The Erwinia chrysanthemi 3937 PhoQ Sensor Kinase Regulates Several Virulence Determinants
http://www.100md.com
《细菌学杂志》
Laboratory of Plant Pathology, Faculty of Agriculture, Institute for Molecular Biology and Biotechnology, Shizuoka University, Shizuoka 422-8529, Japan,Scottish Crop Research Institute (SCRI), Invergowrie, Dundee DD2 5DA, United Kingdom
ABSTRACT
The PhoPQ two-component system regulates virulence factors in Erwinia chrysanthemi, a pectinolytic enterobacterium that causes soft rot in several plant species. We characterized the effect of a mutation in phoQ, the gene encoding the sensor kinase PhoQ of the PhoPQ two-component regulatory system, on the global transcriptional profile of E. chrysanthemi using cDNA microarrays and further confirmed our results by quantitative reverse transcription-PCR analysis. Our results indicate that a mutation in phoQ affects transcription of at least 40 genes, even in the absence of inducing conditions. Enhanced expression of several genes involved in iron metabolism was observed in the mutant, including that of the acs operon that is involved in achromobactin biosynthesis and transport. This siderophore is required for full virulence of E. chrysanthemi, and its expression is governed by the global repressor protein Fur. Changes in gene expression were also observed for membrane transporters, stress-related genes, toxins, and transcriptional regulators. Our results indicate that the PhoPQ system governs the expression of several additional virulence factors and may also be involved in interactions with other regulatory systems.
INTRODUCTION
Erwinia chrysanthemi is often described as a necrotrophic, intracellular, gram-negative plant pathogen that causes soft rot in different hosts. However, recently a biotrophic phase during the infection process of this and other closely related pathogens has been postulated (58). The pectinolytic equipment of E. chrysanthemi is the major determinant of virulence (5, 12, 45). When E. chrysanthemi infects a plant, it confronts an acidic environment in the intercellular apoplastic fluid (22). Upon perception of a variety of signals within the host, sequential synthesis of pectinolytic enzymes (pectate lyases and polygalacturonases) is induced in the bacterial cells (53). The activity of these enzymes leads to host cell lysis and subsequent alkalinization of the apoplast. Therefore, the ability of the neutrophilic E. chrysanthemi to either grow or survive in the acidic intercellular environment (acid tolerance) is important for virulence. Additional factors that contribute to its virulence are expression of siderophore-mediated iron transport systems (54), resistance to antimicrobial peptides (37), and hrp genes (6, 59). The expression of these functions essential for pathogenicity is regulated in a coordinate manner in response to different stimuli such as the presence of pectic substances, acid pH, iron and magnesium limitation, and the presence of antibacterial molecules. Two-component systems (TCSs), which comprise a sensor protein histidine kinase and a cytoplasmic response regulator that together form a signal transduction pathway via histidyl aspartyl phosphorelay, play a crucial role in this regulation. The E. chrysanthemi genome is predicted to contain genes that encode 30 TCSs (Glasner et al., unpublished results). In E. chrysanthemi as well as in a number of other bacteria, the PhoPQ TCS is activated by magnesium limitation (19, 43, 44, 48). Recently, for the best-characterized Salmonella enterica serovar Typhimurium PhoPQ system (1, 7, 15, 18, 23, 24), it was shown that antimicrobial peptides also serve as direct signals for the activation of PhoQ (4, 8). The phospho-PhoP transcriptionally activates or represses a large number of loci involved in diverse cellular functions (17, 24, 25, 29, 42), and, therefore, it has been placed on a par with global regulators such as Crp, Fnr, and ArcA (62). The E. chrysanthemi phoQ and phoP mutants have been previously partially characterized independently by two groups, and these studies have demonstrated the involvement of the E. chrysanthemi PhoPQ system in regulating synthesis of pectinolytic enzymes, its growth at acidic pH, and/or alkalization of external pH in planta, as well as its resistance to antimicrobial peptides (28, 35, 36). The PhoPQ system can participate in complex regulatory cascades that allow the integration of multiple signals and serve to expand the signal transduction network so as to allow some genes to respond to various signals in the environment (56). For example, in Salmonella, resistance to the antimicrobial peptide polymixin is mediated by the PmrA-PmrB TCS via the PhoP-dependent PmrD protein at low Mg2+. High levels of extracellular Fe3+ can activate the PmrA/B TCS independent of the PhoPQ TCS (26). In addition it has been shown that noncognate pairs of sensor kinase and response regulator may be involved in cross talks. It has been estimated that cross talk occurs in about 3% (21/692) of such pairs in Escherichia coli (60). Although such interactions have so far not been reported in E. chrysanthemi, studies with the E. chrysanthemi PhoPQ system have suggested the likelihood of a cross talk between this system and as yet unknown systems: i.e., an alternate sensor kinase may activate PhoP and/or PhoQ may activate additional response regulators, based on the behavior of the mutants with respect to magnesium concentration and pH. At high Mg2+ and acidic pH, resistance to the antibacterial peptide magainin II was severely reduced specifically in the phoQ mutant but not in the phoP mutant compared to that in the wild-type (WT) strain (28). Similarly, differences were observed with respect to virulence for the mutants at high Mg2+ concentrations (the phoQ mutant was more virulent), although at low (10 μM) magnesium concentrations, both mutants showed equally diminished virulence (35). Furthermore, in transcriptomic analysis of E. chrysanthemi phoP and phoQ mutants, it was found that the set of genes differentially regulated in the two mutants showed only minimal overlap (Venkatesh et al., unpublished results). E. chrysanthemi 3937 is widely used as a model system for functional genomic studies, and its complete genome sequence (http://asap.ahabs.wisc.edu/annotation/php/ASAP1.htm) is now available, allowing transcriptomic analyses.
Here we report results of our analysis of the global transcriptional profile of E. chrysanthemi using an E. chrysanthemi phoQ mutant, which indicate that the PhoPQ system interacts with other regulators in this important phytopathogen. Our analysis has identified a role for the system in iron-related metabolic processes, suggesting that the PhoPQ system may also regulate this important virulence determinant in planta.
MATERIALS AND METHODS
Bacterial strains, growth conditions, and biological replicates. The wild-type strain (33) and a phoQ mutant strain (kanamycin cassette insertion) of Erwinia chrysanthemi 3937 (28) were used in our study. Mutation at the phoQ locus had been confirmed by reverse transcription-PCR (RT-PCR) analysis and by Southern hybridization (28). The bacterial strains were grown in M63 minimal medium (3) at pH 7.0 containing glycerol (0.2% [wt/vol]) as the carbon source and a reduced concentration of Mg2+ (500 μM). Both strains were grown aerobically at 28°C until the bacterial optical density at 600 nm reached 2.00 for RNA isolation. For detection of siderophore activity by the chrome azurol S (CAS) assay, M63 minimal medium containing 30 mM phosphate was used since higher concentrations of phosphate interfere with the CAS assay.
Detection of siderophore activity. Siderophore activity in culture supernatants was estimated by the liquid CAS assay described by Schwyn and Neilands (55). Five-hundred-microliter portions of culture supernatants were mixed with 500 μl of CAS assay solution, and after 10 min of incubation at room temperature, the change in color was visually examined. Appropriate controls were set up with minimal medium instead of the culture supernatant.
RNA isolation and preparation of labeled cDNA. The bacterial cells were incubated for at least 30 min on ice with 1% (vol/vol) phenol (pH 4.3) and 20% (vol/vol) ethanol for stabilization of RNA. Total RNA was isolated using a QIAGEN RNeasy RNA isolation kit as described by the manufacturer (QIAGEN, Hilden, Germany). RNA was quantified using a NanoDrop ND-100 spectrophotometer (NanoDrop Technologies, Wilmington, DE), and the quality was assessed on an Agilent Bioanalyzer 2100 electrophoresis system (Agilent Technologies Inc., Palo Alto, CA).
In a reverse transcription reaction, 12 μg of total RNA was labeled in a 45-μl labeling reaction with 10 μl of a combination of short oligonucleotide (11-mer) primers (100 ng/ml) (14), 1 μl of spike RNA mix, 1.8 μl of 25x amino allyldeoxynucleoside triphosphate (dNTP) mixture, 4.5 μl of dithiothreitol (0.1 M), and 2 μl of reverse transcriptase in reaction buffer (Superscript II; Invitrogen). Primers and RNA were heated to 70°C for 10 min and cooled on ice followed by incubation for 2 h at 42°C. To denature the remaining RNA, 15 μl of 1 M NaOH and 15 μl of 0.5 M EDTA (pH 8.0) were added and incubated for 10 min at 65°C. The reaction was neutralized with 15 μl of 1 M HCl. Purification of cDNA was done as described by the supplier (QIAGEN, Hilden, Germany). The Cy dye ester was added to 10 μl of cDNA in amber tubes (QIAGEN, Hilden, Germany) in a total volume of 13 μl, containing 2.0 μl of 1 M sodium carbonate and 1.0 μl of the appropriate Cy dye (suspended in dimethyl sulfoxide) and incubated for 1 h at room temperature in the dark. To the labeled cDNA, 3.0 μl of 4.0 M hydroxylamine hydrochloride was added and incubation was extended for another 30 min in the dark. The labeled targets were combined and mixed with distilled water and further diluted with 500 μl of PB buffer (QIAGEN, Hilden, Germany) before application to a QIAGEN MiniElute column for purification. Finally, cDNA was eluted in 10 μl of elution buffer. The labeling efficiency was estimated with 1 μl of the cDNA using a NanoDrop ND-100 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
Hybridization to DNA microarrays and image analysis. Erwinia chrysanthemi microarrays, based on the genome sequence of strain 3937, have been developed at the Scottish Crop Research Institute (SCRI), Dundee, United Kingdom, through Agilent Technologies, Inc. The details of the microarray experimental protocol will be published elsewhere and will be available on request.
Labeled cDNA was made up to a volume of 80 μl with distilled water and denatured by heating at 98°C for 3 min. The denatured cDNA was mixed with 25 μl of control target and 105 μl of 2x hybridization buffer (Agilent Technologies, Inc.), and hybridization was carried out for 17 h at 60°C. After hybridization, the arrays were washed at room temperature in the dark for 1 min each with wash solution I (6x SSPE [1x SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA, pH 7.7] 0.005% N-lauroylsarcosine), followed by wash solution II (0.06x SSPE, 0.005% N-lauroylsarcosine). Arrays were scanned using an ArrayWoRx Auto scanner (Applied Precision) with appropriate exposure settings for Cy3 (595 nm) and Cy5 (685 nm) at a 9.756-μm resolution, generating separate TIFF images. Intensity data were acquired from images using SoftWoRx Tracker software (Applied Precision). Images were imported and aligned with clone position information (Agilent GAL file) using automated and manual grid alignment features. Median spot and individual median background (annulus setting) intensity values were extracted for each wavelength and imported into analysis software.
Normalization and analysis. The data from duplicate experiments were normalized and analyzed using GeneSpring version 7.2 gene expression software (Agilent Technologies, Inc.). Replicate spots in each array were averaged and data sets for each array were normalized using the Lowess algorithm. Genes were considered as upregulated in the phoQ mutant if their level of transcription was at least 1.5 times higher than the corresponding WT. Genes were considered downregulated in the phoQ mutant if they were determined to be present in the mutant at an expression level no more than two-thirds of that observed in the corresponding WT strain. The statistical significance of differential expression in the phoQ mutant was judged by the Student's t test at a P value of 0.05.
Real-time qRT-PCR. For independent confirmation of gene expression results obtained with oligonucleotide microarrays, total RNA was isolated from the WT and phoQ mutant of E. chrysanthemi 3937. RNA preparations for quantitative reverse transcription (qRT)-PCR were independent from those used for array hybridizations. After random decamer-primed first-strand cDNA was synthesized using the Omniscript RT kit (QIAGEN, Japan), real-time PCR was performed in an MX3000p Muitiplex quantitative PCR system (Stratagene) using the SYBR Premix Ex.Taq RT-PCR kit (TaKaRa, Japan). Primers were designed using the Primer Quest software (http://scitools.idtdna.com/biotools/primer_quest/primer_quest.asp), and the sequences are listed in Table 1. Primer specificity was assessed by using the dissociation curve protocol on the MX3000p Multiplex quantitative PCR system (Stratagene). The efficiency of all primer pairs was verified. PCR amplification conditions were as follows: denaturing at 95°C for 30 s, annealing at 55°C for 60 s, and extension at 72°C for 30 s for 40 cycles. All PCR experiments were performed in triplicates, and standard deviations were calculated.
The fluorescence intensity of SYBR green in each end-point of the annealing phase was detected, and the cycle threshold (CT) of each sample was calculated. The calculated CT data were used for quantitative analysis by the comparative CT method. For each amplification run, the calculated threshold cycle (CT) for each gene amplification was normalized to the CT of the 16S rRNA gene amplified from the corresponding sample before calculating the difference (fold) between the wild type and mutant using the following formula: fold change = , where CT for gene j = (CT,J – CT,16S rRNA)mutant – (CT,J – CT,16S rRNA)wild type.
Flagellar motility assay. Bacteria were inoculated with a toothpick on a minimal agar plate (M63 with glycerol, 500 μM MgSO4, 0.3% agar). The plates were then incubated at 28°C, and a halo corresponding to the spreading of bacteria from the point of inoculation was observed. To calculate the percentage of motility, the number of motile and nonmotile cells was counted in five replicate suspensions using a phase-contrast microscope (Iponacology IS-2000, Japan).
Stress assay. To test oxidative stress resistance, bacteria were inoculated into 5 ml of minimal medium (M63 with 0.2% [wt/vol] glycerol, 500 μM MgSO4) and grown for 16 h at 28°C. Under such conditions, bacterial cultures reached the stationary growth phase. H2O2 at 10 mM, which corresponds to a lethal concentration, was then added to the culture, and the cells were incubated for an additional 2 h at 28°C. The number of CFU was determined, by serial dilution and plating on LB agar, just before H2O2 treatment, just after the beginning of the challenge, and then every 20 min during 2 h. Survival of the H2O2-treated cells was normalized to the number of CFU at the beginning of the challenge. This experiment was repeated three times. Testing for osmotic stress resistance was similarly carried out by incubating stationary-phase cells in the presence of 1 M NaCl.
RESULTS
Transcriptome analysis. The availability of the genome sequence of Erwinia chrysanthemi 3937 allowed the development and utilization of cDNA microarrays to characterize the constitutive role of the phoQ gene at the transcriptome level. In order to identify loci that may be controlled by phoQ, the transcriptional profile of an E. chrysanthemi 3937 mutant strain lacking a functional phoQ was compared with that of the wild-type strain, E. chrysanthemi 3937. Total RNA was isolated from these strains, and cDNA was synthesized and labeled with the fluorescent dye Cy3 or Cy5 and hybridized to an E. chrysanthemi 3937 microarray (for details, see Materials and Methods) that contained 4,750 gene-specific spots (excluding duplicates or triplicates of a single gene). The results presented below are based on the mean of four normalized expression values derived from two biological replicates and two dye-swap experiments.
Our analysis revealed differences in expression levels for a number of genes, the majority of which were not previously known to be regulated by the PhoPQ regulatory system (Fig. 1). A total of 87 open reading frames (ORFs) appeared to be upregulated at least 1.5-fold in the phoQ mutant, and 26 ORFs appeared to be downregulated by the same degree (Fig. 2). Of these, only 40% (35 genes) of upregulated and 19% (5 genes) of downregulated ORFs were statistically significant, when a Student's t test was performed with a P value of less than 0.05 (Table 2). Microarray data were complemented with quantitative real-time PCR data for a subset of genes whose expression levels differed significantly between the mutant and wild-type strains. Based on the results, it appears that phoQ may predominantly effect transcriptional repression at several loci. The differentially expressed genes (Table 2) have been classified according to their putative function as follows.
Genes with unknown function and conserved hypothetical products. Among the 40 genes whose expression is altered, the function of 5 (upregulated) is not known (Table 2). No significant matches exist for these genes in the currently available databases. Seven of the differentially expressed genes (five upregulated and two downregulated) are at present thought to encode hypothetical products. Putative functions could be assigned to six of the seven genes, based on comparisons at the amino acid level using BLAST. For convenience, these genes have been discussed below in the functional classification sections.
The remaining 27 genes encode proteins with known functions, and these have been broadly classified based on these functions.
Iron assimilation. At least 37% of the genes that demonstrated enhanced expression in the mutant are involved in iron metabolism and include genes responsible for siderophore-mediated assimilation of iron (Table 2). Of the eight genes (acsF, acr, acsD, acsE, yhcA, acsC, acsB, and acsA) that have been reported (20) to constitute the achromobactin biosynthesis operon (acs) in Erwinia chrysanthemi, the transcript levels of five, namely acsA, acsC, acsD, acsE, and acsF, were significantly higher in the phoQ mutant. Each of these five gene products catalyzes an individual step in the biosynthesis of achromobactin (9, 20). A qRT-PCR analysis showed that the level of the acsE gene transcript was the highest (317-fold), followed by those of acsA and acsD (11-fold) and acsC and acsF (about 2-fold). Although it appeared from microarray analysis that the transcript levels of the other three genes (acsB, yhcA, and acr) were only marginally higher in the mutant strain, when compared to the WT, qRT-PCR analysis revealed the transcript levels of these three genes to be significantly higher in the mutant (Table 3). The proteins encoded by yhcA and acr participate in transport of the siderophore-iron complex, and acsB is involved in achromobactin biosynthesis.
Transcript levels of genes encoding proteins of the Ton complex (exbB, exbD, and tonB), which shuttles the iron-siderophore (ferriachromobactin) complex across the outer membrane, were also higher in the mutants. While transcript levels of exbB and exbD were significantly higher, that of tonB was only slightly higher (1.37-fold, P = 0.026). Of the two auxiliary proteins coded for by exbB and exbD, the former showed high transcript levels in the mutant (47-fold). In the mutant, the transcript levels of cbrA, encoding the periplasmic component of a permease complex that transports the liganded form of achromobactin (38), and of fhuC, encoding a cytoplasmic membrane-associated ATP-binding protein permease that is necessary for the transport of ferrichrome compounds from the periplasm across the cytoplasmic membrane into the cytoplasm, were about 12- and 15-fold higher, respectively, based on qRT-PCR data (Table 3).
Another gene that may have a role in iron transport and whose expression levels were higher in the phoQ mutant is yqjH (identification no. [ID] 16680). Analysis using BLAST shows that the predicted product is about 40% similar to iron-utilization proteins present in some proteobacteria (Pseudomonas, Bordetella, and Ralstonia) and in some actinomycetales (Kineococcus, Arthrobacter, Streptomyces, Corynebacterium, and Brevibacterium). The YqjH homolog in Escherichia coli (27% identical to E. chrysanthemi YqjH) is involved in the utilization of the catechol siderophore vibriobactin (51). Other genes that may function in iron transport are ycdO (ID 15840), whose predicted product is similar to a predicted periplasmic lipoprotein from Yersinia pestis (gi|45436217; 81% identical and 87% similar over a stretch of 376 amino acids [aa]) involved in iron transport, and a yet to be annotated gene (ID 18178) whose predicted product is highly similar to a putative periplasmic substrate-binding transport protein (gi|49612694) from E. carotovora subsp. atroseptica. The closest match (58% identity, 74% similarity) to the E. carotovora protein is an ABC transporter that binds iron in Agrobacterium tumefaciens.
Another iron-related transcript whose level was elevated in the mutant is bfd (ID 15304), which encodes a bacterioferritin-associated ferredoxin. This gene is thought to have a role in the release of iron from (or delivery of iron to) bacterioferritin. The qRT-PCR expression data for these genes are summarized in Table 3. These data suggest that phoQ may have a regulatory role in the uptake of iron via the achromobactin system in E. chrysanthemi.
General metabolism. Three induced transcripts belong to genes encoding key enzymes that participate in reactions contributing to maintenance of overall metabolism within the cell. Among these is gltA, which encodes citrate synthase, an important tricarboxylic acid cycle enzyme. Of the other two, one encodes a tRNA guanosine methyl transferase (trmD; ID 20494) that specifically methylates guanosine in several tRNAs and prevents frameshifting. The other (ID 17546) encodes a protein that is 79% identical to a putative methionine synthase (gi|49610578) from E. carotovora subsp. atroseptica and about 66% identical to vitamin B12-independent methionine synthases from Chromohalobacter salexigens and Rhodopseudomonas palustris. This enzyme catalyzes the last step in the biosynthesis of methionine.
Ion transport. The mutant showed elevated transcription of the delta subunit of ATPase (atpH; ID 14789) and of two other genes (ID 15807 and 16441) whose predicted protein products show identity to ATP-binding cassettes from other bacteria. The gene with ID 15807 shows up to 70% identity to ABC transporters that traffic anions such as nitrate, phosphate, and sulfonate in some proteobacteria such as Pseudomonas, Xylella, and Zymomonas and probably has a similar function in E. chrysanthemi. No significant match for this gene exists in E. carotovora subsp. atroseptica. For the other gene (ID 16441), no close homologs are present within Enterobacteriaceae. The nearest matches are present in Ralstonia, Desulfitobacterium, Methanosarcina, and Bacillus.
Conversely, transcripts of cusA involved in copper efflux were reduced in the mutant. Another downregulated gene (ID 16479) is highly similar (75 to 90%) to those coding for outer membrane lipoproteins from several bacteria, including E. carotovora subsp. atroseptica (80% similarity), and encodes an outer membrane beta barrel porin that may be involved in drug efflux.
Stress response. The stress-responsive transcripts whose levels were higher in the mutant are sodC (ID 17092), encoding a Cu-Zn superoxide dismutase; proW, encoding the ABC transporter component of OusB; and fliC, which encodes flagellin, the filament structural protein of flagella.
Among the downregulated genes encoding conserved hypothetical products, the product of one (ID 15573) is 70% identical to a putative exported protein found in Erwinia carotovora subsp. atroseptica and between 60 and 70% identical to conserved hypothetical proteins present in Brucella spp. Interestingly, this protein product also shows a slightly lower level of identity (about 50%) with hypothetical signal peptide proteins from Ralstonia solanacearum and Photobacterium profundum. The closest matches to this protein show a cystatin-like fold present in cystein protease inhibitors.
Toxins. The mutant showed elevated transcript levels of a cluster of genes (ID 16663, 16664, and 16665) that are predicted to encode proteins similar to the cytolytic (cyt) delta-endotoxins of Bacillus thuringiensis (Glasner et al., unpublished). To our knowledge, this is the first report of transcription of genes within this cluster. qRT-PCR analysis revealed at least a 10-fold increase of transcript levels for genes 16664 and 16665 in the mutant. It would be interesting to see whether these proteins function as toxins/virulence factors in E. chrysanthemi. Although a gene encoding a protein product similar to the type 1AA cytolytic endotoxin is also present in the E. chrysanthemi genome, this gene was not induced in the mutant. A gene (ID 16850) that encodes a conserved hypothetical product that shows a conserved epimerase domain and shares about 60% identity with hemolysins from Yersinia pestis and Escherichia coli was downregulated in the mutant.
Transcriptional regulators. At least five transcriptional regulators showed altered expression in the mutant. Of these, four were upregulated. These are glnL, encoding the sensor kinase component of the nitrogen-responsive two-component regulatory system; pecS, encoding the global transcriptional repressor of pectate lyase and other macerating enzymes such as cellulase; vfmE, a transcriptional regulator that governs multiple virulence factors in E. chrysanthemi; and an unknown transcriptional repressor (ID 20250), perhaps of the Cro/CI family, based on homology. The phoP gene, the regulatory counterpart of phoQ, was slightly repressed in the mutant strain compared to the WT strain.
We carried out preliminary in silico analysis for two of the regulators for which little information exists in the current databases.
vfmE. The vfmE gene is a member of the AraC family of transcriptional regulators. The vfmE product is 187 amino acids long, and its closest match is a 326-aa-long transcriptional regulator of the AraC family from Desulfovibrio vulgaris (38% identical and 61% similar over a stretch of 102 aa) with a yet-to-be-assigned function. Apparently, no homologs exist in E. carotovora subsp. atroseptica. In an E. chrysanthemi 3937 vfmE mutant, production of pectinases, cellulase, and protease was reduced by a factor of 10-fold (S. Reverchon, personal communication).
The unknown transcriptional regulator (ID 20250) is probably a member of the Cro/CI family of transcriptional repressors. The E. chrysanthemi genome is reported to encode at least 28 members that belong to this family (Glasner et al., unpublished). A BLAST analysis of the predicted protein sequence against the nonredundant database at NCBI revealed the closest match (56% identity and 74% similarity) to be an immunity repressor protein of the Cro/CI family from the Salmonella phage. The closest protein matches from bacteria are also repressors and show less homology (<35% identity and <60% similarity).
In addition to validation by qRT-PCR, we carried out the following phenotypic assays to further evaluate our microarray results.
Siderophore detection in culture supernatant. In the CAS assay for the detection of siderophores, the phoQ mutant produced more siderophore than the WT (Fig. 2) in the presence of 500 μM MgSO4 when iron was supplied. When the medium was starved for iron, no differences were observed between the WT and mutant strains with respect to siderophore production. No color change was apparent for the WT and mutant strains at a high Mg2+ concentration.
Flagellar motility assay. The strains were tested for motility on 0.3% minimal agar plates. After 24 h of incubation, the mutant migrated relatively extensively through the soft agar medium, whereas the wild-type strain migrated to a lesser extent (Fig. 3). The percentage of motile cells determined in cell suspensions was slightly higher in the phoQ mutant (92% ± 1%) than in the WT strain (87% ± 3%).
Stress assay. To determine whether E. chrysanthemi 3937 lacking in PhoQ showed increased resistance to oxidative stress, we analyzed the survival of the phoQ mutant under oxidative stress generated by the addition of 10 mM H2O2 to the M63 minimal medium. We found that the survival of the phoQ mutant was only marginally higher than that of the WT strain (Fig. 4A). A much higher difference was observed between the WT and phoQ mutant strains with respect to osmotic stress tolerance (Fig. 4B).
DISCUSSION
Bacterial cells face a variety of adverse conditions in the natural environment, such as nutrient limitation, temperature and acid shock, osmotic stress, and depletion of metal ions. Such adverse changes in the environment are associated with the encounter of pathogenic bacteria with a potential host and therefore serve as cues to augment a pathogenic life cycle. This recognition of a host by the pathogen triggers modifications in cellular metabolism necessary for survival of the pathogen and successful colonization of its host and often involves a complex interplay of multiple regulatory factors.
Two-component regulatory systems constitute an important part of this regulatory network and mediate adaptive responses to a number of external stimuli. The PhoPQ system responds to Mg2+ limitation. Previous studies have demonstrated that E. chrysanthemi phoQ mutants impaired in the sensor kinase differed from the phoP mutants impaired for the cognate response regulator (28, 35). These studies suggest that an alternate sensor kinase may activate PhoP and/or PhoQ may activate additional response regulators. We carried out a transcriptomic analysis with both PhoP and PhoQ mutants. Interestingly, there was only minimal overlap for the two regulons (data not shown), further suggesting that the sensor and regulator of the PhoPQ system may not be mutually exclusive. Here we report the transcriptomic profile of the phoQ mutant that showed a more drastic phenotype change than the phoP mutant when grown in the presence of 500 μM MgSO4. We chose this concentration because the phoQ mutant surprisingly showed much higher maceration ability on Chinese cabbage than either the wild type or the phoP mutant. At 10 μM and 10 mM concentrations, the mutant behaved as expected: that is, it showed much reduced macerating capacity in the former and higher capacity in the latter (Venkatesh et al., unpublished data).
When grown in the presence of 500 μM MgSO4, the absence of a functional phoQ was correlated with derepression of several genes. The unexpected induction of transcription of a number of genes in the mutant strain lacking a functional PhoQ sensor kinase may reflect the inherent potential of phoQ to affect transcriptional repression of these (or at least at some of these) loci. Although it is possible that at this concentration of Mg2+, an alternate sensor kinase might phosphorylate PhoP, we did not find convincing evidence for such a phenomenon. Only five genes were downregulated in the mutant, including phoP, which encodes the cognate cytoplasmic regulator operating with the PhoQ sensor kinase (Table 2). Functions are yet to be assigned for some of the differentially expressed genes. Since many of the loci are also known to be governed by other regulators, it appears that the PhoQ sensor kinase, and thereby the PhoPQ TCS, has overlapping functions with other regulatory systems.
Microarray analysis showed that several genes related to iron uptake were induced in the mutant. In fact, some earlier studies have demonstrated that under Mg2+-limiting conditions (which activate PhoPQ TCS), E. coli and S. enterica serovar Typhimurium take up large amounts of ferrous iron (27, 32). In line with this, Chamnongpol and Groisman (10) suggested a probable role for the PhoPQ TCS in resistance to high Fe2+ based on the observations that an inactive PhoPQ system rendered Salmonella hypersensitive to Fe2+ iron. The toxicity was due to the uptake of Fe2+ by the nonspecific transporter CorA, whose activity was altered in response to the PhoP-regulated determinants mgtA and mgtB. However, the mechanism of iron uptake in the present context is different since, as discussed below, it is mediated by siderophores that are specific for the Fe3+ form of iron and occurs in the absence of added iron. This uptake of iron was nontoxic, as both the WT and the mutant strain grew normally in the medium tested (data not shown). Interestingly, different levels of induction were evident for the iron assimilation genes—those for the production of the siderophore achromobactin were strongly derepressed, and genes involved in iron uptake functions were weakly derepressed (Table 3). A similar pattern of regulation has been reported in an E. coli fur mutant for the enterobactin biosynthesis and uptake during iron limitation (41). The acs operon in E. chrysanthemi is governed by the Fur repressor, and Franza et al. (21) showed the constitutive derepression of the acs::lacZ fusion transcript in an E. chrysanthemi fur mutant background leading to the uptake of iron. In fact, in our study except for one unknown ORF (ID 18178), all of the differentially expressed genes with iron-related functions have been previously identified as members of the fur regulon from other bacteria. Furthermore, the induction of some non-iron-related genes could be related to common regulation by fur or could be a consequence of induction of other loci by fur. For example, fur controls the expression of fliC and thereby couples iron status to motility. Expression of the flagella is inhibited when bacteria have an excess of the metal so as to prevent the departure of the cells toward other environments with less available iron (40, 57). It has been recently shown that the FliC homolog of Erwinia carotovora subsp. carotovora (64% identity, 79% similarity to E. chrysanthemi FliC) is essential for motility (30). An increased level of yet another transcript, namely that of gltA, encoding citrate synthase correlates well with the fact that citrate is the precursor for achromobactin. However, not all fur-regulated genes were differentially expressed. Thus, it appears that a mutation in phoQ results in the derepression of a subset of genes of the fur regulon. These data provide indirect evidence for cross talk between the PhoPQ TCS and the Fur repressor. The physiological implication for such a phenomenon, if operative in planta, awaits investigation since in the presence of 10 μM MgSO4 some iron-regulated genes such as acsA, acsE, and cbrA showed increased transcript levels in the WT whereas some iron-related genes such as acsF and yqjH showed a decrease in the transcript level (qRT-PCR; data not shown).
Other than the transporters involved in iron uptake and related functions, lack of a functional phoQ resulted in the induction of expression of three additional transporters and the repression of two (Table 2). The differential induction of membrane transporters may reflect altered ion fluxes between the bacterial cells and the surrounding medium to maintain homeostasis. Two transcripts (proW and trmD) that are commonly induced under osmotic stress were also higher in the mutant and may contribute to its higher survivability than the WT strain (Fig. 4B) in the presence of 1 M NaCl. ProW is the ABC transporter component of OusB that is involved in adaptation to osmotic stress in Erwinia chrysanthemi (11). Consistent with the increased transcript levels of sodC (ID 17092), the phoQ mutant showed slightly higher survivability in the presence of hydrogen peroxide (Fig. 4A). The Salmonella sodC gene encodes the periplasmic, phage-associated Cu-Zn superoxide dismutase (SodC1) that is present only in some highly virulent strains and modulates interaction with the host to promote bacterial survival (2).
The fact that achromobactin, like other siderophores, is an important virulence factor (47, 50) presents an interesting dimension. The expression of this virulence factor appears to be coupled to the expression of other virulence factors such as the PecS protein that can repress the production of the degradative enzymes (i.e., pectate lyases and cellulases) (5, 52) and can induce the synthesis of polygalacturonase enzymes (31, 46), and it also has a role in rendering resistance to the products of oxidative burst via an extracellular blue pigment called indigoidine (52). In order to examine whether the effect of the phoQ mutation extends to genes controlled by PecS, we carried out qRT PCR analysis for some such genes. Our results showed that the transcript levels of genes encoding pectate lyases PelA, PelB, PelC, PelD, and PelE were much less in the phoQ mutant than in the wild-type strain. Microarray analyses for these genes also indicate downregulation (Table 3). Our results are consistent with previously published results of Llama Palacios et al. (35) and Haque and Tsuyumu (28) that in phoQ mutants of E. chrysanthemi 3937, the production of pectate lyases is drastically reduced (up to 90% of WT). Our qRT-PCR results for PecS-regulated genes support the reported contention that the exclusion of genes with low microarray expression scores from further analysis may not be justified (13). It has been shown that microarray analyses underestimate the change (fold) of regulated genes in comparison with the quantitative real-time PCR assay either because of overestimating the uninduced transcripts, perhaps due to nonspecific hybridization, or underestimating the level of induction, perhaps due to probe saturation effects (61).
In addition, the transcriptional regulator vfmE, which controls multiple virulence factors in E. chrysanthemi (S. Reverchon, personal communication), and the Cro/CI family transcriptional regulator involved in functions that prevent a lytic life cycle (and therefore may have a role in increasing survival within a host) might also act as virulence determinants. The possible interplay of these regulatory factors with phoQ, as suggested by their enhanced expression (Table 2) in the absence of a functional phoQ, may be interesting to pursue.
In conclusion, E. chrysanthemi produces two siderophores, chrysobactin (49) and achromobactin, in response to iron limitation in planta. These are structurally different and require different levels of iron limitation for their induction (16, 38, 39). Additionally, their production is temporally separated so that achromobactin synthesis precedes chrysobactin synthesis (20). Therefore, it is reasonable that, in addition to the coordinate control of the global regulatory protein Fur, specific mechanisms of regulation may be operating at these loci via effectors, for example, depending on the nature of the stimulus. In fact, Franza and coworkers (21) hinted at the possible involvement of such additional molecules for the chrysobactin system, but not for the achromobactin system, which was thought to be only fur dependent. However, since their investigation was carried out in a normal PhoQ background, it is possible that an additional effect was not evident for the achromobactin system. Our analysis on a transcriptomic scale suggests that an analogous mechanism apparently operates for the achromobactin system as well, at least under certain conditions. Given that pleiotropic effects caused by phoQ mutation in E. chrysanthemi include derepression of some iron metabolism genes, a phenotype also shared with fur mutants of E. chrysanthemi, it seems plausible that at least some of the genetic components of both regulons might be shared. The mechanism to actively lower the availability of Mg2+ is analogous to the sequestration of iron, and both mechanisms are recognized as nonspecific host defense strategies. Therefore, it is possible that the PhoPQ system has been recruited to regulate adaptations to both magnesium- and iron-limiting environments, albeit to different extents. We speculate that a product of the PhoP-PhoQ two-component system effects the repression of the achromobactin synthesis by fur and thereby prevents the needless squandering of energy when iron levels are sufficient. From the present study, the mediation of iron assimilation (and thereby a possible cross talk with the fur regulon) emerges as a novel role for the PhoQ sensor kinase. This concept, however, needs experimental validation.
To summarize, microarray data support the involvement of the PhoQ sensor kinase in a wide variety of cellular functions in E. chrysanthemi, including iron metabolism, synthesis of virulence factors, proton pumping, stress tolerance, and DNA repair functions, although results from this study alone are not sufficient to understand the precise role of the phoQ gene in these functions. Further investigation is required to understand the complex nature of the results presented here that hint toward possible additional roles for the PhoPQ system in planta. In many bacteria, the production of virulence factors is regulated by more than one mechanism to facilitate coordinate regulation (34). The PhoPQ system may operate as part of the multiple regulatory mechanisms that govern virulence in E. chrysanthemi. A better understanding of the PhoPQ system will further our knowledge of the mechanisms of pathogenicity of E. chrysanthemi and related organisms and may help us devise more appropriate disease control strategies.
ACKNOWLEDGMENTS
We thank Manjurul Haque, Shizuoka University, for the gift of the PhoP and PhoQ mutant strains that were used in this study. We gratefully acknowledge the help of Jacques Schrenzel and Brian Gettler from the University Hospitals of Geneva, Switzerland, for their assistance in designing the microarray probes. We thank Jenny Morris, SCRI, for technical help. We sincerely thank the anonymous referees for their valuable comments and suggestions.
REFERENCES
Adams, P., R. Fowler, N. Kinsella, G. Howell, M. Farris, P. Coote, and C. D. Connor. 2001. Proteomic detection of PhoPQ- and acid-mediated repression of Salmonella motility. Proteomics 1:597-607.
Ammendola, S., M. Ajello, P. Pasquali, J. S. Kroll, P. R. Langford, G. Rotilio, P. Valenti, and A. Battistoni. 2005. Differential contribution of sodC1 and sodC2 to intracellular survival and pathogenicity of Salmonella enterica serovar Choleraesuis. Microbes Infect. 7:698-707.
Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl. 1987. Current protocols in molecular biology. John Wiley & Sons, New York, N.Y.
Bader, M. W., S. Sanowar, M. E. Daley, A. R. Schneider, U. Cho, W. Xu, R. E. Klevit, H. LeMoual, and S. I. Miller. 2005. Recognition of antimicrobial peptides by a bacterial sensor kinase. Cell 22:461-472.
Barras, F., F. Van Gijsegem, and A. K. Chatterjee. 1994. Extracellular enzymes and pathogenesis of soft-rot Erwinia. Annu. Rev. Phytopathol. 32:201-234.
Bauer, D. W., A. J. Bogdanove, S. V. Beer, and A. Collmer. 1994. Erwinia chrysanthemi hrp genes and their involvement in soft rot pathogenesis and elicitation of the hypersensitive response. Mol. Plant-Microbe Interact. 7:573-581.
Bijlsma, J. J. E., and E. A. Groisman. 2005. The PhoP/PhoQ system controls the intramacrophage type three secretion system of Salmonella enterica. Mol. Microbiol. 57:85-96.
Bishop, J. L., and B. B. Finlay. 2006. Friend or foe Antimicrobial peptides trigger pathogen virulence. Trends Mol. Med. 12:3-6.
Challis, L. G. 2005. A widely distributed bacterial pathway for siderophore biosynthesis independent of nonribosomal peptide synthetases. ChemBioChem 6:601-611.
Chamnongpol, S., and E. A. Groisman. 2002. Mg2+ homeostasis and avoidance of metal toxicity. Mol. Microbiol. 44:561-571.
Choquet, G., N. Jehan, C. Pissavin, C. Blanco, and M. Jebbar. 2005. OusB, a broad-specificity ABC-type transporter from Erwinia chrysanthemi, mediates uptake of glycine betaine and choline with a high affinity. Appl. Environ. Microbiol. 71:3389-3398.
Collmer, A., and N. T. Keen. 1986. The role of pectic enzymes in plant pathogenesis. Annu. Rev. Phytopathol. 24:383-409.
Dallas, B. P., N. G. Gottardo, M. J. Firth, A. H. Beesley, K. Hoffmann, P. A. Terry, J. R. Freitas, J. M. Boag, A. J. Cummings, and U. R. Kees. 2005. Gene expression levels assessed by oligonucleotide microarray analysis and quantitative real-time RT-PCR—how well do they correlate BMC Genomics 6:59. [Online.]
Dellagi, A., J. Helibronn, A. O. Avrova, M. Montesano, E. T. Palva, H. E. Stewart, I. K. Toth, D. E. Cooke, G. D. Lyon, and P. R. Birch. 2000. A potato gene encoding a WRKY-like transcription factor is induced in interactions with Erwinia carotovora subsp. atroseptica and Phytophthora infestans and is coregulated with class I endochitinase expression. Mol. Plant-Microbe Interact. 13:1092-1101.
Detweiler, C. S., D. B. Cunanan, and S. Falkow. 2001. Host microarray analysis reveals a role for the Salmonella response regulator phoP in human macrophage cell death. Proc. Natl. Acad. Sci. USA 98:5850-5855.
Expert, D., C. Sauvage, and J. B. Neilands. 1992. Negative transcriptional control of iron transport in Erwinia chrysanthemi involves an iron-responsive two-factor system. Mol. Microbiol. 6:2009-2017.
Fields, P. I., E. A. Groisman, and F. Heffron. 1989. A Salmonella locus that controls resistance to microbicidal proteins from phagocytic cells. Science 243:1059-1062.
Fields, P. I., R. V. Swanson, C. G. Haidaris, and F. Heffron. 1986. Mutants of Salmonella typhimurium that cannot survive within the macrophage are avirulent. Proc. Natl. Acad. Sci. USA 83:5189-5193.
Flego, D., R. Marits, A. R. Eriksson, V. Koiv, M. B. Karlsson, R. Heikinheimo, and E. T. Palva. 2000. A two-component regulatory system, pehR-pehS, controls endopolygalacturonase production and virulence in the plant pathogen Erwinia carotovora subsp. carotovora. Mol. Plant-Microbe Interact. 13:447-455.
Franza, T., B. Mahe, and D. Expert. 2005. Erwinia chrysanthemi requires a second iron transport route dependent of the siderophore achromobactin for extracellular growth and plant infection. Mol. Microbiol. 55:261-275.
Franza, T., C. Sauvage, and D. Expert. 1999. Iron regulation and pathogenicity in Erwinia chrysanthemi 3937: role of the fur repressor protein. Mol. Plant-Microbe Interact. 12:119-128.
Grignon, C., and H. Sentenac. 1991. pH and ionic conditions in the apoplast. Annu. Rev. Plant Physiol. 42:103-128.
Groisman, E. A. 2001. The pleiotropic two-component regulatory system PhoP-PhoQ. J. Bacteriol. 183:1835-1842.
Groisman, E. A., E. Chiao, C. J. Lipps, and F. Heffron. 1989. Salmonella typhimurium phoP virulence gene is a transcriptional regulator. Proc. Natl. Acad. Sci. USA 86:7077-7081.
Guina, T., E. C. Yi, H. Wang, M. Hackett, and S. I. Miller. 2000. A PhoP-regulated outer membrane protease of Salmonella enterica serovar Typhimurium promotes resistance to alpha-helical antimicrobial peptides. J. Bacteriol. 182:4077-4086.
Gunn, J. S., and S. I. Miller. 1996. PhoP-PhoQ activates transcription of pmrAB, encoding a two-component regulatory system involved in Salmonella typhimurium antimicrobial peptide resistance. J. Bacteriol. 178:6857-6864.
Hantke, K. 1987. Selection procedure for deregulated iron transport mutants (fur) in Escherichia coli K12: fur not only affects iron metabolism. Mol. Gen. Genet. 210:135-139.
Haque, M. M., and S. Tsuyumu. 2005. Virulence, resistance to magainin II and expression of pectate lyase are controlled by the PhoP-PhoQ two component regulatory system responding to pH and magnesium in Erwinia chrysanthemi 3937. J. Gen. Plant Pathol. 71:47-53.
Hitchen, P. G., J. L. Prior, P. C. Oyston, M. Panico, B. W. Wren, R. W. Titball, H. R. Morris, and A. Dell. 2002. Structural characterization of lipo-oligosaccharide (LOS) from Yersinia pestis: regulation of LOS structure by the PhoPQ system. Mol. Microbiol. 44:1637-1650.
Hossain, M. M., S. Shibata, S. I. Aizawa, and S. Tsuyumu. 2005. Motility is an important determinant for pathogenesis of Erwinia carotovora subsp. carotovora. Physiol. Mol. Plant Pathol. 66:134-143.
Hugouvieux-Cotte-Pattat, N., V. E. Shevchik, and W. Nasser. 2002. PehN, a polygalacturonase homologue with a low hydrolase activity, is coregulated with the other Erwinia chrysanthemi polygalacturonases. J. Bacteriol. 184:2664-2673.
Kot, E., and A. Bezkorovainy. 1993. Effects of Mg2+ and Ca2+ on Fe2+ uptake by Bifidobacterium thermophilum. Int. J. Biochem. 25:1029-1033.
Kotoujansky, A., M. Lemattre, and P. Boistard. 1982. Utilization of a thermosensitive episome bearing transposon Tn10 to isolate Hfr donor strains of Erwinia carotovora subsp. chrysanthemi. J. Bacteriol. 150:122-131.
Litwin, C. M., and S. B. Calderwood. 1993. Role of iron in regulation of virulence genes. Clin. Microbiol. Rev. 6:137-149.
Llama-Palacios, A., E. Lopez-Solanilla, and P. Rodríguez-Palenzuela. 2005. Role of the PhoP-PhoQ system in the virulence of Erwinia chrysanthemi strain 3937: involvement in sensitivity to plant antimicrobial peptides, survival at acid pH, and regulation of pectolytic enzymes. J. Bacteriol. 187:2157-2162.
Llama-Palacios, A., E. Lopez-Solanilla, C. Poza-Carrion, F. García-Olmedo, and P. Rodríguez-Palenzuela. 2003. The Erwinia chrysanthemi phoP-phoQ operon plays an important role in growth at low pH, virulence and bacterial survival in plant tissue. Mol. Microbiol. 49:347-357.
Lopez-Solanilla, E., A. Llama-Palacios, A. Collmer, F. García-Olmedo, and P. Rodríguez-Palenzuela. 2001. Relative effects on virulence of mutations in the sap, pel, and hrp loci of Erwinia chrysanthemi. Mol. Plant-Microbe Interact. 14:386-393.
Mahe, B., C. Masclaux, L. Rauscher, C. Enard, and D. Expert. 1995. Differential expression of two siderophore-dependent-iron acquisition pathways in Erwinia chrysanthemi 3937: characterization of a novel ferrisiderophore permease of the ABC transporter family. Mol. Microbiol. 18:33-43.
Masclaux, C., and D. Expert. 1995. Signaling potential of iron in plant microbe interactions: the pathogenic switch of iron transport in Erwinia chrysanthemi. Plant J. 7:121-128.
McCarter, L., and M. Silverman. 1989. Iron regulation of swarmer cell differentiation of Vibrio parahaemolyticus. J. Bacteriol. 171:731-736.
McHugh, J. P., F. Rodriguez-Quinones, H. Abdul-Tehrani, D. A. Svistunenko, R. K. Poole, C. E. Cooper, and S. C. Andrews. 2003. Global iron-dependent gene regulation in Escherichia coli. A new mechanism for iron homeostasis. J. Biol. Chem. 278:29478-29486.
Miller, S. I., A. M. Kukral, and J. J. Mekalanos. 1989. A two-component regulatory system (phoP phoQ) controls Salmonella typhimurium virulence. Proc. Natl. Acad. Sci. USA 86:5054-5058.
Minagawa, S., H. Ogasawara, A. Kato, K. Yamamoto, Y. Eguchi, T. Oshima, H. Mori, A. Ishihama, and R. Utsumi. 2003. Identification and molecular characterization of the Mg2+ stimulon of Escherichia coli. J. Bacteriol. 185:3696-3702.
Monsieurs, P., S. Keersmaecker, W. W. Navarre, M. W. Bader, F. deSmet, M. McClelland, F. C. Fang, B. deMoor, J. Vanderleyden, and K. Marchal. 2005. Comparison of the PhoPQ regulon in Escherichia coli and Salmonella typhimurium. J. Mol. Evol. 60:462-474.
Nachin, L., and F. Barras. 2000. External pH: an environmental signal that helps to rationalize pel gene duplication in Erwinia chrysanthemi. Mol. Plant-Microbe Interact. 13:882-886.
Nasser, W., V. E. Shevchik, and N. Hugouvieux-Cotte-Pattat. 1999. Analysis of three clustered polygalacturonase genes in Erwinia chrysanthemi 3937 revealed an anti-repressor function for the PecS regulator. Mol. Microbiol. 34:641-650.
Neilands, J. B., and S. A. Leong. 1986. Siderophores in relation to plant disease. Annu. Rev. Plant Physiol. 37:187-208.
Newcombe, J., J. C. Jeynes, E. Mendoza, J. Hinds, G. L. Marsden, R. A. Stabler, M. Marti, and J. J. McFadden. 2005. Phenotypic and transcriptional characterization of the meningococcal PhoPQ system, a magnesium-sensing two-component regulatory system that controls genes involved in remodeling the meningococcal cell surface. J. Bacteriol. 187:4967-4975.
Persmark, M., D. Expert, and J. B. Neilands. 1989. Isolation, characterization and synthesis of chrysobactin, a compound with a siderophore activity from Erwinia chrysanthemi. J. Biol. Chem. 264:3187-3193.
Ratledge, C., and L. G. Dover. 2000. Iron metabolism in pathogenic bacteria. Annu. Rev. Microbiol. 54:881-941.
Redford, P., P. L. Roesch, and R. A. Welch. 2003. degS is necessary for virulence and is among extraintestinal Escherichia coli genes induced in murine peritonitis. Infect. Immun. 71:3088-3096.
Reverchon, S., C. Rouanet, D. Expert, and W. Nasser. 2002. Characterization of indigoidine biosynthetic genes in Erwinia chrysanthemi and role of this blue pigment in pathogenicity. J. Bacteriol. 184:654-665.
Robert-Baudouy, J., W. Nasser, G. Condemine, S. Reverchon, V. E. Shevchik, and N. Hugovieux-Cotte-Pattat. 2000. Pectic enzymes of Erwinia chrysanthemi regulation and role in pathogenesis. Plant-Microbe Interact. 5:221-368.
Sauvage, C., and D. Expert. 1994. Differential regulation by iron of Erwinia chrysanthemi pectate lyases: pathogenicity of iron transport regulatory (cbr) mutants. Mol. Plant-Microbe Interact. 7:71-77.
Schwyn, B., and J. B. Neilands. 1987. Universal chemical assay for the detection and determination of siderophores. Anal. Biochem. 160:47-56.
Soncini, F. C., and E. A. Groisman. 1996. Two-component regulatory systems can interact to process multiple environmental signals. J. Bacteriol. 178:6796-6801.
Stojiljkovic, I., A. J. Bumer, and K. Hantke. 1994. Fur regulon in Gram-negative bacteria. J. Mol. Biol. 236:531-545.
Toth, I. K., and P. R. J. Birch. 2005. Rotting softly and stealthily. Curr. Opin. Plant Biol. 8:424-429.
Toth, I. K., K. Bell, M. C. Holeva, and P. R. J. Birch. 2003. Soft rot erwiniae: from genes to genomes. Mol. Plant Pathol. 4:17-30.
Yamamoto, K., K. Hirao, T. Oshima, H. Aiba, R. Utsumi, and A. Ishihama. 2005. Functional characterization in vitro of all two-component signal transduction systems from Escherichia coli. J. Biol. Chem. 280:1448-1456.
Yuen, T., E. Wurmbach, R. L. Pfeffer, B. J. Ebersole, and S. C. Sealfon. 2002. Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays. Nucleic Acids Res. 30:e48. [Online.] http://nar.oxfordjournals.org.
Zwir, I., D. Shin, A. Kato, K. Nishino, T. Latifi, F. Solomon, J. M. Hare, H. Huang, and E. A. Groisman. 2005. Dissecting the PhoP regulatory network of Escherichia coli and Salmonella enterica. Proc. Natl. Acad. Sci. USA 102:2862-2867.(Balakrishnan Venkatesh, L)
ABSTRACT
The PhoPQ two-component system regulates virulence factors in Erwinia chrysanthemi, a pectinolytic enterobacterium that causes soft rot in several plant species. We characterized the effect of a mutation in phoQ, the gene encoding the sensor kinase PhoQ of the PhoPQ two-component regulatory system, on the global transcriptional profile of E. chrysanthemi using cDNA microarrays and further confirmed our results by quantitative reverse transcription-PCR analysis. Our results indicate that a mutation in phoQ affects transcription of at least 40 genes, even in the absence of inducing conditions. Enhanced expression of several genes involved in iron metabolism was observed in the mutant, including that of the acs operon that is involved in achromobactin biosynthesis and transport. This siderophore is required for full virulence of E. chrysanthemi, and its expression is governed by the global repressor protein Fur. Changes in gene expression were also observed for membrane transporters, stress-related genes, toxins, and transcriptional regulators. Our results indicate that the PhoPQ system governs the expression of several additional virulence factors and may also be involved in interactions with other regulatory systems.
INTRODUCTION
Erwinia chrysanthemi is often described as a necrotrophic, intracellular, gram-negative plant pathogen that causes soft rot in different hosts. However, recently a biotrophic phase during the infection process of this and other closely related pathogens has been postulated (58). The pectinolytic equipment of E. chrysanthemi is the major determinant of virulence (5, 12, 45). When E. chrysanthemi infects a plant, it confronts an acidic environment in the intercellular apoplastic fluid (22). Upon perception of a variety of signals within the host, sequential synthesis of pectinolytic enzymes (pectate lyases and polygalacturonases) is induced in the bacterial cells (53). The activity of these enzymes leads to host cell lysis and subsequent alkalinization of the apoplast. Therefore, the ability of the neutrophilic E. chrysanthemi to either grow or survive in the acidic intercellular environment (acid tolerance) is important for virulence. Additional factors that contribute to its virulence are expression of siderophore-mediated iron transport systems (54), resistance to antimicrobial peptides (37), and hrp genes (6, 59). The expression of these functions essential for pathogenicity is regulated in a coordinate manner in response to different stimuli such as the presence of pectic substances, acid pH, iron and magnesium limitation, and the presence of antibacterial molecules. Two-component systems (TCSs), which comprise a sensor protein histidine kinase and a cytoplasmic response regulator that together form a signal transduction pathway via histidyl aspartyl phosphorelay, play a crucial role in this regulation. The E. chrysanthemi genome is predicted to contain genes that encode 30 TCSs (Glasner et al., unpublished results). In E. chrysanthemi as well as in a number of other bacteria, the PhoPQ TCS is activated by magnesium limitation (19, 43, 44, 48). Recently, for the best-characterized Salmonella enterica serovar Typhimurium PhoPQ system (1, 7, 15, 18, 23, 24), it was shown that antimicrobial peptides also serve as direct signals for the activation of PhoQ (4, 8). The phospho-PhoP transcriptionally activates or represses a large number of loci involved in diverse cellular functions (17, 24, 25, 29, 42), and, therefore, it has been placed on a par with global regulators such as Crp, Fnr, and ArcA (62). The E. chrysanthemi phoQ and phoP mutants have been previously partially characterized independently by two groups, and these studies have demonstrated the involvement of the E. chrysanthemi PhoPQ system in regulating synthesis of pectinolytic enzymes, its growth at acidic pH, and/or alkalization of external pH in planta, as well as its resistance to antimicrobial peptides (28, 35, 36). The PhoPQ system can participate in complex regulatory cascades that allow the integration of multiple signals and serve to expand the signal transduction network so as to allow some genes to respond to various signals in the environment (56). For example, in Salmonella, resistance to the antimicrobial peptide polymixin is mediated by the PmrA-PmrB TCS via the PhoP-dependent PmrD protein at low Mg2+. High levels of extracellular Fe3+ can activate the PmrA/B TCS independent of the PhoPQ TCS (26). In addition it has been shown that noncognate pairs of sensor kinase and response regulator may be involved in cross talks. It has been estimated that cross talk occurs in about 3% (21/692) of such pairs in Escherichia coli (60). Although such interactions have so far not been reported in E. chrysanthemi, studies with the E. chrysanthemi PhoPQ system have suggested the likelihood of a cross talk between this system and as yet unknown systems: i.e., an alternate sensor kinase may activate PhoP and/or PhoQ may activate additional response regulators, based on the behavior of the mutants with respect to magnesium concentration and pH. At high Mg2+ and acidic pH, resistance to the antibacterial peptide magainin II was severely reduced specifically in the phoQ mutant but not in the phoP mutant compared to that in the wild-type (WT) strain (28). Similarly, differences were observed with respect to virulence for the mutants at high Mg2+ concentrations (the phoQ mutant was more virulent), although at low (10 μM) magnesium concentrations, both mutants showed equally diminished virulence (35). Furthermore, in transcriptomic analysis of E. chrysanthemi phoP and phoQ mutants, it was found that the set of genes differentially regulated in the two mutants showed only minimal overlap (Venkatesh et al., unpublished results). E. chrysanthemi 3937 is widely used as a model system for functional genomic studies, and its complete genome sequence (http://asap.ahabs.wisc.edu/annotation/php/ASAP1.htm) is now available, allowing transcriptomic analyses.
Here we report results of our analysis of the global transcriptional profile of E. chrysanthemi using an E. chrysanthemi phoQ mutant, which indicate that the PhoPQ system interacts with other regulators in this important phytopathogen. Our analysis has identified a role for the system in iron-related metabolic processes, suggesting that the PhoPQ system may also regulate this important virulence determinant in planta.
MATERIALS AND METHODS
Bacterial strains, growth conditions, and biological replicates. The wild-type strain (33) and a phoQ mutant strain (kanamycin cassette insertion) of Erwinia chrysanthemi 3937 (28) were used in our study. Mutation at the phoQ locus had been confirmed by reverse transcription-PCR (RT-PCR) analysis and by Southern hybridization (28). The bacterial strains were grown in M63 minimal medium (3) at pH 7.0 containing glycerol (0.2% [wt/vol]) as the carbon source and a reduced concentration of Mg2+ (500 μM). Both strains were grown aerobically at 28°C until the bacterial optical density at 600 nm reached 2.00 for RNA isolation. For detection of siderophore activity by the chrome azurol S (CAS) assay, M63 minimal medium containing 30 mM phosphate was used since higher concentrations of phosphate interfere with the CAS assay.
Detection of siderophore activity. Siderophore activity in culture supernatants was estimated by the liquid CAS assay described by Schwyn and Neilands (55). Five-hundred-microliter portions of culture supernatants were mixed with 500 μl of CAS assay solution, and after 10 min of incubation at room temperature, the change in color was visually examined. Appropriate controls were set up with minimal medium instead of the culture supernatant.
RNA isolation and preparation of labeled cDNA. The bacterial cells were incubated for at least 30 min on ice with 1% (vol/vol) phenol (pH 4.3) and 20% (vol/vol) ethanol for stabilization of RNA. Total RNA was isolated using a QIAGEN RNeasy RNA isolation kit as described by the manufacturer (QIAGEN, Hilden, Germany). RNA was quantified using a NanoDrop ND-100 spectrophotometer (NanoDrop Technologies, Wilmington, DE), and the quality was assessed on an Agilent Bioanalyzer 2100 electrophoresis system (Agilent Technologies Inc., Palo Alto, CA).
In a reverse transcription reaction, 12 μg of total RNA was labeled in a 45-μl labeling reaction with 10 μl of a combination of short oligonucleotide (11-mer) primers (100 ng/ml) (14), 1 μl of spike RNA mix, 1.8 μl of 25x amino allyldeoxynucleoside triphosphate (dNTP) mixture, 4.5 μl of dithiothreitol (0.1 M), and 2 μl of reverse transcriptase in reaction buffer (Superscript II; Invitrogen). Primers and RNA were heated to 70°C for 10 min and cooled on ice followed by incubation for 2 h at 42°C. To denature the remaining RNA, 15 μl of 1 M NaOH and 15 μl of 0.5 M EDTA (pH 8.0) were added and incubated for 10 min at 65°C. The reaction was neutralized with 15 μl of 1 M HCl. Purification of cDNA was done as described by the supplier (QIAGEN, Hilden, Germany). The Cy dye ester was added to 10 μl of cDNA in amber tubes (QIAGEN, Hilden, Germany) in a total volume of 13 μl, containing 2.0 μl of 1 M sodium carbonate and 1.0 μl of the appropriate Cy dye (suspended in dimethyl sulfoxide) and incubated for 1 h at room temperature in the dark. To the labeled cDNA, 3.0 μl of 4.0 M hydroxylamine hydrochloride was added and incubation was extended for another 30 min in the dark. The labeled targets were combined and mixed with distilled water and further diluted with 500 μl of PB buffer (QIAGEN, Hilden, Germany) before application to a QIAGEN MiniElute column for purification. Finally, cDNA was eluted in 10 μl of elution buffer. The labeling efficiency was estimated with 1 μl of the cDNA using a NanoDrop ND-100 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
Hybridization to DNA microarrays and image analysis. Erwinia chrysanthemi microarrays, based on the genome sequence of strain 3937, have been developed at the Scottish Crop Research Institute (SCRI), Dundee, United Kingdom, through Agilent Technologies, Inc. The details of the microarray experimental protocol will be published elsewhere and will be available on request.
Labeled cDNA was made up to a volume of 80 μl with distilled water and denatured by heating at 98°C for 3 min. The denatured cDNA was mixed with 25 μl of control target and 105 μl of 2x hybridization buffer (Agilent Technologies, Inc.), and hybridization was carried out for 17 h at 60°C. After hybridization, the arrays were washed at room temperature in the dark for 1 min each with wash solution I (6x SSPE [1x SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA, pH 7.7] 0.005% N-lauroylsarcosine), followed by wash solution II (0.06x SSPE, 0.005% N-lauroylsarcosine). Arrays were scanned using an ArrayWoRx Auto scanner (Applied Precision) with appropriate exposure settings for Cy3 (595 nm) and Cy5 (685 nm) at a 9.756-μm resolution, generating separate TIFF images. Intensity data were acquired from images using SoftWoRx Tracker software (Applied Precision). Images were imported and aligned with clone position information (Agilent GAL file) using automated and manual grid alignment features. Median spot and individual median background (annulus setting) intensity values were extracted for each wavelength and imported into analysis software.
Normalization and analysis. The data from duplicate experiments were normalized and analyzed using GeneSpring version 7.2 gene expression software (Agilent Technologies, Inc.). Replicate spots in each array were averaged and data sets for each array were normalized using the Lowess algorithm. Genes were considered as upregulated in the phoQ mutant if their level of transcription was at least 1.5 times higher than the corresponding WT. Genes were considered downregulated in the phoQ mutant if they were determined to be present in the mutant at an expression level no more than two-thirds of that observed in the corresponding WT strain. The statistical significance of differential expression in the phoQ mutant was judged by the Student's t test at a P value of 0.05.
Real-time qRT-PCR. For independent confirmation of gene expression results obtained with oligonucleotide microarrays, total RNA was isolated from the WT and phoQ mutant of E. chrysanthemi 3937. RNA preparations for quantitative reverse transcription (qRT)-PCR were independent from those used for array hybridizations. After random decamer-primed first-strand cDNA was synthesized using the Omniscript RT kit (QIAGEN, Japan), real-time PCR was performed in an MX3000p Muitiplex quantitative PCR system (Stratagene) using the SYBR Premix Ex.Taq RT-PCR kit (TaKaRa, Japan). Primers were designed using the Primer Quest software (http://scitools.idtdna.com/biotools/primer_quest/primer_quest.asp), and the sequences are listed in Table 1. Primer specificity was assessed by using the dissociation curve protocol on the MX3000p Multiplex quantitative PCR system (Stratagene). The efficiency of all primer pairs was verified. PCR amplification conditions were as follows: denaturing at 95°C for 30 s, annealing at 55°C for 60 s, and extension at 72°C for 30 s for 40 cycles. All PCR experiments were performed in triplicates, and standard deviations were calculated.
The fluorescence intensity of SYBR green in each end-point of the annealing phase was detected, and the cycle threshold (CT) of each sample was calculated. The calculated CT data were used for quantitative analysis by the comparative CT method. For each amplification run, the calculated threshold cycle (CT) for each gene amplification was normalized to the CT of the 16S rRNA gene amplified from the corresponding sample before calculating the difference (fold) between the wild type and mutant using the following formula: fold change = , where CT for gene j = (CT,J – CT,16S rRNA)mutant – (CT,J – CT,16S rRNA)wild type.
Flagellar motility assay. Bacteria were inoculated with a toothpick on a minimal agar plate (M63 with glycerol, 500 μM MgSO4, 0.3% agar). The plates were then incubated at 28°C, and a halo corresponding to the spreading of bacteria from the point of inoculation was observed. To calculate the percentage of motility, the number of motile and nonmotile cells was counted in five replicate suspensions using a phase-contrast microscope (Iponacology IS-2000, Japan).
Stress assay. To test oxidative stress resistance, bacteria were inoculated into 5 ml of minimal medium (M63 with 0.2% [wt/vol] glycerol, 500 μM MgSO4) and grown for 16 h at 28°C. Under such conditions, bacterial cultures reached the stationary growth phase. H2O2 at 10 mM, which corresponds to a lethal concentration, was then added to the culture, and the cells were incubated for an additional 2 h at 28°C. The number of CFU was determined, by serial dilution and plating on LB agar, just before H2O2 treatment, just after the beginning of the challenge, and then every 20 min during 2 h. Survival of the H2O2-treated cells was normalized to the number of CFU at the beginning of the challenge. This experiment was repeated three times. Testing for osmotic stress resistance was similarly carried out by incubating stationary-phase cells in the presence of 1 M NaCl.
RESULTS
Transcriptome analysis. The availability of the genome sequence of Erwinia chrysanthemi 3937 allowed the development and utilization of cDNA microarrays to characterize the constitutive role of the phoQ gene at the transcriptome level. In order to identify loci that may be controlled by phoQ, the transcriptional profile of an E. chrysanthemi 3937 mutant strain lacking a functional phoQ was compared with that of the wild-type strain, E. chrysanthemi 3937. Total RNA was isolated from these strains, and cDNA was synthesized and labeled with the fluorescent dye Cy3 or Cy5 and hybridized to an E. chrysanthemi 3937 microarray (for details, see Materials and Methods) that contained 4,750 gene-specific spots (excluding duplicates or triplicates of a single gene). The results presented below are based on the mean of four normalized expression values derived from two biological replicates and two dye-swap experiments.
Our analysis revealed differences in expression levels for a number of genes, the majority of which were not previously known to be regulated by the PhoPQ regulatory system (Fig. 1). A total of 87 open reading frames (ORFs) appeared to be upregulated at least 1.5-fold in the phoQ mutant, and 26 ORFs appeared to be downregulated by the same degree (Fig. 2). Of these, only 40% (35 genes) of upregulated and 19% (5 genes) of downregulated ORFs were statistically significant, when a Student's t test was performed with a P value of less than 0.05 (Table 2). Microarray data were complemented with quantitative real-time PCR data for a subset of genes whose expression levels differed significantly between the mutant and wild-type strains. Based on the results, it appears that phoQ may predominantly effect transcriptional repression at several loci. The differentially expressed genes (Table 2) have been classified according to their putative function as follows.
Genes with unknown function and conserved hypothetical products. Among the 40 genes whose expression is altered, the function of 5 (upregulated) is not known (Table 2). No significant matches exist for these genes in the currently available databases. Seven of the differentially expressed genes (five upregulated and two downregulated) are at present thought to encode hypothetical products. Putative functions could be assigned to six of the seven genes, based on comparisons at the amino acid level using BLAST. For convenience, these genes have been discussed below in the functional classification sections.
The remaining 27 genes encode proteins with known functions, and these have been broadly classified based on these functions.
Iron assimilation. At least 37% of the genes that demonstrated enhanced expression in the mutant are involved in iron metabolism and include genes responsible for siderophore-mediated assimilation of iron (Table 2). Of the eight genes (acsF, acr, acsD, acsE, yhcA, acsC, acsB, and acsA) that have been reported (20) to constitute the achromobactin biosynthesis operon (acs) in Erwinia chrysanthemi, the transcript levels of five, namely acsA, acsC, acsD, acsE, and acsF, were significantly higher in the phoQ mutant. Each of these five gene products catalyzes an individual step in the biosynthesis of achromobactin (9, 20). A qRT-PCR analysis showed that the level of the acsE gene transcript was the highest (317-fold), followed by those of acsA and acsD (11-fold) and acsC and acsF (about 2-fold). Although it appeared from microarray analysis that the transcript levels of the other three genes (acsB, yhcA, and acr) were only marginally higher in the mutant strain, when compared to the WT, qRT-PCR analysis revealed the transcript levels of these three genes to be significantly higher in the mutant (Table 3). The proteins encoded by yhcA and acr participate in transport of the siderophore-iron complex, and acsB is involved in achromobactin biosynthesis.
Transcript levels of genes encoding proteins of the Ton complex (exbB, exbD, and tonB), which shuttles the iron-siderophore (ferriachromobactin) complex across the outer membrane, were also higher in the mutants. While transcript levels of exbB and exbD were significantly higher, that of tonB was only slightly higher (1.37-fold, P = 0.026). Of the two auxiliary proteins coded for by exbB and exbD, the former showed high transcript levels in the mutant (47-fold). In the mutant, the transcript levels of cbrA, encoding the periplasmic component of a permease complex that transports the liganded form of achromobactin (38), and of fhuC, encoding a cytoplasmic membrane-associated ATP-binding protein permease that is necessary for the transport of ferrichrome compounds from the periplasm across the cytoplasmic membrane into the cytoplasm, were about 12- and 15-fold higher, respectively, based on qRT-PCR data (Table 3).
Another gene that may have a role in iron transport and whose expression levels were higher in the phoQ mutant is yqjH (identification no. [ID] 16680). Analysis using BLAST shows that the predicted product is about 40% similar to iron-utilization proteins present in some proteobacteria (Pseudomonas, Bordetella, and Ralstonia) and in some actinomycetales (Kineococcus, Arthrobacter, Streptomyces, Corynebacterium, and Brevibacterium). The YqjH homolog in Escherichia coli (27% identical to E. chrysanthemi YqjH) is involved in the utilization of the catechol siderophore vibriobactin (51). Other genes that may function in iron transport are ycdO (ID 15840), whose predicted product is similar to a predicted periplasmic lipoprotein from Yersinia pestis (gi|45436217; 81% identical and 87% similar over a stretch of 376 amino acids [aa]) involved in iron transport, and a yet to be annotated gene (ID 18178) whose predicted product is highly similar to a putative periplasmic substrate-binding transport protein (gi|49612694) from E. carotovora subsp. atroseptica. The closest match (58% identity, 74% similarity) to the E. carotovora protein is an ABC transporter that binds iron in Agrobacterium tumefaciens.
Another iron-related transcript whose level was elevated in the mutant is bfd (ID 15304), which encodes a bacterioferritin-associated ferredoxin. This gene is thought to have a role in the release of iron from (or delivery of iron to) bacterioferritin. The qRT-PCR expression data for these genes are summarized in Table 3. These data suggest that phoQ may have a regulatory role in the uptake of iron via the achromobactin system in E. chrysanthemi.
General metabolism. Three induced transcripts belong to genes encoding key enzymes that participate in reactions contributing to maintenance of overall metabolism within the cell. Among these is gltA, which encodes citrate synthase, an important tricarboxylic acid cycle enzyme. Of the other two, one encodes a tRNA guanosine methyl transferase (trmD; ID 20494) that specifically methylates guanosine in several tRNAs and prevents frameshifting. The other (ID 17546) encodes a protein that is 79% identical to a putative methionine synthase (gi|49610578) from E. carotovora subsp. atroseptica and about 66% identical to vitamin B12-independent methionine synthases from Chromohalobacter salexigens and Rhodopseudomonas palustris. This enzyme catalyzes the last step in the biosynthesis of methionine.
Ion transport. The mutant showed elevated transcription of the delta subunit of ATPase (atpH; ID 14789) and of two other genes (ID 15807 and 16441) whose predicted protein products show identity to ATP-binding cassettes from other bacteria. The gene with ID 15807 shows up to 70% identity to ABC transporters that traffic anions such as nitrate, phosphate, and sulfonate in some proteobacteria such as Pseudomonas, Xylella, and Zymomonas and probably has a similar function in E. chrysanthemi. No significant match for this gene exists in E. carotovora subsp. atroseptica. For the other gene (ID 16441), no close homologs are present within Enterobacteriaceae. The nearest matches are present in Ralstonia, Desulfitobacterium, Methanosarcina, and Bacillus.
Conversely, transcripts of cusA involved in copper efflux were reduced in the mutant. Another downregulated gene (ID 16479) is highly similar (75 to 90%) to those coding for outer membrane lipoproteins from several bacteria, including E. carotovora subsp. atroseptica (80% similarity), and encodes an outer membrane beta barrel porin that may be involved in drug efflux.
Stress response. The stress-responsive transcripts whose levels were higher in the mutant are sodC (ID 17092), encoding a Cu-Zn superoxide dismutase; proW, encoding the ABC transporter component of OusB; and fliC, which encodes flagellin, the filament structural protein of flagella.
Among the downregulated genes encoding conserved hypothetical products, the product of one (ID 15573) is 70% identical to a putative exported protein found in Erwinia carotovora subsp. atroseptica and between 60 and 70% identical to conserved hypothetical proteins present in Brucella spp. Interestingly, this protein product also shows a slightly lower level of identity (about 50%) with hypothetical signal peptide proteins from Ralstonia solanacearum and Photobacterium profundum. The closest matches to this protein show a cystatin-like fold present in cystein protease inhibitors.
Toxins. The mutant showed elevated transcript levels of a cluster of genes (ID 16663, 16664, and 16665) that are predicted to encode proteins similar to the cytolytic (cyt) delta-endotoxins of Bacillus thuringiensis (Glasner et al., unpublished). To our knowledge, this is the first report of transcription of genes within this cluster. qRT-PCR analysis revealed at least a 10-fold increase of transcript levels for genes 16664 and 16665 in the mutant. It would be interesting to see whether these proteins function as toxins/virulence factors in E. chrysanthemi. Although a gene encoding a protein product similar to the type 1AA cytolytic endotoxin is also present in the E. chrysanthemi genome, this gene was not induced in the mutant. A gene (ID 16850) that encodes a conserved hypothetical product that shows a conserved epimerase domain and shares about 60% identity with hemolysins from Yersinia pestis and Escherichia coli was downregulated in the mutant.
Transcriptional regulators. At least five transcriptional regulators showed altered expression in the mutant. Of these, four were upregulated. These are glnL, encoding the sensor kinase component of the nitrogen-responsive two-component regulatory system; pecS, encoding the global transcriptional repressor of pectate lyase and other macerating enzymes such as cellulase; vfmE, a transcriptional regulator that governs multiple virulence factors in E. chrysanthemi; and an unknown transcriptional repressor (ID 20250), perhaps of the Cro/CI family, based on homology. The phoP gene, the regulatory counterpart of phoQ, was slightly repressed in the mutant strain compared to the WT strain.
We carried out preliminary in silico analysis for two of the regulators for which little information exists in the current databases.
vfmE. The vfmE gene is a member of the AraC family of transcriptional regulators. The vfmE product is 187 amino acids long, and its closest match is a 326-aa-long transcriptional regulator of the AraC family from Desulfovibrio vulgaris (38% identical and 61% similar over a stretch of 102 aa) with a yet-to-be-assigned function. Apparently, no homologs exist in E. carotovora subsp. atroseptica. In an E. chrysanthemi 3937 vfmE mutant, production of pectinases, cellulase, and protease was reduced by a factor of 10-fold (S. Reverchon, personal communication).
The unknown transcriptional regulator (ID 20250) is probably a member of the Cro/CI family of transcriptional repressors. The E. chrysanthemi genome is reported to encode at least 28 members that belong to this family (Glasner et al., unpublished). A BLAST analysis of the predicted protein sequence against the nonredundant database at NCBI revealed the closest match (56% identity and 74% similarity) to be an immunity repressor protein of the Cro/CI family from the Salmonella phage. The closest protein matches from bacteria are also repressors and show less homology (<35% identity and <60% similarity).
In addition to validation by qRT-PCR, we carried out the following phenotypic assays to further evaluate our microarray results.
Siderophore detection in culture supernatant. In the CAS assay for the detection of siderophores, the phoQ mutant produced more siderophore than the WT (Fig. 2) in the presence of 500 μM MgSO4 when iron was supplied. When the medium was starved for iron, no differences were observed between the WT and mutant strains with respect to siderophore production. No color change was apparent for the WT and mutant strains at a high Mg2+ concentration.
Flagellar motility assay. The strains were tested for motility on 0.3% minimal agar plates. After 24 h of incubation, the mutant migrated relatively extensively through the soft agar medium, whereas the wild-type strain migrated to a lesser extent (Fig. 3). The percentage of motile cells determined in cell suspensions was slightly higher in the phoQ mutant (92% ± 1%) than in the WT strain (87% ± 3%).
Stress assay. To determine whether E. chrysanthemi 3937 lacking in PhoQ showed increased resistance to oxidative stress, we analyzed the survival of the phoQ mutant under oxidative stress generated by the addition of 10 mM H2O2 to the M63 minimal medium. We found that the survival of the phoQ mutant was only marginally higher than that of the WT strain (Fig. 4A). A much higher difference was observed between the WT and phoQ mutant strains with respect to osmotic stress tolerance (Fig. 4B).
DISCUSSION
Bacterial cells face a variety of adverse conditions in the natural environment, such as nutrient limitation, temperature and acid shock, osmotic stress, and depletion of metal ions. Such adverse changes in the environment are associated with the encounter of pathogenic bacteria with a potential host and therefore serve as cues to augment a pathogenic life cycle. This recognition of a host by the pathogen triggers modifications in cellular metabolism necessary for survival of the pathogen and successful colonization of its host and often involves a complex interplay of multiple regulatory factors.
Two-component regulatory systems constitute an important part of this regulatory network and mediate adaptive responses to a number of external stimuli. The PhoPQ system responds to Mg2+ limitation. Previous studies have demonstrated that E. chrysanthemi phoQ mutants impaired in the sensor kinase differed from the phoP mutants impaired for the cognate response regulator (28, 35). These studies suggest that an alternate sensor kinase may activate PhoP and/or PhoQ may activate additional response regulators. We carried out a transcriptomic analysis with both PhoP and PhoQ mutants. Interestingly, there was only minimal overlap for the two regulons (data not shown), further suggesting that the sensor and regulator of the PhoPQ system may not be mutually exclusive. Here we report the transcriptomic profile of the phoQ mutant that showed a more drastic phenotype change than the phoP mutant when grown in the presence of 500 μM MgSO4. We chose this concentration because the phoQ mutant surprisingly showed much higher maceration ability on Chinese cabbage than either the wild type or the phoP mutant. At 10 μM and 10 mM concentrations, the mutant behaved as expected: that is, it showed much reduced macerating capacity in the former and higher capacity in the latter (Venkatesh et al., unpublished data).
When grown in the presence of 500 μM MgSO4, the absence of a functional phoQ was correlated with derepression of several genes. The unexpected induction of transcription of a number of genes in the mutant strain lacking a functional PhoQ sensor kinase may reflect the inherent potential of phoQ to affect transcriptional repression of these (or at least at some of these) loci. Although it is possible that at this concentration of Mg2+, an alternate sensor kinase might phosphorylate PhoP, we did not find convincing evidence for such a phenomenon. Only five genes were downregulated in the mutant, including phoP, which encodes the cognate cytoplasmic regulator operating with the PhoQ sensor kinase (Table 2). Functions are yet to be assigned for some of the differentially expressed genes. Since many of the loci are also known to be governed by other regulators, it appears that the PhoQ sensor kinase, and thereby the PhoPQ TCS, has overlapping functions with other regulatory systems.
Microarray analysis showed that several genes related to iron uptake were induced in the mutant. In fact, some earlier studies have demonstrated that under Mg2+-limiting conditions (which activate PhoPQ TCS), E. coli and S. enterica serovar Typhimurium take up large amounts of ferrous iron (27, 32). In line with this, Chamnongpol and Groisman (10) suggested a probable role for the PhoPQ TCS in resistance to high Fe2+ based on the observations that an inactive PhoPQ system rendered Salmonella hypersensitive to Fe2+ iron. The toxicity was due to the uptake of Fe2+ by the nonspecific transporter CorA, whose activity was altered in response to the PhoP-regulated determinants mgtA and mgtB. However, the mechanism of iron uptake in the present context is different since, as discussed below, it is mediated by siderophores that are specific for the Fe3+ form of iron and occurs in the absence of added iron. This uptake of iron was nontoxic, as both the WT and the mutant strain grew normally in the medium tested (data not shown). Interestingly, different levels of induction were evident for the iron assimilation genes—those for the production of the siderophore achromobactin were strongly derepressed, and genes involved in iron uptake functions were weakly derepressed (Table 3). A similar pattern of regulation has been reported in an E. coli fur mutant for the enterobactin biosynthesis and uptake during iron limitation (41). The acs operon in E. chrysanthemi is governed by the Fur repressor, and Franza et al. (21) showed the constitutive derepression of the acs::lacZ fusion transcript in an E. chrysanthemi fur mutant background leading to the uptake of iron. In fact, in our study except for one unknown ORF (ID 18178), all of the differentially expressed genes with iron-related functions have been previously identified as members of the fur regulon from other bacteria. Furthermore, the induction of some non-iron-related genes could be related to common regulation by fur or could be a consequence of induction of other loci by fur. For example, fur controls the expression of fliC and thereby couples iron status to motility. Expression of the flagella is inhibited when bacteria have an excess of the metal so as to prevent the departure of the cells toward other environments with less available iron (40, 57). It has been recently shown that the FliC homolog of Erwinia carotovora subsp. carotovora (64% identity, 79% similarity to E. chrysanthemi FliC) is essential for motility (30). An increased level of yet another transcript, namely that of gltA, encoding citrate synthase correlates well with the fact that citrate is the precursor for achromobactin. However, not all fur-regulated genes were differentially expressed. Thus, it appears that a mutation in phoQ results in the derepression of a subset of genes of the fur regulon. These data provide indirect evidence for cross talk between the PhoPQ TCS and the Fur repressor. The physiological implication for such a phenomenon, if operative in planta, awaits investigation since in the presence of 10 μM MgSO4 some iron-regulated genes such as acsA, acsE, and cbrA showed increased transcript levels in the WT whereas some iron-related genes such as acsF and yqjH showed a decrease in the transcript level (qRT-PCR; data not shown).
Other than the transporters involved in iron uptake and related functions, lack of a functional phoQ resulted in the induction of expression of three additional transporters and the repression of two (Table 2). The differential induction of membrane transporters may reflect altered ion fluxes between the bacterial cells and the surrounding medium to maintain homeostasis. Two transcripts (proW and trmD) that are commonly induced under osmotic stress were also higher in the mutant and may contribute to its higher survivability than the WT strain (Fig. 4B) in the presence of 1 M NaCl. ProW is the ABC transporter component of OusB that is involved in adaptation to osmotic stress in Erwinia chrysanthemi (11). Consistent with the increased transcript levels of sodC (ID 17092), the phoQ mutant showed slightly higher survivability in the presence of hydrogen peroxide (Fig. 4A). The Salmonella sodC gene encodes the periplasmic, phage-associated Cu-Zn superoxide dismutase (SodC1) that is present only in some highly virulent strains and modulates interaction with the host to promote bacterial survival (2).
The fact that achromobactin, like other siderophores, is an important virulence factor (47, 50) presents an interesting dimension. The expression of this virulence factor appears to be coupled to the expression of other virulence factors such as the PecS protein that can repress the production of the degradative enzymes (i.e., pectate lyases and cellulases) (5, 52) and can induce the synthesis of polygalacturonase enzymes (31, 46), and it also has a role in rendering resistance to the products of oxidative burst via an extracellular blue pigment called indigoidine (52). In order to examine whether the effect of the phoQ mutation extends to genes controlled by PecS, we carried out qRT PCR analysis for some such genes. Our results showed that the transcript levels of genes encoding pectate lyases PelA, PelB, PelC, PelD, and PelE were much less in the phoQ mutant than in the wild-type strain. Microarray analyses for these genes also indicate downregulation (Table 3). Our results are consistent with previously published results of Llama Palacios et al. (35) and Haque and Tsuyumu (28) that in phoQ mutants of E. chrysanthemi 3937, the production of pectate lyases is drastically reduced (up to 90% of WT). Our qRT-PCR results for PecS-regulated genes support the reported contention that the exclusion of genes with low microarray expression scores from further analysis may not be justified (13). It has been shown that microarray analyses underestimate the change (fold) of regulated genes in comparison with the quantitative real-time PCR assay either because of overestimating the uninduced transcripts, perhaps due to nonspecific hybridization, or underestimating the level of induction, perhaps due to probe saturation effects (61).
In addition, the transcriptional regulator vfmE, which controls multiple virulence factors in E. chrysanthemi (S. Reverchon, personal communication), and the Cro/CI family transcriptional regulator involved in functions that prevent a lytic life cycle (and therefore may have a role in increasing survival within a host) might also act as virulence determinants. The possible interplay of these regulatory factors with phoQ, as suggested by their enhanced expression (Table 2) in the absence of a functional phoQ, may be interesting to pursue.
In conclusion, E. chrysanthemi produces two siderophores, chrysobactin (49) and achromobactin, in response to iron limitation in planta. These are structurally different and require different levels of iron limitation for their induction (16, 38, 39). Additionally, their production is temporally separated so that achromobactin synthesis precedes chrysobactin synthesis (20). Therefore, it is reasonable that, in addition to the coordinate control of the global regulatory protein Fur, specific mechanisms of regulation may be operating at these loci via effectors, for example, depending on the nature of the stimulus. In fact, Franza and coworkers (21) hinted at the possible involvement of such additional molecules for the chrysobactin system, but not for the achromobactin system, which was thought to be only fur dependent. However, since their investigation was carried out in a normal PhoQ background, it is possible that an additional effect was not evident for the achromobactin system. Our analysis on a transcriptomic scale suggests that an analogous mechanism apparently operates for the achromobactin system as well, at least under certain conditions. Given that pleiotropic effects caused by phoQ mutation in E. chrysanthemi include derepression of some iron metabolism genes, a phenotype also shared with fur mutants of E. chrysanthemi, it seems plausible that at least some of the genetic components of both regulons might be shared. The mechanism to actively lower the availability of Mg2+ is analogous to the sequestration of iron, and both mechanisms are recognized as nonspecific host defense strategies. Therefore, it is possible that the PhoPQ system has been recruited to regulate adaptations to both magnesium- and iron-limiting environments, albeit to different extents. We speculate that a product of the PhoP-PhoQ two-component system effects the repression of the achromobactin synthesis by fur and thereby prevents the needless squandering of energy when iron levels are sufficient. From the present study, the mediation of iron assimilation (and thereby a possible cross talk with the fur regulon) emerges as a novel role for the PhoQ sensor kinase. This concept, however, needs experimental validation.
To summarize, microarray data support the involvement of the PhoQ sensor kinase in a wide variety of cellular functions in E. chrysanthemi, including iron metabolism, synthesis of virulence factors, proton pumping, stress tolerance, and DNA repair functions, although results from this study alone are not sufficient to understand the precise role of the phoQ gene in these functions. Further investigation is required to understand the complex nature of the results presented here that hint toward possible additional roles for the PhoPQ system in planta. In many bacteria, the production of virulence factors is regulated by more than one mechanism to facilitate coordinate regulation (34). The PhoPQ system may operate as part of the multiple regulatory mechanisms that govern virulence in E. chrysanthemi. A better understanding of the PhoPQ system will further our knowledge of the mechanisms of pathogenicity of E. chrysanthemi and related organisms and may help us devise more appropriate disease control strategies.
ACKNOWLEDGMENTS
We thank Manjurul Haque, Shizuoka University, for the gift of the PhoP and PhoQ mutant strains that were used in this study. We gratefully acknowledge the help of Jacques Schrenzel and Brian Gettler from the University Hospitals of Geneva, Switzerland, for their assistance in designing the microarray probes. We thank Jenny Morris, SCRI, for technical help. We sincerely thank the anonymous referees for their valuable comments and suggestions.
REFERENCES
Adams, P., R. Fowler, N. Kinsella, G. Howell, M. Farris, P. Coote, and C. D. Connor. 2001. Proteomic detection of PhoPQ- and acid-mediated repression of Salmonella motility. Proteomics 1:597-607.
Ammendola, S., M. Ajello, P. Pasquali, J. S. Kroll, P. R. Langford, G. Rotilio, P. Valenti, and A. Battistoni. 2005. Differential contribution of sodC1 and sodC2 to intracellular survival and pathogenicity of Salmonella enterica serovar Choleraesuis. Microbes Infect. 7:698-707.
Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl. 1987. Current protocols in molecular biology. John Wiley & Sons, New York, N.Y.
Bader, M. W., S. Sanowar, M. E. Daley, A. R. Schneider, U. Cho, W. Xu, R. E. Klevit, H. LeMoual, and S. I. Miller. 2005. Recognition of antimicrobial peptides by a bacterial sensor kinase. Cell 22:461-472.
Barras, F., F. Van Gijsegem, and A. K. Chatterjee. 1994. Extracellular enzymes and pathogenesis of soft-rot Erwinia. Annu. Rev. Phytopathol. 32:201-234.
Bauer, D. W., A. J. Bogdanove, S. V. Beer, and A. Collmer. 1994. Erwinia chrysanthemi hrp genes and their involvement in soft rot pathogenesis and elicitation of the hypersensitive response. Mol. Plant-Microbe Interact. 7:573-581.
Bijlsma, J. J. E., and E. A. Groisman. 2005. The PhoP/PhoQ system controls the intramacrophage type three secretion system of Salmonella enterica. Mol. Microbiol. 57:85-96.
Bishop, J. L., and B. B. Finlay. 2006. Friend or foe Antimicrobial peptides trigger pathogen virulence. Trends Mol. Med. 12:3-6.
Challis, L. G. 2005. A widely distributed bacterial pathway for siderophore biosynthesis independent of nonribosomal peptide synthetases. ChemBioChem 6:601-611.
Chamnongpol, S., and E. A. Groisman. 2002. Mg2+ homeostasis and avoidance of metal toxicity. Mol. Microbiol. 44:561-571.
Choquet, G., N. Jehan, C. Pissavin, C. Blanco, and M. Jebbar. 2005. OusB, a broad-specificity ABC-type transporter from Erwinia chrysanthemi, mediates uptake of glycine betaine and choline with a high affinity. Appl. Environ. Microbiol. 71:3389-3398.
Collmer, A., and N. T. Keen. 1986. The role of pectic enzymes in plant pathogenesis. Annu. Rev. Phytopathol. 24:383-409.
Dallas, B. P., N. G. Gottardo, M. J. Firth, A. H. Beesley, K. Hoffmann, P. A. Terry, J. R. Freitas, J. M. Boag, A. J. Cummings, and U. R. Kees. 2005. Gene expression levels assessed by oligonucleotide microarray analysis and quantitative real-time RT-PCR—how well do they correlate BMC Genomics 6:59. [Online.]
Dellagi, A., J. Helibronn, A. O. Avrova, M. Montesano, E. T. Palva, H. E. Stewart, I. K. Toth, D. E. Cooke, G. D. Lyon, and P. R. Birch. 2000. A potato gene encoding a WRKY-like transcription factor is induced in interactions with Erwinia carotovora subsp. atroseptica and Phytophthora infestans and is coregulated with class I endochitinase expression. Mol. Plant-Microbe Interact. 13:1092-1101.
Detweiler, C. S., D. B. Cunanan, and S. Falkow. 2001. Host microarray analysis reveals a role for the Salmonella response regulator phoP in human macrophage cell death. Proc. Natl. Acad. Sci. USA 98:5850-5855.
Expert, D., C. Sauvage, and J. B. Neilands. 1992. Negative transcriptional control of iron transport in Erwinia chrysanthemi involves an iron-responsive two-factor system. Mol. Microbiol. 6:2009-2017.
Fields, P. I., E. A. Groisman, and F. Heffron. 1989. A Salmonella locus that controls resistance to microbicidal proteins from phagocytic cells. Science 243:1059-1062.
Fields, P. I., R. V. Swanson, C. G. Haidaris, and F. Heffron. 1986. Mutants of Salmonella typhimurium that cannot survive within the macrophage are avirulent. Proc. Natl. Acad. Sci. USA 83:5189-5193.
Flego, D., R. Marits, A. R. Eriksson, V. Koiv, M. B. Karlsson, R. Heikinheimo, and E. T. Palva. 2000. A two-component regulatory system, pehR-pehS, controls endopolygalacturonase production and virulence in the plant pathogen Erwinia carotovora subsp. carotovora. Mol. Plant-Microbe Interact. 13:447-455.
Franza, T., B. Mahe, and D. Expert. 2005. Erwinia chrysanthemi requires a second iron transport route dependent of the siderophore achromobactin for extracellular growth and plant infection. Mol. Microbiol. 55:261-275.
Franza, T., C. Sauvage, and D. Expert. 1999. Iron regulation and pathogenicity in Erwinia chrysanthemi 3937: role of the fur repressor protein. Mol. Plant-Microbe Interact. 12:119-128.
Grignon, C., and H. Sentenac. 1991. pH and ionic conditions in the apoplast. Annu. Rev. Plant Physiol. 42:103-128.
Groisman, E. A. 2001. The pleiotropic two-component regulatory system PhoP-PhoQ. J. Bacteriol. 183:1835-1842.
Groisman, E. A., E. Chiao, C. J. Lipps, and F. Heffron. 1989. Salmonella typhimurium phoP virulence gene is a transcriptional regulator. Proc. Natl. Acad. Sci. USA 86:7077-7081.
Guina, T., E. C. Yi, H. Wang, M. Hackett, and S. I. Miller. 2000. A PhoP-regulated outer membrane protease of Salmonella enterica serovar Typhimurium promotes resistance to alpha-helical antimicrobial peptides. J. Bacteriol. 182:4077-4086.
Gunn, J. S., and S. I. Miller. 1996. PhoP-PhoQ activates transcription of pmrAB, encoding a two-component regulatory system involved in Salmonella typhimurium antimicrobial peptide resistance. J. Bacteriol. 178:6857-6864.
Hantke, K. 1987. Selection procedure for deregulated iron transport mutants (fur) in Escherichia coli K12: fur not only affects iron metabolism. Mol. Gen. Genet. 210:135-139.
Haque, M. M., and S. Tsuyumu. 2005. Virulence, resistance to magainin II and expression of pectate lyase are controlled by the PhoP-PhoQ two component regulatory system responding to pH and magnesium in Erwinia chrysanthemi 3937. J. Gen. Plant Pathol. 71:47-53.
Hitchen, P. G., J. L. Prior, P. C. Oyston, M. Panico, B. W. Wren, R. W. Titball, H. R. Morris, and A. Dell. 2002. Structural characterization of lipo-oligosaccharide (LOS) from Yersinia pestis: regulation of LOS structure by the PhoPQ system. Mol. Microbiol. 44:1637-1650.
Hossain, M. M., S. Shibata, S. I. Aizawa, and S. Tsuyumu. 2005. Motility is an important determinant for pathogenesis of Erwinia carotovora subsp. carotovora. Physiol. Mol. Plant Pathol. 66:134-143.
Hugouvieux-Cotte-Pattat, N., V. E. Shevchik, and W. Nasser. 2002. PehN, a polygalacturonase homologue with a low hydrolase activity, is coregulated with the other Erwinia chrysanthemi polygalacturonases. J. Bacteriol. 184:2664-2673.
Kot, E., and A. Bezkorovainy. 1993. Effects of Mg2+ and Ca2+ on Fe2+ uptake by Bifidobacterium thermophilum. Int. J. Biochem. 25:1029-1033.
Kotoujansky, A., M. Lemattre, and P. Boistard. 1982. Utilization of a thermosensitive episome bearing transposon Tn10 to isolate Hfr donor strains of Erwinia carotovora subsp. chrysanthemi. J. Bacteriol. 150:122-131.
Litwin, C. M., and S. B. Calderwood. 1993. Role of iron in regulation of virulence genes. Clin. Microbiol. Rev. 6:137-149.
Llama-Palacios, A., E. Lopez-Solanilla, and P. Rodríguez-Palenzuela. 2005. Role of the PhoP-PhoQ system in the virulence of Erwinia chrysanthemi strain 3937: involvement in sensitivity to plant antimicrobial peptides, survival at acid pH, and regulation of pectolytic enzymes. J. Bacteriol. 187:2157-2162.
Llama-Palacios, A., E. Lopez-Solanilla, C. Poza-Carrion, F. García-Olmedo, and P. Rodríguez-Palenzuela. 2003. The Erwinia chrysanthemi phoP-phoQ operon plays an important role in growth at low pH, virulence and bacterial survival in plant tissue. Mol. Microbiol. 49:347-357.
Lopez-Solanilla, E., A. Llama-Palacios, A. Collmer, F. García-Olmedo, and P. Rodríguez-Palenzuela. 2001. Relative effects on virulence of mutations in the sap, pel, and hrp loci of Erwinia chrysanthemi. Mol. Plant-Microbe Interact. 14:386-393.
Mahe, B., C. Masclaux, L. Rauscher, C. Enard, and D. Expert. 1995. Differential expression of two siderophore-dependent-iron acquisition pathways in Erwinia chrysanthemi 3937: characterization of a novel ferrisiderophore permease of the ABC transporter family. Mol. Microbiol. 18:33-43.
Masclaux, C., and D. Expert. 1995. Signaling potential of iron in plant microbe interactions: the pathogenic switch of iron transport in Erwinia chrysanthemi. Plant J. 7:121-128.
McCarter, L., and M. Silverman. 1989. Iron regulation of swarmer cell differentiation of Vibrio parahaemolyticus. J. Bacteriol. 171:731-736.
McHugh, J. P., F. Rodriguez-Quinones, H. Abdul-Tehrani, D. A. Svistunenko, R. K. Poole, C. E. Cooper, and S. C. Andrews. 2003. Global iron-dependent gene regulation in Escherichia coli. A new mechanism for iron homeostasis. J. Biol. Chem. 278:29478-29486.
Miller, S. I., A. M. Kukral, and J. J. Mekalanos. 1989. A two-component regulatory system (phoP phoQ) controls Salmonella typhimurium virulence. Proc. Natl. Acad. Sci. USA 86:5054-5058.
Minagawa, S., H. Ogasawara, A. Kato, K. Yamamoto, Y. Eguchi, T. Oshima, H. Mori, A. Ishihama, and R. Utsumi. 2003. Identification and molecular characterization of the Mg2+ stimulon of Escherichia coli. J. Bacteriol. 185:3696-3702.
Monsieurs, P., S. Keersmaecker, W. W. Navarre, M. W. Bader, F. deSmet, M. McClelland, F. C. Fang, B. deMoor, J. Vanderleyden, and K. Marchal. 2005. Comparison of the PhoPQ regulon in Escherichia coli and Salmonella typhimurium. J. Mol. Evol. 60:462-474.
Nachin, L., and F. Barras. 2000. External pH: an environmental signal that helps to rationalize pel gene duplication in Erwinia chrysanthemi. Mol. Plant-Microbe Interact. 13:882-886.
Nasser, W., V. E. Shevchik, and N. Hugouvieux-Cotte-Pattat. 1999. Analysis of three clustered polygalacturonase genes in Erwinia chrysanthemi 3937 revealed an anti-repressor function for the PecS regulator. Mol. Microbiol. 34:641-650.
Neilands, J. B., and S. A. Leong. 1986. Siderophores in relation to plant disease. Annu. Rev. Plant Physiol. 37:187-208.
Newcombe, J., J. C. Jeynes, E. Mendoza, J. Hinds, G. L. Marsden, R. A. Stabler, M. Marti, and J. J. McFadden. 2005. Phenotypic and transcriptional characterization of the meningococcal PhoPQ system, a magnesium-sensing two-component regulatory system that controls genes involved in remodeling the meningococcal cell surface. J. Bacteriol. 187:4967-4975.
Persmark, M., D. Expert, and J. B. Neilands. 1989. Isolation, characterization and synthesis of chrysobactin, a compound with a siderophore activity from Erwinia chrysanthemi. J. Biol. Chem. 264:3187-3193.
Ratledge, C., and L. G. Dover. 2000. Iron metabolism in pathogenic bacteria. Annu. Rev. Microbiol. 54:881-941.
Redford, P., P. L. Roesch, and R. A. Welch. 2003. degS is necessary for virulence and is among extraintestinal Escherichia coli genes induced in murine peritonitis. Infect. Immun. 71:3088-3096.
Reverchon, S., C. Rouanet, D. Expert, and W. Nasser. 2002. Characterization of indigoidine biosynthetic genes in Erwinia chrysanthemi and role of this blue pigment in pathogenicity. J. Bacteriol. 184:654-665.
Robert-Baudouy, J., W. Nasser, G. Condemine, S. Reverchon, V. E. Shevchik, and N. Hugovieux-Cotte-Pattat. 2000. Pectic enzymes of Erwinia chrysanthemi regulation and role in pathogenesis. Plant-Microbe Interact. 5:221-368.
Sauvage, C., and D. Expert. 1994. Differential regulation by iron of Erwinia chrysanthemi pectate lyases: pathogenicity of iron transport regulatory (cbr) mutants. Mol. Plant-Microbe Interact. 7:71-77.
Schwyn, B., and J. B. Neilands. 1987. Universal chemical assay for the detection and determination of siderophores. Anal. Biochem. 160:47-56.
Soncini, F. C., and E. A. Groisman. 1996. Two-component regulatory systems can interact to process multiple environmental signals. J. Bacteriol. 178:6796-6801.
Stojiljkovic, I., A. J. Bumer, and K. Hantke. 1994. Fur regulon in Gram-negative bacteria. J. Mol. Biol. 236:531-545.
Toth, I. K., and P. R. J. Birch. 2005. Rotting softly and stealthily. Curr. Opin. Plant Biol. 8:424-429.
Toth, I. K., K. Bell, M. C. Holeva, and P. R. J. Birch. 2003. Soft rot erwiniae: from genes to genomes. Mol. Plant Pathol. 4:17-30.
Yamamoto, K., K. Hirao, T. Oshima, H. Aiba, R. Utsumi, and A. Ishihama. 2005. Functional characterization in vitro of all two-component signal transduction systems from Escherichia coli. J. Biol. Chem. 280:1448-1456.
Yuen, T., E. Wurmbach, R. L. Pfeffer, B. J. Ebersole, and S. C. Sealfon. 2002. Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays. Nucleic Acids Res. 30:e48. [Online.] http://nar.oxfordjournals.org.
Zwir, I., D. Shin, A. Kato, K. Nishino, T. Latifi, F. Solomon, J. M. Hare, H. Huang, and E. A. Groisman. 2005. Dissecting the PhoP regulatory network of Escherichia coli and Salmonella enterica. Proc. Natl. Acad. Sci. USA 102:2862-2867.(Balakrishnan Venkatesh, L)