins in vitro but that it is not able to bind to the non-effector protein XopA.

Figure 6.

HpaB interacts with AvrBs3 and AvrBs1 in vitro.
A. GST, GST-AvrBs3, GST-AvrBs1 and GST-XopA were immobilized on glutathione sepharose beads and incubated with E. coli lysates containing HpaB-c-myc. Total cell lysates (total protein) and proteins that were eluted from the beads (eluate) were analysed by SDS-PAGE and immunoblotting, using an anti-c-myc antibody. The blot was also reacted with an anti-GST antibody to visualize GST-fusion proteins. Signals corresponding to HpaB-c-myc and GST-fusion proteins are indicated by arrows.
B. HpaB binds to the first 50 amino acids of AvrBs3. GST, GST-AvrBs31−50 and GST-AvrBs3Δ2 were immobilized on glutathione sepharose beads and incubated with E. coli lysates containing HpaB-c-myc. Total cell lysates (total protein) and eluted proteins (eluate) were analysed as described in (A).

To test whether the HpaB binding site is located within the N-terminal 50 amino acids of AvrBs3, which harbour the signal for HpaB-dependent secretion (Fig. 4A), we performed a GST pull-down assay with HpaB-c-myc and GST-AvrBs31−50, which contains the first 50 amino acids of AvrBs3 fused to GST. Figure 6B shows that HpaB-c-myc bound to GST-AvrBs31−50, but not to GST alone. Because there was no interaction between HpaB-c-myc and GST-AvrBs3Δ2 (Fig. 6B), we conclude that the HpaB-binding site is located within the N-terminal 50 amino acids of AvrBs3.

Discussion

hpaB encodes an essential pathogenicity factor

In this study, we found that HpaB is involved in the exit control of type III-dependent protein secretion in X. campestris pv. vesicatoria. Infection studies revealed that deletion of hpaB still allows a partial HR induction in resistant plants but abolishes disease symptom formation and bacterial growth in susceptible plants. Thus, HpaB is an essential pathogenicity factor. The fact that hpaB, which is coexpressed with the TTS system, is conserved among different xanthomonads and the plant pathogens R. solanacearum and B. pseudomallei suggests that it is also involved in other plant–pathogen interactions. Consistent with this hypothesis is the finding that a Xanthomonas axonopodis pv. glycines hpaB mutant is no longer pathogenic (Kim et al., 2003).

HpaB promotes the secretion of several effector proteins

HpaB from X. campestris pv. vesicatoria most likely acts inside the bacterial cell. The finding that overexpression of hpaB leads to a reduction of type III secretion (Fig. 2B) and bacterial pathogenicity suggests that HpaB is involved in the regulation of type III-dependent protein export. This hypothesis is supported by the fact that deletion of hpaB severely reduces in vitro secretion of five different effector proteins analysed to date. Furthermore, preliminary studies suggest that six additional effectors that were recently identified in X. campestris pv. vesicatoria are affected in type III-dependent export in ΔhpaB mutant strains (J. Stuttmann, F. Thieme and D. Büttner, unpubl. data). We therefore conclude that HpaB is involved in the type III-dependent export of a large set of effector proteins. Consistent with this hypothesis is the finding that ΔhpaB mutants are no longer pathogenic in susceptible plants. Obviously, there is a threshold of the amount of effector proteins that need to be translocated into the host cell in order to promote bacterial growth and disease symptom formation.

Interestingly, deletion of hpaB does not affect the in vitro secretion of HrpF and XopA (Fig. 3A), which are presumably part of the extracellular components of the TTS system. It is therefore tempting to speculate that the secretion deficiency of ΔhpaB mutants is not attributed to defects in the TTS system. In agreement with this assumption, the expression of TTS components as well as the morphology of the extracellular Hrp pilus are unaltered in ΔhpaB mutants compared to the wild-type strain (E. Weber, T. Ojanen-Reuhs, R. Koebnik and U. Bonas, unpubl. data). In summary, our data suggest that HpaB specifically promotes the type III-dependent effector protein export but that it is not required for the secretion of non-effectors. To our knowledge, a protein with such a role in type III-dependent secretion has to date not been described for any other plant pathogenic bacterium.

The signal for HpaB-dependent secretion in AvrBs3 also harbours the HpaB-binding site

Type III secretion signals have been localized in the N-terminal 15–20 amino acids of proteins from plant and animal pathogenic bacteria (e.g. Lloyd et al., 2001; 2002; Guttman et al., 2002; Petnicki-Ocwieja et al., 2002; Schechter et al., 2004). In the case of AvrBs3 from X. campestris pv. vesicatoria, the minimal secretion signal is not yet known. Domain swapping experiments revealed that a secretion signal is located within the first 50 amino acids of AvrBs3 (Fig. 4A). This region also contains the signal for HpaB-dependent secretion. The finding that the N-terminal 50 amino acids of AvrBs3 provide a binding site for HpaB, as shown by GST pull-down assays (Fig. 6), corroborates these conclusions. It is therefore tempting to speculate that binding of HpaB to AvrBs3 is a prerequisite for efficient secretion of this effector protein. Intriguingly, the signal for HpaB-dependent secretion is not essential for secretion of the AvrBs3 protein per se because the N-termini of HrpF and XopA, respectively, can restore secretion of the AvrBs3Δ2 reporter in ΔhpaB mutant strains. Thus, it is the signal for HpaB-dependent secretion itself that determines the need of AvrBs3 for HpaB.

Does HpaB act as a TTS chaperone?

HpaB is a small, acidic and leucine-rich protein and thus exhibits typical features of TTS chaperones (Feldman and Cornelis, 2003). TTS chaperones bind to translocon or effector proteins and often contribute to the stability and/or secretion of their interaction partners (Parsot et al., 2003). As HpaB binds to AvrBs3 as well as to the non-related effector protein AvrBs1 (Fig. 6) and promotes the secretion of both effectors, it might act as a TTS chaperone. Similarly to the TTS chaperones ShcA, ShcM and ShcV from P. syringae (van Dijk et al., 2002; Badel et al., 2003; Wehling et al., 2004), HpaB does not seem to stabilize its interaction partners. This is in contrast to most known TTS chaperones from animal pathogenic bacteria (Feldman and Cornelis, 2003; Parsot et al., 2003).

TTS chaperones of effector proteins from animal pathogens were grouped into two main classes (Page and Parsot, 2002; Parsot et al., 2003). Class IA chaperones bind to one effector or to a conserved binding site in several effectors. Members of this class were recently also identified in the plant pathogenic bacteria P. syringae and E. amylovora (van Dijk et al., 2002; Gaudriault et al. 2002;Badel et al., 2003; Shan et al., 2004; Wehling et al., 2004). By contrast, class IB chaperones interact with multiple effectors that do not share any sequence homology (Page and Parsot, 2002; Parsot et al., 2003). Known examples for the chaperone class IB are Spa15 from Shigella flexneri and its homologue InvB from Salmonella enterica (Page et al. 2002;Ehrbar et al., 2003; 2004; Lee and Galan, 2003). It is tempting to speculate that also HpaB from X. campestris pv. vesicatoria acts as a class IB chaperone because it binds to two effector proteins that do not share any amino acid sequence similarity. Thus, the HpaB-binding site appears to be not conserved on the amino acid level. Experiments are needed to clarify whether HpaB also interacts with additional effector proteins. The finding that HpaB promotes the secretion of at least five effectors prompts us to speculate that it has a broad substrate specificity.

HpaB prevents the translocation of non-effectors

Surprisingly, we found that the N-termini of the non-effectors HrpF and XopA target the AvrBs3Δ2 reporter for translocation in a ΔhpaB mutant, but not in the wild-type. This indicates that HrpF and XopA do not only contain a secretion but also a translocation signal, but that translocation is normally inhibited by HpaB. To our knowledge, this is the first hint that non-effectors can in principle be delivered across the plant plasma membrane. The mechanism by which HpaB prevents the translocation of non-effector fusions remains to be investigated. As we could not detect a direct interaction between HpaB and XopA (Fig. 6A), it is possible that HpaB binds to non-effectors via a to date unknown linker protein. Alternatively, HpaB might confer a competitive advantage to effectors for their type III-dependent export and thus indirectly prevent the delivery of non-effectors into the plant cell. Such a scenario has been described for several TTS chaperones from animal pathogenic bacteria (Boyd et al., 2000; Wulff-Strobel et al., 2002; Thomas and Finlay, 2003).

Genomic approaches will foster the analysis of the complete repertoire of type III effectors in X. campestris pv. vesicatoria (Büttner et al., 2003). It will then be possible to address the question whether HpaB acts as a general exit control protein as suggested by the data presented here. It is puzzling why overexpression of hpaB leads to a reduction of in vitro type III secretion (Fig. 2B). One possible explanation is that elevated levels of HpaB interfere with the availability of another factor that is crucial for secretion. Protein interaction studies should clarify whether HpaB interacts with components of the secretion machinery as was recently reported for the TTS chaperones CesT from enteropathogenic E. coli and FlgN from Salmonella typhimurium (Gauthier and Finlay, 2003; Thomas et al., 2004).

Experimental procedures

Bacterial strains and growth conditions

Bacterial strains and plasmids used in this study are described in Table 2. E. coli cells were cultivated at 37°C in Luria-Bertani (LB) or Super medium (Qiagen, Hilden, Germany). X. campestris pv. vesicatoria strains were grown at 30°C in NYG medium (Daniels et al., 1984) or in minimal medium A (Ausubel et al., 1996) supplemented with sucrose (10 mM) and casamino acids (0.3%). Plasmids were introduced into E. coli by electroporation and into X. campestris pv. vesicatoria by conjugation, using pRK2013 as a helper plasmid in triparental matings (Figurski and Helinski, 1979). Antibiotics were added to the media at the following final concentrations: ampicillin, 100 µg ml−1; kanamycin, 25 µg ml−1; rifampicin, 100 µg ml−1; spectinomycin, 100 µg ml−1; tetracycline, 10 µg ml−1.

Table 2. Published strains and plasmids used in this study.
 Relevant characteristicsReference or source
  1. Ap, ampicillin; Km, kanamycin; Rif, rifampicin; Sm, spectinomycin; Tc, tetracycline; r, resistant.

X. campestris pv. vesicatoria
 85-10Pepper-race 2; wild-type; carries avrBs1; Rifr Canteros (1990)
 75-3Tomato race 1; wild-type; Rifr Minsavage et al. (1990)
 85*85-10 derivative containing the hrpG* mutation Wengelnik et al. (1999)
 82*82-8 derivative containing the hrpG* mutation; carries avrBs3 Wengelnik et al. (1999)
 85-10ΔhrpG hrpG deletion mutant of 85-10 Wengelnik et al. (1999)
E. coli
 DH5αFrecA hsdR17(rk,mk+) Φ80dLacZ ΔM15Bethesda Research Laboratories, Bethesda, MD
 DH5αλpirFrecA hsdR17(rk,mk+) Φ80dLacZ ΔM15 [λpir] Ménard et al. (1993)
 BL21(DE3)FompT hsdS20(rbmb) galStratagene, Heidelberg, Germany
Plasmids
 pBlueskript(II) KSPhagemid, pUC derivative; AprStratagene
 pUC119ColE1 replicon; Apr Vieira and Messing (1987)
 pC3003pUC19 plasmid containing a triple c-myc tag; AprJ. Kämper
 pDSK602Broad-host-range vector; contains triple lacUV5 promoter; Smr Murillo et al. (1994)
 pDSK604Derivative of pDSK602 with modified polylinker Escolar et al. (2001)
 pGEX-2TKMptac GST lacIq pBR322 ori; Apr, derivative of pGEX-2TK with polylinker of pDSK604Stratagene Escolar et al. (2001)
 pOK1 sacB sacQ mobRK2 oriR6K; Smr Huguet et al. (1998)
 pRK2013ColE1 replicon, TraRK+ Mob+ Kmr Figurski and Helinski (1979)
 pDSM100pDSK602 expressing AvrBs1-c-myc Escolar et al. (2001)
 pDSM440pDSK602 expressing AvrBsT-c-myc under the control of its own promoter Escolar et al. (2001)
 pDS300FpDSK602 expressing AvrBs3; FLAG-tagged Van den Ackerveken et al. (1996)
 pUS356FpUC119 containing avrBs3Δ2 fused to a FLAG epitope-coding sequence Szurek et al. 2002)
 pDS356FpDSK602 expressing AvrBs3Δ2, which is deleted in amino acids 2–153; FLAG-tagged Szurek et al. (2002)
 pDhrpFN356pDSK602 expressing a fusion protein between the N-terminal 386 amino acids from HrpF and AvrBs3Δ2 Büttner et al. (2002)
 pL6xopC356pLAFR6 expressing a fusion protein between the N-terminal 200 amino acids of XopC and AvrBs3Δ2 under control of the xopC promoter Noël et al. (2003)
 pL6xopJ356pLAFR6 expressing a fusion protein between the N-terminal 155 amino acids of XopC and AvrBs3Δ2 under control of the xopJ promoter Noël et al. (2003)
 pDhrpFΔNpDSK602 expressing HrpF deleted in the first 152 amino acids Büttner et al. (2002)
 pG100pGEX-2TKM expressing GST-AvrBs1L. Escolar and U. Bonas, unpubl. data
 pG300pGEX-2TKM expressing GST-AvrBs3 Szurek et al. (2001)

Plant material and plant inoculations

The near-isogenic pepper cultivars ECW, ECW-10R and ECW-30R (Minsavage et al., 1990) were grown and inoculated with X. campestris pv. vesicatoria as described previously (Bonas et al., 1991). Bacteria were hand-infiltrated into the intercellular spaces of fully expanded leaves at concentrations of 4 × 108 cfu ml−1 in 1 mM MgCl2. The appearance of disease symptoms and the HR was scored over a period of 3 days after inoculation. For better visualization of the HR, leaves were bleached in ethanol. For in planta growth curves, bacteria were inoculated at a density of 104 cfu ml−1 into leaves of pepper cultivar ECW. Bacterial growth was determined as described previously (Bonas et al., 1991). Experiments were repeated at least three times.

Sequencing of the hrpE region

The hrpE region was sequenced from X. campestris pv. vesicatoria strain 75-3 (accession number AF056246). Details are available from the authors upon request.

RNA analyses

RNA extraction, cDNA synthesis and RT were performed as described previously (Noël et al., 2001).

Generation of a hpaB deletion mutant

A 420 bp in frame deletion of hpaB (resulting in deletion of amino acids 13–149 of HpaB) was generated by excising a Csp45I fragment from the 3.1 kb hrpE region cloned into the EcoRV/XhoI sites of pBlueskript(II) KS (construct pBhrpE). The resulting 2.7 kb fragment was cloned into the BamHI/SalI sites of the suicide plasmid pOK1. Construct pOKΔhpaB was conjugated into strains 85-10, 85*, 82*, 85*ΔhrcV and 85*ΔhrpF as described (Huguet et al., 1998). Double crossing-overs gave rise to 85-10ΔhpaB, 85*ΔhpaB, 82*ΔhpaB, 85*ΔhrcVΔhpaB and 85*ΔhpaBΔhrpF.

Construction of hpaB expression plasmids

For the generation of a hpaB overexpression construct, hpaB was amplified by PCR with primers hpaB-for (5′-CGAAT TCGTGCATGTCTCACCACAGATC-3′) and hpaB-rev (5′-CGAGCTCGGCGCGTAACCACAGATAGTT-3′) and the resulting PCR product was cloned into the EcoRI/SacI sites of pDSK604, giving pDhpaB. In addition, the hpaB PCR product was cloned into the EcoRI/SacI sites of pC3003, in frame with a triple-c-myc epitope-encoding sequence, thus generating pChpaB. The resulting insert encoding HpaB-c-myc was then introduced into the EcoRI/XhoI sites of pDSK604, giving pDMhpaB.

For complementation of the ΔhpaB mutant, a 2.4 kb EcoRV/SalI fragment spanning the hrpE region was subcloned into the SmaI/SalI sites of pUC119 and then ligated into the EcoRI/HindIII sites of pLAFR3, giving pLhrpE.

Generation of N-terminal domain swapping constructs

To generate the AvrBs31−50-HrpFΔN expression construct, the first 50 codons of avrBs3 were amplified by primers avrBs3-50for (5′-TACTGAATTCATGGATCCCATTCG-3′) and avrBs3-50rev (5′-AGCTAAGCTTCTCGAGGGACATCGTCCG-3′) and cloned into the EcoRI/XhoI sites of pDhrpFΔN (Table 2), giving pD50hrpFΔN.

For the construction of the HrpF1-200-AvrBs3Δ2 expression construct, the first 200 codons of hrpF were amplified by primers HrpF-for (5′-TACTGAATTCGCCTCTATGTCGCTC-3′) and HrpF-200rev (5′-CTGTCGAATTCGATCTTGCCGC CGCACTTG-3′) and cloned into the EcoRI site of plasmid pDS356F (Table 2), giving pDS200F356F.

For the generation of the XopA1-51-AvrBs3Δ2 and the XopA1-100-AvrBs3Δ2 fusion constructs, 680 bp upstream sequence as well as the first 51 codons and the first 100 codons, respectively, of xopA were amplified by PCR using primers xopApromfor (5′-TACTGAATTCGTACCGTTGTTGT TGCGATG-3′) and xopArev51 (5′-AGCTGAATTCCTCGAG GTCCTGCAGAAGC-3′) or xopArev100 (5′-AGCTGAATTC CTCGAGGTTGAGCCCTCCATCG-3′). PCR products were ligated into the EcoRI site of plasmid pUS356F, which contains the avrBs3Δ2 gene (Table 2), giving pU51X356F and pU100X356F respectively. The inserts of both plasmids were then introduced into the EcoRI/HindIII sites of plasmid pLAFR6, giving pL51X356F and pL100X356F respectively. To express XopA1-51-AvrBs3Δ2 under the control of the triple lacUV promoter, the first 51 codons of xopA were amplified by primers xopAfor-1 (5′-TACTGAATTCATGATCAATTC ATTG-3′) and xopArev51 and cloned into the EcoRI site of plasmid pDS356F (Table 2), giving pD51X356F.

Secretion experiments and protein analysis

Secretion experiments were performed as described (Büttner et al., 2002). Equal amounts of bacterial total cell extracts and culture supernatants were analysed by SDS-PAGE and immunoblotting, using polyclonal antibodies against HrpF (Büttner et al., 2002), AvrBs3 (Knoop et al., 1991), HrcN (Rossier et al., 2000) and XopA (Noël, 2001), and monoclonal antibodies directed against the c-myc epitope and GST (Amersham Pharmacia Biotech, Freiburg, Germany) respectively. Horseradish peroxidase-labelled anti-rabbit, anti-mouse and anti-goat antibodies (Amersham Pharmacia Biotech) were used as secondary antibodies. Reactions were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech). Blots were routinely reacted with the antiserum against the intracellular protein HrcN to ensure that no bacterial lysis had occurred.

GST pull-down assays

To construct a GST-XopA fusion protein, xopA was amplified by PCR using primers xopAfor-2 (5′-ACGCGTCGACG GATCCATGATCAATTCATTGAATAC-3′) and xopArev (5′- GGAATTCTTAGTGATGGTGATGGTGATGCTGCA TCGATGC AGTG-3′) and cloned into the BamHI and EcoRI sites of pGEX-2TKM, giving construct pGxopA. To generate a GST-AvrBs31−50 fusion, the first 50 codons of avrBs3 were amplified by PCR using primers avrBs3-50for and avrBs3-50rev and cloned into the EcoRI/XhoI sites of pGEX-2TKM, giving pGavrBs350. For the generation of a GST-AvrBs3Δ2 expression construct, avrBs3Δ2 was excised from construct pUS356F (Table 2), ligated into the EcoRI/HindIII sites of pBluescript(II) KS, thus generating pBS356, and then introduced into the EcoRI/XhoI sites of pGEX-2TKM, giving pG356F.

GST-XopA, GST-AvrBs3, GST-AvrBs1, GST-AvrBs31-50 and GST-AvrBs3Δ2 were expressed in E. coli BL21. Similarly, GST was expressed from vector pGEX-2TK (Amersham Pharmacia Biotech; Table 2) and HpaB-c-myc from construct pDMhpaB in E. coli BL21. Bacterial pellets from 15 ml of cultures were resuspended in phosphate-buffered saline (PBS) and broken by sonication. Insoluble cell debris were removed by centrifugation and soluble proteins were immobilized on a glutathione resin according to the manufacturer's instructions (Amersham Pharmacia Biotech). A total of 40 µl of GST lysate, 400 µl of GST-AvrBs3 lysate, 300 µl of GST-AvrBs1 lysate, 20 µl of GST-XopA lysate, 400 µl of GST-AvrBs31-50 lysate and 400 µl of GST-AvrBs3Δ2 lysate were incubated with 30 µl of resin according to protein stabilities and/or expression levels. Unbound E. coli proteins were removed by washing twice with PBS. After incubation with 650 µl of HpaB-c-myc lysate for 4 h at 4°C, unbound proteins were removed by centrifugation and the resin was washed four times with PBS containing 1% Triton X-100. HpaB-c-myc was eluted with 30 µl of 10 mM reduced glutathione at room temperature for 2 h. Five microlitres of total protein lysates and 20 µl of solution of eluted proteins were analysed by SDS-PAGE and Western blotting, using anti-c-myc and anti-GST antibodies.

Acknowledgements

We are grateful to T. Lahaye for critical reading of the manuscript. We thank A. Landgraf and J. Stuttmann for providing XopA-AvrBs3Δ2 expression constructs. This work was funded by grants from the Deutsche Forschungsgemeinschaft (BO 790/7-1) and the Bundesministerium für Bildung und Forschung (GenoMik) to U.B. and the Fonds der Chemischen Industrie to D.B.

Ancillary