Regulation of the Pseudomonas aeruginosa toxA, regA and ptxR genes by the iron-starvation sigma factor PvdS under reduced levels of oxygen

Cystic fibrosis (CF) is a genetic disease characterized by airway obstruction, chronic bacterial infection, and excessive host inflammatory response (Hassett et al., 2002; Sadikot et al., 2005; Yoon et al., 2002). Pseudomonas aeruginosa is the predominant opportunistic pathogen associated within the thickened mucus layer of the CF lung (Baltch, 1994; Hassett et al., 2002). As disease progresses, P. aeruginosa accelerates lung deterioration and establishes a chronic infection with frequent exacerbations (Baltch, 1994; Davis et al., 1996). P. aeruginosa produces an arsenal of virulence factors that facilitate its survival within the CF lung and are important for pathogenesis (van Delden, 2004).

Exotoxin A (ETA) is considered one of the most powerful extracellular virulence factors produced by P. aeruginosa (Iglewski & Kabat, 1975). The 68 kDa ETA protein, encoded by toxA, is an ADP-ribosyl transferase that irreversibly inhibits protein synthesis in eukaryotic cells causing cell death (Hamood et al., 2004; Iglewski & Kabat, 1975). The regulation of ETA production is an intricate process that involves environmental factors and several regulators (Hamood et al., 2004). Environmental factors include cation concentration, temperature, oxygen and iron levels (Hamood et al., 2004; Liu, 1973). The most extensively analysed of these factors is iron which represses ETA production at the transcriptional level (Frank & Iglewski, 1988; Grant & Vasil, 1986; Hamood et al., 2004; Lory, 1986). Therefore, to achieve maximum ETA levels in vitro, P. aeruginosa is usually grown in iron-deficient medium at 32 °C with good aeration (Frank & Iglewski, 1988; Grant & Vasil, 1986; Hamood et al., 2004; Lory, 1986).

The regulation of toxA expression involves at least three positive regulators, including RegA, PtxR and the iron-starvation (alternative) sigma factor PvdS (Frank & Iglewski, 1988; Hamood et al., 2004; Wick et al., 1990). The 29 kDa protein RegA is essential for toxA expression (Frank & Iglewski, 1988; Hamood et al., 2004; Wick et al., 1990). Neither ETA nor toxA mRNA were detected in a regA-deficient mutant of the P. aeruginosa strain PA103 (Wick et al., 1990). At this time, the mechanism through which RegA regulates toxA expression has not been completely determined (Frank & Iglewski, 1988; Hamood & Iglewski, 1990; Hamood et al., 2004; Raivio et al., 1996; Wick et al., 1990). The LysR-type transcriptional activator PtxR is not essential, but does increase toxA expression by four- to fivefold, although the exact mechanism of this regulation is unknown (Hamood et al., 1996, 2004). PtxR also co-regulates the expression of several genes controlled by quorum sensing, and of the pvc operon (Carty et al., 2006; Stintzi et al., 1999). PvdS enhances toxA expression and directs the transcription of genes required for production of the siderophore pyoverdine (e.g. pvdD, pvdE, pvdF) and two extracellular proteases (Cunliffe et al., 1995; Ochsner et al., 1996). The iron-starvation (IS) box (consensus TAAAT, followed by the CGT triplet 16–17 nt downstream) is the DNA signature recognized by the PvdS sigma factor (Rombel et al., 1995; Visca et al., 2002; Wilson et al., 2001). Confirmed or potential IS boxes have been detected in the promoter region of PvdS-regulated genes of P. aeruginosa PAO, including toxA, regA, ptxR, pvdD, pvdEF and several other genes (Hunt et al., 2002; Ochsner et al., 2002; Wilson et al., 2001). Evidence suggests PvdS regulates toxA expression through regA (Hamood et al., 2004; Ochsner et al., 1996). Repression of toxA and the pyoverdine genes occurs through the negative regulation of pvdS expression by iron and the Fur repressor protein (Hamood et al., 2004; Visca, 2004). Moreover, PvdS activity is controlled at the post-translational level by the FpvR anti-sigma factor through a surface signalling mechanism which is triggered by pyoverdine binding to the outer-membrane receptor FpvA (Beare et al., 2003; Lamont et al., 2002; Visca et al., 2002).

Within the thickened CF mucus layer in the lung alveoli, iron availability is limited and oxygen is reduced, forming a hypoxic gradient that ranges from microaerobic [10 % environmental oxygen (EO)] to anaerobic conditions (Hassett et al., 2002; Worlitzsch et al., 2002). Traditionally described as an obligate aerobe, P. aeruginosa survives the reduced oxygen conditions within the CF mucus by utilizing nitrate within the airway surface fluid as an alternative electron acceptor (Hassett et al., 2002; Yoon et al., 2002). Sufficient levels of nitrate within the mucus layer allow for favourable growth of P. aeruginosa, which leads to chronic colonization (Hassett et al., 2002; Yoon et al., 2002). Such a unique microniche within the CF mucus is likely to affect the production of virulence factors by P. aeruginosa. Notably, local production of ETA and pyoverdine, and expression of corresponding genes has been documented in sputa of CF patients chronically infected with P. aeruginosa (Haas et al., 1991; Hunt et al., 2002; Jaffar-Bandjee et al., 1995).

We have shown that reduced oxygen levels obtained under static culture conditions considerably increase toxA transcription and ETA production in iron-deficient medium (Gaines et al., 2005). Preliminary analysis suggested that the influence of reduced oxygen on ETA production does not occur directly through regA or pvdS (Gaines et al., 2005). Therefore, the influence of reduced oxygen on ETA synthesis in P. aeruginosa may not occur through pvdS or regA. In this study, we examined the effect of reduced oxygen on the toxA regulatory circuit. We also examined how PvdS would affect the regulation of regA, toxA, and ptxR under different conditions of EO availability.

Bacterial strains, plasmids and growth media.
All strains and plasmids used in this study are listed in Table 1. For general growth experiments, including preparation of overnight cultures, plasmid DNA extraction and electroporation, P. aeruginosa or Escherichia coli strains were grown in Luria–Bertani (LB) broth (Miller, 1972). To examine gene expression, ETA synthesis and pyoverdine production, P. aeruginosa strains were grown in iron-depleted medium (TSB-DC). TSB-DC is a Chelex 100-treated trypticase soy broth dialysate containing 1 % (v/v) glycerol, 500 mM monosodium glutamate, 2.5 mM MgSO₄ and 0.5 mM CaCl₂ (Ohman et al., 1980). Iron-sufficient conditions were obtained by adding 25 µg ml^–1 FeCl₃ to TSB-DC. Antibiotics were added to the growth medium at the following concentrations: for P. aeruginosa – carbenicillin 300 µg ml^–1, gentamicin 60 µg ml^–1, streptomycin 300 µg ml^–1, tetracycline 50 µg ml^–1; and for E. coli – carbenicillin 100 µg ml^–1, kanamycin 50 µg ml^–1.

Table 1. Strains and plasmids used in this study r, resistant; Gm, gentamicin; Cb, carbenicillin; Km, kanamycin; Sm, streptomycin; Tc, tetracycline.

Growth conditions.
P. aeruginosa containing different plasmids was routinely grown overnight in LB broth with aeration at 37 °C. Cells were pelleted, washed, resuspended in fresh TSB-DC, and inoculated into TSB-DC as previously described (Gaines et al., 2005). Individual 125 ml flasks were incubated at 32 °C under aerobic conditions with shaking at 250 r.p.m. (referred to as A-sh, ∼21 % O₂ atmosphere) or, alternatively, under microaerobic static conditions (referred to as M-st, ∼10 % O₂ atmosphere) as previously described (Gaines et al., 2005). M-st conditions were achieved using GasPak Jars with Campy Pak Plus envelopes (Becton Dickinson). Throughout the 24 h growth cycle, individual flasks for each time point were removed from their incubation condition, and samples of the cultures were obtained for analysis. Each growth curve experiment was repeated at least three times.

DNA manipulations.
Routine genetic manipulations were conducted following standard protocols as described by Sambrook & Russell (2001). Plasmid DNA was extracted using the Wizard Plus Minipreps DNA Purification System (Promega). Plasmid DNA was introduced into PAO or PAO : : pvdS by electroporation (Smith & Iglewski, 1989). To construct the ptxR–P2 expression plasmid, a 430 bp fragment of the ptxR–ptxS intergenic region containing the putative ptxR P2 promoter (Vasil et al., 1998) together with the region encoding the first 55 aa of PtxR was obtained from pJAC24 by PCR. The blunt-ended PCR product was cloned into the SmaI site of pUC18, producing pJAC73. The orientation of the ptxR promoter in pJAC73 was determined by restriction digestion with BamHI. The 430 bp fragment was then obtained from pJAC73 by EcoRI/BamHI digestion and cloned in-frame with the lacZ gene in the previously described translational vector pSW205 (West et al., 1994). Construction of this recombinant plasmid (pJH2) was confirmed by sequencing. Plasmid pIN10 was constructed by cloning the 1.5 kb AvaI–PstI fragment from pDF18-202 (Frank et al., 1989) into the SalI–PstI sites of pUC18. The fragment carries intact regA from the P. aeruginosa strain PA103 and approximately 500 bp of the regA upstream region. In the resulting recombinant plasmid (pIN9), regA is constitutively expressed from the lac promoter. The 1.8 kb stability fragment that allows ColEI plasmids to replicate stably in P. aeruginosa was cloned into the PstI site of pIN9 generating pIN10.

Purification of PvdS.
PvdS was purified by affinity chromatography from E. coli M15 carrying the expression plasmid pPvdSF, in which the pvdS coding sequence is cloned in-frame with the C-terminus FLAG peptide (Leoni et al., 2000). E. coli M15/pPvdSF was grown overnight in LB broth supplemented with 0.4 % glucose (v/v). A 5 ml aliquot of the overnight culture was subcultured into 500 ml LB broth supplemented with 0.4 % glucose and grown at 37 °C for 2.5 h to approximately 0.5 OD₆₀₀. Cultures were then induced with 500 µM IPTG for 3 h at 37 °C. Cells were centrifuged and pellets stored at –80 °C.

Analysis of the different cell fractions indicated that the recombinant PvdS (PvdSF) was localized to inclusion bodies (data not shown). The insoluble fraction was lysed and purified using CelLytic B and the inclusion bodies solubilized using CelLytic IB according to the manufacturer's recommendations (Sigma-Aldrich). Proteins were refolded by dialysis in 6 M urea for 24 h at 4 °C with an additional 250 ml of 25 mM Tris/HCl (pH 7.4) added at 6, 12 and 18 h. PvdSF was purified using a 1 ml anti-FLAG M2 affinity gel chromatography column as instructed by the manufacturer (Sigma-Aldrich). Proteins in the eluates were estimated by the Bradford protein assay kit with BSA as the standard (Pierce Biotechnology). Proteins (10 µg) from the eluates were separated by SDS-PAGE, transferred to membrane and probed for PvdSF using anti-FLAG M2 antibody (Sigma-Aldrich). Eluates containing PvdSF were combined, concentrated and exchanged into DNA binding buffer for a final PvdSF concentration of 0.582 µg µl^–1.

SDS-PAGE and immunoblotting.
Samples from A-sh and M-st PAO cultures were collected at 4 h intervals, and total bacterial protein concentration was determined using the DC protein assay kit (Bio-Rad) with BSA as the standard. Comparable protein samples (40 µg of total bacterial proteins for PvdS detection or 10 µg for RpoD and RpoA detection) were separated in duplicate on SDS-PAGE gels prepared according to Laemmli (1970). After electrophoresis, gels were either stained with Coomassie brilliant blue to assess protein resolution, or electrotransferred to nitrocellulose membrane (Hybond C extra; Amersham) and probed for PvdS, RpoD and RpoA quantification using a mouse polyclonal anti-PvdS antiserum (Ambrosi et al., 2005) and commercial monoclonal anti-RpoD or anti-RpoA antibodies (Neoclone). Immune complexes were detected using secondary anti-mouse antibodies conjugated with either alkaline phosphatase (Promega) or horseradish peroxidase (Calbiochem). Membranes were developed with the 5-bromo-4-chloro-3-indoyl-phosphate or nitro blue tetrazolium chloride reagents for colorimetric determinations (Promega), or with the Amersham ECL chemiluminescent reagents (Amersham Biosciences), followed by exposure to X-ray film (Kodak).

Electrophoretic mobility shift assays (EMSA).
DNA fragments of 211, 251, 354 and 129 bp carrying the PvdS-specific IS box were obtained by PCR from the promoter regions of toxA, regA, ptxR and pvdEF, respectively. Oligonucleotide primers were designed using Primer Express 1.0 software (Applied Biosystems) (Table 2). PCR products were purified from agarose gels using the Qiaex II Gel Extraction Kit (Qiagen). Purified DNA fragments were end-labelled with [γ-³²P]ATP using T4 polynucleotide kinase (Sambrook & Russell, 2001). To determine optimum concentrations required to produce a specific PvdS–core RNA polymerase (RNAP)–DNA complex, several preliminary EMSA were performed using different concentrations of PvdS and E. coli RNAP with the pvdEF control probe. Where indicated, purified PvdSF (2.91 µg, equivalent to 130 pmol, in 5 µl) and core RNAP (0.70 µg, equivalent to 1.8 pmol, in 1 µl) (Epicentre) were preincubated for 30 min at 37 °C in DNA-binding buffer [10 mM Tris/HCl (pH 7.4), 1 mM EDTA, 10 mM KCl, 1 mM DTT/5 % glycerol with 50 µg ml^–1 BSA and 5 µg ml^–1 poly(d[I-C])] for a total reaction volume of 20 µl to reconstitute the holoenzyme (Dupuy & Matamouros, 2006; Sambrook & Russell, 2001). Radiolabelled probe (0.1 pmol, corresponding to 10⁵–10⁷ c.p.m., depending on the probe) was added to the mixture and incubated for 1 h at room temperature. Reactions were stopped with the addition of 5 µl of 0.8 µg ml^–1 heparin-loading dye (1 % bromophenol blue, 50 % glucose) and allowed to incubate for 5 min at room temperature. Reactions were separated by 5 % PAGE in 0.5x Tris-borate-EDTA buffer for 16 h at 4 °C. Gels were then dried and exposed to X-ray film.

Table 2. Primer sets utilized in this study and location of IS boxes in PvdS-regulated promoters

β-Galactosidase and pyoverdine assays and sandwich ELISA.
Assays for β-galactosidase were performed as previously described (Miller, 1972), and activity was expressed in units (U). Triplicate samples were obtained for each growth condition and time point throughout the growth cycle. The OD₆₀₀ of the culture was used to compensate for growth differences.

Pyoverdine levels were determined throughout the growth cycle for PAO grown in TSB-DC at 32 °C. Supernatant fractions collected every 4 h were appropriately diluted in TSB-DC at pH 7.3. Pyoverdine levels were quantified spectrophotometrically at A₄₀₅ according to Stintzi et al. (1996). Normalization of pyoverdine levels for growth was obtained by dividing A₄₀₅ by OD₆₀₀.

Sandwich ELISA for ETA quantification was done as previously described by Gaines et al. (2005). Values were standardized by dividing the amount of ETA (pg µl^–1) from each supernatant fraction by the OD₆₀₀ of the culture from which that fraction was obtained.

Reverse transcriptase PCR (RT-PCR).
Oligonucleotide primers for RT-PCR are listed in Table 2. PAO cultures were grown until early stationary phase in TSB-DC under A-sh and M-st conditions. Total RNA was extracted by the hot phenol/chloroform method as previously described (Carty et al., 2003). Residual DNA was removed from the RNA with RQI DNase I (1 U µl^–1) for 1.5 h at 37 °C in the presence of RNase inhibitor (RNasin; Promega) and the RNA was purified using the RNeasy Mini Kit (Qiagen). Reverse transcription was performed using 1 µg of RNA, 250 ng of random hexamers (Promega), 10 mM dNTPs, and 6 U of StrataScript reverse transcriptase (Stratagene). After a 2 h incubation at 42 °C, reactions were stopped by heating at 94 °C for 5 min. PCR was performed on 250 ng of cDNA using 3 µM concentration of each of the appropriate primers, 10 mM dNTPs and 1.25 U of GoTaq DNA Polymerase (Promega) per reaction. PCR to detect rpsL or pvdS messages was conducted as described by Sobel et al. (2003); an initiation cycle of 94 °C for 5 min, 20 cycles (rpsL) or 29 cycles (pvdS) including 30 s at 95 °C, 1 min at 60 °C, 1 min at 72 °C, and a final elongation cycle at 72 °C for 7 min. The amount of product was assessed on 1.5 % agarose gels and visualized with GelStar stain (Cambrex).

Statistical analysis.
Statistics were calculated using InStat (Graph Pad Software). ANOVA and the Student-t test were used to determine significant differences among the various conditions.

Purified PvdS binds to the promoter regions of toxA, regA, and ptxR genes of P. aeruginosa PAO
Previous studies have shown that the iron-starvation sigma factor PvdS complexed with core RNAP binds the IS box within the promoters of different pyoverdine genes (Leoni et al., 2000; Wilson et al., 2001). Potential IS boxes were also identified in the upstream regions of toxA, regA and ptxR (Table 2) (Hunt et al., 2002; Ochsner et al., 2002; Wilson et al., 2001). However, specific binding of the PvdS-RNAP holoenzyme to these upstream regions has not yet been demonstrated. It is possible that PvdS regulates the expression of these three genes at different levels, either indirectly (through a regulatory cascade) or directly (by binding to their upstream regions and promoting transcription initiation). Therefore, we examined the binding of the PvdS-RNAP to the toxA, regA and ptxR promoters. PvdSF was overexpressed in E. coli M15(pPvdSF) and purified by immunoaffinity chromatography using the FLAG-M2 antibody as previously described (Leoni et al., 2000). The specific interaction of the PvdSF-RNAP holoenzyme with the promoter regions of the PvdS regulated genes was examined by electrophoretic mobility shift assays (EMSA) using radiolabelled DNA fragments encompassing the promoter regions of toxA, regA and ptxR. The pvdEF promoter was used as the positive control for DNA binding by PvdSF-RNAP, as previously demonstrated by Wilson et al. (2001). The specificity of the holoenzyme binding to the DNA was enhanced by the addition of heparin, a polyanion that prevents non-specific binding of RNAP to DNA (Leoni et al., 2000; Pfeffer et al., 1977). Binding reactions were conducted as described in Methods and primarily consisted of E. coli core RNAP, purified PvdSF, probe and heparin. Titration of core RNAP with purified PvdSF was done to determine optimal binding conditions (data not shown). As expected, the PvdSF–RNAP holoenzyme produced two gel shift bands with the pvdEF probe, likely due to the presence of two divergently oriented IS boxes which could generate 1 : 1 and 1 : 2 DNA–protein complexes (Fig. 1a). DNA–protein complexes were detected with the PvdSF-RNAP and the toxA, regA and ptxR probes (Fig. 1b, c, d). Despite repeated experiments, the core RNAP also produced smaller sized complexes with the regA and ptxR probes (Fig. 1c, d), and minor bands were also observed with the pvdEF and toxA probes (Fig. 1a, b). However, additional unique gel shift bands were detected when any of the probes was incubated with the PvdSF–RNAP complex (Fig. 1). Leoni et al. (2000) previously indicated that the PvdS–RNAP holoenzyme bound the pvdA promoter with low efficiency. As shown in Fig. 1, we detected the same phenomenon in so far as most of the probe was detected in the free, unbound form. However, purified PvdSF alone did not specifically bind any of the probes, which indicates the requirement of PvdSF–RNAP holoenzyme formation for DNA recognition (Fig. 1). These results suggest that PvdS regulates, at least in part, the expression of toxA, and possibly regA and ptxR. However, additional experiments, including in vitro transcription of toxA, regA, and ptxR in the presence and absence of PvdS, will be required to confirm that the formation of these complexes enhances the transcription of these genes.

Table 2). The PvdSF–RNAP complex was incubated with the DNA fragments (18 : 1 molar ratio) that carry (a) the pvdEF intergenic region (positive control), (b) the toxA upstream region, (c) the regA upstream region and (d) the ptxR upstream region. EMSA were conducted in the presence of heparin (0.8 µg µl–1). Arrows indicate specific PvdS–RNAP–DNA complexes.

The increase in toxA expression by P. aeruginosa PAO under reduced EO conditions requires PvdS
Evidence based on the close correlation between pvdS, regA, ptxR and toxA expression indicates that PvdS regulates toxA through regA and ptxR (Ochsner et al., 1996; Vasil et al., 1998). Moreover, binding of the PvdS–RNAP complex to both the toxA and regA upstream regions (Fig. 1) suggests that PvdS regulates toxA expression both directly and indirectly, through regA and ptxR. However, we recently showed that under reduced EO, toxA expression differs from that of regA and pvdS (Gaines et al., 2005). Under microaerobic and anaerobic conditions, toxA expression in PAO increased while regA and pvdS expression decreased (Gaines et al., 2005). Thus, we addressed the following questions: (i) does toxA expression differ from regA and pvdS expression microaerobically, and (ii) if so, does it occur throughout the growth cycle? (iii) Does the increase in toxA expression under microaerobic conditions still depend on PvdS? (iv) If PvdS regulates toxA expression microaerobically, does this regulation occur through regA? (v) How does PvdS regulate ptxR expression throughout the growth cycle aerobically and microaerobically? To answer these questions, we examined the role of pvdS on the expression of toxA, regA, ptxR and the pyoverdine genes throughout the growth cycle and under two different levels of EO, A-sh and M-st, using a pvdS deletion mutant of PAO.

We examined toxA expression using the toxA–lacZ fusion plasmid pSW228 in P. aeruginosa PAO and its pvdS deletion mutant PAO : : pvdS (Table 1). In agreement with our previous report (Gaines et al., 2005), we confirmed that the level of toxA expression in PAO under M-st conditions was significantly higher (P<0.001) compared with A-sh (data not shown). Comparison of toxA expression between PAO and PAO : : pvdS showed that under A-sh and M-st conditions and between the 12- and 24 h time points, the level of toxA expression in PAO : : pvdS was significantly lower (P<0.001) than that in PAO (Fig. 2a, b). This indicates that, under A-sh and M-st conditions, maximal toxA expression in PAO requires functional PvdS. Despite the loss of PvdS, toxA expression in PAO : : pvdS during the 10–24 h time frame was slightly higher under M-st conditions in comparison with A-sh (Fig. 2c), suggesting that reduced EO causes a limited increase in toxA expression even in PAO : : pvdS. To further confirm these results, we measured the amount of ETA secreted by PAO and PAO : : pvdS under A-sh and M-st conditions. Cells were grown to mid-stationary phase, and the amount of ETA within the supernatant fraction was determined by sandwich ELISA as previously described (Gaines et al., 2005). Both PAO and PAO : : pvdS produced more ETA under M-st conditions than under A-sh conditions (data not shown). In addition, under both conditions, PAO : : pvdS produced significantly less ETA (P<0.001, M-st; P<0.01, A-sh) than the wild-type PAO (data not shown).

(17K): Fig. 2. Effect of reduced EO levels on toxA expression throughout the growth cycle of PAO and PAO : : pvdS. PAO and PAO : : pvdS carrying the toxA–lacZ fusion plasmid pSW228 were grown in TSB-DC medium under A-sh or M-st conditions as described in Methods. Six hours after inoculation, triplicate samples were obtained every 2 h, and the level of β-galactosidase activity was determined. (a) toxA expression in PAO and PAO : : pvdS under A-sh conditions; (b) toxA expression in PAO and PAO : : pvdS under M-st conditions; (c) toxA expression in PAO : : pvdS under both conditions. Filled symbols, PAO; open symbols, PAO : : pvdS; • and , A-sh conditions; and , M-st conditions. Values represent the means of three independent experiments (±SEM). *, P<0.001

Under microaerobic conditions, regA expression in P. aeruginosa PAO is PvdS-dependent
The transcriptional activator RegA is considered essential for toxA expression in P. aeruginosa (Storey et al., 1991; Vasil et al., 1989). Under A-sh conditions, PvdS regulates the expression of both toxA and regA (Ochsner et al., 1996). We compared the role of PvdS in regA expression under A-sh and M-st conditions using plasmid pRL88, which contains a regA–lacZ translational fusion (Storey et al., 1990). Opposite to toxA expression, between 8 and 20 h, regA expression in wild-type PAO under A-sh conditions was significantly higher (P<0.001) than that under M-st conditions (Fig. 3). Moreover, regA expression in PAO : : pvdS was abrogated under both A-sh and M-st conditions (Fig. 3). These results suggest that reduced EO negatively influences regA expression in PAO, and that PvdS is essential for regA expression regardless of the level of EO.

(15K): Fig. 3. Effect of reduced EO on regA expression in PAO and PAO : : pvdS. PAO and PAO : : pvdS carrying the regA–lacZ fusion plasmid pRL88 were grown in TSB-DC medium under A-sh or M-st conditions as described in Methods. Beginning 8 h post-inoculation, triplicate samples were obtained every 4 h, and the level of β-galactosidase activity was determined. Filled symbols, PAO; open symbols, PAO : : pvdS; • and , A-sh conditions; and , M-st conditions. Values represent the means of three independent experiments (±SEM).

We also explored the role of regA in the regulation of ETA production by PvdS. We examined the effect of plasmid pIN10, in which regA is constitutively expressed from the lac promoter, on ETA production by PAO : : pvdS grown in iron-deficient (TSB-DC) and -sufficient (TSB-DC with 25 µg ml^–1 FeCl₃) media under A-sh and M-st conditions. Under A-sh conditions, pIN10 significantly enhanced ETA production in PAO : : pvdS and deregulated it to a great extent with respect to iron (data not shown). Under A-sh conditions, the amount of ETA produced by PAO : : pvdS(pIN10) was about 25 pg µl^–1 OD₆₀₀^–1 in iron-deficient medium but 20 pg µl^–1 OD₆₀₀^–1 in iron-sufficient medium (data not shown). Detection of the limited repressive effect of iron in the absence of functional PvdS, and upon regA overexpression, suggests that this effect of iron is independent of PvdS and RegA. We obtained similar results under M-st conditions (data not shown). These results suggest that PvdS regulates toxA expression through regA both aerobically and microaerobically. To confirm these results further, we used sandwich ELISA to examine the effect of pPVD31 (in which pvdS is expressed from the tac promoter) (Ochsner et al., 1996) on ETA production in the PAO regA mutant 6424 : : regA–lacZ (Bailey & Manoil, 2002; Jacobs et al., 2003) that was obtained from the University of Washington Genome Center, . In iron-deficient medium, strain 6424 : : regA–lacZ(pPVD31) produced no detectable ETA under A-sh and M-st conditions (data not shown). Ochsner et al. (1996) previously described similar results using semiquantitative immunoblot analysis of ETA production. Despite constitutive expression of regA, ETA is still slightly repressed by iron.

The growth of P. aeruginosa PAO under M-st conditions reduces pyoverdine production and the expression of pyoverdine genes
Previous studies showed that, under A-sh conditions, pyoverdine production by P. aeruginosa is stringently controlled by PvdS (Cunliffe et al., 1995; Hunt et al., 2002; Ochsner et al., 1995). Therefore, we examined whether reduced EO alters pyoverdine production and if PvdS stringently controls pyoverdine production under these conditions. P. aeruginosa PAO was grown in TSB-DC under A-sh and M-st conditions, and samples were taken every 4 h to determine the level of pyoverdine in the supernatant fractions as previously described (Stintzi et al., 1996). Between the 8 and 24 h time points, wild-type PAO produced significantly (P<0.001) lower levels of pyoverdine under M-st conditions than under A-sh conditions (Fig. 4a). As previously demonstrated (Leoni et al., 1996; Ochsner et al., 1995; Visca et al., 1992), PAO : : pvdS produced no pyoverdine under either condition, and pyoverdine production by wild-type PAO was fully repressed by iron (data not shown).

(17K): Fig. 4. (a) Effect of reduced EO on pyoverdine production by PAO in iron-deficient medium. Beginning 4 h post-inoculation, samples were obtained every 4 h, and the supernatant fractions were separated. Pyoverdine levels were determined by measuring the A405 of the supernatant fraction. Values were standardized by dividing the amount of pyoverdine in each fraction by the OD600 of the culture from which the fraction was obtained. *, P<0.001. (b) Effect of reduced EO on pvdD expression throughout the growth cycle of PAO and PAO : : pvdS. PAO or PAO : : pvdS carrying the pvdD–lacZ fusion plasmid ppvdD were grown in TSB-DC medium at 32 °C under A-sh or M-st conditions as described in Methods. Triplicate samples were obtained every 4 h beginning 8 h post-inoculation, and the level of β-galactosidase activity was determined. Filled symbols, PAO; open symbols, PAO : : pvdS; • and , A-sh conditions; and , M-st conditions. Values represent the means of three independent experiments (±SEM).

To examine whether the reduction in pyoverdine production is the result of a decrease in the expression of the pyoverdine synthesis genes, we analysed in detail the expression of one of these genes, namely pvdD, using the pvdD–lacZ transcriptional fusion plasmid ppvdD (Rombel et al., 1995). Between the 8 and 20 h time points, pvdD expression in PAO under M-st conditions was significantly (P<0.001) reduced in comparison with A-sh conditions (Fig. 4b). Expression of pvdD was fully repressed in PAO : : pvdS, irrespective of EO (Fig. 4b). Using a pvdE–lacZ fusion plasmid, we also observed that pvdE expression in PAO was significantly lower under M-st conditions than under A-sh conditions (data not shown). These results suggest that reduced EO affects the expression of regA, pvdD and pvdE by a mechanism(s) different from the one governing toxA expression.

PvdS influences the expression from the ptxR–P2 promoter under A-sh but not M-st conditions
The LysR transcriptional activator PtxR modulates ETA production in P. aeruginosa by increasing toxA expression by four- to fivefold (Hamood et al., 1996). We have previously demonstrated that, similar to toxA expression, ptxR expression increases under reduced EO (Gaines et al., 2005). Thus, we examined the role of PvdS in ptxR expression under A-sh and M-st conditions using the previously described ptxR–lacZ fusion plasmid pJAC24, which carries both P1 and P2 ptxR promoters (Colmer & Hamood, 1999). Under A-sh conditions, the pattern of ptxR expression in PAO : : pvdS was essentially similar to that in wild-type PAO, reaching a peak at 6 h and gradually declining thereafter (data not shown). There were no differences in ptxR expression between wild-type and the PAO : : pvdS mutant under M-st conditions (data not shown). However, under M-st conditions, ptxR expression in PAO : : pvdS was significantly higher than that under A-sh conditions at 10 to 24 h time points (data not shown).

Available evidence suggests that ptxR is expressed from two separate promoters, P1 and P2 (Vasil et al., 1998). Based on the results of RNase protection experiments, we previously suggested that ptxR expression from P2 is regulated by PvdS under microaerobic conditions (Vasil et al., 1998). Therefore, we determined the effect of both reduced EO and PvdS on the expression of ptxR–P2 throughout the growth cycle of PAO. These experiments were conducted using the ptxR–P2–lacZ fusion plasmid pJH2, which carries only the P2 promoter (Table 1). Under M-st conditions and in both strains, ptxR–P2 expression was significantly higher than that under A-sh conditions at several time points (Fig. 5a, b). Throughout the growth cycle and under A-sh and M-st conditions, differences in ptxR–P2 expression between PAO and PAO : : pvdS were limited and inconsistent (Fig. 5c, d). These results suggest that unlike its effect on toxA and regA expression, PvdS slightly affects ptxR expression. In addition, the enhancement in ptxR expression by reduced EO is unlikely to occur through PvdS.

(25K): Fig. 5. Effect of reduced EO on ptxR–P2 expression throughout the growth cycle of PAO and PAO : : pvdS. PAO and PAO : : pvdS carrying the ptxR–P2–lacZ fusion plasmid pJH2, which contains only the putative P2 promoter, were grown in TSB-DC under A-sh or M-st conditions as described in Methods. Starting 4 h post-inoculation triplicate samples were obtained every 2 h, and the level of β-galactosidase activity was determined. (a) ptxR–P2 expression in PAO under A-sh and M-st conditions; (b) ptxR–P2 expression in PAO : : pvdS under both conditions; (c) ptxR–P2 expression in both strains under A-sh conditions; (d) ptxR–P2 expression in both strains under M-st conditions. Filled symbols, PAO; open symbols, PAO : : pvdS; • and , A-sh conditions; and , M-st conditions. Values represent the means of three independent experiments (±SEM). *, P<0.001; , P<0.05

Reduced EO represses PvdS expression throughout the growth cycle of P. aeruginosa PAO
The above results suggest that PvdS regulates the expression of toxA, regA and the pyoverdine genes under both A-sh and M-st conditions, and ptxR expression under A-sh conditions only (Figs 2–5). We have previously shown that anaerobic conditions reduce pvdS expression in PAO (Gaines et al., 2005). Thus, to gain further insight into the role of PvdS in such regulation, the expression profile of pvdS was examined throughout the growth cycle of P. aeruginosa PAO under A-sh and M-st conditions. Gene fusions and RT-PCR assay were used to monitor pvdS transcription. As shown in Fig. 6(a), and similar to regA, pvdD and pvdE expression, the level of pvdS expression in PAO under A-sh conditions was four- to sixfold higher at the 10 to 16 h time points than that under M-st conditions. Decreased pvdS transcription under M-st conditions was further confirmed by RT-PCR. The RT-PCR analysis was carried out on total RNA extracted from early stationary phase (10 h) P. aeruginosa PAO cultures. As a control, the level of the housekeeping gene rpsL that encodes the ribosomal subunit S12 was determined (Sobel et al., 2003). As shown in Fig. 6(b), the amount of pvdS transcript produced under A-sh conditions was markedly (∼ sixfold) higher than that produced under M-st conditions, in substantial agreement with the results of the pvdS–lacZ transcriptional fusion analysis. We also monitored the intracellular levels of PvdS throughout the growth cycle of PAO by Western blot analysis under both A-sh and M-st conditions. As controls, the levels of the vegetative σ⁷⁰ factor RpoD and the α subunit (RpoA) of the RNAP were also determined (Fujita et al., 1993). As shown in Fig. 6(c), PvdS production correlates well with the transcriptional profile. The level of PvdS under M-st conditions was markedly lower than that under A-sh conditions (Fig. 6c). Levels of RpoD and RpoA were essentially the same under both conditions and throughout the growth cycle (Fig. 6c). These results suggest that lower EO reduces pvdS transcription and the intracellular levels of PvdS in PAO. Consequently, the expression of toxA, regA and the pyoverdine genes (that are controlled by the PvdS-dependent RNAP) is reduced. The failure of reduced EO to affect the levels of the two main components of the vegetative transcriptional machinery, RpoA and RpoD, indicates that low EO specifically reduces PvdS expression and hence the level of PvdS-dependent RNAP.

(32K): Fig. 6. Effect of reduced EO on PvdS expression. (a) PAO carrying the pvdS–lacZ fusion plasmid, pMP220 : : PpvdS, was grown under A-sh (solid line) or M-st (broken line) conditions as described in Methods. Triplicate samples were obtained at 6, 10, 12 and 16 h, and the level of β-galactosidase activity was determined. Values represent the mean of three independent experiments (±SEM). *, P<0.001. (b) Total RNA extracted from early stationary phase PAO cultures under A-sh and M-st conditions was subjected to RT-PCR analysis. The PCR products were separated on a 1.5 % agarose gel and stained with GelStar Stain (Cambrex) for visualization. Transcription of the constitutively expressed rpsL encoding ribosomal subunit S12 was used as a control. (c) Whole-cell lysates (40 µg for PvdS or 10 µg for both RpoA and RpoD) obtained at indicated time points from A-sh (A) and M-st (M) cultures were probed with anti-PvdS polyclonal mouse antiserum or with monoclonal anti-RpoD and anti-RpoA antibodies. Whole-cell extract (40 or 10 µg) from PAO : : pvdS under A-sh condition was used as a negative control. Purified PvdS (10 ng) was used as a positive control. Immunoblots are representative of one of three independent experiments.

In certain P. aeruginosa infection sites, the environmental conditions are likely to be microaerobic (Allen et al., 1997; Worlitzsch et al., 2002; Xiong et al., 2000). In particular, within the mucus layer of the CF lung, EO forms a hypoxic gradient (microaerobic to anaerobic) (Hassett et al., 2002), which could influence the expression of extracellular virulence factors by P. aeruginosa. We have previously reported that reduced EO levels significantly increase toxA expression in PAO, while reduced EO decreased rather than increased pvdS expression (Gaines et al., 2005; this study). Therefore, in this study we have addressed the impact of PvdS and other toxA regulators on the expression of toxA under microaerobic (M-st) conditions. We selected aerobic (A-sh) conditions as a control since most laboratory studies have examined the interaction between PvdS and PvdS-regulated genes under aerobic conditions (Cunliffe et al., 1995; Hunt et al., 2002; Leoni et al., 2000; Ochsner et al., 1996; Wilderman et al., 2001; Wilson et al., 2001). Based on the results of our analyses, we propose the following models to explain the regulation of different genes under A-sh and M-st conditions.

Regulation of toxA, regA, ptxR and the pyoverdine genes by PvdS under A-sh conditions
Under aerobic conditions and in iron-deficient medium, toxA, regA, pvdS and the pyoverdine genes are highly expressed (Figs 2–4, 6). In addition, the expression of regA, toxA and the pyoverdine genes requires functional PvdS (Figs 2–4, 7a). With respect to the pyoverdine genes, previous studies showed that PvdS directly regulates the expression of at least pvdD and pvdF (Ochsner et al., 2002; Visca, 2004; Wilson et al., 2001). With respect to toxA, our results strongly suggest that PvdS regulates its expression through regA (data not shown; Fig. 7a). The expression of regA from the exogenous lac promoter complemented the defect of PAO : : pvdS in ETA production (data not shown). Similar to its effect on the pyoverdine genes, PvdS regulates regA expression by binding directly to the regA upstream region (Fig. 1). More specifically, PvdS may regulate the expression from the regA P2 promoter. Computer analysis showed that the potential PvdS binding site within the regA upstream region is located –1 bp 5' to +8 bp 3' of the T2 transcription start site (Hunt et al., 2002). Whether PvdS binding at this site affects regA transcription is not known. However, if further analysis confirms that PvdS binds to this sequence, it would be consistent with the notion that regA expression at mid log and early stationary phases of growth occurs through the iron-responsive regA P2 promoter (Frank & Iglewski, 1988; Storey et al., 1990). The iron-insensitive regA P1 promoter influences the expression at earlier stages of growth (Frank & Iglewski, 1988; Storey et al., 1990). Previous studies indicated that iron-Fur represses regA and toxA expression through PvdS (Ochsner et al., 1996; Visca, 2004). Thus, it would be logical if PvdS regulates the iron-responsive regA P2 promoter (Fig. 7a). What is not clear at this time is the relevance of the PvdS–RNAP binding to the toxA upstream region (Fig. 1). Besides the physical interaction of the PvdS–RNAP with the toxA upstream region, the toxA upstream region carries a potential PvdS binding site that contains more conserved nucleotides than the PvdS binding site within the regA upstream region (Hunt et al., 2002). However, we have excluded the possibility that an increased level of PvdS bypasses regA and enhances toxA expression directly by binding to the toxA upstream region. As Ochsner et al. (1996) previously reported in PAOΔregA, and we could confirm in this study for the 6424 : : regA–lacZ mutant, constitutive pvdS expression from the tac promoter does not complement the defect in toxA expression in a regA-defective background (data not shown).

(19K): Fig. 7. Schematic diagram depicting the regulation of toxA, regA, ptxR, and the pyoverdine genes by PvdS in low iron medium under A-sh (a) and M-st (b) conditions. (a) Heavy solid lines indicate strong activation by increased levels of PvdS; thin solid line indicates mild activation by increased levels of PvdS. (b) Broken lines indicate limited activation by reduced levels of PvdS. ? indicates a potential regulator that is induced by reduced EO and enhances toxA and ptxR expression independently of PvdS.

With respect to ptxR, under A-sh conditions PvdS has a limited effect on ptxR expression (data not shown; Figs 5, 7a). Under A-sh conditions and in comparison with PAO, the level of ptxR expression from the P1/P2 promoter (pJAC24) in PAO : : pvdS was significantly lower at the 8 h time point only (data not shown). In addition, throughout the growth cycle, and under A-sh conditions differences in ptxR expression from the P2 promoter (pJH2) between the two strains were inconsistent (Fig. 5c). Although statistically insignificant, ptxR expression from either P1/P2 or P2 was lower in PAO : : pvdS than in PAO between the 16 and 24 h time points (Fig. 5c). Whether this limited effect is due to PvdS binding is not known at this time. The ptxR upstream region contains a well conserved IS box located –153 to –130 bp 5' of the T2 transcription start site (Table 2), and a second putative box located –25 to –2 bp 5' of it (Colmer-Hamood et al., 2006). Both sites are included within the 354 bp fragment to which PvdS–RNAP binds (Table 2; Fig. 1d). We ruled out the possibility that PvdS regulates the expression of the ptxR P1 promoter. Using a ptxR–P1–lacZ fusion plasmid, we could confirm no difference in ptxR expression between wild-type PAO and PAO : : pvdS (data not shown).

Regulation of toxA, regA, ptxR and the pyoverdine genes by PvdS under M-st conditions
Under these conditions, the expression of regA, pvdD, pvdE and pvdS was significantly reduced (Figs 3, 4, 6a; data not shown). The abolition of regA, pvdD and pvdE expression indicates that these genes are strictly controlled by PvdS under M-st as well as A-sh condition (Figs 3, 4, 7b). As discussed above and similar to the A-sh conditions, this control is likely to be accomplished through direct binding of the PvdS–RNAP complex to the upstream region of each gene (Fig. 1). Previous studies suggested that microaerobic conditions reduce PvdS production in P. aeruginosa PAO (Ochsner et al., 1996). Using RNase protection experiments, Ochsner et al. (1996) previously showed that, under microaerobic conditions, pvdS transcription in PAO is reduced by about 10-fold (in comparison with aerobic conditions). We confirmed that the effect of reduced EO occurs at the level of pvdS transcription. Compared to A-sh conditions, pvdS expression under M-st conditions was reduced by about sixfold throughout the growth cycle of PAO (Fig. 6a). Also, PvdS expression from the exogenous tac promoter eliminated the effect of reduced EO on pyoverdine production. In fact, PAO(pPVD31) produced comparable levels of pyoverdine under A-sh and M-st conditions in the presence and absence of iron (data not shown). We also showed that the effect is specific and independent of the vegetative transcriptional machinery. As shown in Fig. 6(c), while the relative concentration of PvdS was reduced by low EO levels, the relative concentration of RpoD and RpoA was not affected. However, the mechanism(s) by which low EO reduces the expression of pvdS and other genes under M-st conditions is still not known. Certain factors that regulate PvdS production in P. aeruginosa including iron-Fur, the anti sigma factor FpvR and the quorum sensing regulator VqsR have been identified (Beare et al., 2003; Juhas et al., 2004; Lamont et al., 2002; Visca, 2004). Since we examined pvdS expression in iron-deficient medium only, iron-Fur is less likely to be involved in the reduced pvdS expression under M-st conditions. Available evidence suggests that FpvR regulates pvdS expression post-transcriptionally (Beare et al., 2003; Visca, 2004). Our analysis suggests that reduced EO affects primarily pvdS transcription, and consequently PvdS synthesis (Fig. 6). However, a role for FpvR in PvdS synthesis cannot be dismissed. Reduced EO may regulate PvdS expression through vqsR. Juhas et al. (2004) previously showed that a vqsR deletion in PAO reduced the transcription of more than 20 iron-regulated genes including pvdS, as well as the pyoverdine and pyochelin synthesis genes. We do not know at this time if the growth of PAO under M-st conditions reduces vqsR transcription.

Unlike regA, pvdD, pvdE and pvdS, the expression of toxA under M-st conditions is increased (Fig. 2). This increase is less likely to occur through pvdS whose expression is significantly reduced under M-st conditions (Figs 6a, 7b). Indeed, the increase in toxA expression under M-st conditions was detected even in the absence of functional PvdS. Between the 10 and 22 h time points, toxA expression in PAO : : pvdS under M-st conditions ranged from 1.5 to 3.7-fold higher than that under A-sh conditions (Fig. 2c), a difference comparable to that seen between the 16 and 24 h time points in wild-type PAO (direct comparison not shown). In comparison with PAO, toxA expression in PAO : : pvdS was significantly reduced under both A-sh and M-st conditions (Fig. 2a, b). Again, whether PvdS binding to the toxA upstream region regulates toxA expression under these conditions is still unknown. Similar to the observation under A-sh conditions, PvdS still regulates toxA expression through regA under M-st conditions. As for A-sh conditions, constitutive PvdS expression from the exogenous tac promoter (in pPVD31) failed to complement the defect of the 6424 : : regA–lacZ mutant in ETA production under M-st conditions (data not shown). Thus, under M-st conditions, toxA expression in P. aeruginosa may be increased by a regA- and pvdS-independent mechanism (Fig. 7b). However, both regA and pvdS are required for maximum levels of ETA production by PAO (Fig. 7b). Although we did not examine the effect of iron on toxA expression in the present study, results of our previous study suggest that iron does not influence the increase in toxA expression under M-st conditions (Gaines et al., 2005). In iron-sufficient medium, toxA expression in PAO was significantly reduced under both A-sh and M-st conditions (Gaines et al., 2005). Despite this reduction, however, the level of toxA expression under M-st conditions was relatively higher than that under A-sh conditions (Gaines et al., 2005).

Besides RegA and PvdS, other P. aeruginosa regulators that control toxA expression (directly or indirectly) are Vfr and LasR (Gambello et al., 1993; Hamood et al., 1996; Suh et al., 2002; West et al., 1994). In comparison with PAO, the level of toxA expression in PAOΔvfr was significantly reduced (West et al., 1994). In addition, Vfr specifically binds to a Vfr consensus sequence within the toxA upstream region, suggesting that Vfr may regulate toxA expression directly (Kanack et al., 2006). However, under M-st condition, the level of toxA expression in PAOΔvfr was higher than that under A-sh conditions (i.e. vfr deletion did not interfere with the increase in toxA expression). Under A-sh conditions and at the 16 h time point, the level of β-galactosidase activity produced by PAO/pSW228 was 1377.5±88.68, while that of PAOΔvfr/pSW228 was 352.5±30.17 (P=0.0023). Under M-st conditions, PAO/pSW228 produced 3014.0±79.30 units of β-galactosidase activity with PAOΔvfr/pSW228 producing 729.0±23.24 (P=0.0002). Thus, similar to PvdS, Vfr is essential for efficient toxA expression, but is not the factor through which the increase in toxA expression under M-st conditions occurs. Using PAO and its lasR deletion mutant, we also ruled out a role for LasR in the observed increase in toxA expression under microaerobic conditions. Under A-sh conditions, and in comparison with PAO/pSW228, the level of β-galactosidase activity produced by PAOΔlasR/pSW228 was about 2.6-fold less (1060.5±60.80 vs 411.5±12.20; P=0.0025). Under M-st conditions, the level of β-galactosidase activity produced by both strains was increased by two- to threefold compared to A-sh conditions (2295.5±107.16 for PAO/pSW228 and 1315.8±66.78 for PAOΔlasR/pSW228).

It is possible that toxA and ptxR expression is induced through the denitrification (anaerobic respiration) process in which nitrate functions as the terminal electron acceptor. We previously demonstrated that the most significant increase in toxA and ptxR expression occurs when PAO is grown anaerobically in the presence of nitrate (Gaines et al., 2005). Schreiber et al. (2007) recently showed that the denitrification process in P. aeruginosa involves a regulatory network that includes the oxygen regulator Anr, the nitric oxide regulator Dnr and the nitrate responsive two-component regulatory systems NarXL. Whether any of these regulators enhance toxA and/or ptxR expression anaerobically or microaerobically in the presence of nitrate is yet to be determined.

Similar to toxA expression, ptxR expression in PAO was increased under M-st conditions. This increase occurred whether ptxR was expressed from the P1/P2 promoters or the P2 promoter alone (data not shown; Fig. 5d) (Gaines et al., 2005). In addition, the increase was detected in both PAO and PAO : : pvdS, thus appearing independent of PvdS (data not shown; Figs 5d, 7b). We hypothesize that reduced EO increases toxA and ptxR expression through a single mechanism or as a general effect on the expression of several genes (Fig. 7b). Due to the nature of the PAOΔptxR mutation, we cannot confirm or deny a role for ptxR in regulating ETA production under reduced EO. We previously reported that upon its initial characterization PAOΔptxR produced a significantly lower level of ETA than PAO (Hamood et al., 1996). However, upon continuous subculturing, PAOΔptxR produced levels of ETA comparable to that produced by PAO (Hamood et al., 1996). This phenomenon could be due to either a mutation in another gene that suppresses the ptxR mutation or to a compensation of PtxR function by another LysR protein. This effect is specific to ETA as it differs from the effect of ptxR on the production of other virulence factors such as pyocyanin (Carty et al., 2006). Besides the genes described in this study, other P. aeruginosa genes whose expression varies under reduced EO in iron-deficient medium are not known. Previous studies that described the effect of reduced EO on the expression of P. aeruginosa genes were conducted in iron-sufficient media (Bragonzi et al., 2005; Eschbach et al., 2004). In addition, Ochsner et al. (2002) described a transcriptome analysis of PAO and PAO : : pvdS in iron-deficient medium under A-sh conditions. Furthermore, Palma et al. (2003) compared PAO global transcription in iron-deficient and iron-sufficient media under A-sh conditions. Therefore, a comprehensive transcriptome analysis of PAO and PAO : : pvdS in iron deficient medium under M-st conditions is essential to understand the relationship between pvdS, regA, toxA and ptxR.

In conclusion, our results show that reduced levels of EO enhance toxA, and to a lesser extent ptxR expression, but reduce the expression of regA, pvdD, pvdE and pvdS (Fig. 7). The enhancement in toxA and ptxR expression does not appear to occur through PvdS. In addition, our results show that the PvdS–RNAP complex binds to the ptxR, regA and toxA upstream regions. While the present evidence supports a direct interaction between PvdS and the promoters of regA, ptxR and the pyoverdine genes, the effect of this interaction on the expression of these genes is yet to be determined. In addition, the environmental conditions under which PvdS interacts directly with the toxA promoter are still unknown.

We thank Iain Lamont and Michael Vasil for the Pseudomonas aeruginosa strains and plasmids. This work was supported by grant AI-33386 to A. N. H. from the National Institutes of Health. J. M. G. was supported in part by a fellowship from the Cystic Fibrosis Foundation. P. V. was supported by grants from the Ministry for Health of Italy (Ricerca Corrente 2006 to National Institute for Infectious Diseases L. Spallanzani), Ministry of University and Research of Italy (PRIN-2006) and Fondazione per la Ricerca sulla Fibrosi Cistica (Progetti-2007).

Edited by: P. Cornelis

Footnotes

†Present address: Department of Microbiology and Immunology, East Carolina University, Greenville, NC 27858, USA.