|Home | About | Journals | Submit | Contact Us | Français|
To investigate the role of iron uptake mediated by the siderophore pyoverdine in the virulence of the plant pathogen Pseudomonas syringae pv. tabaci 6605, three predicted pyoverdine synthesis-related genes, pvdJ, pvdL, and fpvA, were mutated. The pvdJ, pvdL, and fpvA genes encode the pyoverdine side chain peptide synthetase III l-Thr-l-Ser component, the pyoverdine chromophore synthetase, and the TonB-dependent ferripyoverdine receptor, respectively. The ΔpvdJ and ΔpvdL mutants were unable to produce pyoverdine in mineral salts-glucose medium, which was used for the iron-depleted condition. Furthermore, the ΔpvdJ and ΔpvdL mutants showed lower abilities to produce tabtoxin, extracellular polysaccharide, and acyl homoserine lactones (AHLs), which are quorum-sensing molecules, and consequently had reduced virulence on host tobacco plants. In contrast, all of the mutants had accelerated swarming ability and increased biosurfactant production, suggesting that swarming motility and biosurfactant production might be negatively controlled by pyoverdine. Scanning electron micrographs of the surfaces of tobacco leaves inoculated with the mutant strains revealed only small amounts of extracellular polymeric matrix around these mutants, indicating disruption of the mature biofilm. Tolerance to antibiotics was drastically increased for the ΔpvdL mutant, as for the ΔpsyI mutant, which is defective in AHL production. These results demonstrated that pyoverdine synthesis and the quorum-sensing system of Pseudomonas syringae pv. tabaci 6605 are indispensable for virulence in host tobacco infection and that AHL may negatively regulate tolerance to antibiotics.
Phytopathogenic bacteria employ a variety of virulence mechanisms to overcome the defense systems of plants. Pseudomonas syringae pv. tabaci 6605 is a gram-negative bacterium that causes wildfire disease on host tobacco plants. Previously, we demonstrated that flagellin, a component of the flagellar filament of this organism, is a major elicitor of the hypersensitive reaction and is posttranslationally modified by glycosylation (26, 28-30). A genetic region composed of three open reading frames (ORFs), namely, fgt1, fgt2, and orf3, was previously identified in a flagellum gene cluster. fgt1 and fgt2 encode flagellin glycosyltransferase, and orf3 shows significant homology to the 3-oxoacyl-(acyl carrier protein [ACP]) synthase III in the fatty acid elongation cycle, required for the synthesis of acyl homoserine lactones (AHLs) (11, 30, 31). Analysis of an orf3 deletion (Δorf3) mutant revealed that orf3 played no role in the glycosylation of flagellin, although the virulence of the Δorf3 mutant on tobacco plants was remarkably reduced.
Many virulence factors of bacteria have been reported to be under the regulation of a cell density-dependent system called quorum sensing. AHLs are synthesized by the coupling of the homoserine lactone ring from S-adenosylmethionine and acyl chains from the acyl-ACP by PsyI in P. syringae (9, 11). P. syringae pv. tabaci 6605 secretes three types of AHLs as signal molecules: N-hexanoyl-l-homoserine lactone, N-(3-oxohexanoyl)-l-homoserine lactone, and N-octanoyl-l-homoserine lactone (31). Our previous study indicated that the Δorf3 mutant and the quorum-sensing molecule-defective ΔpsyI mutant had significantly reduced abilities to produce AHLs and to take up iron (31). Furthermore, a scanning electron micrograph revealed little extracellular polymeric substance matrix surrounding the inoculated Δorf3 and ΔpsyI mutants on the tobacco leaf surface, indicating a lack of biofilm development (31). Iron acquisition has been reported to affect biofilm formation (2), and iron uptake is also involved in biofilm development under the regulation of quorum sensing in P. syringae pv. tabaci 6605 (2, 3, 11, 31).
Iron is indispensable for the growth of almost all organisms, and the ability to acquire iron is thought to be an important factor in virulence (27). Because the concentration of Fe(III) in the environment is quite low, owing to its insolubility under environmental conditions, the fluorescent Pseudomonas group produces a yellow-green Fe(III)-chelating siderophore called pyoverdine in order to acquire iron effectively (24). The genes required for pyoverdine synthesis are well characterized in the Pseudomonas aeruginosa strain PAO1 (33), and pyoverdine biosynthesis mutants of this pathogen exhibit reduced virulence (20, 33). For the regulation of iron homeostasis, it was reported that Fur (ferric uptake regulator) is a global regulator that controls the expression of siderophore-mediated iron uptake in P. syringae pv. tabaci 11528 (6).
In the present study, two genes predicted to be involved in pyoverdine synthesis, encoding the pyoverdine side chain peptide synthetase III l-Thr-l-Ser component (pvdJ) and the pyoverdine chromophore synthetase (pvdL), and the TonB-dependent ferripyoverdine receptor gene (fpvA) were disrupted in order to elucidate the roles of the pyoverdine-mediated iron acquisition system in the virulence of P. syringae pv. tabaci 6605. By use of these mutants, several important virulence factors, flagellum-dependent motility, the production of tabtoxin and extracellular polysaccharide (EPS), and biofilm formation were investigated. Although the ΔpvdJ and ΔpvdL mutants had reduced ability to produce EPS, the antibiotic tolerance of these mutants was drastically increased. The correlations among the pyoverdine-mediated iron uptake system, quorum-sensing regulation, and multidrug efflux are also discussed.
All bacterial strains used in this study are listed in Table Table1.1. Pseudomonas syringae pv. tabaci 6605 strains were maintained as described previously (30). For tabtoxin detection or swarming assays, each strain was grown in mineral salts-glucose (MG) medium (44 mM Na2HPO4, 22 mM KH2PO4, 17.8 mM NH4Cl, 11.3 mM NaCl, 0.1 mM CaCl2, 1 mM MgSO4, and 0.2% glucose). Escherichia coli strains were grown at 37°C in Luria-Bertani (LB) medium.
Tobacco plants (Nicotiana tabacum L. cv. Xanthi NC) were grown at 25°C with a 12-h photoperiod. The spray inoculation and infiltration methods have been described in a previous paper (30). For dip inoculation, bacterial strains were suspended at a density of 2 × 108 CFU/ml in 10 mM MgSO4 and 0.02% Silwet L77 (OSI Specialties, Danbury, CT), and tobacco leaves were soaked in that solution for 20 min. After being inoculated by either procedure, the leaves were incubated in a growth cabinet for 10 days at 23°C. The bacterial population in dip-inoculated leaves was measured as described previously (30).
To generate the siderophore synthesis-defective (ΔpvdL and ΔpvdJ) mutants and the TonB-dependent siderophore receptor-defective (ΔfpvA) mutant, the whole genetic region of fpvA (2,229 bp) and the N-terminal regions of pvdL and pvdJ (2,340 bp and 1,500 bp, respectively) were isolated by the Zero Blunt Topo PCR (Invitrogen, Carlsbad, CA) and pGEM-T Easy vector (Promega, Madison, WI) cloning systems. Three sets of PCR primers (FpvAF and FpvAR, PvdJF and PvdJR, and PvdLF and PvdLR) were designed based on the registered sequences of each gene of P. syringae pv. phaseolicola 1448A (GenBank accession number, CP000058), as shown in Table Table2.2. The constructs were digested with StuI, AfaI, or AgeI to delete 483, 196, or 975 internal bp for generation of the ΔfpvA, ΔpvdJ, or ΔpvdL mutant, respectively. Each resulting plasmid was introduced into pK18mobsacB at the EcoRI site (25). Deletion mutants were obtained based on conjugation and homologous recombination according to methods previously reported (30, 32).
For complementation of the ΔfpvA and ΔpvdL mutants, 2,381-bp or 14,827-bp fragments containing the predicted promoter region and the complete fpvA or pvdL gene were first cloned using two sets of primers (FpvA-proF, 5′-AAGAGCGCGAATTGCGTGCACA-3′; FpvAR and PvdL-proF, 5′-GGCAGCTTTGTGTATGGCGT-3′; PvdLR1, 5′-ATTTCATCGACAATCAGCGC-3′) and then inserted into the NotI and KpnI sites of pBSL118, a transposon vector, to generate the complemented ΔfpvAC and ΔpvdLC strains (1). The vectors were introduced into the ΔfpvA and ΔpvdL mutants by conjugation using the E. coli λpir strain S17-l.
Leaves spray inoculated with the wild-type (WT) or mutant strains and incubated for 10 days at 23°C were observed by scanning electron microscopy. The detailed procedure is described in previous reports by Taguchi et al. (30, 31).
Bacteria cultured overnight in LB medium containing 10 mM MgCl2 at 25°C were resuspended in 10 mM MgSO4 and adjusted to an optical density at 600 nm (OD600) of 1.0. Three-microliter aliquots were inoculated into the centers of MG plates containing 0.25% agar for the swimming assay and into the centers of SWM plates (0.5% peptone, 0.3% yeast extract, and 0.5% agar) or MG plates containing 0.4% agar for the swarming assay (15, 30, 31). Motility was observed after 48 h of incubation at 25°C.
Each bacterial strain was grown on an MG plate containing 1.5% agar for 48 h at 27°C. The detailed procedure has been described in previous papers (12, 31). Total EPS was calculated relative to total cellular protein when the EPS value from the ΔpvdL mutant was defined as 1.
Production of tabtoxin was detected by an agar plate diffusion test using E. coli DH5α as an indicator strain (13). An overnight culture of E. coli was centrifuged, and cells were resuspended in 0.9% NaCl at an OD600 of 0.1. Then 20 ml of MG with 0.5% agar was mixed with 2 ml of the E. coli suspension. An MG-glutamine plate containing 17 μM glutamine was used for the control experiment. Five microliters of an overnight culture of the WT or of each mutant in MG medium was spotted onto sterilized filter paper on the plates, followed by 48 h of incubation at 25°C.
Each bacterium was incubated on a modified KB plate (2% peptone, 1.6 nM MgSO4·7H2O, 8.6 mM K2HPO4, 1% glycerol, and 1.5% agar) for 24 h at 27°C. The blue fluorescence was visualized under UV light at 300 nm.
The concentration of pyoverdine produced by the WT and each mutant after overnight culture in MG medium was estimated spectrophotometrically as the absorbance at 405 nm. Each value was normalized for differences in cell density (35).
Each bacterial strain was grown on an MG plate containing 1.5% agar for 48 h at 27°C, scraped off, and suspended in 10 mM MgSO4. After adjustment to an OD600 of 1.0, 10-μl aliquots were spotted onto the film (Parafilm; Alcan Packaging, Neenah, WI) (35).
Each strain was grown for 24 h in KB medium, and 1 ml of an overnight-cultured bacterial suspension was diluted in 15 ml of KB medium with 0.5% agar. A paper disc (diameter, 5 mm; 3MM paper; Whatman PLC, Brentford, United Kingdom) containing an antibiotic (10 μl of chloramphenicol at 17.5 μg/μl or spectinomycin at 25 μg/μl) was overlaid on the plate. The growth inhibition zone was compared after incubation for 24 h at 27°C (31).
For the survival assay, each bacterial strain was grown in KB liquid culture for 24 h, and the concentration was adjusted to an OD600 of 0.1 with 10 mM MgSO4. Chloramphenicol was added at a final concentration of 100 μg/ml, and 10 μl of the bacterial suspension was plated onto KB plates to determine viable cell numbers at the indicated time intervals.
To examine the effects of pyoverdine and carbonyl cyanide m-chlorophenylhydrazone (CCCP) on the antibiotic susceptibility of each bacterial strain, commercially available pyoverdine or CCCP (both from Sigma-Aldrich, Inc., St. Louis, MO) at a final concentration of 0.5 ng/μl or 20 μM, respectively, was added to KB plates containing the ΔpvdL or ΔpsyI mutant, and inhibition of the growth of each strain was examined. After 24 h of incubation, inhibition zones were compared.
Each strain was grown in LB medium with 10 mM MgCl2 at 27°C to an OD600 of 0.3 and was incubated in MMMF medium [50 mM potassium phosphate buffer, 7.6 mM (NH4)2SO4, 1.7 mM MgCl2, and 1.7 mM NaCl (pH 5.7), supplemented with 10 mM (each) mannitol and fructose] for 1 h. The total RNA was extracted using a High Pure RNA isolation kit (Roche, Mannheim, Germany) according to the manufacturer's instructions. Semiquantitative reverse transcription-PCR (RT-PCR) analysis was conducted using 2 μg of total RNA (10). According to the registered sequences of P. syringae pv. phaseolicola 1448A, specific sets of primers (mexE-F and mexE-R, mexF-F and mexF-R, tolC-F and TolC-R, algT-F and algT-R, and epsD-F and epsD-R) were designed as shown in Table Table2.2. PCR was performed with one denaturation cycle of 2 min at 95°C, followed by 32 cycles of 30 s at 95°C, 30 s at 52°C, and 50 s at 72°C. Ten microliters of the PCR product was loaded onto a 2% agarose gel. The RT-PCR was also carried out without avian myeloblastosis virus reverse transcriptase as a negative control.
The nucleotide sequences of the fpvA, pvdL, and pvdJ genes have been deposited in the DDBJ, EMBL, and GenBank nucleotide sequence databases under accession numbers AB488783, AB488784, and AB488785, respectively.
To identify the genes involved in the pyoverdine-mediated iron uptake system in P. syringae pv. tabaci 6605, a gene cluster predicted to be involved in pyoverdine synthesis was identified by a BLAST search of the P. syringae pv. phaseolicola 1448A genome sequence (Fig. (Fig.1A).1A). We selected three genes to mutate, as shown in Table Table1.1. The pvdJ gene was hypothesized to encode the pyoverdine side chain peptide synthetase III l-Thr-l-Ser component, and pvdL may be required for synthesis of the pyoverdine chromophore. The fpvA gene is reported to produce the TonB-dependent ferripyoverdine receptor. Each gene was mutated as described in Materials and Methods (Fig. (Fig.1B1B).
The three mutants obtained, the ΔpvdJ, ΔpvdL, and ΔfpvA mutants, were able to grow in KB medium, and their growth rates were similar to that of the WT (data not shown). The total concentrations of pyoverdine in the WT and in these mutants grown in MG medium as the iron-depleted medium were estimated spectrophotometrically (35). As shown in Fig. Fig.1C,1C, the ΔfpvA mutant had a lower ability to produce pyoverdine, and the ΔpvdJ and ΔpvdL mutants were almost completely unable to synthesize pyoverdine under this experimental condition.
Pyoverdines of all Pseudomonas species have been reported to be composed of three parts: a conserved dihydroxyquinoline chromophore, an acyl side chain, and a variable peptide chain (33). The dihydroxyquinoline is responsible for the fluorescence of the pyoverdine as the characteristic pigmentation of the fluorescent group in Pseudomonas species (23). To investigate the effects of defects in pyoverdine biosynthesis, the WT, the ΔpvdJ, ΔpvdL, and ΔfpvA mutants, and the complemented ΔpvdL and ΔfpvA strains were grown on MG plates, and pigment production was examined (1). As shown in Fig. Fig.2A,2A, the ΔpvdJ and ΔpvdL mutants produced no pigment detectable under UV light at 300 nm, while the WT and the ΔfpvA mutant secreted considerable amounts of pigments. Both complemented strains had partially restored abilities to produce pigments.
Because iron acquisition is reported to play an important role in the surface movement of bacteria (18, 31), the motilities of the WT and pyoverdine-related mutants were investigated on an MG plate containing 0.25% agar for swimming and on an MG plate containing 0.4% agar and an SWM plate containing 0.5% agar for swarming. All of the strains showed similar swimming motilities under this condition (data not shown), whereas mutant strains, especially the ΔpvdJ and ΔpvdL mutants, exhibited a hyperswarming phenotype (Fig. (Fig.2B).2B). The WT strain showed more-active swarming motility on the SWM plate containing 0.5% agar than on the MG plate containing 0.4% agar, and its swarming motility was nearly lost on the MG plate containing 0.4% agar after the addition of FeCl3 at a final concentration of 50 μM. However, all mutant strains had higher swarming ability than the WT even on the MG plate containing FeCl3. This surface movement depends on flagellar motility, because the ΔfliC mutant, which is defective in the production of flagellin protein, was unable to spread on the 0.4% agar surface (data not shown). The swarming motilities of the two complemented strains, the ΔfpvAC and ΔpvdLC strains, were again reduced (Fig. (Fig.2B2B).
To compare the production of biosurfactants, which affect swarming motility, the drop-collapsing test was carried out. Results are shown in Fig. Fig.2C.2C. The ΔpvdJ and ΔpvdL mutants exhibited enhanced biosurfactant production, suggesting that pyoverdine might inhibit movement on a semisolid surface via biosurfactant production. The production of biosurfactants was reduced in the complemented strains (the ΔpvdJC and ΔpvdL-C strains) (Fig. (Fig.2C2C).
The virulence of the WT and the ΔpvdJ, ΔpvdL, and ΔfpvA mutants on host tobacco leaves was examined (Fig. 3A and B). The pyoverdine synthesis-defective ΔpvdJ and ΔpvdL mutants were less virulent than the WT by both the infiltration and the spray inoculation method. The ΔfpvA mutant retained a level of virulence similar to that of the WT. Bacterial populations of the WT and mutants were calculated at 2 and 6 days after dip inoculation (Fig. (Fig.3E).3E). The results demonstrated that the pyoverdine synthesis-defective ΔpvdJ and ΔpvdL mutants produced approximately 10% of the population of the WT on host tobacco leaves. The bacterial growth of the ΔfpvA mutant and the WT showed no significant difference. The virulence of the ΔfpvA and ΔpvdL complemented strains was restored completely by the infiltration method but only partially by the spray inoculation method (Fig. 3C and D).
P. syringae pv. tabaci produces a monocyclic β-lactam named tabtoxin (13). This dipeptide toxin induces chlorosis in host tobacco leaves to release the toxic tabtoxinine-β-lactam, which is an inhibitor of glutamine synthesis, by cleavage of the peptide bond. Because the bacterial virulence of the ΔpvdJ and ΔpvdL mutants was lower than that of the WT, as shown in Fig. 3A to E, the amounts of tabtoxin secreted were compared by an agar plate diffusion assay with E. coli as the indicator strain. Each bacterial solution was placed on a paper disc on an MG agar plate that had been seeded with E. coli and was incubated for 24 h at 27°C. As shown in Fig. Fig.3F,3F, inhibition zones were observed around the WT, the ΔfpvA mutant, and the ΔpsyI mutant, a mutant defective in the production of AHL, but not around the ΔpvdJ and ΔpvdL mutants. Because there was no inhibition zone surrounding any strain on the MG-glutamine plates (data not shown), the inhibition zone-producing toxin was determined to be tabtoxin. These results demonstrated that pyoverdine might control tabtoxin production independently of AHL.
The surfaces of tobacco leaves inoculated with each strain for 8 days were observed by scanning electron microscopy (Fig. (Fig.4A).4A). The WT bacteria were crowded around extracellular polymeric matrixes on the leaf surface, as reported previously (30). In contrast, mucoid material was hardly detected around the pyoverdine synthesis-defective ΔpvdJ and ΔpvdL mutants. A peculiar structure with a large amount of dried material was observed around the ΔfpvA mutant.
The amount of EPS was quantified on an MG plate containing 1.5% agar after 48 h of incubation at 27°C. As shown in Fig. Fig.4B,4B, the abilities of all mutants to produce EPS were considerably decreased. In particular, the ΔpvdL mutant secreted only small amounts of EPS. The abilities of both of the complemented strains to produce EPS were again reduced (Fig. (Fig.4C4C).
Because it was reported that P. syringae pathovars produce alginate as major EPS molecules, the expression of the algT gene, which encodes a positive transcription factor in alginate synthesis (17), was investigated. As shown in Fig. Fig.4D,4D, there was no significant difference between the algT expression of the WT and that of each mutant. Therefore, we examined the expression level of another gene, epsD, which is involved in EPS synthesis in Streptococcus thermophilus (21), by RT-PCR. The tyrosine kinase encoded by epsD in this bacterium phosphorylates phosphogalactosyltransferase (EpsE) and initiates EPS synthesis by a tyrosine phosphorylation regulatory system; these genes are reported to be well conserved among bacteria (21). Therefore, we investigated the expression of the epsD orthologue in P. syringae pv. tabaci. As shown in Fig. Fig.4D,4D, the expression of epsD was reduced in the ΔpvdL and ΔfpvA mutants. From these results, taken together with the scanning electron microscopy observations, the pyoverdine-mediated iron uptake system appeared to be correlated closely with biofilm formation.
It has been reported that EPS production is regulated by quorum sensing and that an EPS-deficient mutant is more susceptible to environmental stresses (7, 30). Because EPS production was decreased in the pyoverdine synthesis-defective ΔpvdL mutant, as shown in Fig. Fig.4B,4B, the tolerance to antibiotics was compared by a growth inhibition test. Unexpectedly, the growth inhibition zone was dramatically shrunk, suggesting that the ΔpvdL mutant had higher tolerance to antibiotics (Fig. (Fig.5A5A).
We previously found that AHL-defective (ΔpsyI and Δorf3) mutants had enhanced antibiotic tolerance (31), suggesting that this phenomenon might be correlated with quorum sensing. Therefore, the production of AHL by the ΔpvdL, ΔpvdJ, and ΔfpvA mutants was compared with that by the WT strain. Figure Figure5B5B shows the detection of AHL molecules by TLC analysis using Chromobacterium violaceum CV026, indicating the lower ability of the ΔpvdL mutant to produce AHL. However, the ability of the complemented ΔpvdL strain to produce AHL was restored.
For more-precise analysis of growth inhibition by antibiotics, each strain was treated with chloramphenicol for 10 to 30 min, and the viabilities of the strains were compared. As shown in Fig. Fig.5C,5C, most WT cells were killed immediately by chloramphenicol, and only 20% of the WT cells survived after 10 to 30 min of treatment, whereas more than 90% of the ΔpvdL and ΔpsyI mutants cells were still alive after 30 min of treatment. These results demonstrated that inhibition of the growth of P. syringae pv. tabaci 6605 might be dependent on AHL and/or pyoverdine chromophore production. Antibiotic susceptibility was partially complemented by the addition of exogenous pyoverdine prepared from Pseudomonas fluorescens to the ΔpvdL and ΔpsyI mutants (Fig. (Fig.5D).5D). The predicted structures of pyoverdine from P. syringae pv. tomato DC3000 and P. fluorescens were deduced from the molecular organization of the side chain peptide synthetase, and the peptide chains of pyoverdine from P. fluorescens and P. syringae are very different (24). This might be the reason why exogenous pyoverdine did not complement the mutants completely.
There is a possibility that the enhanced antibiotic tolerance of the ΔpvdL and ΔpsyI mutants is a result of an increase in drug efflux. Therefore, we investigated the effect of the uncoupler CCCP as an inhibitor of the multidrug efflux pump. As shown in Fig. Fig.5E,5E, the addition of CCCP decreased antibiotic tolerance, suggesting the participation of drug efflux in this phenomenon.
Among the numerous multidrug efflux pumps in gram-negative bacteria, transporters of the resistance-nodulation-division family (RND superfamily) have a predominant role in multidrug resistance (22). The RND superfamily is commonly responsible for the secretion of small molecules such as antibiotics (24). Therefore, the expression of three genes in the RND superfamily, mexE, mexF, and tolC, was examined by RT-PCR. As shown in Fig. Fig.6,6, the expression of these genes in the ΔpvdL and ΔpsyI mutants was significantly elevated over that in the WT strain.
Because iron is insoluble at biological pHs, the concentration of iron in the environment is very low. However, many bacteria have developed an iron acquisition mechanism utilizing a chelator, called a siderophore, that has a high affinity for iron(III). The structures of more than 50 pyoverdines from different strains of Pseudomonas have been determined as siderophores (33). They are composed of a highly conserved dihydroxyquinoline fluorescent chromophore attached to a variable peptide chain, which is synthesized by nonribosomal peptide synthetases. Pyoverdine has three iron-binding ligands; one is an o-dihydroxy aromatic group of quinoline in the chromophore, and the others are situated in the peptide chain (5). The chemical structure of the pyoverdine from P. syringae pv. tomato DC3000, whose genome sequencing has been completed, was deduced from the molecular organization of nonribosomal peptide synthetase genes (24). Gene clusters involved in pyoverdine biosynthesis and iron uptake in this bacterium are highly homologous to those in P. aeruginosa, although fpvI is not present in P. syringae. It has been reported that the production of pyoverdine and its receptors in P. aeruginosa depends on two extracytoplasmic sigma factors, PvdS and FpvI, respectively, and under iron-rich conditions, the central regulator Fur works as a repressor of iron uptake genes via these sigma factors (23). P. syringae pv. phaseolicola 1448A has two fpvA genes, the first of which contains a PvdS-dependent promoter, suggesting that this gene might be regulated by PvdS (24). PvdS has been reported to control the production of the virulence factors endoprotease and exotoxin A in P. aeruginosa (35). To reveal the role of pyoverdine in bacterial virulence, the regulation of pyoverdine-related gene expression must be investigated.
In this study, we investigated the functions of the siderophore pyoverdine and the ferripyoverdine receptor in P. syringae pv. tabaci 6605 using genetically defective mutants. The ΔpvdJ and ΔpvdL mutants had remarkably reduced abilities to produce pyoverdine in MG medium, whereas this ability was not drastically decreased for the ΔfpvA mutant, as shown in Fig. Fig.1C.1C. From the results obtained in this study, we have schematically depicted the pyoverdine-mediated regulatory network of virulence expression in P. syringae pv. tabaci 6605 (Fig. (Fig.7).7). Mutations of pvdL resulted in the reduction of both tabtoxin (Fig. (Fig.3F)3F) and AHL (Fig. (Fig.5B)5B) production; thus, the ΔpdvL mutant exhibited activation of multidrug efflux pump-related genes, such as mexE, mexF and tolC (Fig. (Fig.6),6), tolerance to antibiotics (Fig. 5A and C), increased swarming motility (Fig. (Fig.2B)2B) including enhancement of biosurfactant production (Fig. (Fig.2C),2C), and repression of EPS production (Fig. (Fig.4B)4B) including epsD expression (Fig. (Fig.4D).4D). Although we generated two pyoverdine-defective mutant strains, the ΔpvdL and ΔpvdJ mutants, the effect of the pvdJ mutation was weaker than that of the pvdL mutation, and the ΔpvdJ mutant showed reduced production of tabtoxin and weakly enhanced biosurfactant production and swarming motility. On the other hand, the production of tabtoxin and AHLs by the ΔfpvA mutant was not impaired, and only production of EPS and epsD expression were reduced. In P. syringae pv. phaseolicola 1448A, two fpvA genes were repeatedly clustered as PSPPH1927 and PSPPH1928. The number of fpvA genes in P. syringae pv. tabaci 6605 is not known, and another fpvA gene may compensate for the function of the disrupted fpvA gene. These results indicate that pyoverdine may function as a signal molecule to activate the production of tabtoxin and AHLs and to repress expression of the mexE, mexF, and tolC genes and tolerance to antibiotics independently of FpvA. As mentioned above, the effects of the mutation were prominent in the ΔpvdL mutant but limited in the ΔpvdJ mutant, although neither mutant was able to produce pyoverdines. These results suggest that the dihydroxyquinoline chromophore, which is synthesized in the ΔpvdJ but not in the ΔpdvL mutant, is important and has some signaling activity. Thus, pyoverdine, especially the dihydroxyquinoline chromophore, seems to be required for the production of tabtoxin and AHLs. The reductions in the levels of tabtoxin production in the ΔpdvL and ΔpvdJ mutant strains may cause the weak symptoms observed for inoculated tobacco leaves. Furthermore, the reduction in the level of AHL production in the ΔpvdL mutant might lead to the transcriptional activation of multidrug efflux pump-related genes, consequently increasing tolerance to antibiotics. Although the mechanism by which the ΔpvdL mutant activates biosurfactant production, swarming motility, and expression of the mexE, mexF, and tolC genes is still unclear, PvdS and Fur may be involved in this regulation.
Recently, genomewide screening of the genes required for swarming motility was performed using a systematic and comprehensive collection of gene-disrupted E. coli K-12 mutants (14). The mutations in most of the genes required for enterobactin, a major siderophore in this bacterium, and other genes encoding iron uptake proteins significantly affected swarming motility. Microarray analysis using gene expression profiles of Salmonella enterica serovar Typhimurium during swarming and in liquid media indicated that iron metabolism was strongly induced in bacteria grown on swarming agar plates under reduced nutritional conditions (34). In P. aeruginosa, lower nutritional conditions may induce swarming and biosurfactant production, although excess iron reduces swarming motility (4, 8). These results are consistent with our findings in this study that the ΔpvdJ and ΔpvdL mutants of P. syringae pv. tabaci 6605 had enhanced swarming ability but that exogenous iron reduced it (Fig. (Fig.2B).2B). Previously, we demonstrated that a homologue of the 3-oxoacyl-(ACP) synthase III gene (orf3) located between the flagellin gene fliC and the flagellar glycosyltransferase gene fgt2 is required for the production of AHL and fatty acids in P. syringae pv. tabaci 6605 (31). The Δorf3 mutation enhanced biosurfactant production and swarming motility. Although the structure and biosynthetic pathway of biosurfactant production have not been elucidated, we detected candidate glycolipids in the ethyl acetate extracts of this mutant on the developed TLC plate using orcinol reagent. The pyoverdine synthesis-defective ΔpvdJ and ΔpvdL mutants also overproduced biosurfactants, as shown in Fig. Fig.2C,2C, indicating that biosurfactant production is controlled by pyoverdine. In P. syringae pv. tabaci 6605, swarming ability is dependent on flagella (30). These results and previous reports indicate that iron starvation induces swarming motility by biosurfactant overproduction, because P. syringae pv. tabaci 6605 uses flagellar movement under more-favorable nutrient conditions.
Both the ΔpvdJ and ΔpvdL mutants showed reduced virulence toward the host tobacco leaf, as evidenced by the fact that necrotic lesions were hardly observed in areas inoculated with these mutants (Fig. 3A and B). Tabtoxin induces chlorosis in the host tobacco plant (13). As shown in Fig. Fig.3F,3F, we found that pyoverdine-defective mutants had reduced abilities to produce tabtoxin. For P. aeruginosa, it has been reported that the siderophore pyoverdine regulates the production of major causes of disease, such as exotoxin A, an endoprotease, and pyoverdine itself (16). Tabtoxin production in P. syringae pv. tabaci 6605 might also be under the control of pyoverdine.
Recent studies of P. aeruginosa PAO1 have reported that iron acquisition and biofilm formation are linked and that a pyoverdine-mediated iron acquisition system is necessary for a mature biofilm to form a mushroom-like structure (2, 3). Lactoferrin, which has iron chelator activity, inhibits bacterial biofilm maturation, and the pyoverdine-defective mutant formed only a thin layer of bacterial aggregates (2). In P. syringae pv. tabaci 6605, the polymeric extracellular substance was hardly observed around the pyoverdine-defective mutants on inoculated tobacco leaves, as observed in scanning electron micrographs (Fig. (Fig.4A).4A). The expression of epsD, which is involved in EPS biosynthesis via a tyrosine phosphorylation regulatory system, was remarkably reduced in the ΔpvdL and ΔfpvA mutants, as shown in Fig. Fig.4D.4D. Previously, it was reported that pyoverdine controls the production of virulence factors in P. aeruginosa and that pyoverdine itself acts as a signaling molecule like the AHLs in the quorum-sensing system (16). In this study, EPS production seemed to be regulated by iron uptake via pyoverdine and the TonB-dependent receptor. However, the regulation of EPS biosynthesis must be investigated in detail.
The absence of pyoverdine impaired mature biofilm formation not only by animal pathogens but also by phytopathogenic bacteria. There was a dried and flat biofilm around the ΔfpvA mutant on the inoculated tobacco surface. The shape of this biofilm resembled those of the Δorf3 and ΔpsyI mutants, in which the abilities to form AHL and take up iron have been reduced (31). A previous paper reported that a fur (ferric uptake regulator) deletion mutant in P. syringae pv. tabaci 11528 showed decreased production of AHLs and that Fur regulates the psyI and psyR genes at the transcriptional level (6). Although the molecular mechanisms by which pyoverdine and iron control biofilm development still remain obscure, pyoverdine-mediated iron acquisition might be linked with the regulation of quorum sensing in phytopathogenic bacteria.
The mechanisms for increased tolerance to antibiotics in the mutants defective in AHL production (ΔpsyI mutant) and pyoverdine chromophore production (ΔpvdL mutant) are not known yet. However, the active multidrug efflux pumps are probable candidates (22) to explain this phenotype. In gram-negative bacteria, the RND superfamily pumps play the most important role in multidrug resistance (22). As shown in Fig. Fig.6,6, the expression of genes related to the RND multidrug efflux pump was significantly induced. What is the role of the multidrug efflux pump in iron uptake? A pyoverdine precursor is synthesized in the cytoplasm and transported into the periplasm through ATP-binding cassette transporters (PvdE), whereas the mechanism for secretion of pyoverdine is not clear (24). It is not yet known what compounds are secreted by RND-type multidrug efflux pumps. Pyoverdine itself is also one of the candidates for the compounds secreted by the RND multidrug efflux pumps. Further experiments with these genes will be necessary to elucidate their relationship to pyoverdine synthesis, AHL production, and antibiotic tolerance. Our previous report and current results indicate that iron acquisition mediated by pyoverdine is indispensable for biofilm formation and virulence in P. syringae pv. tabaci 6605 (31). It might be possible to control bacterial virulence by altering the iron uptake and quorum sensing systems using pyoverdine and AHLs.
We thank the Leaf Tobacco Research Laboratory of Japan Tobacco Inc. for providing P. syringae pv. tabaci 6605.
This work was supported in part by the Program for Promotion of Basic Research Activities for Innovative Bioscience (PROBRAIN).
Published ahead of print on 23 October 2009.