INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The characteristic ability of many phytopathogenic bacteria to elicit the hypersensitive response (HR) in nonhost plants or to be pathogenic in host plants is dependent on hrp and hrc genes (2). hrc genes represent a subset of the hrp genes that have been renamed to reflect their conservation among the type III protein secretion systems of both plant and animal pathogens (6). Among these, hrcC has been particularly well studied; it encodes an outer membrane protein that is essential for type III protein secretion and has a primary role in protein translocation across the outer membrane (2, 7, 32). The hrcC genes of Erwinia amylovora and Pseudomonas syringae are flanked by four small genes, which together form the hrpC operon. These four genes, hrpF, hrpG, hrpT, and hrpV, appear to be characteristic of group I hrp clusters, such as those of P. syringae and E. amylovora, and they are absent from the group II hrp clusters of Ralstonia solanacearum and Xanthomonas campestris pv. vesicatoria (9, 19). Group I and II hrp clusters also differ notably in their regulatory components, with group I hrp genes being activated by an alternate sigma factor and group II hrp genes being activated by an AraC homolog (2).

Essential activities in type III secretion can be ascribed to many of the Hrc proteins, such as HrcC, but less is known about the functions of the Hrp proteins. Notable exceptions are the HrpA, -L, -R, -S, and -Z proteins of P. syringae. HrpA is a Hrp-specific pilin (26). HrpR and -S show similarity with ⁵⁴-dependent promoter-enhancer-binding proteins, and both are required to activate the ⁵⁴-dependent production of HrpL, a sigma factor in the ECF (extracytoplasmic factor) family which activates the expression of other hrp genes and many avr genes (17). HrpZ is a harpin, a type of protein first reported from E. amylovora (31), which can elicit an apparent programmed cell death when infiltrated into the leaves of tobacco and several other plants (15). HrpZ is secreted in culture in a hrp-dependent manner from P. syringae (15), but the protein does not appear to be the physiological elicitor of the HR: mutations in hrmA, an avr-like gene, completely block the ability of a functional cluster of P. syringae hrp genes to function in Escherichia coli to elicit the HR, but they have no effect on HrpZ secretion (1, 3). Avr (avirulence) proteins appear to be the actual elicitors of the HR, and there is compelling evidence that many of these function inside plant cells following delivery by the P. syringae Hrp system (11, 22, 28, 30). Whether HrpZ has a primary role as an extracellular component of the Avr protein delivery system is unknown, but its secretion in culture provides an assay for the functioning of the Hrp secretion pathway in P. syringae.

In the accompanying paper, we have shown that mutations in the P. syringae pv. syringae 61 hrpF, hrpG, hrpT, and hrpV genes result in altered plant reaction phenotypes, with the effects of each mutation being quantitatively different (9). Unexpectedly, expression of hrpV in trans reduced the ability of wild-type P. syringae pv. syringae 61 to elicit the HR, suggesting that HrpV may be a negative regulator of the Hrp regulon. To test this hypothesis and to investigate further the functions of the other genes in the hrpC operon, we have determined the effects of mutations in these genes on the production of several Hrp marker proteins and on the secretion of the HrpZ harpin. Subsequently, we also investigated the place of HrpV inhibition in the HrpRS-HrpL regulatory cascade.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. Most of the bacterial strains and plasmids used in this study are described in the accompanying paper (9). Additional strains used were P. syringae pv. syringae 61-2074 (hrpL::TnphoA), P. syringae pv. syringae 61-2088 (hrcU::TnphoA), and P. syringae pv. syringae 61-2095 (hrpS::TnphoA) (16). E. coli was routinely grown in Luria broth or Terrific broth at 37°C (4). Pseudomonas strains were grown in King's B (KB) medium (20) or LM medium (13) at 28 to 30°C, but for certain experiments the hrp-derepressing minimal medium of Huynh et al. (18), adjusted to pH 5.5, was used at 25°C. Antibiotics were used in selective media at the following concentrations (micrograms per milliliter): ampicillin, 100; kanamycin, 50; tetracycline, 10; spectinomycin 50; gentamicin 10; and nalidixic acid, 20.

Recombinant DNA techniques. Restriction endonuclease digestion, agarose gel electrophoresis, DNA fragment preparation, plasmid extraction, DNA ligation, and transformation by the CaCl₂ procedure were performed according to standard procedures (4). Plasmids were introduced into bacteria by transformation, electroporation (Gene Pulsar; Bio-Rad, Richmond, Calif.), or triparental mating (10).

Construction of pCPP2385, pCPP2389, and pCPP2383. pCPP2385, which contains a constitutively expressed hrpL, was constructed by inserting an Sp^r fragment (7) downstream of hrpL in pCPP2311. This provided a selectable marker that would be effective in P. syringae for pCPP2311, which contains hrpL subcloned downstream of the lac promoter in pUCP18 (27). pCPP2389 constitutively expresses hrpRS and was made by subcloning a 2.2-kb BamHI-BglII fragment derived from pHIR11 downstream of the lac promoter in pCPP30 (5). The 190-bp promoter-active SacI-HincII fragment upstream of hrpA was cloned into a promoterless uidA gene in pCPP45 (5) to create the hrpA'::uidA transcriptional fusion in plasmid pCPP2383.

Preparation of anti-HrcC and anti-HrcQ_B antibodies. HrcC and HrcQ_B were purified from E. coli NovaBlue (DE3) (Novagen, Madison, Wis.) carrying pNCHU316 and pNCHU366, respectively, according to previously described procedures (7). Before immunoblot analysis the antisera were preabsorbed with cell lysate mixtures of E. coli and P. syringae pv. syringae 61 mutants according to the following modifications of previously used procedures (15). Five-milliliter cultures of E. coli and P. syringae pv. syringae 61 mutants (P. syringae pv. syringae 61-N393 and P. syringae pv. syringae 61-N322 for incubation with HrcC and HrcQ_B antibodies, respectively) were grown overnight to stationary phase, washed several times with the same volume of 1× phosphate-buffered saline (0.058 M Na₂HPO₄, 0.017 M NaH₂PO₄ · H₂O, 0.068 M NaCl), sonicated, incubated at 100°C for 10 min, and then allowed to cool to room temperature. Two hundred microliters of the corresponding antiserum and sodium azide at a final concentration of 0.02% were added into the cell lysate and incubated at room temperature for 2 h. The mixtures were centrifuged at 20,000 × g for 30 min, and the preabsorbed antisera were collected from the supernatant.

Immunoblot analysis of the expression of HrpZ, HrcC, HrcJ, and HrcQ_B proteins. Wild-type P. syringae pv. syringae 61 and hrpF, hrpG, hrcC, hrpT, and hrpV mutants were grown in 5 ml of KB broth at 30°C to an optical density at 600 nm (OD₆₀₀) of 0.5. The cells were collected by centrifugation, washed once in 5 ml of Hrp minimal medium, resuspended in 5 ml of the same medium, and incubated with shaking for 5 h. The cell and supernatant fractions were then separated by centrifugation. HrpZ expression was analyzed as described previously (24) or as outlined below. The supernatants were precipitated with trichloroacetic acid at a final concentration of 5%, washed with acetone, dissolved in 30 µl of 10 mM Tris buffer (pH 8.0), and boiled for 5 min after an equal volume of 2× loading buffer (0.625 M Tris [pH 6.8], 2% sodium dodecyl sulfate [SDS], 10% glycerol, 2% -mercaptoethanol) was added. The cell pellets were washed with 10 mM Tris buffer (pH 8.0), resuspended in 125 µl of 10 mM Tris buffer, and boiled with an equal volume of 2× loading buffer for 5 min. A 20-µl sample of each fraction was subjected to SDS-10% polyacrylamide gel electrophoresis in a 0.75-mm-thick gel in a Mighty Small apparatus (Hoefer Scientific Instruments, San Francisco, Calif.). The prestained protein markersranging from 175.0 to 6.5 kDawere from Bio-Rad. After separation, the protein bands were transferred to Immobilon-P membranes (Millipore Inc., Bedford, Mass.) in a TE70 semidry transfer unit (Hoefer Scientific Instruments) for 40 to 60 min. The membranes were probed with polyclonal anti-HrpZ, anti-HrcJ, anti-HrcC, anti-HrcQ_B, or anti-NPTII antibodies individually. Anti-NPTII antibodies (5 Prime3 Prime, Inc., Boulder, Colo.) were used as an internal control to standardize the protein concentration of each lane. Immunodetection of the bands was performed by alkaline phosphatase-based chromogenic and chemiluminescent assays with Sigma Fast BCIP (5-bromo-4-chloro-3-indolylphosphate)-nitroblue tetrazolium tablets (0.15 mg of BCIP/ml and 0.3 mg of nitroblue tetrazolium/ml) or 0.25 mM disodium 2-chloro-5(4-methoxyspiro{1,2-dioxetane-3,2'-(5'-chloro)-tricyclo(3.3.1.1^3,7)decan}-4-yl)-1-phenyl phosphate (CDP-Star; Tropix Inc., Bedford, Mass.), respectively. Membranes were exposed for 2 to 15 min to OMAT X-ray film (Kodak, Rochester, N.Y.) for visualization of chemiluminescence.

Measurement of -glucuronidase activity. Transcription of the hrpZ operon was determined by measuring the expression of -glucuronidase from the hrpA'::uidA transcriptional fusion plasmid pCPP2383. P. syringae pv. syringae 61 derivatives carrying pCPP2383 were grown in KB medium to stationary phase and then transferred to Hrp minimal medium at an initial OD₆₀₀ of 0.5. The bacteria were harvested 0, 4, and 10 h later for determination of -glucuronidase activity and total protein concentration. -Glucuronidase activity was estimated with 4-methylumbelliferyl--D-glucuronide and a TKO 100 minifluorometer (Hoefer Scientific Instruments) according to the procedure outlined in the operating instructions. Total protein concentration was determined with the Bio-Rad protein assay reagent (Bio-Rad Laboratories, Hercules, Calif.). Specific activity is presented in nanomoles of methylumbelliferone production per milligram of total protein per minute.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

nptII cartridge insertions in hrpF, hrpG, hrpT, and hrpV result in three classes of mutants that are altered in their HrpZ production and HrpZ secretion phenotypes. Functionally nonpolar nptII cartridge insertions were constructed previously in each gene in the P. syringae pv. syringae 61 hrpC operon (9). To determine the effects of mutations in hrpF, hrpG, hrpT, and hrpV on HrpZ production and secretion, we performed an immunoblot analysis of HrpZ levels in the cell and supernatant fractions of cultures that were in logarithmic growth in Hrp minimal medium 5 h after being shifted from logarithmic growth in complex KB medium (Fig. 1A). Three classes of mutants could be discerned: (i) hrpF and hrpG mutants were unable to secrete HrpZ, (ii) the hrpT mutant had strongly reduced HrpZ production, and (iii) the hrpV mutant overproduced HrpZ and secreted at least some of the protein. NptII expressed from its cognate promoter in the cartridge inserted in each mutant was used as a control for constitutive gene expression and for nonspecific leakage of cytoplasmic proteins to the medium.

View larger version (39K): [in this window] [in a new window]   FIG. 1.   Effects of nptII cartridge insertions in the P. syringae pv. syringae 61 (Pss61) hrpF, hrpG, hrpT, and hrpV genes on the production and secretion of HrpZ. (A) Bacteria in early-logarithmic-phase growth in KB medium were harvested by centrifugation and resuspended at an OD600 of 0.5 in Hrp minimal medium. After a further 5 h of incubation, the cultures were fractionated into cell (C) and supernatant (S) fractions by centrifugation and assayed for the accumulation of HrpZ and NptII by SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis with appropriate antibodies and chemiluminescent immunodetection. (B) Bacteria were grown to stationary phase in KB medium and then handled as described above.

Effects on HrcC and HrcQ_B accumulation of nptII cartridge insertions in hrpF, hrpG, hrpT, and hrpV further indicate that HrpV is a negative regulator of hrp expression. To determine the effects of nptII cartridge insertions within the hrpC operon on the production of other components of the Hrp system, immunoblot analysis of HrcC and HrcQ_B (encoded by the hrpU operon) was performed on the cell fractions of various mutants grown to early logarithmic phase in complex medium and then in minimal medium (Fig. 2). Again, three classes of mutants could be discerned: (i) hrpF and hrpG mutants (with nptII insertions upstream of hrcC) produced higher levels of HrcC but wild-type levels of HrpQ_B, (ii) the hrpT mutant produced no detectable HrcC or HrpQ_B, and (iii) the hrpV mutant produced higher levels of both HrcC and HrcQ_B than did wild-type P. syringae pv. syringae 61. NptII levels provided an internal control against protein loading variability that would account for these differences. These observations indicated that in early-logarithmic-phase growth in Hrp minimal medium, certain P. syringae pv. syringae 61 hrpC operon mutantsP. syringae pv. syringae 61-N491 (hrpF::nptII) and P. syringae pv. syringae 61-N492 (hrpG::nptII)accumulate more HrcC through expression of the nptII promoter than through expression of the native promoter, and they further support the notion that overproduction of HrpV represses hrp expression whereas disruption of hrpV increases hrp expression.

View larger version (59K): [in this window] [in a new window]   FIG. 2.   Effects of nptII cartridge insertions in the P. syringae pv. syringae 61 (Pss61) hrpF, hrpG, hrpT, and hrpV genes on hrcC and hrcQB expression. Cell fractions of cultures grown for 5 h in Hrp minimal medium, as described in the legend to Fig. 1A, were assayed for the accumulation of HrcC and HrcQB by SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis with appropriate antibodies and chemiluminescent immunodetection.

Constitutive expression of hrpV in trans represses production of proteins encoded by three of the major hrp operons. To directly test the effects of altered levels of hrpV expression on regulation of the hrp system, we examined the accumulation of the products of three of the major hrp secretion operons in P. syringae pv. syringae 61 cells that were either wild type, deficient in hrpV, or overexpressing hrpV (Fig. 3). Overexpression of hrpV was achieved by transforming hrpV::nptII mutant P. syringae pv. syringae 61-N407 with pCPP2371, which has the hrpV gene expressed from the nptII promoter in vector pML122. HrcC, HrcQ_B, and HrcJ, which are products of the hrpC, hrpU, and hrpZ operons, respectively, were equally affected by these changes in hrpV expression. That is, accumulation of the three Hrc proteins was higher in P. syringae pv. syringae 61-N407 than in the wild type and almost undetectable in P. syringae pv. syringae 61-N407(pCPP2371).

View larger version (35K): [in this window] [in a new window]   FIG. 3.   Effects of mutating or overexpressing hrpV on the accumulation of Hrc proteins encoded by three different hrp operons. Bacteria grown for 5 h in Hrp minimal medium, as described in the legend to Fig. 1A, were assayed for the accumulation of HrcJ, HrcC, and HrcQB by SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis with appropriate antibodies and chemiluminescent immunodetection. Pss61, P. syringae pv. syringae 61.

Constitutive expression of hrpL or hrpRS blocks the repressive effects of hrpV overexpression. To determine at what level in the hrp regulatory cascade HrpV acts, hrpV was overexpressed in cells containing constitutively expressed hrpL or hrpRS, and the resulting levels of HrcJ production were determined (Fig. 4). HrcJ was analyzed because it is encoded by the hrpZ operon but its accumulation is not subject to apparent posttranscriptional regulation, as has been observed with HrpZ (7). Constitutive expression of hrpL or hrpRS results in constitutive expression of hrp genes under the normally repressive conditions of growth in complex media (34). To rigorously test the possibility that HrpV was an anti-sigma factor acting on HrpL we introduced plasmids into P. syringae pv. syringae 61-2074 (hrpL::TnphoA) that were designed to produce much higher levels of HrpV than HrpL. hrpL was expressed from the lac promoter in pUCP18, which is relatively weak in Pseudomonas spp., while hrpV was expressed from the nptII promoter in pML122, which is much stronger (21). This combination of copy number and promoter strength resulted in high production of HrpV, which was visible as a unique band of the predicted size on a Coomassie blue-stained SDS-polyacrylamide gel (data not shown). Nevertheless, HrpV did not reduce HrcJ production in the presence of HrpL, indicating that HrpV acts upstream of HrpL in the regulatory cascade (Fig. 4A). Similarly, hrpV failed to repress the accumulation of HrcJ that resulted from constitutive expression of hrpRS (Fig. 4B). Constitutive expression of hrpL or hrpRS also overcame the repressive effects of hrpV overexpression by pCPP2371 when the bacteria were grown in Hrp minimal medium rather than LM medium (data not shown). It is also important to note that hrpV mutant P. syringae pv. syringae 61-N407 produced no detectable HrcJ in Hrp-repressive LM medium (Fig. 4B). These observations suggest that the step at which hrpV interferes with hrp gene expression lies upstream of the hrpRS and hrpL induction cascade.

View larger version (27K): [in this window] [in a new window]   FIG. 4.   Prevention of hrpV-dependent inhibition of HrcJ accumulation by constitutive expression of hrpL and hrpRS. All bacteria were grown in LM medium and assayed for the accumulation of HrcJ by immunoblot analysis, as described in the legend to Fig. 1, but with chromogenic immunodetection. Cultures denoted hrpL carried a chromosomal hrpL::TnphoA mutation; those denoted hrpV carried a hrpV::nptII mutation. (A) Lanes: 1, P. syringae pv. syringae 61-2074; 2, P. syringae pv. syringae 61-2074(pCPP2385); 3, P. syringae pv. syringae 61-2074(pCPP2385/pCPP2371); 4, P. syringae pv. syringae 61-N407(pCPP2385). (B) Lanes: 1, P. syringae pv. syringae 61; 2, P. syringae pv. syringae 61(pCPP2389); 3, P. syringae pv. syringae 61(pCPP2389/pCPP2371); 4, P. syringae pv. syringae 61(pCPP2389/pML122); 5, P. syringae pv. syringae 61-N407; 6, P. syringae pv. syringae 61-N407(pCPP2389).

Constitutive expression of hrpV represses transcription of the hrpZ operon. The previous experiments were based on the differential accumulation of HrpZ and three Hrc proteins, particularly HrcJ, which is in the hrpZ operon. To verify that overexpression of HrpV represses transcription of the hrpZ operon, as opposed to acting posttranscriptionally, we constructed plasmid pCPP2383 (hrpA'::uidA), in which the promoter of the hrpZ operon is transcriptionally fused to a uidA gene, and analyzed -glucuronidase activity in P. syringae pv. syringae 61 derivatives that carried the plasmid and were altered in their production of HrpV and other components of the Hrp system (Fig. 5). Deletion of hrpV resulted in levels of hrpA'::uidA expression significantly higher than that in the wild type, whereas overexpression of hrpV resulted in a reduction in hrpA'::uidA expression to a level equivalent to that of hrpL and hrpS mutants. P. syringae pv. syringae 61-2088, a hrcU::TnphoA mutant that fails to export HrpZ out of the cytoplasm (7), expressed hrpA'::uidA at wild-type levels. Thus, the effects of hrpV on the accumulation of two products of the hrpZ operon, HrpZ and HrpJ, that were observed as described above can be attributed to the transcription of the operon.

View larger version (27K): [in this window] [in a new window]   FIG. 5.   Effects on hrpA'::uidA expression of altering the levels of expression of hrpV, hrpL, and hrpS. Bacteria were grown in KB medium to stationary phase and transferred to Hrp minimal medium at an initial OD600 of 0.5. -Glucuronidase activity and total protein concentration were measured at the indicated times as described in the text. The values represent the means of three replicates. The vertical lines indicate standard errors, and where they are absent, the limits were within the symbol dimensions. Except where noted, all strains carried pCPP2383 (hrpA'::uidA). The strains tested were P. syringae pv. syringae 61-N407 (hrpV), P. syringae pv. syringae 61-2088 (hrcU), P. syringae pv. syringae 61 (WT), P. syringae pv. syringae 61-N407(pCPP2371) (hrpV/hrpV++), P. syringae pv. syringae 61-2095 (hrpS), P. syringae pv. syringae 61(pCPP2371) (hrpV++), P. syringae pv. syringae 61-2074 (hrpL), and P. syringae pv. syringae 61 without pCPP2383 (WT w/o hrpA'::uidA). MU, methylumbelliferone.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

By examining the effects of nptII cartridge insertions in the four small genes that flank hrcC in the hrpC operon, we have learned that hrpF, hrpG, and hrpT have roles in HrpZ secretion, whereas hrpV encodes a negative regulator of hrp gene expression. These observations raise several questions regarding the plant reaction and biochemical phenotypes of these genes, the functions of their products in Hrp secretion and regulation, and the global control of the hrp regulon.

To properly interpret the phenotypes of these mutants, it is necessary to consider the effects of the nptII cartridge that was used to construct them. The cartridge lacks a transcription terminator, thus permitting nptII and downstream genes to be driven by the cartridge promoter and enabling the construction of mutants in complex media that support robust bacterial growth but poor hrp expression. The efficacy of this approach for testing the functions of individual genes is supported by the observation that appropriate single-gene subclones can restore wild-type HR elicitation activity to nptII-marked hrpF, hrpG, hrpT, and hrcC mutants (9). These subclones can similarly restore the ability to secrete HrpZ (data not shown). However, constitutive expression of downstream genes can have unanticipated effects, as we have found here with hrpV and will discuss further below.

It is also noteworthy that the plant reaction phenotypes of some of these hrpC operon mutations, reported in the accompanying paper (9), are not as strong as their HrpZ secretion phenotypes. We anticipated that any mutation preventing HrpZ secretion would have a strong Hrp phenotype, as was observed with hrpF and hrcC. However, the hrpG and hrpT mutants retained some ability to elicit the HR in tobacco and disease in beans. Since the physiological elicitors of the HR are now thought to be Avr proteins that are transferred into plant cells by the Hrp system, one explanation is that these mutations do not completely inhibit that process. At present, this possibility is not testable because there is no assay for P. syringae Avr protein secretion in culture or for quantitative Avr protein transfer in planta.

The functions of HrpF, HrpG, and HrpT in the Hrp secretion pathway are unknown. Although they are not present in animal-associated bacteria, they are conserved among different pathovars of P. syringae (9) and are found in other phytopathogenic bacteria, such as E. amylovora (19), suggesting that they have specific functions in plant pathogenesis. The amino acid sequence of HrpT reveals that it is a putative lipoprotein, and it may have a role as a HrcC chaperone. This notion is based on two observations: (i) homologs of HrcC that are involved in type II secretion require a similarly small lipoprotein chaperone for insertion into the outer membrane (14) and (ii) group II hrp clusters lack a lipoprotein like HrpT, and their HrcC proteins lack the C-terminal region that is thought to interact with these chaperones, whereas the HrcC proteins of group I hrp systems possess this region (8). However, the biological function of HrpT is yet to be determined in these bacteria.

HrpV appears to be a negative regulator of the HrpR/S-HrpL activator cascade and is the first negative regulator reported for the P. syringae Hrp system. However, in assessing its role it is important to note that overexpression of hrpV eliminates expression of the Hrp regulon, whereas deletion of hrpV results only in moderate increases in hrp expression. One explanation for this would be the presence of a functionally equivalent copy of hrpV elsewhere in the genome. This would be analogous to the two recently reported negative regulators, YscM1 and YscM2, of the Yop virulon (which includes the type III secretion system) in Yersinia enterocolitica (29). To seek a second copy of hrpV, we probed P. syringae pv. syringae 61 total DNA with PCR-amplified hrpV genes in a DNA gel blot at moderate stringency, but we observed no extra hybridizing bands (data not shown).

We can postulate three potential functions for HrpV as a negative regulator. First, HrpV may be a negative-feedback regulator preventing overproduction of HrcC. The location of hrpV at the end of the hrpC operon is consistent with this hypothesis, as is the observation that the X. campestris pv. vesicatoria HrcC (HrpA1) protein induces the phage shock protein operon when expressed in E. coli (32).

A second potential function for HrpV is to delay the expression of other hrp operons until the channel-forming HrcC multimers have formed in the outer membrane. This model is based on the concept that the type III protein secretion system has evolved through the recruitment of two separate translocators: a flagellum export-derived system for translocation across the inner membrane (encoded by the hrpJ and hrpU operons) and HrcC for translocation across the outer membrane. We have previously shown that P. syringae pv. syringae 61 hrcC mutants accumulate some HrpZ in the periplasm, whereas mutants affected in the hrpJ and hrpU operons accumulate HrpZ only in the cytoplasm (7). This suggests that premature expression of the hrpJ and hrpU operons could result in deleterious localization of some Hrp and Avr proteins in the periplasm. Supporting this hypothesis is the observation that hrcC (hrpA1) in X. campestris pv. vesicatoria is activated by HrpG, which is higher in the regulatory cascade than HrpX, which in turn activates the other hrp genes (33). If HrpV preferentially represses either the hrpC operon or the hrpJ and hrpU operons, this must be determined by factors in addition to HrpL that affect the expression of these operons.

A third potential function for HrpV is to increase expression of the Hrp regulon upon contact with host cells through removal of the protein from the bacterial cell. This would be analogous to the host contact-dependent secretion of LcrQ, a negative regulator of yop and ysc expression, by Yersinia spp. (23). If HrpV can be secreted, this does not appear to happen in culture, because mutations blocking Hrp secretion do not inhibit hrp expression, as would be expected with accumulation of a normally secreted negative regulator. Furthermore, there is no consensus that P. syringae hrp expression is higher in planta than in minimal media that mimic plant intercellular fluids (25, 35). However, any increase in hrp expression in planta may be hard to detect if it is transient or of moderate magnitude. Support for the hypothesis that HrpV can be secreted is found in the puzzling observation that HrpV overexpression has a relatively minor effect on elicitation of the HR in tobacco leaves (9) whereas it virtually abolishes hrp expression in culture. This would be explainable if the HrpV pool could be secreted in planta but not in culture. Regardless of which hypothesis is correct, it is important to note that the hrpV mutation does have an effect on the pathogenesis of P. syringae pv. syringae 61 in beans (9).

Our data indicate that HrpV acts upstream of HrpRS. Several aspects of HrpRS remain puzzling. There is controversy regarding whether HrpS alone or both HrpR and HrpS are required to activate hrp expression (12, 34) and whether the hrpRS operon is expressed at a higher level in planta than in culture under Hrp-derepressing conditions (25, 35). The observation that the N-terminal domains associated with regulation in related proteins are missing from HrpR and HrpS suggests that the levels of active HrpRS may be key in regulating hrp expression, and our data provide indirect evidence that high levels of HrpV inhibit HrpRS production. Key questions for the future are whether HrpV directly controls expression of the hrpR promoter, whether bacteria in planta secrete HrpV and thereby increase hrp gene expression, and why hrpV mutants overexpressing the Hrp system are impaired in pathogenesis. Given the complex functions of the Hrp pathway in protein secretion and pathogenesis, it is not surprising that there is at least one negative regulator controlling the system.