GbdR Regulates Pseudomonas aeruginosa plcH and pchP Transcription in Response to Choline Catabolites

Abstract

Pseudomonas aeruginosa hemolytic phospholipase C, PlcH, can degrade phosphatidylcholine (PC) and sphingomyelin in eukaryotic cell membranes and extracellular PC in lung surfactant. Numerous studies implicate PlcH in P. aeruginosa virulence. The phosphorylcholine released by PlcH activity on phospholipids is hydrolyzed by a periplasmic phosphorylcholine phosphatase, PchP. Both plcH gene expression and PchP enzyme activity are positively regulated by phosphorylcholine degradation products, including glycine betaine. Here we report that the induction of plcH and pchP transcription by glycine betaine is mediated by GbdR, an AraC family transcription factor. Mutants that lack gbdR are unable to induce plcH and pchP in media containing glycine betaine or choline and in phosphatidylcholine-rich environments, such as lung surfactant or mouse lung lavage fluid. In T broth containing choline, the gbdR mutant exhibited a 95% reduction in PlcH activity. In electrophoretic mobility shift assays, a GbdR-maltose binding protein fusion bound specifically to both the plcH and pchP promoters. Promoter mapping, alignment of GbdR-regulated promoter sequences, and analysis of targeted promoter mutants that lack GbdR-dependent induction of transcription were used to identify a region necessary for GbdR-dependent transcriptional activation. GbdR also plays a significant role in plcH and pchP regulation within the mouse lung. Our studies suggest that GbdR is the primary regulator of plcH and pchP expression in PC-rich environments, such as the lung, and that pchP and other genes involved in phosphorylcholine catabolism are necessary to stimulate the GbdR-mediated positive feedback induction of plcH.

The Pseudomonas aeruginosa plcH gene encodes a hemolytic phospholipase C (PLC) with high specificity for phosphatidylcholine (PC) and sphingomyelin (SM) (4). Studies have shown PlcH to have both pathogenic activities and proinflammatory capability (3, 21, 45). In addition, PlcH is important for P. aeruginosa virulence in mice (29), rabbits (46), insects (12), plants (28, 29), and Candida albicans (10). Purified PlcH is capable of inducing pulmonary inflammation and inhibits the oxidative burst of neutrophils (40, 45). Evidence also suggests that P. aeruginosa PlcH-mediated degradation of the PC in lung surfactant contributes to lung infections (16).

The enzymatic activity of PlcH on either PC or SM releases phosphorylcholine (ChoP), which can be dephosphorylated in the periplasm by a phosphatase, PchP, to yield choline (20). Choline is actively transported into the cytoplasm (19, 33), where it can be used in a variety of different ways. P. aeruginosa BetAB catalyzes the conversion of choline to glycine betaine (GB), which is both a potent osmoprotectant (9) and a source of carbon, nitrogen, and energy (13). During catabolism of GB, it is successively demethylated to form dimethylglycine (DMG), sarcosine (monomethylglycine), and finally glycine (7, 43). Microarray studies of cystic fibrosis (CF) patient sputum samples showed that in the lungs of CF patients, P. aeruginosa induces plcH and genes involved in both choline and diacylglycerol catabolism, which supports the hypothesis that PlcH is important for P. aeruginosa colonization of the lung (39).

PlcH activity provides the P. aeruginosa cell with ChoP, a source of both phosphate and choline, and both of these products participate in the regulation of plcH transcription (36). Shortridge et al. (36) have shown that plcH transcription is induced by phosphate limitation in a PhoB-dependent manner. The plcH gene is also transcriptionally induced in response to choline and its downstream catabolic products, GB and DMG (32, 36). Induction of plcH by GB and DMG is PhoB independent, and there is evidence for distinct transcriptional start sites for phosphate limitation-induced and choline-induced transcripts (32, 36, 41). The inability of choline to induce plcH in a mutant defective for betB, required for conversion of choline to GB, provided evidence that choline is not a direct activator of plcH transcription, but rather that GB and DMG are the inducers (31). The transcription factor responsible for this induction has not previously been described.

Here we show that GbdR, an AraC family transcription factor that is required for the induction of the GB and DMG catabolic genes in response to GB and DMG (43), also regulates the expression of plcH and pchP in a similar manner. Through deletion mapping, electrophoretic mobility shift assays, and targeted promoter mutagenesis, we have identified residues that are required for GB- and DMG-dependent induction of plcH and pchP. We show that GbdR is critical for induction of plcH and pchP in response to bovine surfactant and mouse bronchoalveolar lavage fluid (BALF). Analysis of P. aeruginosa transcripts during acute murine lung infection also indicated that GbdR is important for the induction of plcH and pchP in vivo.

MATERIALS AND METHODS

Strains, media, and growth conditions.

P. aeruginosa PAO1 and PA14 wild-type (WT) strains and deletion strains, as well as Escherichia coli strains (Table 1), were maintained on Luria broth (LB) medium. For transcriptional induction studies, cells were grown overnight in morpholinepropanesulfonic acid (MOPS) minimal medium (24) amended with 25 mM sodium pyruvate and 5 mM d-glucose prior to transfer to inducing medium. When necessary, gentamicin was added to a final concentration of 10 μg/ml for E. coli, 50 μg/ml for P. aeruginosa in LB medium, and 25 μg/ml for P. aeruginosa in MOPS medium.

TABLE 1.

Strain and plasmid list

Strain or plasmid	Strain no.	Description
Strains
P. aeruginosa strains
PAO1	DH20	P. aeruginosa WT
ΔgbdR	DH543	In-frame PA5380 deletion in PAO1 (43)
ΔgbdR att::gbdR	DH1008	Complementation of the gbdR deletion at the att site
ΔgbcAB	DH841	In-frame PA5410 and PA5411 deletions in PAO1 (43)
ΔdgcA	DH1178	In-frame PA5398 deletion in PAO1 (43)
ΔplcHR	DH860	In-frame plcHR deletion in PAO1 (36)
pchP::Tn5	DH503	PAO1 transposon mutant 10802
betB::Tn5	DH491	PAO1 transposon mutant 104
PA14	DH122	P. aeruginosa WT (31)
ΔgbdR	DH466	In-frame PA5380 deletion in PA14 (43)
E. coli strains
S17/Δpir	DH522
Ec-PA5380KO	DH540	DH522 with pPA5380KO; Gmr
Plasmids
pUCP22		High-copy-number Pseudomonas stabilized vector; Gmr (34)
pMQ30		Suicide vector; Gmr (35)
pMQ80		High-copy-number Pseudomonas-yeast shuttle vector (35)
pMW5		pUCP22 containing lacZYA from pRS415
pMW22		plcH promoter fragment (positions −374 to −13) in pMW5
pMW23		plcH promoter fragment (positions −248 to −13) in pMW5
pMW24		plcH promoter fragment (positions −177 to −13) in pMW5
pMW25		plcH promoter fragment (positions −86 to −13) in pMW5
pMW26		MBP-GbdR expression vector (pMalc derivative)
pMW47		GFP reporter vector based on modified pMQ80
pMW71		pchP promoter fragment (positions −369 to −15) in pMW47
pMW72		pchP promoter fragment (positions −275 to −15) in pMW47
pMW73		pchP promoter fragment (positions −141 to −15) in pMW47
pMW84		pchP promoter fragment (positions −75 to −15) in pMW47

PLC activity assays.

PLC activity was measured using the artificial substrate ρ-nitrophenyl-phosphorylcholine (NPPC) by the method described by Kurioka and Matsuda (15). Bacteria were grown overnight in 5 ml of tryptone broth or tryptone broth plus 5 mM choline overnight at 37°C. The reaction buffer was 100 mM Tris-HCl (pH 7.2), 25% glycerol, and 20 mM NPPC. NPPC hydrolysis was detected by measuring the absorbance at 410 nm. Hemolytic activity was analyzed on agar plates containing 5% defibrinated sheep blood. To eliminate plcH induction in response to phosphate limitation, 500 μl of 1 M potassium phosphate (pH 7.0) solution was top spread and the plates were allowed to dry prior to inoculation with P. aeruginosa strains. After 24 h of incubation at 37°C, hemolytic activity was detected by clear halos, which were photographed after removal of the bacteria from the plate by scraping with a coverslip.

RNA isolation, RT, and qRT-PCR.

For in vitro gene induction experiments using reverse transcription-PCR (RT-PCR), cells were grown overnight in MOPS with 20 mM pyruvate and 5 mM glucose. Cells were harvested by centrifugation, resuspended in MOPS with 20 mM pyruvate and 10 mM of the inducing carbon source, and grown for 2 h at 37°C. RNAs were isolated from ∼10⁷ cells by use of an RNeasy kit (Qiagen). Pyruvate served as a growth substrate to allow growth without detectable catabolite repression of GB-induced genes (7; data not shown). During isolation, RNAs were treated with an on-column DNase I treatment (Qiagen) for 30 min at room temperature. The resulting RNAs were subjected to PCR to verify the absence of contaminating DNA before being quantified using a Nanodrop spectrophotometer. cDNA was synthesized using Superscript III (Invitrogen) from 300 ng of starting RNA, with a 5′-NSNSNSNSNS-3′ primer instead of random hexamers. The regimen for cDNA synthesis was 25°C for 5 min, 52°C for 60 min, and 70°C for 15 min. Primers used are listed in Table S1 in the supplemental material. The PCR regimen was 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s. Quantitative real-time RT-PCR (qRT-PCR) was conducted with Power SYBR green with AmpliTaq Gold DNA polymerase according to the manufacturer's instructions (Applied Biosystems).

Analysis of plcH promoter activities by use of lacZYA fusions.

The pMW5 reporter construct was generated by amplifying the lacZYA operon from plasmid pRS415 (37), using specific primers with KpnI and EcoRI sites, and digesting the product with KpnI and EcoRI, followed by ligation into similarly digested pUCP22 (34). This placed the lacZYA operon in an orientation opposite that of the pUCP22 lacZ alpha fragment and the corresponding promoter, leaving the majority of the polycloning region intact. This construct shows very little background β-galactosidase activity in P. aeruginosa.

We constructed a deletion series of the plcH promoter by using PCR to amplify products from the pAES110 plcH::lacZYA fusion construct (32). Primers used for these constructs are detailed in Table S1 in the supplemental material. Briefly, each PCR product was gel purified, digested with XbaI and BamHI, and ligated into similarly digested pMW5. Ligation mixtures were transformed into E. coli DH5α cells (Invitrogen) by chemical transformation. Plasmid preparations from E. coli clones were transformed into P. aeruginosa by electroporation (5) and selected for growth on gentamicin. After overnight growth in MOPS-pyruvate-glucose medium as described above, cells were pelleted and resuspended in MOPS with 25 mM pyruvate and 2 mM of the inducing compound and grown under inducing conditions for 4 h at 37°C unless otherwise specified. The β-galactosidase assays were conducted according to the method of Miller (23).

Analysis of plcH reporter expression in Survanta (Abbott) was done using the specified P. aeruginosa strains carrying the pMW22 reporter plasmid. Survanta was diluted in MOPS with no additional carbon source (1:100). Additions of phosphorylcholine, choline, or GB (all from Sigma) were such that the final concentration of each additive was 0.25 mM. This concentration approximates that which would be derived if all PC in Survanta were hydrolyzed and every subsequent enzymatic step proceeded to completion.

Analysis of the pchP promoter by use of promoter fusions to gfp.

The pchP promoter deletion series was constructed by amplifying fragments from PAO1 genomic DNA. Primers for these constructs are detailed in Table S1 in the supplemental material. pMW47 was generated by amplification of the gfp-mut3 gene from pMQ80 (35) with adaptor primers. This product was digested with HindIII and EcoRI and ligated to a similarly cut pMQ80 backbone. This left the gfp gene in reverse orientation to the arabinose-inducible promoter region of pMQ80 and allowed yeast-based cloning of promoters in front of gfp. Briefly, each PCR product was gel purified and transformed along with KpnI-linearized pMW47 into Saccharomyces cerevisiae according to the method of Shanks et al. (35). Yeast cells were plated onto synthetic complete plates lacking uracil, and colonies were picked after 48 h at 30°C. Plasmid DNA was isolated from yeast according to the method of Shanks et al. (35). Transformation of the plasmids into E. coli and mobilization into P. aeruginosa were done as described for the plcH promoter fusion constructs. Cells were induced as described for the plcH promoter studies. Fluorescence was measured in black-walled 96-well dishes, using a Bio-Tek Instruments model Fix-800I fluorometer with an excitation wavelength of 485 nm and a 528-nm emission filter. Background autofluorescence of PAO1 was subtracted by measuring the fluorescence of identically grown cells carrying a Gm^r plasmid but no gfp gene.

EMSA.

We constructed a maltose binding protein (MBP) fusion to GbdR (pMW26) by using the pMALc vector system (NEB). E. coli DH5α carrying the pMW26 plasmid was grown overnight in LB plus 120 μg/ml carbenicillin. The overnight culture was transferred to four 250-ml flasks containing 50 ml of LB-carbenicillin and shaken at 220 rpm for 4 h. IPTG (isopropyl-β-d-thiogalactopyranoside) was added to 0.2 mM, and the cells were induced for 5 h. Cells were pelleted, lysed in column buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA) with 3 mg/ml lysozyme, clarified by centrifugation, and filtered through a 0.22-μm membrane before being applied to an amylose column. The column was washed with eight column volumes of column buffer, and protein was eluted in column buffer with 10 mM maltose. The fractions containing the fusion protein were dialyzed against 20 mM Tris-HCl, pH 7.5, 125 mM KCl, 1 mM MgCl₂, 5% glycerol at 4°C in a dialysis cassette (10,000-molecular-weight cutoff; Pierce). The full-length fusion protein was used in electrophoretic mobility shift assays (EMSAs), as the Xa cleavage product was insoluble (data not shown). We verified the ability of MBP-GbdR to function in vivo by the restoration of growth on GB to the ΔgbdR mutant by a plasmid expressing the pMAL-GbdR fusion (data not shown).

DNA fragments for EMSA analysis were created by PCR amplification, verification of the presence of a single band, and spot dialysis of the product. DNA was diluted to the specified concentration from this preparation. For labeled DNA, one of the primers contained a covalently linked 5′-biotin tag (IDT). EMSA experiments were conducted per the manufacturer's instructions (Pierce LightShift), with the following changes. First, buffer conditions were optimized to be similar to those used for other Pseudomonas AraC family proteins (22). This resulted in a final binding reaction buffer containing 1× Pierce binding buffer (10 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM dithiothreitol), 0.1 mM GB, 0.2 μg/μl poly(dI-dC), 3 mM MgCl₂, 7% glycerol, 1 mM EDTA, and 50 μg/ml bovine serum albumin. To reduce the formation of insoluble precipitates with the pchP probe, bovine serum albumin was increased to 250 μg/ml. Where included, unlabeled competitor DNA was present at 0.6 pmol/μl, labeled probe DNA was present at 1 fmol/μl, and MBP-GbdR was added to a final concentration of 0 to 3.0 μM. Binding reactions were carried out at 37°C for 40 min. Binding reactions were run in nondenaturing 5% Tris-borate-EDTA gels (Bio-Rad), transferred to BioDyne nylon membranes (Pierce), and detected using the supplied streptavidin-horseradish peroxidase protocol (Pierce).

In vitro P. aeruginosa gene induction in response to bovine surfactant and murine BALF.

For in vitro surfactant and BALF experiments, P. aeruginosa was pregrown in MOPS medium as described for qRT-PCR. For the surfactant study, cells from overnight cultures were pelleted and resuspended in either MOPS plus 10 mM pyruvate or a 1:50 dilution of Survanta in MOPS plus 10 mM pyruvate, resulting in a phospholipid concentration of 0.5 mg/ml total phospholipid. BALF for in vitro P. aeruginosa gene induction studies was harvested from uninfected adult male C57BL/6 mice. Mice were anesthetized with intraperitoneal sodium pentobarbital (70 to 90 mg/kg of body weight). Tracheas were cannulated with 2 cm of 22-gauge polyethylene tubing attached to a 23-gauge needle, 1 ml of cold Dulbecco's phosphate-buffered saline (DPBS) was instilled into the lungs, and the BALF was collected. The BALF was sterile filtered and frozen at −80°C until use. For the BALF experiment, cells were resuspended in DPBS plus 10 mM pyruvate or into BALF plus 10 mM pyruvate. For both analyses, cells were shaken at 37°C for 2 h and RNAs were prepared as described above.

Mouse lung infection.

Male C57BL/6 mice (Jackson Labs) of 8 to 12 weeks of age were inoculated with 1 × 10⁸ CFU of P. aeruginosa PAO1 via oropharyngeal aspiration following brief anesthesia with isoflurane (1). The infection proceeded for 24 hours, followed by nonsurvival surgery and acquisition of BALF by the same methods described above for uninfected animals. Samples were immediately treated with RNAprotect bacterial reagent (Qiagen) per the manufacturer's instructions. RNA isolation was performed as described above, with the following change: 95 to 160 ng total RNA was used to generate cDNA. The amplification conditions were 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 56°C for 1 min, and 72°C for 1 min, using an Applied Biosystems 7500 Fast instrument. Only one PCR product was obtained for all primers and all samples, based on melting curve analysis. All procedures that involved mice were approved by the University of Vermont Institutional Animal Care and Use Committee.

RESULTS

plcH and pchP induction in response to choline, GB, and DMG is dependent on the GbdR transcription factor.

In previously published studies, we showed that the GbdR transcription factor controls the expression of gbcAB and dgcAB, which are required for GB and DMG degradation, respectively, in response to both GB and DMG (43). Because plcH expression, PlcH activity, and PchP activity are similarly induced by these two osmoprotectants (18, 36), we hypothesized that GbdR may be the GB- and DMG-dependent regulator of plcH and pchP induction. To test this hypothesis, PlcH activity was first assessed in ΔgbdR mutant cultures by both the NPPC hydrolysis assay and assessing hemolysis on blood agar plates. Cultures of P. aeruginosa PAO1 WT, PAO1 ΔgbdR, PAO1 ΔplcHR, and PAO1 ΔgbdR complemented with a functional copy of gbdR at the att site were grown in T broth with choline, and the supernatants were analyzed for NPPC hydrolysis activity (Fig. 1A). In WT supernatants, the addition of choline to the medium led to a 4.2-fold increase in NPPC hydrolysis activity; no increase in NPPC cleavage activity was observed upon addition of choline in either the ΔgbdR or ΔplcHR strain. Complementation of the gbdR mutant partially restored the induction of PlcH activity in response to choline (Fig. 1A). In the blood agar hemolysis assay, zones of clearing were observed around the WT but not the ΔplcH and ΔgbdR mutants. Complementation of the ΔgbdR mutant restored hemolytic activity (Fig. 1B).

FIG. 1.

Role of GbdR transcription factor in controlling PlcH activity and plcH transcript levels. (A and B) PAO1 WT, ΔplcHR and ΔgbdR mutants, and a ΔgbdR mutant complemented with gbdR at the att locus (ΔgbdR att::gbdR) were assayed for choline-induced NPPC hydrolysis activity in cultures grown in T broth (TB) (black bars) or T broth containing choline (white bars) (A) and in a hemolysis assay using phosphate-supplemented blood agar plates (B). (C and D) qRT-PCR quantification of plcH and pchP transcript levels normalized to rplU levels in P. aeruginosa strains PAO1 (C) and PA14 (D). Transcript induction is reported as the transcript level measured under inducing conditions (GB) divided by the transcript level under noninducing conditions (pyruvate). Error bars represent standard deviations between levels in three replicate cultures, and results were comparable between at least three separate experiments.

The regulation of PlcH enzyme activity by GbdR suggested the possibility that GbdR regulates the induction of plcH transcription. plcH transcription has been shown to be induced in response to choline, GB, and DMG (36), and induction by choline is not observed in a betB mutant strain which cannot convert choline to GB, indicating that GB and/or its catabolic product DMG is the compound responsible for plcH induction in choline-containing media (31). We have shown that either GB or DMG can stimulate GbdR-dependent transcription of the GB catabolic genes even in a mutant (ΔgbcAB) that is unable to convert GB to DMG, suggesting that both GB and DMG can serve as inducing compounds (43). To determine if GbdR also regulates plcH induction by GB and DMG, we used qRT-PCR to measure plcH transcripts in WT and ΔgbdR strains (Fig. 1C). plcH mRNA levels were assessed in both MOPS minimal medium with pyruvate, in which plcH transcript levels are very low, and in the same medium amended with GB. The plcH transcript levels are reported as relative changes between uninduced (pyruvate) and induced (GB) conditions. The relative transcript level for each strain was normalized to the rplU transcript, which has been shown to remain constant relative to total RNA under a variety of conditions (14). In the presence of GB, the plcH transcript was induced 2.4- ± 0.09-fold in the WT (Fig. 1C), while no induction was observed in the ΔgbdR strain (0.88- ± 0.06-fold change). Incorporation of gbdR at the att chromosomal site almost fully complemented the phenotype of the ΔgbdR strain (2.09- ± 0.08-fold induction). Induction with DMG showed a similar pattern of plcH transcriptional activation, in a gbdR-dependent manner (data not shown). Similar results were obtained with P. aeruginosa strain PA14, with a 6.69- ± 0.44-fold induction of plcH transcript upon induction with GB and no induction (0.64- ± 0.03-fold) in the ΔgbdR mutant (Fig. 1D). These data support the hypothesis that plcH transcription is regulated by GbdR.

Phosphorylcholine, a product of PlcH-mediated cleavage of PC and SM, is hydrolyzed by the periplasmic phosphorylcholine phosphatase, PchP, to yield choline. While to our knowledge induction conditions for the pchP transcript have not previously been reported in the literature, functional assays have been used to demonstrate induction of enzyme activity in response to choline, GB, and DMG, similar to the regulation observed for plcH (18, 20). We hypothesized that pchP induction was also dependent on GbdR. Similar to the results with plcH, the pchP transcript was induced in response to GB compared to pyruvate in PAO1 WT (4.5- ± 0.7-fold) (Fig. 1B) but not in the ΔgbdR strain (0.71- ± 0.1-fold). Integration of gbdR at the att chromosomal site partially rescued the GB-dependent induction (3.1- ± 0.3-fold). A similar pattern of induction of pchP by GB in a GbdR-dependent manner was observed in P. aeruginosa strain PA14 (Fig. 1D). These data support the regulation of pchP by GbdR in a manner similar to that of plcH.

Our previous studies with GbdR demonstrated that either GB or DMG could induce GbdR-dependent transcription of GB and DMG catabolic genes, and therefore we hypothesized that both GB and DMG would be capable of inducing plcH and pchP. To test this, we analyzed plcH and pchP transcript levels in strains unable to catabolize either compound (43). In a P. aeruginosa PAO1 ΔgbcAB strain, which is unable to convert GB to DMG, both GB and DMG induced the plcH transcript (3.62- ± 0.87-fold and 3.23- ± 0.64-fold, respectively). In a P. aeruginosa PAO1 ΔdgcA mutant, which is unable to convert DMG to sarcosine, DMG was sufficient to induce plcH transcription 2.88- ± 0.35-fold. Together, these data indicate that both GB and DMG are capable of stimulating plcH transcription.

Deletion mapping to determine the region necessary for GbdR-dependent induction in the plcH and pchP promoters.

To determine the segment of each promoter that was necessary for GbdR-dependent activation of plcH and pchP transcription, plcH and pchP promoter fragments were cloned upstream of the lacZYA and gfp-mut3 reporter genes, respectively (Fig. 2A and B). The resulting constructs were transformed into both WT and ΔgbdR backgrounds. Transcriptional activation was assessed in cells grown either in medium with pyruvate as the sole source of carbon or in medium containing pyruvate plus GB. For the full-length plcH fragment (plcH-A) and two of the truncated fragments (plcH-B and plcH-C), transcriptional activation was observed in the WT strain but not in the ΔgbdR mutant (Fig. 2A). The smallest promoter fragment (plcH-D) failed to show GB-dependent induction (Fig. 2A). This suggests that the promoter region necessary for GB-dependent transcriptional induction is between positions −177 and −86 or overlaps the −86 site.

FIG. 2.

Promoter truncation analysis of the plcH and pchP promoters. (A) Schematic map of the truncation constructs for the plcH promoter, indicating their positions relative to the translational start site, and their corresponding abilities to induce lacZ transcription in medium with GB. The fusion constructs (A to D) were analyzed in the WT and ΔgbdR (Δ) strain backgrounds in the absence (−) and presence (+) of GB. (B) Schematic map of the truncation constructs for the pchP promoter, indicating their positions relative to the translational start site, and corresponding gfp fluorescence driven by the truncated pchP-reporter constructs, measured in arbitrary fluorescence units after subtraction of the PAO1 WT (not expressing gfp) autofluorescence baseline. Error bars represent standard deviations between levels in three replicate cultures, and results were comparable between at least three separate experiments.

For pchP, a promoter truncation series fused to the gfp reporter gene was constructed. The three longest constructs retained GB-dependent induction of green fluorescent protein fluorescence and did not show induction in the absence of gbdR (Fig. 2B), while the shortest construct (D) showed no induction by GB. This suggests that nucleotides within the region between −141 and −75 bp are required for GB-dependent transcriptional induction in the pchP promoter.

Evidence for GbdR binding to the plcH and pchP promoters.

Our studies above identified the regions of the plcH and pchP promoters that were necessary for GbdR-dependent induction of reporter activity in response to GB. To test the hypothesis that GbdR interacts directly with the plcH promoter, we used a 164-bp fragment of the plcH promoter (corresponding to fragment C from the promoter deletion analysis) in an EMSA. GbdR was purified as an MBP fusion (MBP-GbdR); the MBP-GbdR expression construct was sufficient to restore growth on GB in the ΔgbdR mutant, indicating that the fusion protein was functional. MBP-GbdR specifically shifted the plcH promoter with increasing concentrations of protein (Fig. 3A, lanes 2 and 3). This shift could be competed with unlabeled plcH promoter (Fig. 3A, lane 4). We also tested whether MBP-GbdR binds to the pchP promoter by using a 260-bp fragment of the pchP promoter (corresponding to fragment B of our promoter deletion analysis). MBP-GbdR specifically shifted the pchP promoter with increasing concentrations of protein, leading to a larger proportion of the fragment shifting and to an additional shifted band (Fig. 3B, lanes 2 to 4), although the efficiency of binding was lower than that for plcH (Fig. 3A, lane 8). This shift could be competed with unlabeled pchP promoter (Fig. 3B, lane 5). The lower affinity of our MBP-GbdR fusion for the pchP promoter under these conditions is consistent with our promoter fusion data showing that the induction of the pchP transcript is less than that seen with the plcH promoter. The MBP-GbdR protein bound to both the plcH and pchP promoters in the presence (Fig. 3A) and absence (data not shown) of GB. Combined with the promoter deletion analyses and the induction studies, this in vitro binding is strong support for direct GbdR binding and regulation at the plcH and pchP promoters.

FIG. 3.

EMSA to examine interactions between MBP-GbdR and plcH and pchP promoter fragments. (A) The WT plcH promoter fragment was incubated alone or with MBP-GbdR (0.2 and 2 μM) (lanes 1 to 3); to the reaction mix in lane 4, unlabeled plcH probe was added in excess. Similar incubations were performed with a probe spanning the same region but with two nucleotide substitutions (plcH**) (lanes 5 to 7). In lane 8, the shift of the pchP promoter probe is shown for comparison. (B) The WT pchP promoter probe (lanes 1 to 5) and the mutated promoter probe (pchP**) (lanes 6 to 8) were incubated with MBP-GbdR at the concentrations shown (μM). An excess of unlabeled pchP probe was added to the reaction mix in lane 5.

Identification and mutation of conserved residues in the plcH and pchP promoters that are necessary for GbdR induction.

The promoter regions necessary for plcH and pchP induction by GB were aligned with each other by use of LALIGN (11). One region of alignment corresponding to positions −174 to −138 of the plcH promoter was also aligned via the LALIGN pairwise method to the promoter regions of other genes known to be controlled by GbdR, including gbcB, involved in GB catabolism, and PA5396, the first gene in the putative operon that contains the DMG catabolic genes. The alignment of these promoters is shown in Fig. 4A. The transcriptional start sites for the phosphate- and choline-dependent plcH transcripts, as demonstrated by Vasil and coworkers (41), are shown with dotted arrows (Fig. 4A).

FIG. 4.

Role of two conserved base pairs in the plcH and pchP promoters in GbdR-dependent induction. (A) Schematic of the plcH promoter showing the conserved promoter region, with base pairs denoting the distance from the translational start site. The phosphate- and choline-dependent transcriptional start sites are shown with dotted arrows (41). The alignment shows a common promoter region present near four genes controlled by GbdR. The asterisks mark the two base pairs changed to adenines in the mutant constructs. (B) The nonmutated promoter fusions were analyzed in the WT and ΔgbdR (Δ) strain backgrounds in the presence (+) and absence (−) of GB. The mutated constructs with the 2-bp alteration were analyzed in the WT background (WT**). Error bars represent standard deviations between levels in three replicate cultures, and results were comparable between at least three separate experiments.

To test the hypothesis that the aligned regions represented the motif recognized by GbdR, we mutated GC residues (marked with asterisks in Fig. 2) to adenines in both the plcH and pchP promoters. In both cases, promoter fusion constructs that were identical to plcH-C and pchP-C (Fig. 2A and B), respectively, except for this 2-bp mutation, completely lacked GbdR-dependent induction of transcription in medium with GB (Fig. 4B). The 2-bp changes made in the plcH and pchP promoters were not associated with either the −10/−35 sites or the ribosome binding site, consistent with the observation that basal transcription of plcH or pchP in strains carrying the mutated promoters is not different from that in strains carrying the nonmutated versions (Fig. 4B). Mutation of the adjacent C residue at −169 to an adenine in the plcH promoter led to abrogation of all GbdR-dependent induction, similar to the mutations described above, while mutation of the G residue at −161 led to a 59% decrease in induction of GbdR-dependent transcription levels (data not shown).

To determine if the conserved promoter residues represented part of the region required for GbdR binding to the promoters, we tested the ability of either the mutated plcH (plcH**) or pchP (pchP**) promoter to bind MBP-GbdR in an EMSA (Fig. 3A and B). The plcH** mutant promoter (Fig. 3A, lanes 5 to 7) failed to show a shift at any tested concentration of MBP-GbdR. The mutant pchP promoter (pchP**) could be shifted by the addition of 3.0 μM MBP-GbdR, but at a much lower affinity than the WT promoter (Fig. 3B, lanes 6 to 8). The defects of these mutant promoter fragments in associating with MBP-GbdR suggest that these two base pairs are important for GbdR binding.

GbdR controls plcH and pchP induction in response to surfactant and BALF and in the mouse lung.

The major component of mammalian lung surfactant is PC. As such, this compound represents a sizable pool of PlcH substrate within the lung. To test whether GbdR controls plcH and pchP gene induction in response to surfactant, we grew PAO1 WT and ΔgbdR strains in MOPS medium containing 10 mM pyruvate or the same medium amended with the bovine-derived surfactant Survanta. RNAs were harvested from cells, and qRT-PCR was used to determine plcH, pchP, and rplU transcript levels; plcH and pchP levels were normalized to levels of rplU. As shown in Fig. 5A, P. aeruginosa WT induced plcH and pchP 6.4-fold and 8.3-fold, respectively, in the presence of Survanta over levels in medium without surfactant addition. No induction was observed in the ΔgbdR mutant. Similar induction experiments were performed with mouse BALF. P. aeruginosa strains were resuspended in sterile filtered BALF or DPBS, with 10 mM pyruvate added to both. In BALF-pyruvate, the normalized transcript levels of plcH and pchP were 10.2- and 4.6-fold higher, respectively, than levels in cells incubated in buffer with pyruvate. No induction upon incubation in BALF was observed for the ΔgbdR strain (Fig. 5B).

FIG. 5.

GbdR-dependent regulation of plcH and pchP transcripts in Survanta and murine BALF and during mouse lung infection. (A) RNAs from WT (black bars) and ΔgbdR (white bars) cells grown in medium with and without surfactant were analyzed by qRT-PCR to determine relative plcH and pchP transcript levels; transcript levels were normalized to those of the rplU control transcript. The levels of induction represent transcript levels in cells grown in medium with Survanta divided by levels in cells grown with pyruvate as the sole source of carbon. (B) qRT-PCR with cells grown in filter-sterilized murine BALF from uninfected mice. Induction is reported as in panel A. (C) qRT-PCR was conducted on RNAs isolated from BALF of mice infected with either PAO1 WT or PAO1 ΔgbdR for 24 h. Transcript levels are reported as plcH or pchP transcript levels relative to levels of the rplU transcript. There were seven mice infected for each group, and standard errors are shown.

To test the importance of GbdR for the regulation of plcH and pchP in P. aeruginosa within the mouse lung, we infected C57BL/6 mice with 1 × 10⁸ CFU of either PAO1 or the ΔgbdR mutant via oropharyngeal aspiration (n = 7 per group). After 24 h, the BALF was harvested and total RNA isolated and analyzed by quantitative PCR. plcH and pchP transcript signals were normalized to that of rplU. The results of the transcript analyses are shown in Fig. 5C. Both plcH and pchP showed significantly higher transcript levels in WT cells than in ΔgbdR cells. These data support an important role for GbdR in regulation of plcH and pchP in vivo.

Examination of the plcH positive feedback induction loop.

Based on the work presented here and data and interpretations published by others (8, 41), we hypothesize that PlcH activity initiates a series of catalytic steps leading to GbdR-dependent induction of plcH and pchP, establishing a positive feedback loop in which choline phosphate is degraded to GB, which further induces plcH expression (Fig. 6A). To establish the role of each predicted member of the proposed positive feedback loop, we measured the expression of the plcH promoter fusion in the PAO1 WT and in mutant strains, including the ΔplcHR, pchP::Tn, betB::Tn, and ΔgbdR strains, in the presence of MOPS-Survanta medium (containing approximately 0.25 mg/ml PC) (Fig. 6B). As shown above (Fig. 5A), plcH and pchP transcription in this medium was induced in the WT but not in the gbdR mutant. There was no induction of the plcH transcriptional fusion in the plcH, pchP, or betB mutant, suggesting that PC catabolism was necessary for induction of plcH by GbdR (Fig. 6B). To explore the role of choline phosphate (ChoP), choline, and GB in the pathway, we added 0.25 mM of each compound to the medium and assessed plcH transcription activity. The addition of ChoP resulted in plcH induction in a ΔplcHR strain but not in the pchP::Tn or betB::Tn strain, underlining the importance of the PchP phosphatase in ChoP hydrolysis and resultant choline catabolite-dependent gene induction. The addition of choline restored the induction of plcH in the ΔplcHR and pchP::Tn strains, but choline was unable to induce the plcH promoter fusion in the betB::Tn strain, as reported previously (31). GB restored plcH induction in all strains but the ΔgbdR mutant. These data highlight the steps critical for plcH induction in lung surfactant as well as establish the role of each enzyme in the initiation of the plcH positive feedback induction system.

FIG. 6.

PlcH is the initiator of a positive feedback loop during P. aeruginosa exposure to lung surfactant (Survanta). (A) Model for the GbdR-dependent, GB-induced PlcH positive feedback loop. Thick solid arrows represent catalytic activity conducted by the adjacent protein, and thin solid arrows represent transport events. Dashed arrows represent induction events. (B) PAO1 WT and mutants defective in plcH, pchP, betB, and gbdR carrying the pMW22 plcH reporter construct were exposed to MOPS medium with diluted Survanta (1%). Similar cultures were amended with phosphorylcholine (ChoP), choline, or GB to a final concentration of 0.25 mM. Induction is reported as the Miller units for the listed induction conditions divided by the Miller units for comparable cultures grown in MOPS-pyruvate medium. Error bars represent standard deviations between levels in three replicate cultures, and results are comparable between at least three separate experiments.

DISCUSSION

In this study, we identified GbdR as the transcription factor required for induction of the plcH and pchP transcripts in the presence of GB and DMG. Using promoter deletion and mutation analysis, we identified residues required for GbdR-dependent induction and demonstrated direct binding to the plcH and pchP promoters by EMSA. GbdR is required for induction of plcH and pchP in bovine lung surfactant and mouse BALF. In addition, using an acute model of mouse lung infection, we showed that GbdR plays an important role in induction of plcH and pchP in vivo.

The two known stimuli of plcH regulation, low phosphate and the presence of GB, regulate plcH in different ways. Transcriptional induction of the plcH gene in response to limiting phosphate conditions is controlled by PhoB and is hypothesized to function as part of a phosphate-scavenging system (36). The acquisition of phosphate from PC and SM will eventually turn off plcH induction via the Pho system. The regulation of plcH transcription by GbdR in response to GB represents a positive feedback regulatory scheme. As PlcH activity releases more choline from PC and SM, conversion of choline to GB will lead to induction of plcH transcription by GbdR. By adding each metabolite in the proposed plcH induction pathway (Fig. 6A) to strains capable or incapable of each predicted catabolic step, we were able to fully validate the role of each intermediate in the PlcH-initiated positive feedback loop during P. aeruginosa interaction with host-derived phospholipids (Fig. 6B). The positive feedback regulation, leading to increased PlcH production and increased host damage, may in part explain the striking phenotypes of plcH mutants in a variety of animal model systems (12, 27, 29, 46).

While P. aeruginosa produces multiple PLCs and phosphatases, PlcH and PchP appear to be of major importance for inducing GbdR induction of virulence-regulated genes (2, 26). In PC-rich Survanta, experiments using the plcH reporter system supported the hypothesis that PlcH is critical for ChoP acquisition (Fig. 6B), suggesting that other described PLCs (2, 26) do not play a significant role in PC hydrolysis under these conditions. The phosphorylcholine phosphatase (PchP) was identified biochemically and subsequently cloned (18, 20). We are unaware of any published reports that verify the requirement of PchP for hydrolysis of ChoP in an environment similar to that used in our assay. Although a number of other putative phosphatases are produced by P. aeruginosa, some of which are highly induced during growth on Survanta (M. J. Wargo and D. A. Hogan, unpublished data), we established that, in the time frame of our assay, PchP is required for plcH induction in Survanta (Fig. 6B).

plcH transcript accumulation is likely impacted by many different factors. Because there are several catabolic steps involved in generating GB for induction of plcH by GbdR in PC- and SM-rich environments such as the lung, induction of plcH by GB and DMG could be regulated at a number of points. In particular, while we have discussed PlcH as a single enzyme for the sake of simplicity, PlcH can be present freely or in complex with PlcR₁ or PlcR₂, two chaperones encoded by overlapping reading frames of the plcR gene (6). These complexes have different biochemical activities from that of the free enzyme (17). Interestingly, it has been shown that choline induction fails to robustly induce the plcR transcript, while phosphate starvation is capable of inducing both plcH and plcR (17). plcH transcript stability can also vary in response to iron levels (25). The availability of the choline-derived GB-inducing signal could also be affected by choline phosphatase activity, choline uptake, or choline conversion of choline into GB. betAB gene induction by choline is independent of GbdR (30, 42, 43). Previous work has shown only very slight alterations (<2-fold changes) in gbdR transcript levels upon growth in media containing choline, GB, or DMG (45); mechanisms for regulating GbdR activity have not been explored. Due to the complexity of the plcH positive feedback loop, there are a variety of opportunities for the development of strategies for the inhibition of this pathway in vivo.

Although these analyses elucidate the major GB- and DMG-dependent regulator of plcH and pchP and show the role of GbdR during lung infection, several questions remain to be answered. First, it is important to understand how the phosphate and GB/DMG signals are integrated at both the plcH and pchP promoters by assessing the interaction between GbdR and PhoB. Second, understanding the mechanism of catabolite repression (34) of these loci and other genes related to GB and DMG catabolism will enable us to better understand the regulation of virulence factors by cellular metabolic processes. Finally, understanding the intricacies of GbdR-dependent regulation in P. aeruginosa during infection could shed light on the respective roles of GB utilization as a carbon and nitrogen source, an osmoprotectant (38, 44), and an inducer of virulence-related genes such as plcH in different environments.