INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Pseudomonas syringae is a plant pathogen whose individual strains are classified into pathovars largely on the basis of host specificity. In incompatible or nonhost plants, P. syringae elicits the plant defense-associated hypersensitive response (HR), a rapid, localized, active death of plant cells that are in contact with bacteria (15, 33). As is characteristic of the common gram-negative plant-pathogenic bacteria, elicitation of the HR in nonhosts or pathogenesis in hosts is dependent on hrp genes (3, 36). Nine of these have recently been renamed hrc to indicate that they encode conserved components of a type III (host contact-dependent) secretion pathway that animal pathogens such as Yersinia spp. and plant pathogens such as Pseudomonas and Xanthomonas spp. apparently use to introduce pathogen proteins into host cells (3, 9, 14). Genes encoding the type III pathway are clustered on plasmids or in pathogenicity islands containing related virulence functions (21). Cosmid pHIR11, cloned from P. syringae pv. syringae 61, carries all the genes necessary for nonpathogenic bacteria such as Pseudomonas fluorescens and Escherichia coli to elicit the HR in tobacco (but not to cause disease) (26). These include genes encoding positive regulatory factors, the type III secretion pathway, and HrpZ and HrmA, two proteins thought to travel the pathway (4, 19, 23, 26, 48).

Three classes of proteins that are secreted by plant-pathogenic bacteria and have strong effects on plants have been extensively studied. (i) Pectic enzymes, especially pectate lyase (Pel) isozymes, cleave -1,4-galacturonosyl linkages in plant cell wall pectic polymers, resulting in tissue maceration and death of the constituent cells due to osmotic fragility (12). Host-promiscuous, macerating pathogens such as Erwinia chrysanthemi and Erwinia carotovora secrete copious amounts of several Pel isozymes by the type II (Sec-dependent) pathway (7). However, Pel production by P. syringae seems to have little role in pathogenesis (8). (ii) Harpins, such as the Erwinia amylovora HrpN and P. syringae HrpZ proteins, are glycine-rich, cysteine-lacking proteins that are secreted in culture when the Hrp (type III) system is expressed and possess heat-stable HR elicitor activity when infiltrated into the leaf intercellular spaces of tobacco and several other plants (2, 23, 47). (iii) Avr proteins are so named because their presence in an Hrp⁺ bacterium triggers the HR defense in plants carrying a corresponding R gene, thus rendering the pathogen avirulent. Avr proteins are not secreted in culture and have no apparent effect when infiltrated into the intercellular spaces of leaves. There is now strong but indirect evidence that many Avr proteins are transferred to the interior of plant cells by the Hrp systems of Pseudomonas and Xanthomonas spp. and that at least one pair of avr-R gene products (AvrPto-Pto) physically interact within the plant cell cytoplasm (19, 35, 43, 45, 46). According to a current model for P. syringae-plant interactions, multiple Avr proteins are transferred into plant cells where they either collectively promote parasitism or individually betray the parasite to the host R gene surveillance system (2).

The activity of the P. syringae HrpZ harpin in HR elicitation is puzzling in many ways. A nonpolar hrpZ mutation causes a strong reduction in the HR phenotype of E. coli(pHIR11) but only a weak reduction in the HR phenotype of P. syringae (1). This suggests that P. syringae pv. syringae carries at least one other gene outside of the region cloned in pHIR11 whose product functions similarly to HrpZ. Furthermore, it appears that the Avr-like HrmA, and not HrpZ, is responsible for the HR elicited by nonpathogenic bacteria carrying pHIR11 (1, 4). Finally, nonoverlapping fragments of HrpZ possess elicitor activity, and expression of the gene in trans in wild-type bacteria reduces rather than enhances HR elicitation (1).

P. syringae pv. tomato DC3000 offers several experimental advantages over P. syringae pv. syringae 61 for searching for a second harpin. DC3000 is a pathogen of both tomato and the model plant Arabidopsis thaliana; its hrpZ locus also has been cloned and characterized; the bacterium has been shown to secrete, in an Hrp-dependent manner, four proteins in addition to HrpZ; and the region flanking the hrp cluster, which contains the avrE locus, has been partially characterized (10, 38, 41, 49). We report here the cloning, mutagenesis, and analysis of the P. syringae pv. tomato DC3000 hrpW gene, which encodes a protein that is secreted by the Hrp pathway, is capable of eliciting the HR, and, surprisingly, has features of both harpins and Pels.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. E. coli strains were routinely grown in LM (22) or Terrific broth (42) at 37°C. The E. coli strains primarily used for plasmid constructions were DH5 and DH5F'IQ (Life Technologies, Grand Island, N.Y.). For standard DNA manipulations, the pBluescript II vectors from Stratagene (La Jolla, Calif.) were used. P. syringae pv. tomato DC3000 (41) and P. fluorescens 55 (26) were grown in King's B broth (32) or in hrp-derepressing fructose minimal medium (27) at 30°C. Antibiotics were used at the following concentrations (µg/ml): ampicillin, 100; gentamicin, 10; kanamycin, 50; rifampin, 100; spectinomycin, 50; and tetracycline, 20.

DNA manipulations. DNA manipulations and PCR were performed according to standard protocols (28, 42). Oligonucleotide primers for sequencing or PCR were purchased from Life Technologies. PCR was performed with Pfu polymerase (Stratagene). All DNA sequencing was done at the Cornell Biotechnology Center with an Automated DNA Sequencer, model 373A (Applied Biosystems, Foster City, Calif.). DNA sequence was analyzed with Genetics Computer Group version 7.3 (17) and DNASTAR (Madison, Wis.) software packages.

Plant assays. Tobacco (Nicotiana tabacum L. `Xanthi') and tomato (Lycopersicum esculentum Mill. `Moneymaker') plants were grown and inoculated with bacteria as described previously (19). For virulence assays, bacterial suspensions containing 10⁴ cells/ml were infiltrated into tomato leaves and monitored daily over a 5-day period for symptom development and bacterial multiplication.

Isolation of DNA flanking the hrp cluster in P. syringae pv. tomato DC3000 and B728a. A library of total DNA from P. syringae pv. tomato DC3000, partially digested with Sau3A, was constructed in cosmid vector pCPP47 (8). Hybridization at high stringency with ³²P-labeled PstI fragments containing hrpK and hrpR from P. syringae pv. syringae 61 yielded several cosmids. A 6.5-kb EcoRI fragment from pCPP2357 (hybridizing with hrpR) was subcloned into pML123 (34), producing pCPP2373. pCPP2374 and pCPP2375 were constructed by partially digesting pCPP2373 with MfeI and inserting an EcoRI fragment carrying the Sp^r fragment from pHP45 into transcription units IV and V, respectively (40). A P. syringae pv. syringae B728a cosmid library was provided by D. W. Bauer, and approximately 5,000 colonies were probed with a ³²P-labeled P. syringae pv. syringae 61 hrcC fragment. Restriction mapping and DNA sequencing were used to identify cosmid pCPP2346, which contains hrpW.

DNA gel blots. Total DNA (2 µg) was digested with EcoRI and separated by electrophoresis on 0.5% agarose gels. DNA was transferred to Immobilon-N Membrane (Millipore Co., Bedford, Mass.) and hybridized at 62°C for 8 h in HYB-9 DNA Hybridization Solution (GENTRA Systems, Research Triangle Park, N.C.) with a 1.3-kb PCR-amplified hrpW fragment that was labeled with ³²P by using the Prime-It II kit (Stratagene). The membranes were washed four times in 1.0% sodium dodecyl sulfate (SDS) and 1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) followed by two washes in 1.0% SDS and 0.2× SSC. Membranes were exposed to OMAT X-ray film (Eastman Kodak, Rochester, N.Y.) for 4 to 12 h.

HrpZ and HrpW secretion assays. P. fluorescens 55 strains carrying pHIR11 and pML123 constructs containing the hrpW region were grown overnight in 2 ml of King's B plus appropriate antibiotics. The cultures were washed once with hrp gene-derepressing fructose minimal medium and resuspended in 15 ml of the same medium to an optical density at 600 nm of 0.4. Cultures were grown with moderate shaking at room temperature. After 12 h, the cells were collected by centrifugation in a Sorvall SS-34 rotor at 3,000 × g and resuspended up to 500 µl with water plus 20 µl of 100 mM phenylmethylsulfonyl fluoride (PMSF). Twenty microliters of 100 mM PMSF was added to the culture supernatants, and the supernatants were centrifuged at 25,000 × g for 30 min to pellet the remaining cells. The supernatant proteins were concentrated 100-fold with Centricon-10 concentrators (Amicon, Inc., Beverly, Mass.). SDS-polyacrylamide gel electrophoresis (PAGE) and immunoblot analyses, using previously obtained anti-HrpZ antibodies (23, 49) and the Western light chemiluminescent detection system (Tropix, Bedford, Mass.) and OMAT X-ray film (Eastman Kodak), were performed as previously described (1).

Preparation of HrpW and derivatives. A MfeI-XhoI fragment from pCPP2357 containing hrpW was ligated to the EcoRI-XhoI sites of pET21(+) (Novagen, Madison, Wis.) to make pCPP2417. The complete coding sequence for HrpW was PCR amplified from pCPP2417 with the primers 5'-ATGAGGATCCAGCATCGGCATCACACCC-3' (named W1) and 5'-ATGAAAGCTTAAGCTCGGTGTGTTGGGT-3' (named W2), which contained BamHI and HindIII sites, respectively. DNA encoding the N-terminal 186 amino acids of HrpW was PCR amplified from pCPP2417 with the W1 primer and the primer 5'-ATGAAAGCTTGCCACCGCCTGTTGCAGT-3', which contained a HindIII site. DNA encoding the C-terminal 236 amino acids of HrpW was PCR amplified from pCPP2368 with the primer 5'-ATGAGGATCCGAGGGTGGCGTAACACCG-3', which contained a BamHI site and the W2 primer. Amplified products corresponding to full-length HrpW and the N-terminal and C-terminal portions of HrpW were directionally cloned into the BamHI and HindIII sites of pQE30 (Qiagen) resulting in pCPP2377, pCPP2378, and pCPP2379, respectively. Procedures used to isolate His-tagged proteins with Ni-nitrilotriacetic acid (NTA) spin columns (Qiagen) were as described previously (1). E. coli ABLE K (Stratagene) grown on M9 medium (42) and supplemented with glucose (0.2%), casamino acids (0.02%), and thiamine (1 µg/ml) was used to obtain His₆-HrpW because of apparent toxicity. SDS-PAGE and immunoblot analyses, using previously obtained anti-HrpZ and anti-HrpW antibodies (23, 49) and the Western light chemiluminescent detection system (Tropix) and OMAT X-ray film (Eastman Kodak), were performed as described previously (1).

Preparation of calcium pectate beads and pectate binding assays. One hundred millimolar CaCl₂ was added dropwise to vigorously stirring 0.2% (wt/vol) sodium pectate (Sigma, St. Louis, Mo.) dissolved in 100 mM Tris (pH 8.0) or 50 mM MES (morpholineethanesulfonic acid) (pH 5.6) to make calcium pectate beads. The beads were pelleted by low-speed centrifugation and resuspended in 2 volumes of buffer. For binding assays, 50 to 500 µg of protein was mixed with 500 µl of pectate bead mix and incubated for 30 min to 12 h at room temperature. The beads were pelleted by centrifugation, washed several times in 500 µl of buffer, and resuspended to 500 µl in buffer. Aliquots of 20 µl were taken from the bead mix prior to centrifugation, from the wash buffer at each step, and from the washed beads and then analyzed with SDS-PAGE to determine protein distribution. Calcium alginate beads were similarly prepared from medium-viscosity sodium alginate (Sigma), and binding assays were performed as described above for pectate. Tests for the ability of either calcium chloride or pectate to precipitate HrpW were conducted with and without the addition of 0.2% (wt/vol) hydrated beads of agarose (Agarose IV; Amresco, Solon, Ohio).

Construction of hrpZ and hrpW marker exchange mutations in P. syringae pv. tomato DC3000. To construct the hrpZ mutation, a 603-bp ClaI fragment internal to hrpZ was deleted from pCPP2334, a LITMUS 28 (New England Biolabs) derivative that contains hrpA and hrpZ, producing pCPP2336. An nptII derivative lacking a transcriptional terminator was PCR amplified from pCPP2988 (1) with the primers 5'-CCATCGATGGTGGTGGCGATAGCTAGACTGG-3' and 5'-CCATCGATGGTCTCGTGATGGCAGGTTG-3' and cloned into the unique ClaI site of pCPP2336 in the correct orientation. A BglII-HindIII fragment from the resulting construct, pCPP2338, carrying the hrpZ mutation was exchanged for the BglII-HindIII fragment, carried in pCPP2340, producing pCPP2342. A 5.3-kb EcoRI fragment from pCPP2342 that carried the hrpZ mutation was cloned into the broad-host-range plasmid, pRK415 (30), producing pCPP2344. An 8.5-kb EcoRI fragment from pCPP2375, which carried hrpW interrupted with an Sp^r fragment, was subcloned into pRK415, producing pCPP2376. Separately, pCPP2376 and pCPP2344 were electroporated into P. syringae pv. tomato DC3000. Loss of the plasmid and retention of the marker were achieved as previously described (1).

Nucleotide sequence accession numbers. The nucleotide sequences reported in this paper have been deposited in GenBank under accession numbers AF005221 (P. syringae pv. tomato hrpW) and AF037983 (P. syringae pv. syringae hrpW).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

hrpW expressed in trans eliminates the ability of P. fluorescens(pHIR11) to elicit the HR. To identify any harpin-like genes in the P. syringae pv. tomato DC3000 DNA flanking the hrp-hrc gene cluster, we isolated cosmid pCPP47 derivatives containing inserts hybridizing with hrpK or hrpR. These two genes define the left and right borders of the cluster. Subclones were constructed in pML123 and screened for two potential harpin phenotypes: (i) the ability to promote tobacco HR elicitation activity in P. fluorescens cells carrying pCPP2274, a hrpZ pHIR11 derivative (19), and (ii) interference with the HR elicitation activity of P. fluorescens cells carrying wild-type pHIR11 (1). No subclones had the first phenotype, but subclone pCPP2373 had the second (data not shown). pCPP2373 contains a 6.5-kb EcoRI fragment from cosmid pCPP2357 (Fig. 1A), which hybridized with hrpR and has transcriptional units IV and V, which were previously identified and partially sequenced by Lorang and Keen (38). To determine which transcriptional unit in pCPP2373 was responsible for eliminating the HR elicitation activity of P. fluorescens(pHIR11), an Sp^r fragment was inserted into MfeI sites in transcriptional units IV and V to construct pCPP2374 and pCPP2375, respectively (Fig. 1A). Both plasmids were transformed into P. fluorescens(pHIR11) cells, which were then infiltrated into tobacco leaves. Only pCPP2374 blocked HR elicitation, indicating that transcriptional unit V encoded a protein with one of the characteristics of HrpZ. HrpZ was produced in all strains but was not secreted in strains expressing transcriptional unit V, indicating that the block in HR elicitation was not due to negative regulation of hrpZ by transcriptional unit V (data not shown).

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Physical map of the P. syringae pv. tomato DC3000 hrpW region and structural features of HrpW. (A) The physical map of the DC3000 genome adjacent to the hrp gene cluster is shown with open arrowheads denoting putative 54 promoters and with filled arrowheads denoting putative HrpL-dependent promoters that control previously defined transcriptional units (38). ORFs in polycistronic operons are marked by divisions in the open arrows denoting the operons. hrpR and hrpS encode regulatory proteins and are located at the right end of the hrp cluster. The region carried in pCPP2373 is indicated above the physical map, and inverted filled triangles mark the location of the Spr insertions in pCPP2374 and pCPP2375. (B) The diagram of HrpW indicates the harpin-like and Pel-like domains, with numbers on the top corresponding to DC3000 and on the bottom to B728a. The DC3000 hrpW PCR subclone-generated His6-tagged harpin domain fragment encompasses amino acids 1 to 186; the His6-tagged Pel domain has amino acids 187 to 425. (C) The sequences of the region in the middle of the HrpW proteins from P. syringae pv. tomato DC3000 (HrpWPto) and P. syringae pv. syringae B728a (HrpWPsy), which contains glycine-rich repeats (denoted by the bar in panel B), are shown in the boxes. Dashes were inserted where necessary to permit alignment.

The DNA sequence of hrpW predicts a protein with both harpin and Pel domains. The complete DNA sequence of transcriptional unit V was determined for P. syringae pv. tomato DC3000, revealing a 1,275-bp open reading frame (ORF) that was designated hrpW. The gene is preceded by a consensus hrp promoter (38) and followed by a rho-independent terminator. The predicted N-terminal sequence of HrpW matches that of EXP-60, one of the five P. syringae pv. tomato DC3000 Hrp-secreted proteins identified by Yuan and He (49). hrpW is flanked by operons transcribed in divergent directions and appears to be in a monocistronic operon. Like harpins, the predicted 42.9-kDa HrpW protein is acidic and glycine rich, lacks cysteine, and is deficient in aromatic amino acids. The predicted protein sequence of HrpW reveals at least two distinct domains (Fig. 1B). A harpin-like domain (amino acids 1 to 189) is rich in glutamine, serine, and glycine. The region from amino acids 119 to 186 contains six imperfect glycine-rich repeats with many acidic and polar residues (Fig. 1C). The TPS/DAT motif in this region is predicted to have -sheet structure with one side of the -sheet having all the threonine and serine residues, and the glycine repeats are predicted to be turns by the Garnier-Robson algorithm (39). The alternating -sheets and turns could form a -barrel structure. Database searches done with BLAST (5) revealed no proteins with significant similarity to the harpin domain.

hrpW appears widely distributed in plant-pathogenic bacteria and is in a region conserved between two P. syringae pathovars. We further examined the distribution of hrpW by DNA gel blot analysis. The P. syringae pv. tomato DC3000 hrpW ORF was amplified by PCR and used as a probe for high-stringency gel blot hybridization with EcoRI-digested DNA from representative necrogenic gram-negative plant pathogens. The hrpW probe hybridized to at least one distinct band for each of the P. syringae pathovars tested: glycinea race 4 U1, papulans Psp19, pisi H27, phaseolicola 343, tabaci ATCC 11528, and syringae strains B728a and 61 (weakly). Hybridization was also observed with Pseudomonas viridiflava PV5, Ralstonia (Pseudomonas) solanacearum CR10 (weakly), and Xanthomonas campestris pathovars amoraciae XC-4 and vesicatoria H44 (Fig. 2). No hybridization was observed with DNA from E. amylovora Ea321, E. carotovora subsp. carotovora Ecc71, or E. chrysanthemi CUCPB5047.

View larger version (67K): [in this window] [in a new window]   FIG. 2.   DNA gel blot analysis of the hybridization of hrpW under high-stringency conditions to total DNA from other bacterial plant pathogens. DNA from the indicated pathogens was isolated, digested with EcoRI, resolved on a 0.5% agarose gel, transferred to an Immobilon-N membrane, and hybridized with a 32P-labeled hrpW subclone at 62°C. Abbreviations: Pto, P. syringae pv. tomato; Psy, P. syringae pv. syringae; Pgy, P. syringae pv. glycinea; Ppp, P. syringae pv. papulans; Ppi, P. syringae pv. pisi; Pph, P. syringae pv. phaseolicola; Pta, P. syringae pv. tabaci; Pvf, P. viridiflava; Rso, Ralstonia solanacearum; Xam, Xanthomonas campestris pv. amoraciae; Xvs, X. campestris pv. vesicatoria; Ea, Erwinia amylovora; Ecc, E. carotovora subsp. carotovora; Ech, E. chrysanthemi.

HrpW and its harpin domain elicit an HR-like necrosis in tobacco leaves. PCR subclones of the P. syringae pv. tomato DC3000 hrpW were constructed in pQE30 to permit production of derivatives of HrpW and the two domain fragments carrying N-terminal His₆ tags. These fusion proteins were partially purified by Ni-NTA chromatography and analyzed by SDS-PAGE and immunoblotting with antibodies raised against P. syringae pv. tomato DC3000 Hrp-secreted proteins (Fig. 3). Anti-HrpW antibodies bound to the full-length HrpW and to both fragments, but binding to the harpin domain fragment was noticeably weaker. Transformants producing HrpW were highly unstable in their maintenance of the plasmid. Thus, HrpW levels were quite low, and Ni-NTA chromatography yielded a preparation that was only partially enriched in HrpW. Nevertheless, the HrpW preparation elicited an HR-like necrosis in tobacco leaves, which visibly differed from the necrosis elicited by the P. syringae pv. syringae 61 HrpZ only in developing ca. 12 h later (Fig. 4). The elicitor activity was heat stable and protease sensitive, and vector control preparations produced no response (data not shown). The partially purified harpin domain fragment also elicited a necrosis that was slightly delayed, and this response, like that elicited by HrpZ, could be inhibited by 1.0 mM lanthanum chloride, a calcium channel blocker (Fig. 4). Thus, the necrosis elicited by the HrpW harpin domain is an active plant response. In contrast, purified E. chrysanthemi PelE, obtained from E. coli JA-221(pPEL748) (29), elicited a black, macerated necrosis that was not inhibited by 1.0 mM lanthanum chloride, 50 µM sodium vanadate, or 100 µM cycloheximide (data not shown). This is consistent with the expectation that pectic enzymes kill by lysis of turgid protoplasts through weakened cell walls rather than by elicitation of cell death programs. The HrpW Pel domain elicited no visible response in the infiltrated tobacco tissue.

View larger version (71K): [in this window] [in a new window]   FIG. 3.   SDS-PAGE and immunoblot analysis of preparations containing the P. syringae pv. tomato HrpW and its harpin domain and Pel domain fragments. (A) His6-tagged full-length HrpW and the two domain fragments were partially purified by Ni-NTA chromatography, separated by SDS-PAGE, and stained with Coomassie blue R250. The arrowhead indicates the full-length HrpW, which is produced in very small amounts. Lanes: 1, Pel domain fragment; 2, harpin domain fragment; 3, HrpW. (B) The same HrpW derivatives were also visualized on immunoblots with anti-HrpW antibodies used in conjunction with the Western light chemiluminescence assay. Lanes: 4, Pel domain fragment; 5, harpin domain fragment; 6, HrpW.

View larger version (141K): [in this window] [in a new window]   FIG. 4.   Elicitation in tobacco leaves of active tissue death indicative of the HR by cell-free preparations containing the P. syringae pv. tomato HrpW and its N-terminal fragment. The protein preparations analyzed in Fig. 3 were infiltrated into tobacco leaves, in some cases with 1.0 mM lanthanum chloride. Leaves were photographed 48 h later. (A) P. syringae pv. syringae 61 HrpZ (0.12 µg/ml); (B) HrpW; (C) harpin domain fragment of HrpW (0.22 µg/ml); (D) HrpZ plus lanthanum chloride; (E) HrpW plus lanthanum chloride; (F) Pel domain fragment of HrpW (1.40 µg/ml).

The HrpW Pel domain binds calcium pectate but lacks detectable Pel activity. HrpW, its harpin domain, and its Pel domain were tested for Pel activity with the A₂₃₀ assay for 4,5-unsaturated pectic products (13). No activity was detected despite trying both polygalacturonic acid and a 31% methylesterified derivative as substrates, CaCl₂ and MnCl₂ as cofactors, and several pH levels. Although the HrpW Pel domain lacks detectable Pel activity, it did bind to calcium pectate beads, even in the presence of 500 mM NaCl, at both pH 5.6, the pH of the plant apoplast, and pH 8.0, the optimal pH for activity of active Pels (Fig. 5). Full-length HrpW was not tested in these binding experiments because of problems in producing the protein in E. coli, as described above. The HrpW Pel domain was solubilized from beads by addition of EDTA or soluble pectate, both of which dissolve the beads. The protein did not precipitate when mixed with either CaCl₂ or sodium pectate individually. The HrpW Pel domain did not bind to calcium alginate beads prepared in the same manner as the pectate beads, indicating the binding is specific to pectate. Boiling the HrpW Pel domain for 10 min did not decrease its subsequent ability to bind pectate; thus the pectate binding and the HR eliciting activities of HrpW are both heat stable. HrpZ and marker proteins bracketing the pI and molecular weight of the HrpW Pel domain were tested for calcium pectate binding, and none bound under these conditions (Fig. 5).

View larger version (50K): [in this window] [in a new window]   FIG. 5.   Binding of the HrpW Pel domain to calcium pectate beads. Purified HrpW Pel domain or marker proteins were mixed with calcium pectate or calcium alginate beads in 50 mM MES (pH 5.6), 500 mM NaCl and washed several times with the same buffer. In panels A through D, equal volumes from the original protein-bead mix (Total), the supernatant from the original mix (Unbound), the last buffer wash (Wash), and the washed beads (Bound) were analyzed by SDS-PAGE to determine if proteins bound to calcium pectate or calcium alginate beads. (A) HrpW Pel domain with calcium pectate beads; (B) HrpW Pel domain with calcium alginate beads; (C) mixed low-molecular-weight, low-pI, and high-pI marker proteins (Pharmacia, Uppsala, Sweden) with calcium pectate beads; (D) HrpZ with calcium pectate beads. In panels E and F, equal volumes from the original protein mix (Total), the supernatant, and the pellet were similarly analyzed to determine protein precipitation. (E) HrpW Pel domain with CaCl2; (F) HrpW Pel domain with soluble pectate.

The ability of a P. syringae pv. tomato DC3000 hrpZ hrpW mutant to elicit the HR is substantially reduced. Marker exchange mutagenesis was used to construct P. syringae pv. tomato mutants CUCPB5094 (hrpZ::nptII), CUCPB5096 (hrpW::Sp^r), and CUCPB5095 (hrpZ::nptII hrpW::Sp^r). The hrpZ::nptII mutation is nonpolar. All mutant constructions were confirmed by DNA gel blotting and immunoblotting (data not shown). Tobacco leaves were infiltrated with P. syringae pv. tomato DC3000 and the three mutant derivatives at two levels of inoculum and then examined 48 h later to determine the percentage of infiltrated tissue that was necrotic (Table 1). Only the hrpZ hrpW mutant was significantly reduced in the frequency with which it elicited collapse of more than 50% of the infiltrated area. We were unable to complement this phenotype since expression of hrpW or hrpZ in trans reduces the HR elicited by wild-type P. syringae pv. tomato DC3000. To determine if this mutant was reduced in virulence, tomato leaves inoculated with the mutant and wild-type DC3000 were monitored for symptom production and bacterial multiplication over a period of 5 days. No difference was observed (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

To identify an anticipated second harpin in P. syringae, we screened DNA in the P. syringae pv. tomato DC3000 hrp gene region for genes with harpin-like phenotypes. hrpW had the expected but paradoxical phenotype of interfering with HR elicitation when expressed in trans. This interference was not due to repression of the hrpZ operon, since HrpZ was expressed. It is more likely related to the blocking of HrpZ secretion, which we subsequently observed. HrpW was found to be identical to the previously identified and partially sequenced transcriptional unit V and to encode the previously identified Hrp-secreted protein EXP-60 (38, 49). In a preliminary report, this protein was referred to as HopPtoA (11). It is now designated HrpW based on its HR elicitor activity, the phenotype of a hrpZ hrpW mutant, and the homology of the protein with the HrpW proteins from E. amylovora strains Ea321 (31) and CFBP1430 (18). The latter report appeared when this article was under revision.

HrpW has several general characteristics of harpins, including amino-acid composition, heat stability, unexpectedly low mobility in SDS-PAGE, and the ability of both full-length and truncated proteins to elicit the HR (1, 2, 23, 47). HrpW also has six glycine-rich repeats that are similar to a repeated sequence found in HrpZ and are reminiscent of the repeat-rich structure of HrpZ (1). The general lack of cysteine residues in harpins is particularly striking in HrpW because comparison with the homologous fungal and bacterial Pels reveals that all eight of the conserved cysteine residues in those proteins have been substituted in HrpW. All of these properties raise the possibility that harpins, like the Salmonella typhimurium FlgM protein, may be in an unfolded state in the absence of their substrates or targets (16). This appears to be important to FlgM because of spatial constraints on the movement of globular proteins through the flagellum. With harpins, an unfolded state is more likely important for penetration into the plant cell wall matrix than for translocation through the Hrp pathway, since several Avr proteins thought to travel the pathway are relatively large and cysteine rich.

The ability of isolated P. syringae HrpZ and HrpW proteins to elicit the HR when infiltrated into tobacco leaf tissue may not directly reflect biological function because the Avr proteins now appear to be both essential and sufficient (once delivered to the plant cytoplasm) for elicitation of the bacterial HR (2). Our data do not permit us to compare the relative efficacy of HrpZ and HrpW to elicit the HR, but it is worth noting that harpins remain the only bacterial proteins known to elicit apparent programmed plant cell death when exogenously applied.

Harpins appear to have a site of action in plant cell walls. HrpZ associates with the walls rather than the membranes of plant cells, and the protein elicits no response from wall-less protoplasts (25). The Pel domain in HrpW binds to pectate (in the presence of calcium), which is the component controlling porosity of the cell wall matrix (6). This binding occurs at a physiological pH that is above the pK of polygalacturonic acid (pectate) and the predicted pI (4.9) of the HrpW Pel domain, which eliminates binding attributable to a net charge difference between the two molecules. The failure of polymannuronic acid (alginate) to bind the HrpW Pel domain provides further evidence for the specificity of pectate binding. The Pel domain does not have elicitor activity, and its presence does not prevent native HrpW from having this activity. This suggests that HrpW elicitor activity resides in an interaction with the plant cell wall, or at least that it can occur despite immobilization in the pectic matrix of the wall. Membership of the homologous E. amylovora HrpW in a newly defined class of fungal and bacterial Pels is discussed in the accompanying paper by Kim and Beer (31). Neither of these HrpW proteins has detectable Pel activity. Pel homologs with no detectable activity in vitro are also found in the pollen and style tissues of several plants, and the conservation of catalytic residues in these proteins suggests a cryptic enzymatic function (24).

Mutation of transcription unit V reduced neither the HR nor the virulence phenotypes of P. syringae pv. tomato DC3000 (38), and our hrpZ hrpW mutant was significantly reduced only in its HR phenotype (although our virulence assay would likely miss a subtle reduction). Because the expression of hrpZ and hrpW in trans interferes with HR elicitation, even in wild-type cells, we were unable to perform standard complementation tests with our hrpZ hrpW double mutant. However, the construction of each mutation was confirmed by sequence analysis, and the full virulence of the double mutant further argues against second-site mutations. One interpretation of the observation that hrpZ mutation strongly reduces HR elicitation by E. coli(pHIR11) and P. fluorescens(pHIR11) (1, 19), whereas hrpZ hrpW double mutation only weakly reduces HR elicitation by P. syringae pv. tomato, is that P. syringae produces additional harpin-like proteins, analogous to the multiple Pel isozymes of E. chrysanthemi and E. carotovora (7).

The presence of sequences hybridizing with hrpW in several other plant-pathogenic bacteria, particularly P. viridiflava and X. campestris, is significant because it suggests that HrpW has some broadly important function, and it implies that P. viridiflava has a Hrp system and that X. campestris produces a harpin. Although the Hrp system of X. campestris has been extensively characterized, no harpin or any other protein has been found to be secreted by the Hrp system in culture. hrpW should be useful as a probe to clone a gene encoding such a protein from X. campestris. The function of harpins and their possible involvement in the delivery of Avr proteins through plant cell walls remains a matter of speculation. However, our finding of a P. syringae harpin with a Pel domain and calcium pectate binding activity provides further evidence that Avr proteins and harpins differ in their sites of action, with many Avr proteins acting inside plant cells and harpins acting outside, probably in the cell wall.