hpaA mutants of Xanthomonas campestris pv. vesicatoria are affected in pathogenicity but retain the ability to induce host-specific hypersensitive reaction

Authors


Ulla Bonas E-mail bonas@genetik.uni-halle.de; Tel. (345) 5526320; Fax (345) 5527151.

Abstract

Xanthomonas campestris pv. vesicatoria is the causal agent of bacterial spot disease on pepper and tomato plants. We reported previously that the main hrp (hypersensitive reaction and pathogenicity) gene cluster in X. c. pv. vesicatoria contains six transcription units, designated hrpA to hrpF. We present here the sequence of the hrpD operon and an analysis of non-polar mutants in each of the six genes. Three genes, hrcQ, hrcR and hrcS, are predicted to encode conserved components of type III protein secretion systems in plant and mammalian pathogenic bacteria. For hrpD5 and hrpD6, homologues have only been found in Ralstonia solanacearum. Interestingly, the hrpD operon contains one gene, hpaA (for hrp-associated), which is specifically required for disease development. hpaA mutants are affected in pathogenicity, but retain in part the ability to induce avirulence gene-mediated, host-specific hypersensitive reaction (HR). In addition, HpaA was found to contain two functional nuclear localization signals, which are important for the interaction with the plant. We propose that HpaA is an effector protein that may be translocated into the host cell via the Hrp secretion pathway.

Introduction

Most Gram-negative phytopathogenic bacteria colonize the intercellular spaces of plants. Whereas some pathogens, such as Erwinia chrysanthemi or E. carotovora, secrete pectic enzymes leading to tissue maceration and necrosis, other bacteria have adopted a more subtle approach: they multiply in the living host tissue for some period before causing severe disease symptoms (Bauer et al., 1994; reviewed in Alfano and Collmer, 1996). In the latter pathogens, essential pathogenicity determinants are encoded by hrp genes (for hypersensitive reaction and pathogenicity genes). Interestingly, hrp mutants have a pleiotropic phenotype, as they are unable to multiply and to cause disease symptoms in susceptible plants, and they are unable to induce the hypersensitive reaction (HR) in resistant or non-host plants. The HR is a rapid, highly localized, defence response, which leads to cell death (necrosis). Most hrp genes are organized in large clusters containing 20 or more genes and have been isolated from many phytopathogenic Gram-negative bacteria, such as Pseudomonas syringae pathovars, Ralstonia solanacearum, Erwinia amylovora and Xanthomonas campestris pathovars (reviewed in Bonas, 1994; He, 1997). Sequence analysis has revealed that a number of hrp genes code for proteins that show high sequence similarity to pathogenicity factors of mammalian bacterial pathogens, such as Yersinia, Shigella and Salmonella spp. (Bogdanove et al., 1996a). These broadly conserved hrp genes have consequently been renamed hrc (for hrp conserved) with an additional capital letter corresponding to the Yersiniaysc homologue (Bogdanove et al., 1996a). The homologous proteins in mammalian pathogens are part of a type III protein secretion system, which secretes and even directly translocates virulence factors into the eukaryotic host cell (reviewed in Cornelis and Wolf-Watz, 1997). Hrp-dependent protein secretion has been demonstrated for the plant pathogens P. syringae, R. solanacearum and E. amylovora. Secreted proteins are, for example, non-specific HR elicitors such as harpins and PopA (Wei et al., 1992; He et al., 1993; Arlat et al., 1994; Bogdanove et al., 1996b), HrpA, the subunit of the P. syringae pv. tomato Hrp pilus (Roine et al., 1997), and DspA, an essential pathogenicity factor of E. amylovora (Gaudriault et al., 1997).

Our research focuses on the interaction between X. c. pv. vesicatoria, the causal agent of bacterial spot disease, and its host plants pepper and tomato. In a susceptible plant, the bacteria multiply in the intercellular space, causing water-soaked lesions, which later become necrotic. In X. c. pv. vesicatoria, a 23 kb chromosomal region contains six hrp transcription units, hrpA to hrpF (Bonas et al., 1991). The unlinked hrpG and hrpX genes regulate the expression of the hrp gene cluster, which is induced in planta and during growth in certain minimal culture media (Schulte and Bonas, 1992a,b; Wengelnik and Bonas, 1996; Wengelnik et al., 1996a,b).

Based on sequence similarities, protein structure predictions and protein localization studies, it appears that the core components of a type III secretion system are present in X. c. pv. vesicatoria (Fenselau et al., 1992; Fenselau and Bonas, 1995; Wengelnik et al., 1996a; Huguet and Bonas, 1997; O. Rossier and U. Bonas, unpublished). Possible substrates of the Hrp secretion machinery are, for example, virulence factors such as suppressors of plant defence responses (Brown et al., 1995) and avirulence proteins, which determine host specificity. Bacteria expressing a given avirulence (avr ) gene are recognized by plants carrying the corresponding resistance gene, often resulting in the induction of the HR. Recently, transient expression of the X. c. pv. vesicatoria avirulence gene avrBs3 in pepper was shown to induce an HR specifically on pepper plants carrying the Bs3 resistance gene (Van den Ackerveken et al., 1996). As avrBs3 function in X. c. pv. vesicatoria depends on hrp genes (Knoop et al., 1991), it has been suggested that the AvrBs3 protein is secreted and translocated into pepper cells via the Hrp secretion system (Van den Ackerveken et al., 1996; Bonas and Van den Ackerveken, 1997).

One of the questions remaining to be addressed is what is the contribution of genes in the hrp region that are not conserved among plant and animal pathogens? To determine the role of individual genes in the hrpD operon, a non-polar mutagenesis was performed. The genetic analysis revealed that a non-polar mutation in one of the genes, hpaA (for hrp associated), affects the interaction with susceptible host plants, but, unlike hrp mutations, only in part affects the ability to induce avr gene-specific HR in resistant plants. Furthermore, we present evidence that the hpaA product may play a role inside the plant cell.

Results

Genetic and sequence analysis of the hrpD operon

The minimal size of the X. c. pv. vesicatoria hrpD operon has been defined genetically in strain 85-10 by transposon Tn3-gus insertion no. 140 (Bonas et al., 1991) and by an additional insertion, no. 506 (Fig. 1). Both insertions result in a hrp phenotype, i.e. no disease symptoms on the susceptible pepper line ECW and no HR induction on pepper line ECW-10R, which carries the corresponding resistance gene Bs1 (Schulte and Bonas, 1992a; Wengelnik and Bonas, 1996).

Figure 1.

. Genomic organization of the Xanthomonas campestris pv. vesicatoria hrpD region. Schematic drawing of the two genomic EcoRV (EV) fragments in plasmids pBCD and pBDE respectively. Location and orientation of the hrp loci are indicated by arrows. Open boxes indicate ORFs. Only relevant restriction sites are shown. Positions and orientations of Tn3-gus insertions are indicated by triangles. All insertions result in a hrp phenotype.

Sequence analysis of the 4 kb hrpD operon revealed the presence of six open reading frames (ORFs); the corresponding genes were designated hrcQ, hrcR, hrcS, hpaA, hrpD5 and hrpD6 for reasons that become obvious below (Fig. 1). Based on codon preference studies and primer extension experiments (see below), we propose that the ATG at position 425 (Fig. 2), which is preceded by a putative ribosome binding site, is used as the translation start site of hrcQ. However, we cannot rule out that the GTG at position 317 is used. The putative translation start codon of hrcR, which is also preceded by a potential ribosome binding site, is 14 nucleotides upstream of the TGA codon of hrcQ. Translation of hrcS and the following three ORFs is probably coupled, as their putative translation start codons overlap with the last nucleotides of the previous ORF. Transposon Tn3-gus insertions nos 140 and 506 are localized 102 nucleotides upstream of the hrcQ translation start codon (position 425 in Fig. 2) and after codon 227 of hrpD5 respectively.

Figure 2.

. hrpD promoter region of X. campestris pv. vesicatoria. The DNA sequence shown starts 110 bp upstream of the transcription start site (marked by an arrow, + 1) and ends at the ATG of hrcQ. The PIP-box and putative ATG or GTG translation start sites are boxed. SD (Shine–Dalgarno) refers to a putative ribosome binding site. Oligonucleotides 85 and 108 were used to determine the transcription initiation start site. The insertion site of Tn3-gus insertion 140, which results in a hrp phenotype, is shown by an open triangle. The asterisk at position 183 refers to the stop codon of hrpC3.

Mapping of the hrpD transcription start site

As Tn3-gus insertion no. 140 abolishes hrp function (Schulte and Bonas, 1992a), it is probably situated in the hrpD promoter. DNA fragments comprising a region upstream of the hrcQ transcription start site were subcloned into the promoter probe plasmid pL6GUSB (Knoop et al., 1991) carrying the uidA gene as reporter. Only the longest construct, containing 433 bp upstream of the hrcQ transcription start site, showed promoter activity under hrp-inducing conditions (data not shown). To define the hrpD promoter more precisely, the transcription start site was determined by primer extension analysis (data not shown) and found to be at the guanidine at position 111 (Fig. 2). The position of the transcription start site is in good agreement with the genetic delimitation of the hrpD operon. Inspection of the DNA sequence revealed no homology to the canonical −10 and −35 promoter elements in sigma 70-dependent promoters of E. coli. However, a PIP-box (plant-inducible promoter) is present 42 bp upstream of the hrpD transcription start site. This sequence motif, TTCGC-N15-TTCGC, has also been identified in the hrpB and hrpC promoters (Fenselau and Bonas, 1995).

The hrpD operon encodes six major pathogenicity determinants

The main features of the hrpD-encoded proteins are summarized in Tables 1 and 2. The first three ORFs encode proteins that share similarity with components of the type III protein secretion pathway in Gram-negative bacteria of both plants and mammals. They have, therefore, been designated HrcQ, HrcR and HrcS respectively (Bogdanove et al., 1996a). Furthermore, these proteins are also related to flagellar proteins of Salmonella spp., E. coli, B. subtilis and Caulobacter crescentus.

Table 1.  . Main features of predicted proteins encoded in the Xanthomonas campestris pv. vesicatoria hrpD operon. Thumbnail image of
Table 2.  . Protein sequence similarity between Xanthomonas campestris pv. vesicatoria hrpD-encoded proteins and other bacterial proteins. The values give percentage similarity/percentage identity using the BESTFIT program.a. Van Gijsegem et al. (1995); b. Huang et al. (1995); c, Lidell and Hutcheson (1994); d. Fields et al. (1994); e. Bergman et al. (1994); f. Venkatesan et al. (1992); g. Sasakawa et al. (1993); h. Groisman and Ochman (1993); i. Ohnishi et al. (1997); j. Malakooti et al. (1994); k. Bischoff and Ordal (1992); l. Bischoff et al. (1992); m. Ramakrishnan et al. (1994); n. Gober et al. (1995); o. Zhuang and Shapiro (1995); p. Bogdanove et al. (1996b); q. Mulholland et al. (1993); r. Keen and Tamaki (1986); s. accession no. L12251; t. Freiberg et al. (1997). *The N-terminus (first 140 amino acids) of HrcQ is homologous to HrcQA, while the C-terminus is homologous to HrcQB. †Only the C-terminus of HrcQ is homologous to FliN of C. crescentus and to the C-terminus of FliY of B. subtilis.Thumbnail image of

The product of hpaA is predicted to be 275 amino acids long. Interestingly, HpaA contains two putative monopartite nuclear localization signals (NLSs), located in the N-terminus and in the C-terminus respectively. In addition, putative casein kinase II phosphorylation sites (S/TxxD/E), a feature commonly associated with NLSs, have been identified upstream of both NLSs (Rihs et al., 1991). The HpaA protein shares sequence similarity only with HrpV of R. solanacearum (45% similarity/27% identity), which contains one putative NLS in its N-terminal domain. Although the overall identity is very low, the two proteins from X. c. pv. vesicatoria and R. solanacearum display a short, highly conserved sequence of 10 hydrophobic amino acids, FHLLLPLILL in HpaA and FHLLLPLMLL in HrpV, the biological significance of which is not known.

The predicted product of hrpD5 (312 amino acids) contains one hydrophobic region and shows similarity to proteins from other plant pathogenic bacteria: HrpW from R. solanacearum; HrpJ3 from P. syringae pv. syringae; and HrpQ from E. amylovora. However, the homology in the last two cases is at the limit of significance.

The product of hrpD6 is hydrophilic and is the smallest protein encoded in the hrpD operon (85 amino acids). It shares weak similarity to HrpX of R. solanacearum, the pectate lyase B precursor of E. chrysanthemi and the chitinase precursor of Streptomyces spp. (42% similarity/24% identity; SWISSPROT accession no. 116350).

Phenotype of non-polar hrpD mutants

To assess the role of individual genes in the hrpD operon, we constructed non-polar mutations in each of the six genes. For this, a new suicide plasmid, pOK1, based on pKNG101 (Kaniga et al., 1991), was created. Fragments containing deleted versions of individual hrpD genes were cloned into pOK1, generating plasmids pOKΔhrcQ to pOKΔhrpD6, and transferred into strain 85-10. The resulting mutants 85-10ΔhrcQ, 85-10ΔhrcR, 85-10ΔhrcS, 85-10ΔhpaA, 85-10ΔhrpD5 and 85-10ΔhrpD6 were inoculated into pepper to determine their phenotype. Non-polarity of the mutations was demonstrated by complementation with the respective wild-type genes. The non-polar mutations in all genes in the hrpD operon, except for hpaA, resulted in the inability to cause disease on the susceptible pepper line ECW and in loss of HR induction on ECW-10R plants, indicating that the corresponding genes are hrp genes. In contrast, inoculation of mutant 85-10ΔhpaA at high density (> 5 × 108 cfu ml−1) into young leaves of ECW-10R pepper plants induced a reduced HR (i.e. a light brown HR at inoculation sites), whereas no water-soaking symptoms were visible on susceptible plants. However, 85-10ΔhpaA grew to a 100-fold higher level than hrp mutants in susceptible pepper ECW plants (data not shown). Strain 85-10ΔhpaA(pFG*), which carries a constitutively active version of hrpG, the hrp regulatory gene (Wengelnik et al., 1996b; K. Wengelnik and U. Bonas, unpublished), also induced a reduced HR on resistant ECW-10R plants and no disease symptoms on susceptible ECW plants.

These results indicate that hpaA is not a true hrp gene. As the hpaA mutant phenotype also differs from that of dsp mutants (disease specific; Barny et al., 1990), which are non-pathogenic but still able to induce a full HR, the gene was designated hpaA (for hrp associated).

Role of hpaA in the function of  X. c. pv. vesicatoria avr genes

It has been demonstrated recently that the X. c. pv. vesicatoria AvrBs3 protein acts inside the plant cell to elicit the HR, rendering this protein a candidate for translocation by the Hrp secretion system into the plant cell (Van den Ackerveken et al., 1996). We were, therefore, curious to test X. c. pv. vesicatoria hpaA mutants expressing avrBs3 for their recognition by ECW-30R pepper plants. For this purpose, pD36 (containing avrBs3 under the control of the triple lac UV5 promoter) was introduced into strain 85-10ΔhpaA. Strain 85-10ΔhpaA(pD36) induced a reduced HR on ECW-30R plants (Table 3) and, surprisingly, did not elicit an HR on ECW-10R plants, although the mutant lacking pD36 was able to do so (Table 3).

Table 3.  . Role of hpaA in the function of X. c. pv. vesicatoria avr genes. Bacteria were infiltrated at 5 × 108 cfu ml−1. Phenotypes were scored 1–2 days after inoculation for HR and 3–5 days after inoculation for disease symptoms (water soaking).a. All pepper lines recognize avrBsT.b. Tomato line Hawaii recognizes avrRxv.c. 85-10 and 75-3: wild type; ΔhpaA : non-polar hpaA mutation.HR, hypersensitive reaction; hr, reduced, light brown hypersensitive reaction; WS, water soaking; −, no reaction; ND, not determined.Thumbnail image of

These results prompted us to test whether a hpaA mutant is able to deliver other avr gene-specific signals (see Table 3). First, mutant strains were tested for delivery of the avrBs3-2 signal. The avirulence gene avrBs3-2 is 97% homologous to avrBs3 and mediates bacterial recognition by tomato cultivars (cv.) containing the Bs4 resistance gene, such as cv. Moneymaker and Hawaii (Bonas et al., 1993; U. Bonas, unpublished). When pDS200F (pDSK602 plasmid harbouring avrBs3-2 under the control of the triple lac UV5 promoter) was introduced into 85-10ΔhpaA, the transconjugants induced an HR, albeit reduced on cv. Moneymaker.

Then, the role of hpaA in the recognition of X. c. pv. vesicatoria 75-3, a tomato pathogen that harbours avrBs1, avrBsT and avrRxv (Minsavage et al., 1990), was tested. All pepper lines recognize avrBsT, and tomato cv. Hawaii recognizes avrRxv (Whalen et al., 1993). A hpaA non-polar mutation was generated in strain 75-3. The mutant strain 75-3ΔhpaA induced a reduced HR on all pepper lines tested and on tomato cv. Hawaii. No disease symptoms were observed with hpaA mutants on all susceptible plant lines tested.

In conclusion, hpaA is not essential for the function of the five avirulence genes avrBs1, avrBs3, avrBs3-2, avrBsT and avrRxv (Table 3). However, the HR is always weaker in the case of a X. c. pv. vesicatoria hpaA mutant when compared with the equivalent wild-type strain.

Site-directed mutagenesis of putative nuclear localization signal sequences in HpaA

The findings described above indicate that the HpaA protein might play a role in modulating secretion, or be secreted itself. The fact that HpaA contains two putative NLSs supports the latter possibility. NLS1 (PRRRRRG) is located in the N-terminus and NLS2 (KPRR) in the C-terminus. This prompted us to investigate whether these sequences are important for HpaA function. Two amino acid residues were modified in each putative NLS by site-directed mutagenesis. NLS1 was modified by replacing the two arginine residues R55 and R57 (PRRRRRG) by threonine, giving PTRTRRG. The second NLS (KPRR) was mutated by replacing the lysine residue K244 by histidine, and the arginine residue R246 was replaced by threonine, giving HPTRR. Mutated hpaA versions, containing either NLS mutation or both, were cloned into pDSK600 under the control of the triple lac UV5 promoter, generating plasmids pDhpaA-NLS1, pDhpaA-NLS2 and pDhpaA-NLS12 respectively. Constructs were tested for their ability to complement the hpaA mutant 85-10ΔhpaA. While single NLS mutations did not affect the ability of hpaA to complement the mutant, the NLS double mutation only allowed partial complementation: reduced disease symptoms on the susceptible pepper ECW. To assess the effect of the different HpaA mutations on the interaction with susceptible plants better, bacterial growth of the transconjugants was determined in pepper ECW (Fig. 3). Strain 85-10ΔhpaA containing plasmids pDhpaA, pDhpaA-NLS1 and pDhpaA-NLS2 grew at the same rate and to similar levels in the plant, but approximately fivefold less well than the wild-type strain. However, the growth of 85-10ΔhpaA(pDhpaA-NLS12) was reduced: this strain grew 30- to 40-fold less well than 85-10, and six- to eightfold less well than the other transconjugants.

Figure 3.

. Growth of X. campestris pv. vesicatoria strain 85-10ΔhpaA complemented with different versions of hpaA. Growth of 85-10, 85-10ΔhpaA (hpaA), 85-10ΔhpaA(pDhpaA) (HpaA), 85-10ΔhpaA(pDhpaA-NLS1) (NLS1), 85-10ΔhpaA(pDhpaA-NLS2) (NLS2) and 85-10ΔhpaA(pDhpaA-NLS12) (NLS12) was determined in susceptible ECW pepper leaves. Values represent the mean of three samples from three different plants. Error bars represent the standard deviation. Results are from one representative experiment.

Hpa2 carries two functional nuclear localization signals

As HpaA function is affected by mutations in both putative NLSs, we analysed whether these sequence motifs can target a reporter protein to the nucleus. Therefore, translational fusions of the entire HpaA protein or derivatives containing different NLS mutations (see above) to the N-terminus of GUS (β-glucuronidase) were constructed. The constructs were introduced into onion epidermal cell layers by particle bombardment, and localization of GUS enzyme activity was determined histochemically. Activity of unmodified GUS was localized in the cytoplasm (Fig. 4A). In contrast, HpaA–GUS fusion proteins targeted the GUS activity to the nucleus (Fig. 4B). HpaA–GUS fusion proteins carrying a mutation in either NLS1 or NLS2 (constructs HpaA-NLS1 or HpaA-NLS2) mainly targeted GUS activity to the nucleus; however, some staining was also observed in the cytoplasm around the nucleus (data shown for HpaA-NLS2 in Fig. 4C). Nuclear localization of GUS activity was not observed when the fusion protein carried both NLS1 and NLS2 mutations (construct HpaA-NLS12). Instead, blue staining was detected throughout the entire cytoplasm (Fig. 4D). These results indicate that both NLS1 and NLS2 of HpaA are able and sufficient to promote the nuclear import of a reporter protein.

Figure 4.

. HpaA contains two functional nuclear localization signals. Histochemical localization of GUS activity after bombardment of onion epidermal cells with constructs expressing (A) GUS; and translational fusions to the N-terminus of GUS with HpaA (B), HpaA (C) carrying a mutation in NLS2, HpaA-NLS2, HpaA (D) carrying mutations in both NLS1 and NLS2, HpaA-NLS12. Eighty transformed onion cells were scored for localization of GUS activity for each construct. GUS activity was revealed 24 h after bombardment.

Discussion

In this study, we report on the characterization of the X. c. pv. vesicatoria hrpD operon. A large part of the X. c. pv. vesicatoria hrp gene cluster (hrpAhrpD ) is practically collinear to that of R. solanacearum (Van Gijsegem et al., 1995; Alfano and Collmer, 1997). From the hrpE locus onwards, however, X. c. pv. vesicatoria harbours novel genes, which have not been identified in other bacterial pathogens so far. The protein encoded by hrpF, which is 4 kb distant from hrpE, is homologous to NolX of the symbiotic bacterium Rhizobium fredii (Meinhardt et al., 1993; Huguet and Bonas, 1997). Genetic fine analysis of the hrpD operon, using systematic non-polar mutagenesis of individual genes, has led to the identification of three classes of genes according to sequence similarity, and phenotypes associated with each gene:

1hrcQ, hrcR and hrcS encode proteins that are highly conserved among animal and plant pathogens and are also related to proteins involved in flagellar assembly. Interestingly, the HrcQ homologue in Salmonella spp., SpaO, is secreted via the type III invasion-associated secretion system (Li et al., 1995; Collazo and Galan, 1996). Furthermore, a mutation in spaO prevents the export of all other type III secreted proteins (Collazo and Galan, 1996), suggesting that the secretion of SpaO is required for the export of others (Galan, 1996). In contrast to HrcQ, the HrcR and HrcS ‘families’ are highly conserved. In Yersinia and Shigella spp., both HrcR and HrcS homologues are necessary for type III secretion of the Yop and Ipa effector proteins respectively (Sasakawa et al., 1993; Bergman et al., 1994). In Yersinia, YscR has been localized to the inner membrane (Fields et al., 1994). The homologous flagellum proteins, FliP (HrcR) and FliQ (HrcS), of S. typhimurium are also membrane bound (Ohnishi et al., 1997) and, recently, FliP was shown to be associated with the flagellar basal body (Fan et al., 1997). These proteins are necessary for formation of the rod, hook and filament of the flagellum and are proposed as components of the type III flagellar export apparatus (Kubori et al., 1997). Based on sequence similarities and mutant phenotypes, we propose that HrcR and HrcS of X. c. pv. vesicatoria are integral inner membrane proteins that are essential core components of the putative Hrp secretion pathway.

2Proteins encoded by two genes, hrpD5 and hrpD6, have equivalents in R. solanacearum but so far no counterparts in animal pathogenic bacteria, and yet they are clearly essential for the interaction of X. c. pv. vesicatoria with the plant. These proteins might therefore be plant-specific components of the type III secretion system, specialized in dealing with the plant environment. In Ralstonia, phoA fusions with the HrpD5 and HrpD6 homologues indicated that the proteins are located on the outside of the bacterial inner membrane (Van Gijsegem et al., 1995). As both HrpD5 and HrpD6 are mostly hydrophilic, one might speculate that they are secreted. Interestingly, HrpD6 shares weak homology with the precursors of pectate lyase of E. chrysanthemi and of chitinase of Streptomyces spp. Whether this homology indicates that the bacterium may degrade the plant cell wall awaits protein localization and enzymatic studies to elucidate a possible lytic function of HrpD6.

3One gene, designated hpaA, represents a new class of genes that is essential for the virulence of X. c. pv. vesicatoria, but not for resistance gene-specific recognition.

What could be the biochemical function of the HpaA protein? X. c. pv. vesicatoria hpaA mutants have an intermediate phenotype in the interaction with the plant. In a susceptible plant, hpaA mutants grow to a certain extent but cause no disease symptoms, indicating that HpaA is an essential factor for disease development. As delivery of avr signals depends on an intact Hrp secretion apparatus (Knoop et al., 1991; O. Rossier and U. Bonas, unpublished) and HR induction is only impaired by hpaA mutants in part, HpaA is most probably not a component of the protein secretion machinery. Hence, HpaA could be an effector protein, which is secreted and translocated into the plant cell, where it is targeted to the nucleus to fulfil its function. This hypothesis is supported by the presence of two functional NLSs and is further sustained by the positive correlation between nuclear localization activity and HpaA activity. Mutations in both NLSs result in a reduction in HpaA activity, i.e. reduced disease symptom development, and also affect the ability to target GUS to the nucleus. HpaA derivatives with a mutation in only one NLS are not affected in function. The presence of some GUS activity in the cytoplasm with these derivatives may reflect a slight reduction in nuclear targeting efficiency. Similar observations have been made in other systems with karyophilic proteins, which contain two independent NLSs, e.g. the polyoma large T antigen (Richardson et al., 1986) and the VirE2 protein of A. tumefaciens (Citovsky et al., 1992).

HpaA is among the few bacterial proteins that contain NLS and are targeted to the eukaryotic host cell nucleus. Recently, the avirulence protein AvrBs3 of X. c. pv. vesicatoria has been shown to carry NLSs that are required for its function, i.e. to induce host-specific HR (Van den Ackerveken et al., 1996). NLSs have also been identified in the VirD2 (Howard et al., 1992) and VirE2 (Citovsky et al., 1992) proteins of A. tumefaciens, which play a role in transfer of the T-DNA into plant cells.

Future challenges are to test for the secretion of HpaA via the Hrp secretion pathway and to determine whether HpaA can be detected in the plant nucleus during infection of pepper by X. campestris pv. vesicatoria. In addition, analysis of transgenic plants expressing hpaA might elucidate how HpaA acts inside the plant cell.

Experimental procedures

Bacterial strains, plasmids and growth conditions

The E. coli strains DH5α (BRL) and DH5αλpir were used for maintenance of the suicide vector (Ménard et al., 1993). The X. c. pv. vesicatoria 85-10 and 75-3 strains have been described previously (Bonas et al., 1989; Minsavage et al., 1990). Plasmids used were pBluescript-KS II (Stratagene), pUC118 (Vieira and Messing, 1987), derivatives of pDSK600 and pDSK602 (Murillo et al., 1994), pLAFR3 (Staskawicz et al., 1987), pLAFR6 (Bonas et al., 1989), pKEx4tr-G (Mindrinos et al., 1994) and pKNG101 (Kaniga et al., 1991). pXV2, pXV4 and pXV9 plasmids contain, the hrpBhrpF, hrpC2hrpF and hrpAhrpE regions of X. c. pv. vesicatoria strain 75-3 in pLAFR3 respectively (Bonas et al., 1991; Wengelnik and Bonas, 1996). Plasmids were introduced into E. coli by electroporation and into X. c. pv. vesicatoria by conjugation, using pRK2013 as a helper plasmid in triparental matings (Figurski and Helinski, 1979; Ditta et al., 1980). E. coli cells were cultivated at 37°C in Luria–Bertani medium, and X. c. pv. vesicatoria strains were grown at 28°C in NYG broth (Daniels et al., 1984) or on NYG 1.5% agar or XVM2 medium (Wengelnik et al., 1996a). Antibiotics were added to the media at the following final concentrations: ampicillin, 100 μg ml−1; kanamycin, 25 μg ml−1; tetracycline, 10 μg ml−1; rifampicin, 100 μg ml−1; spectinomycin, 100 μg ml−1.

Plant material and plant inoculations

Pepper (Capsicum annuum) cultivar ECW and two near-isogenic lines ECW-10R and ECW-30R, which carry the resistance genes Bs1 and Bs3, respectively, have been described by Minsavage et al. (1990). Tomato (Lycopersicon esculentum) cultivars used were Moneymaker and Hawaii 7998, which both recognize the avrBs3-2 gene (Bonas et al., 1993; U. Bonas, unpublished). Tomato cv. Hawaii is also resistant to avrRxv containing strains of X. c. pv. vesicatoria (Whalen et al., 1993). Inoculation of pepper and tomato and reisolation of bacteria from plant tissue were performed as described previously (Bonas et al., 1991).

Sequence analysis

For sequence analysis, two consecutive EcoRV fragments present in pXV9 were subcloned in both orientations into pBluescript-KSII giving pBCD and pBCDI (4.5 kb EcoRV fragment, containing part of hrpC2 to the beginning of hrpD5 ) and pBDE and pBDEI (4.6 kb EcoRV fragment, containing part of hrpD5 and hrpE ) (Fig. 1). Standard molecular techniques were used (Sambrook et al., 1989). A series of nested deletions of plasmids pBCDI and pBDEI was generated with DNase I, and sequencing was performed as described previously (Bonas et al., 1989). Tn3-gus insertion sites in pXV9 and pXV4 were sequenced as described previously (Huguet and Bonas, 1997). Sequence data were analysed using the University of Wisconsin GCG package (version 8.0; 20) and TBLAST (Altschul et al., 1990). Mutations introduced in the NLSs of hpaA were sequenced with the ABI Prism Dye Terminator Cycle sequencing kit (Perkin-Elmer).

Suicide vector pOK1

To construct the suicide vector pOK1, pKNG101 (Kaniga et al., 1991) was digested by SpeI and NotI, thereby liberating the fragment encoding the strAB genes. The ends of the linearized plasmid were rendered blunt and ligated with the HindIII filled-in Ω fragment, conferring spectinomycin resistance (Prentki and Krisch, 1984), giving pOK1.

Generation of non-polar mutants and complementation

For the generation of non-polar mutants 85-10ΔhrcQ, 85-10ΔhrcR, 85-10ΔhrcS, 85-10ΔhpaA, 85-10ΔhrpD5 and 85-10ΔhrpD6, more detailed information can be obtained from the authors.

hrcQ.

In pOKΔhrcQ, hrcQ was deleted from position 66–647 (deletion of 193 amino acids).

hrcR.

In pOKΔhrcR, 520 nucleotides were deleted from the gene (deletion of 169 amino acids), such that only the first 12 nucleotides and the last 109 nucleotides of the gene remain.

hrcS.

In pOKΔhrcS, hrcS was deleted of 111 nucleotides from position 129–240 (deletion of 37 amino acids).

hpaA.

pOKΔhpaA carries a deletion of 534 nucleotides out of 824 and a frameshift. Only the first 207 and last 81 nucleotides of hpaA remain (deletion of 178 amino acids).

hrpD5.

In pOKΔhrpD5, hrpD5 contains the first 96 and the last 400 bp of hrpD5 (deletion of 146 amino acids).

hrpD6.

In pOKΔhrpD6, nucleotides 73–77 were deleted from hrpD6.

Mutated genes were introduced into the genome of 85-10 or 75-3 by homologous recombination in two steps, as described by Kaniga et al. (1991). In all cases, the presence of the correct deletion in the derived strains was verified by Southern analysis or polymerase chain reaction (PCR).

Plasmids containing an individual wild-type hrpD gene were constructed for complementation analysis. Details of the cloning of hrcQ, hrcR, hrcS, hrpD5 and hrpD6 can be obtained from the authors. The hpaA gene was subcloned into pDSK600 in two steps. To amplify the 5′ end of hpaA, PCR was performed using a derivative of pBCD and the oligonucleotide D4Xba (5′-GCTCTAGAGGTGCCATGATCCGTCGCATC-3′), which contains an XbaI site (underlined) and the first five codons of hpaA (bold), and the T7 primer. The 303 bp PCR product was digested with XbaI and EcoRI and cloned into pBluescript KSII (pBhpaAN). The 713 bp EcoRI–EcoRV fragment corresponding to the 3′ end of hpaA was cloned into pBD4N, giving pBD4. The insert of pBhpaA, which contains the entire hpaA gene and the first 96 nucleotides of hrpD5, was cloned as an XbaI–HindIII fragment into pDSK600 (pDhpaA).

Mutations in the nuclear localization signals of hpaA

Amino acid substitutions in the putative NLSs of hpaA were obtained by PCR and resulted in the creation of MluI restriction sites (underlined in primer sequences). The template DNA used was pBD4, and the primers used were: NLS1A (5′-GTAACGCGTCGGTGGAGCAGGGCGCAGACGCGG-3′); NLS1B (5′-CGACGCGTACACGCCGCGGGATCCGGTCGCTGGAC-3′); NLS2A (5′-CGACGCGTCGGATGTTCCGCATTCAGCAGGATCAGTGG-3′); NLS2B (5′-CGACGCGTCGGGTCGACCGTACACATGCCATC-3′). After amplification using primer T7 and one of the NLS ‘A’ primers, the PCR products were digested with XbaI and MluI. After amplification using primer T3 and one of the NLS ‘B’ primers, the PCR products were digested with MluI and HindIII. Digested PCR products 1A and 1B, 2A and 2B were cloned in a three-step ligation into XbaI–HindIII-digested pBluescript-KSII. After sequence analysis, the mutated fragments were cloned into XbaI–HindIII-digested pDSK600, generating pDhpaA-NLS1 and pDhpaA-NLS2. Combination of the two NLS mutations was obtained by exchanging the EcoRI–HindIII 712 bp fragment from pDhpaA-NLS2 into pDhpaA-NLS1, giving pDhpaA-NLS12.

Nuclear localization assay

To obtain a translational fusion of HpaA to the N-terminus of GUS, the entire hpaA coding sequence was amplified from pDhpaA and its derivatives carrying NLS mutations using primers D4Xba (5′-GCTCTAGAGGTGCCATGATCCGTCGCATC-3′) and D4Bam (5′-GATGGATCCTGGGCGAACCTCCTG-3′). The amplification products were digested with XbaI and BamHI and cloned into XbaI–BamHI-digested pKEx4tr-G (Mindrinos et al., 1994), resulting in constructs pKhpaA, pKhpaA-NLS1, pKhpaA-NLS2 and pKhpaA-NLS12, which express HpaA, HpaA-NLS1, HpaA-NLS2 and HpaA-NLS12 respectively. HpaA-GUS constructs were transiently expressed in onion epidermal cells as described by Varagona et al. (1992) and Van den Ackerveken et al. (1996).

Primer extension analysis

X. c. pv. vesicatoria strain 85E(pXV74) (Wengelnik et al., 1996a) was grown for 16 h in NYG or XVM2 or recovered from susceptible pepper plants 3 days after whole-plant infiltration. The mutant strain 85EΔAD, which lacks the region from hrpA1 to hpaA, was included as a negative control (Wengelnik et al., 1996a). Bacterial RNA was extracted as described previously (Aiba et al., 1981). The following oligonucleotides were used (positions refer to Fig. 2): no. 85 (5′-CGGCGTAACCGCACGACGCGCACCCAGCGG-3′, position 352); no. 108 (5′-GGAGGTTTACCGAAAGACAGC-3′, position 217). Primer extension was performed as described previously (Ausubel et al., 1996) using the reverse transcriptase Superscript RNases H Reverse Transcriptase MMLV according to the manufacturer's instructions (Gibco BRL, Life Technologies).

Accession number

The nucleotide sequence containing the hrpD operon can be retrieved from GenBank (accession no. 182811). Position 1 in Fig. 2 corresponds to position 1 in the sequence deposited.

Footnotes

  1. Imperial College, Department of Biology, Prince Consort Road, London SW7 2BB, UK

  2. Institut für Genetik, Martin-Luther-Universität, Halle-Wittenberg, Domplatz 1, 06108 Halle, Germany

Acknowledgements

We thank Stefan Fenselau and Martina Gutschow for sequencing, Stéphanie Broyer for excellent technical assistance, and George Frey for greenhouse work. We are extremely grateful to Guido Van den Ackerveken for helpful suggestions and critical reading of the manuscript. This work was funded in part by an ACC-SV6 grant and EC (BIO4-CT97-2244) grant to U.B. E.H. was supported by a grant from the Ministère de l'Education Nationale, de la Recherche et de la Téchnologie, and K.W. by a grant from the Human Capital and Mobility program of the European Union.

Ancillary