INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

In gram-negative bacteria, the alternate sigma factor ⁵⁴, working in concert with a transcriptional activator that belongs to the NtrC superfamily, activates a variety of genes that are regulated in response to external stimuli (1). For example, in various bacteria, ⁵⁴ is required for expression of the enzymatic pathways responsible for nitrogen utilization, dicarboxylate transport, xylene degradation, and hydrogen utilization (6, 32, 39, 41, 61). ⁵⁴ is also involved in the regulation of virulence-related factors in both plant and animal pathogens, including pilin, flagellin, and alginate synthesis in Pseudomonas aeruginosa (19, 58, 60); capsular expression in Klebsiella pneumoniae (3); and regulation of hrp gene expression and coronatine biosynthesis in Pseudomonas syringae (23, 24).

Our laboratory has developed a bacterial pathogenicity model that utilizes a clinical isolate of P. aeruginosa (strain UCBPP-PA14 [referred to here as PA14]) that elicits severe soft-rot-like symptoms and proliferates when infiltrated into Arabidopsis leaves (52), kills the larvae of the wax moth caterpillar Galleria mellonella (30), causes lethal sepsis in a mouse full-skin-thickness burn model (52), and kills the nematode Caenorhabditis elegans (40, 56, 57). Interestingly, there is significant overlap among the PA14 virulence factors required for pathogenesis in plants, nematodes, insects, and mice. For example, among 21 genes identified as being involved in pathogenesis by screening transposon-induced PA14 mutants in plants and nematodes, 18, 17, 19, and 21 of these genes were required for pathogenicity in Arabidopsis, nematodes, wax moths, and mice, respectively (30, 40, 53, 57).

In other studies, we showed that rpoN is a key virulence factor for the plant pathogen P. syringae (23, 24). Specifically, molecular and genetic analysis showed that the P. syringae rpoN gene is required for expression of the P. syringae hrp gene cluster, a block of contiguous genes, some of which encode components of a type III secretory system (2, 18, 26, 45, 46, 59).

Given the facts that RpoN activates the expression of a wide variety of environmentally regulated genes and is required for virulence in a variety of pathogens, we hypothesized that RpoN would play a central role in the evolution of P. aeruginosa's ability to be a pathogen of evolutionarily disparate hosts. In this study we describe the results of experiments that involved the construction of a P. aeruginosa PA14 rpoN mutant to study the role of ⁵⁴ in P. aeruginosa pathogenesis in a variety of plant and animal hosts. Surprisingly, we report that the P. aeruginosa rpoN gene is not a universal virulence factor required for multihost pathogenesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains, plasmids, and culture conditions. The bacterial strains and plasmids used and constructed in this study are listed in Table 1. Escherichia coli and P. aeruginosa strains were grown at 37°C in L broth, King's A (KA), King's B (KB) (33), or M9 minimal salts media. Nitrogen source utilization tests for PA14 rpoN mutants were performed in M9 salts minimal medium by replacing ammonium chloride with an alternative nitrogen source at 5 mM when required. Bacterial motility was tested on "swarm plates" (35). Pyocyanin assays (17) were carried out in KA broth containing 100 µM FeCl₃ (17, 33). Pyoverdin was assayed on KB plates as described previously (57). C. elegans killing assays were carried out on NG agar ("slow killing" [56]) or PGS agar ("fast killing" [56]) as described elsewhere. Antibiotic concentrations for E. coli strains were as follows: streptomycin, 150 µg/ml; kanamycin, 25 µg/ml; tetracycline, 12 µg/ml; gentamicin, 5 to 10 µg/ml; and spectinomycin, 20 µg/ml. Antibiotic concentrations for P. aeruginosa were as follows: streptomycin, 200 µg/ml; kanamycin, 200 µg/ml; tetracycline, 75 µg/ml; gentamicin, 30 µg/ml; nalidixic acid, 50 µg/ml; rifampin, 100 µg/ml, and carbenicillin, 300 µg/ml.

Bacterial genetics. pJSR1 derivatives were introduced into Pseudomonas strains via triparental matings with MM294/pRK2013 as the donor of transfer functions as described previously (14). Plasmid pSMC21 (7) containing the Aequorea victoria green fluorescent protein (GFP) was introduced into PA14 rpoN::Gen^r by electroporation.

Pyocyanin assays. Pyocyanin was measured at 520 nm in acidic solution by using a modified version of a previously described method (17). Cultures were grown from a 100-fold dilution of a log-phase culture in KA broth modified with 100 µM FeCl₃. After 20 h, 1 ml of the culture was extracted with 2 ml of chloroform and centrifuged for 5 min. The blue chloroform solution was transferred into a new tube containing 1 ml of 0.2 N HCl to extract pyocyanin into the acidic solution. The concentration was determined by measuring the optical density at 520 nm (OD₅₂₀).

Plant material and growth of plants. Arabidopsis ecotypes Llagostera (Ll-0) and Landsberg erecta (La-er) were obtained from the Arabidopsis Biological Resource Center, Columbus, Ohio. Arabidopsis plants were grown in Metro-Mix 2000 in either a climate-controlled greenhouse at 19°C under a 12-h light-dark cycle with supplemental fluorescent illumination or in a Percival AR-60L growth chamber at 20°C and 50% relative humidity.

Arabidopsis pathogenicity assays. Six- to eight-week-old intact or detached Arabidopsis rosette leaves were used for pathogenicity assays. The pathogenicity of PA14 strains was tested by placing detached Ll-0 leaves on a 1.5% water agar surface with their petioles embedded into the agar and inoculating the leaves by growing lawns of PA14 or PA14 rpoN::Gen^r overnight on LB agar medium containing rifampin (PA14) and gentamicin (rpoN::Gen^r), cutting 3-mm-diameter agar cylinders from these plates, and placing the cylinders bacterial side down on one or both sides of the central vein in the top half of a leaf. The growth of PA14 or PA14 rpoN::Gen^r in Arabidopsis leaves was determined by infiltrating the leaves of intact La-er plants with 5 × 10⁴ CFU/cm² of leaf area as described previously (52). The growth of PA14 or PA14 rpoN::Gen^r in detached Arabidopsis leaves was determined by infiltrating Ll-0 leaves with a bacterial suspension in 10 mM MgSO₄ at a density of 5 × 10⁴ CFU/ml for 1 h under a slight vacuum at room temperature in the wells of a six-well microtiter plate. At days 0, 1, 2, 3, and 4, the titer of the bacteria in each of two punches in each of three leaves was determined as described previously (15).

Infection of G. mellonella larvae. The 50% lethal dose (LD₅₀) of PA14 strains in G. mellonella larvae was determined as described previously (30). In brief, overnight cultures grown in KB medium were diluted 1:100, allowed to grow until they reached an OD₆₀₀ of 0.3 to 0.4, and resuspended in 10 mM MgSO₄. After dilution to an OD₆₀₀ of 0.1 with 10 mM MgSO₄, serial 10-fold dilutions were made in 10 mM MgSO₄ containing 1 mg of rifampin/ml and 10 mg of carbenicillin/ml. A 10-µl Hamilton syringe was used to inject 5-µl aliquots into the hindmost left proleg of fifth-instar G. mellonella larvae purchased from Van der Horst Wholesale, St. Mary's, Ohio. Groups of 10 larvae infected with the same dose of bacteria were placed in petri dishes and incubated at 25°C for 60 h. Larvae were scored as dead when they no longer moved upon shaking of the petri dish or poking with a pipette tip.

Mouse full-skin-thickness burn model. Inbred AKR mice that had been subjected to a thermal burn injury were infected with PA14 strains as described previously (54). In brief, 6-week-old male mice were anesthetized by the injection of phenobarbital. The animals were then shaved, and a ventral skin fold was elevated. Two brass blocks preheated to 92 to 95°C were applied to the skin fold for 5 s to deliver a full-skin-thickness burn covering ca. 5% of the body surface area. The burn eschar was injected with 100 µl of bacterial suspension at a titer of 5 × 10⁴ or 5 × 10⁶ bacteria per ml. Bacteria for the inoculation were grown overnight in LB, diluted 1:100, allowed to grow until they reached an OD₆₀₀ of 1.6 to 1.7, pelleted by centrifugation, and resuspended and diluted in 10 mM MgSO_4. The mouse protocol was reviewed and approved by the Animal Care Committee of the Massachusetts General Hospital.

C. elegans killing assay. The killing kinetics of C. elegans strain Bristol N2 by PA14 strains were determined on PGS agar (fast-killing assay) or NG agar (slow-killing assay) as described previously (56).

Nucleic acid manipulations. Routine DNA manipulations such as DNA blots and plasmid DNA isolation were performed as described previously (4). Restriction enzymes, T4 DNA ligase, and calf intestine phosphatase were purchased from Boeringer Mannheim and New England BioLabs and used according to the manufacturers' specifications.

Cloning the P. aeruginosa PA14 rpoN gene and construction of a PA14 rpoN mutant. A cosmid clone, pRPON10, containing a presumptive P. aeruginosa PA14 rpoN gene was identified in a pJSR1 cosmid library (52) by colony hybridization with P. aeruginosa strain PAK rpoN gene on plasmid pKI11 as a hybridization probe. The presumptive PA14 rpoN gene was mapped to a 4.0-kb XhoI-EcoRI fragment and was subcloned into pBSK(+) to produce pPAR4. DNA sequence analysis showed 95% identity out of 141 nucleotides sequenced between the PA14 rpoN gene and the P. aeruginosa PAO1 rpoN gene (reference 31 and data not shown).

Scanning electron and confocal microscopy. To monitor the attachment of PA14 GFP-labeled strains to the epidermal cell walls of Ll-0 leaves and to Ll-0 trichomes by confocal laser microscopy, 4-mm-diameter leaf disks were immersed in bacterial suspensions at a density of ca. 10⁹ cells/ml under weak vacuum to remove possible air bubbles from the surface of the leaf disks and make the leaf surface more accessible to bacteria. Subsequently, leaf disks were incubated for 24 h in the bacterial suspensions with shaking under normal pressure and room temperature. The leaf disks were examined with a confocal laser spectrophotometer (Leica TCS NT) by excitation at 488 nm and by monitoring the emission intensity at 511 nm. To monitor the development of PA14 strains in Ll-0 parenchyma cells by scanning electron microscopy, 4-mm-diameter leaf disks were cut from the central portion of a leaf from either side of the central vein with a cork borer and then immersed in a PA14 or PA14 rpoN::Gen^r suspension at an OD₆₀₀ of 0.02 in 10 mM MgSO₄ in the wells of a 24-well microtiter dish, followed by incubation at room temperature. Leaf disks were fixed at 1, 2, 3, and 4 days postinfection (dpi) in 4% paraformaldehyde, passed through an ethanol series (30, 50, 70, 96, and 100%), and freeze fractured in liquid nitrogen. The plant material was dried in a critical-point drying apparatus (Samdri-PVT-3B; Tousimis), mounted on stubs, coated with a 20- to 25-µm layer of gold-palladium in a Hummer II Sputter Coater (Technics), and studied by using an AMRAY 1000 scanning electron microscope.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

PA14 rpoN::Gen^r exhibits a variety of characteristic RpoN^. phenotypes. As described in Materials and Methods, an interspecies hybridization method was used to identify and clone the P. aeruginosa PA14 rpoN gene. The PA14 rpoN gene was partially sequenced and over this region showed 95% identity to both P. aeruginosa PAO1 and PAK rpoN sequences (data not shown). Also, as described in Materials and Methods, a PA14 rpoN mutant (rpoN::Gen^r) was constructed by inserting a DNA cassette conferring gentamicin resistance into the ClaI site of the PA14 rpoN gene and transferring the disrupted gene into the PA14 genome by homologous recombination. The Gen^r cassette is inserted within the N-terminal 33% of rpoN, before the highly conserved carboxy-terminal region (31, 41).

PA14 rpoN::Gen^r synthesizes reduced levels of pyocyanin. P. aeruginosa secretes a variety of compounds under nutrient-limiting conditions, including the yellow-green siderophore pyoverdin and the blue microbial toxin pyocyanin (10, 43). Pyoverdin may be a virulence factor in burn wound infections (5, 43), and pyocyanin has been implicated in pulmonary artery injury (9, 22). KA Fe³⁺ medium or swarm plates turned noticeably blue when PA14 was grown to stationary phase, whereas no blue color was observed after the growth of PA14 rpoN::Gen^r. Quantitative measurement showed that PA14 rpoN::Gen^r produced 56% ± 13% less pyocyanin than did the wild type and that pyocyanin production was restored to wild-type levels in PA14 rpoN::Gen^r(pPAR4SR). Pyoverdin production appeared to be unaffected in PA14 rpoN::Gen^r. PA14 colonies grown on KB plates express pyoverdin, which can be detected by fluorescence when the colonies are exposed to long-wavelength UV light. Visual inspection revealed no noticeable difference between wild-type and PA14 rpoN::Gen^r colonies.

PA14 rpoN::Gen^r exhibits reduced pathogenicity in mice. As shown in Table 3, PA14 rpoN::Gen^r was significantly less pathogenic than wild-type PA14 in a mouse burn model. A high percentage of lethality was observed with PA14 at both a relatively high dose of 5 × 10⁵ cells (100% lethality) and a relatively low dose of 5 × 10³ cells (~80% lethality). In contrast, only five of eight and one of seven animals died when inoculated with 5 × 10⁵ and 5 × 10³ PA14 rpoN::Gen^r, respectively. The use of two different doses allows a better estimation of the mutation's effects on virulence. In the case of rpoN, even the high dose was not as lethal as wild type at a 100-fold-lower dose, demonstrating a significant reduction in virulence. These burn model data are consistent with previous reports demonstrating that P. aeruginosa rpoN mutants or P. aeruginosa mutants which contain lesions in RpoN-regulated genes exhibit reduced colonization and virulence in a number of other model systems. rpoN mutants showed a reduced ability to colonize in a chronic murine intestinal mucosal model (48) or in human respiratory epithelial xenografts (11) and exhibited reduced virulence in a murine corneal scratch model (51) or in a murine model of acute pneumonia (12). One limitation of these studies, as well as of our own, however, is the fact that rpoN mutants are nonmotile, and it is not known to what extent the nonmotile phenotype of the rpoN mutants contributes to the loss of virulence. Nonmotile mutants of P. aeruginosa strains M2, PA01, and MT1200 exhibited significant reduction in virulence in a mouse burn model (16). The reduction was at the same level or greater than that seen with the PA14 rpoN mutant. Thus, the entire effect on virulence of the rpoN mutant in P. aeruginosa PA14 in the mouse burn model could be due to the loss of motility. RpoN is also an important virulence factor for virulence of Vibrio anguillarum in fish (47) and Vibrio cholerae in an infant mouse colonization model (13, 34). In the case of V. cholerae, it appears that unknown rpoN-regulated genes, in addition to the flagellum genes, are required for colonization.

PA14 rpoN::Gen^r exhibits reduced pathogenicity in Arabidopsis. We have developed several assays to assess the pathogenicity of P. aeruginosa in Arabidopsis leaves, comparing attachment to plant leaf surfaces and monitoring growth rate and symptom development. We previously showed (49) that when Arabidopsis ecotype Ll-0 leaf disks are incubated in a dense bacterial suspension (i.e., ~10⁹ cells/ml) for 24 h, wild-type PA14 cells attached to the entire leaf epidermal cell surface, congregated above most of the stomata, and formed large, multilayered clusters on the trichomes (Fig. 1A and C). In contrast, as shown in Fig. 1B and D, PA14 rpoN::Gen^r cells primarily attached to the grooves formed at cell junctions, and very few of the mutant bacteria were located above the stomata or associated with trichomes.

View larger version (142K): [in this window] [in a new window]   FIG. 1.   P. aeruginosa adheres to leaf surfaces. Confocal scanning microscopy of PA14 or rpoN::Genr expressing GFP in association with Arabidopsis ecotype Ll-O. (A) PA14(pSCM21) adheres to the entire surface of the leaf, including the stomatal (S) openings. Bar, 20 µm. (B) PA14 rpoN::Genr(pSCM21) cells primarily attached to the grooves formed at cell junctions; very few of the mutant bacteria were located above the stomata (S). Bar, 20 µm. (C) PA14(pSCM21) (b) formed large multilayered clusters on trichomes (T). Bar, 10 µm. (D) PA14 rpoN-GFP::Genr(pSCM21) did not attach to the trichome (T) surface and concentrated loosely around them. Bar, 10 µm. Leaf disks were incubated with PA14/pSMC21 (A and C) or PA14 rpoN::Genr/pSMC21(B and D) in bacterial suspensions at a density of ~109 CFU/ml under slight vacuum for 1 h, followed by incubation for 24 h in an orbital shaker at room temperature, and then examined by confocal laser microscopy as described in Materials and Methods.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Growth of PA14 and PA14 rpoN::Genr in Arabidopsis leaves. Detached Ll-0 leaves were vacuum infiltrated with bacterial suspensions as described in Materials and Methods. The initial titers were ca. 5 × 104 CFU/cm2 leaf area. Bacterial titers in the infiltrated leaves were measured immediately after infiltration and at 1, 2, 3, and 4 days after infiltration as described in Materials and Methods. Each value represents the average of three replicates. Similar results were obtained when the leaves of intact 6-week-old La-er plants were infiltrated at a titer of 5 × 104 CFU/cm2 of leaf area and incubated as described in Materials and Methods (not shown). Symbols: , PA14; , PA14 rpoN::Genr.

View larger version (104K): [in this window] [in a new window]   FIG. 3.   Scanning electron micrographic visualization of freeze-fractured infected Arabidopsis Ll-0 leaves depicting the failure of P. aeruginosa PA14 rpoN::Genr cells to attach perpendicularly to and penetrate the cell walls of vessel parenchyma cells. The leaves of intact Ll-0 plants were infiltrated under vacuum, the infected plants were incubated for 72 h, and then examined by scanning electron microscopy as described in Materials and Methods. b, bacterium, W, plant wall; WC, wall convolution. (A) PA14 rpoN::Genr cells parallel to the surface of a smooth cell wall of a parenchyma vessel cell. Bar, 1 µm. (B and C) PA14 cells attached perpendicularly to the surface of a highly convoluted cell wall of a parenchyma vessel cell. Holes (arrows) in the plant cell walls with diameters similar to that of the bacteria are readily visible. Bars, 1 µm.

PA14 rpoN::Gen^r exhibits reduced C. elegans killing. When the nematode C. elegans is fed a lawn of PA14 grown on solid medium, the nematodes die in two characteristic ways depending on the medium on which PA14 is grown (56). On a high-osmolarity but low-phosphate medium (e.g., PGS), the nematodes die rapidly (fast killing) over the course of 24 h. In contrast, nematodes that are fed PA14 grown on a low-osmolarity but high-phosphate medium (NGM) are killed at a slower rate (slow killing), dying after 2 to 3 days. Previous studies have shown that the killing of C. elegans by PA14 under these two different conditions is mediated by distinct molecular mechanisms (40, 56, 57). As shown in Fig. 4, there was significantly less killing of C. elegans by PA14 rpoN::Gen^r compared to killing by wild-type PA14 under both the fast- and slow-killing conditions. The killing of PA14 rpoN::Gen^r in the fast-killing assay was restored to wild-type levels by pPAR4SR, which carries wild-type rpoN, indicating that the loss of pathogenicity phenotype of PA14 rpoN::Gen^r was due to the disruption of the rpoN gene (data not shown). The decrease in fast killing by PA14 rpoN::Gen^r is consistent with the results described above showing that PA14 rpoN::Gen^r synthesizes reduced levels of pyocyanin. Previous results from our laboratory showed that pyocyanin is an important toxin mediating the fast-killing process (40).

View larger version (18K): [in this window] [in a new window]   FIG. 4.   PA14 rpoN::Genr killing of C. elegans under fast- and slow-killing conditions. (A) Fast killing. L4 stage worms were seeded onto bacterial lawns on PGS medium, and mortality was recorded as described earlier (40). Each value is the mean for three replicates. Symbols: , PA14; , PA14 rpoN::Genr. (B) Slow killing. L4 stage worms were seeded onto bacterial lawns on NGM media, and mortality was recorded as described previously (56). Each value is the mean for six replicates. Symbols: , PA14; , rpoN::Genr.

PA14 rpoN::Gen^r does not exhibit reduced pathogenicity in G. mellonella A single PA14 cell is sufficient to kill a greater wax moth caterpillar when PA14 is injected into the hemolymph (30). We found that PA14 and PA14 rpoN::Gen^r were indistinguishable in their ability to kill fifth-instar wax moth larvae (data not shown). The LD₅₀ in both cases was approximately one bacterial cell. This is interesting in light of the decreased virulence of PA14 rpoN::Gen^r in the mouse burn model. Microbial defense in insects shows interesting parallels to innate immunity in mammals and plants (8, 25, 27). Members of the cecropin class of antibacterial peptides have been found in both mammals and insects, and cecropin A in Drosophila melanogaster is regulated by a cascade similar to the one involving NF-B activation in the mammalian inflammatory response (38). The reduction in virulence of PA14 rpoN::Gen^r in mice but not in wax moths suggests that rpoN does not play a significant role in the regulation of virulence factors that affect components of the host innate immune system conserved between insects and mammals.

Conclusions. Previous work with PA14 in our laboratory has shown that PA14 utilizes a common set of virulence factors in evolutionarily disparate hosts (30, 40, 53, 57). Given this and the highly pleiotropic nature of rpoN mutants, we expected that disruption of rpoN would result in a significant impairment in plant, nematode, insect, and mouse pathogenicity. Surprisingly, a major reduction in pathogenicity of the P. aeruginosa PA14 rpoN mutant was observed only in C. elegans killing and in the elicitation of sepsis in the mouse burn model. Although decreased lesion size and a 10-fold reduction in in planta growth was observed in Arabidopsis at early stages of an infection (3 to 4 dpi), at later stages (7 dpi) the P. aeruginosa rpoN mutant elicited disease symptoms that were indistinguishable from those caused by the wild type. This contrasts with the absolute requirement for RpoN function in P. syringae for pathogenicity (23, 24). No effect of the rpoN mutation was observed in G. mellonella killing. Thus, although it has been previously reported that P. aeruginosa rpoN mutants exhibit impaired virulence in mice, the results reported here provide a more complete perspective concerning the role of RpoN in the diverse pathogenic interactions that P. aeruginosa has with several evolutionarily disparate hosts. The major conclusion is that, in contrast to our expectations, rpoN does not appear to regulate any genes that are likely to encode virulence factors universally required for pathogenicity irrespective of the host.