PsrA Is a Positive Transcriptional Regulator of the Type III Secretion System in Pseudomonas aeruginosa

Abstract

The type III secretion system (TTSS) of Pseudomonas aeruginosa is induced in vivo upon contact with eukaryotic cells and in vitro by calcium depletion in culture medium. We have observed a previously identified protein, PsrA, necessary for full activation of TTSS gene expression in P. aeruginosa. Electrophoretic mobility shift assays showed that recombinant PsrA could bind to the exsCEBA promoter region. A mutant with a deletion in the psrA gene was constructed. Using transcriptional fusions, we demonstrated that PsrA is required for the full activation of transcription of the TTSS regulatory operon exsCEBA and effector exoS, although the deletion mutant still responded to calcium depletion, to serum, and to host cell contact. The psrA mutant showed a marked decrease in the secretion of the type III effectors and weak resistance to phagocyte-like PLB-985 cells. The defect in TTSS transcription and secretion in the psrA mutant could be complemented by expression in trans of psrA. PsrA was previously identified as a transcriptional activator of RpoS, a central regulator during stationary phase. We confirmed with our strain that RpoS has a negative effect on TTSS gene expression. Taken altogether, these results suggest that PsrA is a newly identified activator that is involved in the expression of the TTSS by enhancing the exsCEBA transcriptional level.

Pseudomonas aeruginosa is a gram-negative opportunistic pathogen that causes acute and chronic infections in hospitalized individuals, burn victims, and cystic fibrosis patients (2, 31, 35). One of the virulence factors most strongly correlated with severe infection both in animal models and in human patients is the type III secretion system (TTSS) (17, 41). The TTSS is a specialized protein secretion apparatus and is encoded by more than 25 genes. The TTSS is induced by at least three environmental signals: (i) in vitro removal of calcium from the medium, (ii) in vivo contact with eukaryotic host cells, and (iii) the presence of serum (11, 50, 56). Upon activation, the type III secretion apparatus translocates effector molecules into the cytoplasm of host cells, resulting in targeted cell dysfunction or death (3, 10, 12, 23, 24, 34, 42, 47). Four type III secreted toxins in P. aeruginosa have been described. Exoenzyme S (ExoS) and ExoT are highly related proteins with two functional domains (55). They both possess a GTPase-activating protein for Rho GTPase (rho-GAP) domain, which causes cell rounding by effecting depolymerization of actin in targeted host cells (14, 28). They also both contain an ADP-ribosylation domain, the cellular targets of which differ between ExoS and ExoT. ExoS ADP-ribosylates small Ras-like proteins (33, 37, 38), whereas ExoT targets CrkI and CrkII (47). The other two exoenzymes described to date are ExoU (10, 19), a phospholipase (42), and ExoY, an adenylate cyclase (57). It is worth pointing out that the presence of ExoU and ExoS in the genome varies from strain to strain, and these proteins appear to be mutually exclusive (9). In this study, we have used strain CHA, whose genome contains the genes encoding ExoS, ExoT, and ExoY but not that encoding ExoU (3).

Expression of the P. aeruginosa TTSS and effector genes is coordinately regulated by an AraC-like transcriptional activator, ExsA. ExsA binds to a consensus sequence (TNAAAANA) approximately 50 bp upstream of the transcriptional start site of the TTSS genes (22, 54, 56). In recent years, a cascade of proteins has been discovered which may influence posttranscriptionally the activity of ExsA. ExsD, an antiactivator, is a negative regulator that binds ExsA and prevents it from activating TTSS gene transcription (32). ExsC has recently been shown to interact with ExsD and is thought to act as an anti-antiactivator that releases ExsA from inhibition by ExsD (7). ExsE, a small TTSS secreted protein binding tightly to ExsC and acting as a repressor of the TTSS, has been recently identified (39, 49). ExsE sequesters ExsC when the TTSS channel is closed but is secreted when the channel is activated by a signal such as low calcium concentration, leaving ExsC free to interact with ExsD and releasing ExsA, thereby allowing ExsA to activate TTSS expression. It is interesting that, except for exsD, the genes encoding these proteins, exsA, exsC, and exsE, are located in the same operon, exsCEBA, but other genes, such as that for PtrA, pseudomonas type III repressor A, may also affect the TTSS through direct interaction with ExsA. PtrA is highly and specifically inducible by a high copper concentration signal and suppresses the TTSS by inhibiting the function of ExsA through direct interaction, although it is not clear whether PtrA blocks the DNA binding ability of ExsA or its ability to interact with the RNA polymerase (16). Furthermore, mutants containing deletions in cyaAB or vfr (53), aceAB (5, 40), and rtsM, a novel sensor kinase response regulator hybrid, (29) could be restored to normal function by the in trans expression of exsA, indicating another important transcriptional level of regulation of the TTSS through the activity of the exsCEBA promoter (pC). Therefore, it is interesting to study the regulation of exsCEBA itself to explore the regulation of the TTSS.

In this study, we describe a novel regulator of the TTSS, PsrA, previously described as a regulator involved in positive regulation of RpoS and in negative autoregulation by direct binding to the promoters (25-27). We show here that PsrA is required for the full activation of the transcription of exsCEBA and hence of the effector exoS. PsrA may activate the TTSS directly at the level of the pC or indirectly via another unknown TTSS regulator(s) controlled by itself.

MATERIALS AND METHODS

Bacterial strains and plasmids.

All bacterial strains and plasmids used in this study are listed in Table 1. P. aeruginosa strains were maintained on Pseudomonas isolation agar (Difco, Detroit, Mich.) or cultured in Luria broth (LB) with antibiotics as required (300 μg of carbenicillin ml⁻¹, 200 μg of gentamicin ml⁻¹, and 250 μg of tetracycline ml⁻¹). Escherichia coli strains were cultured in LB with antibiotics as required (100 μg ampicillin ml⁻¹, 10 μg of gentamicin ml⁻¹, and 20 μg of tetracycline ml⁻¹). P. aeruginosa strains were stored in Protect bacterial preservers (TSC, Lancashire, United Kingdom) at −80°C. E. coli strains were maintained at −80°C as 50% glycerol stocks. To induce the expression of the TTSS, P. aeruginosa strains were grown in LB supplemented with 5 mM EGTA-20 mM MgCl₂. Noninducing conditions were LB in the presence of 5 mM CaCl₂.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source or reference
Strains
E. coli DH5α	Used for plasmid transformation	Stratagene
E. coli BL21 (DE3)	Used for recombinant protein production	Stratagene
E. coli S17-1	Used for incorporating mating constructs into P. aeruginosa	44
CHA	Mucoid cystic-fibrosis-isolate	48
ΔexsA	exsA::Gmr mutant of CHA	3
CHA(pC::lux)	lux fusion with exsC promoter integrated at the attB site of the CHA chromosome	This study
CHA(pS::lux)	lux fusion with exoS promoter integrated at the attB site of the CHA chromosome	This study
ΔpsrA	CHA containing an in-frame deletion in psrA ORF, corresponding to amino acids 14-194	This study
ΔpsrA(pC::lux)	lux fusion with exsC promoter integrated at the attB site of the ΔpsrA chromosome	This study
ΔpsrA(pS::lux)	lux fusion with exoS promoter integrated at the attB site of the ΔpsrA chromosome	This study
ΔrpoS	CHA containing an in-frame deletion in rpoS ORF corresponding to amino acids 1-207	This study
ΔrpoS(pC::lux)	lux fusion with exsC promoter integrated at the attB site of the ΔrpoS chromosome	This study
ΔpsrA(pPsrA)(pC::lux)	ΔpsrA(pC::lux) complemented by psrA	This study
ΔpsrA(pPsrA)(pS::lux)	ΔpsrA(pS::lux) complemented by psrA	This study
ΔpsrA(pDD2)(pC::lux)	ΔpsrA(pC::lux) containing pDD2 vector	This study
ΔpsrA(pDD2)(pS::lux)	ΔpsrA(pS::lux) containing pDD2 vector	This study
Plasmids
pTOPO	Blunt end cloning vector; Apr	Invitrogen
pEX100Tlink	Allelic replacement suicide vector; Apr (Cbr)	36
PUCGmlox	pUC18-based vector containing the aacC1 gene flanked by lox; Apr, Gmr	36
mini-CTX-exsCp-lux	mini-CTX-lux containing exsC promoter; Tcr	29
mini-CTX-exoSp-lux	mini-CTX-lux containing exoS promoter; Tcr	29
pFLP2	Flp recombinase; Apr (Cbr)	18
pUCP20	Escherichia-Pseudomonas shuttle vector; Apr (Cbr)	52
p1ApC	pUCP20-derived vector containing pC	4
pDD2	pUCP20-derived vector containing exsA gene; Apr	3
psrAp	pTOPO-derived vector containing psrA promoter	This study
pfleQp	pTOPO-derived vector containing fleQ promoter	This study
psrAp-Rec	pET15b-derived vector containing psrA; Apr	This study
vfrp-Rec	pET15b-derived vector containing vfr; Apr	This study
pKO-psrA	pEX100Tlink-derived vector containing up- and downstream regions of psrA and Gmlox; Apr, Gmr	This study
pKO-rpoS	pEX100Tlink-derived vector containing up- and downstream regions of rpoS and Gmlox; Apr, Gmr	This study
pPsrA	pUCP20-derived vector containing psrA ORF as well as its promoter region; Apr (Cbr)	This study

^aAp^r, ampicillin resistance; Gm^r, gentamicin resistance; Tc^r, tetracycline resistance; Cb^r, carbenicillin resistance; Gmlox, aacC1 (gentamicin cassette) flanked by the lox sequence (36).

Strain construction.

All PCR primers employed in this study are listed in Table 2 and are based on the PAO1 genome sequence (http://www.pseudomonas.com) (46). All amplifications were carried out with Pfu Turbo polymerase (Promega Corporation, Madison, Wis.) by using CHA genomic DNA as the template. PCR products were subcloned into pTOPO (Invitrogen, San Diego, Calif.) and sequenced to confirm that no mutations were introduced during amplification. Plasmids were introduced into chemically competent E. coli DH5α.

TABLE 2.

Primer sequences used in this study

Primer	Sequencea	Function
psrA-U1	5′-AGAATTCCGCCGTCGTCATTCGCCGA-3′	To amplify upstream region flanking psrA
psrA-U2	5′-TAAGCTTCCGCATCGAGAATGCGTTCG-3′
psrA-D1	5′-TAAGCTTGGTGATGCACCTGATGGTGC-3′	To amplify downstream region flanking psrA
psrA-D2	5′-AGGATCCCATCGACAGCTCCCTCAATC-3′
rpoS-U1	5′-TAGGATCCAGACCCTGAAAGGCGCC-3′	To amplify upstream region flanking rpoS
rpoS-U2	5′-ACAAGCTTCGCTGGGTTTTCTGGC-3′
rpoS-D1	5′-ATAAGCTTCTCGTCCAGCATGATGCC-3′	To amplify downstream region flanking rpoS
rpoS-D2	5′-ATGAATTCGAAGACCTGCATCGCCG-3′
vfr-R1	5′-ACATATGGTAGCTATTACCCACACA-3′	For recombinant Vfr protein production
vfr-R2	5′-AGGATCCTCAGCGGGTGCCGAAGAC-3′
psrA-R1	5′-ACATATGGCCCAGTCGGAAACCG-3′	For recombinant PsrA protein production
psrA-R2	5′-AGGATCCTCAGGCCTTGGCGGGCGT-3′
psrA-C1	5′-AGAATTCACTGGTGCGTCATCTGCCGC-3′	For fragment containing psrA ORF as well as its promoter region
psrA-C2	5′-AGGATCCTCAGGCCTTGGCGGGCGT-3′
psrAp-F	5′-AGAATTCGGGGCGTCAGCTTCTGCAT-3′	For fragment containing psrA promoter region
psrAp-R	5′-AGAATTCAATGCGTTCGACGGTTTCCG-3′
fleQp-F	5′-AGAATTCCTACCAGATGTTCGGATAAGT-3′	For fragment containing vfr promoter region
fleQp-R	5′-AGAATTCCGAATGGGTCTCGCTCGACC-3′
psrA-1	5′-ATGGCCCAGTCGGAAACCG-3′	Verification of ΔpsrA
psrA-2	5′-TCAGGCCTTGGCGGGCGT-3′
rpoS-1	5′-GCGAGTTGGTCATCATCAAAC-3′	Verification of ΔrpoS
rpoS-2	5′-GCATCAAGTGCAAATACACCG-3′

^aRestriction enzyme sites are underlined.

Plasmids mini-CTX-exsCp-lux (pC::lux) and mini-CTX-exoSp-lux (pS::lux) (29) were mobilized into CHA by mating, and vector backbone sequences were excised by Flp recombinase as previously described (20). Each reporter construct is therefore present as a single, unmarked copy integrated at the chromosomal attB site.

Unmarked, in-frame deletion of amino acids 14 to 194 encoded by psrA or amino acids 1 to 207 encoded by rpoS in CHA(pC::lux) or in CHA(pS::lux) was carried out by allelic exchange (43). Briefly, upstream and downstream regions flanking psrA or rpoS were amplified using the primer pairs psrA-U1-psrA-U2 and psrA-D1-psrA-D2 or rpoS-U1-rpoS-U2 and rpoS-D1-rpoS-D2, respectively. The amplified regions were subcloned in tandem into the gene replacement vector pEX100Tlink (36) to generate pKO-psrA or pKO-rpoS. These constructs were then used to transform E. coli S17-1 and mobilized into CHA(pC::lux) or CHA(pS::lux) by mating. Gentamicin-resistant bacteria were isolated and then resolved as described previously, and relevant sequences were confirmed by PCR (13). The resulting mutants, ΔpsrA(pC::lux), ΔpsrA(pS::lux), and ΔrpoS(pC::lux), have no discernible growth difference compared with the wild type. ΔpsrA and ΔrpoS mutations were confirmed by PCR by using primers psrA-1-psrA-2 and rpoS-1-rpoS-2, respectively.

For the generation of a ΔpsrA complementation construct, a 782-bp PCR fragment (primers, psrA-C1 and psrA-C2) containing the psrA open reading frame (ORF) as well as 80 bp upstream encompassing the promoter region was cloned into pUCP20 (EcoRI-BamHI) to generate pPsrA. In this construct, psrA is under the transcriptional control of its own promoter. We also constructed a ΔpsrA mutant containing the pDD2 plasmid, allowing exsA expression as previously described (3).

Cell extract preparations.

Bacteria were grown overnight and then diluted 1:100 in 200 ml of LB for 3 h of growth. The cultures were centrifuged, and bacterial pellets were resuspended and washed twice in 30 ml of buffer A (25 mM HEPES, pH 7.3, containing 5% glycerol) and then lysed by sonication. Cell debris was removed from the extracts by centrifugation (15,000 × g for 30 min at 4°C).

Heparin-Sepharose affinity chromatography.

Affinity chromatography was performed first using a 5-ml HiTrap heparin column (Amersham Biosciences, Buckinghamshire, United Kingdom) equilibrated with buffer A containing 0.2 M NaCl. Thirty milliliters of ΔexsA strain extract (1 mg/ml) obtained from a 200-ml culture with an optical density at 600 nm (OD₆₀₀) of 1.5 was loaded at 1 ml/min. The column was washed once with 50 ml of buffer A-0.2 M NaCl and once with 50 ml of buffer A-0.3 M NaCl, and then elution was performed at a flow rate of 2 ml/min with 20 ml of buffer A-0.6 M NaCl. The fractions (5 mg) were collected and then analyzed for pC binding activity by electrophoretic mobility shift assays (EMSA). The preliminarily purified fractions (eluted from 0.6 M NaCl) having pC binding activity were pooled, dialyzed against 50 volumes of buffer A-0.2 M NaCl, applied onto a 1-ml HiTrap heparin column equilibrated by the same buffer but containing 0.3 M NaCl, washed with 0.3 M NaCl, and then eluted delicately with a serial 0.42 to 0.6 M NaCl step gradient at a flow rate of 0.5 ml/min; 5-ml fractions (0.1 to 0.3 mg) were collected and then analyzed again for pC binding activity by EMSA.

EMSA.

A 280-bp DNA fragment containing the pC was obtained by agarose gel purification of the product of BamHI restriction of vector pIApC. The fragment was labeled by end filling by the Klenow enzyme and biotin-16-dUTP (Roche, Meylan, France). A 280-bp psrA promoter (psrAp) and a 280-bp fleQ promoter (fleQp) were obtained by PCR with oligonucleotides (psrAp-F-psrAp-R and fleQp-F-fleQp-R), cloned into the TOPO system (Invitrogen), recovered by EcoRI digestion of ppsrAp and pfleQp, respectively, and labeled as pC fragments. The binding reactions were performed as described previously (8), with the following modifications. Briefly, the 20-μl reaction mixtures contained the biotin-labeled DNA fragment (at a concentration of approximately 20 pM) and various protein samples in a solution containing 25 mM HEPES (pH 7.3), 50 mM NaCl, 5% glycerol, and 2 μg of poly(dI-dC), the nonspecific competitor DNA (Amersham Biosciences). The reaction mixtures were incubated for 30 min at 37°C and then were electrophoresed on a 5% native polyacrylamide gel in 0.5× Tris-borate-EDTA for 2 h at 10 V/cm at room temperature. The gel was electrophoretically transferred (1 mA/cm² for 60 min) onto a Hybond-N⁺ membrane (Amersham Biosciences), fixed by UV cross-linking, and then visualized according to the instructions of the manual for the LightShift kit (Pierce, Bonn, Germany).

Expression and purification of rPsrA and rVfr.

psrA and vfr were amplified from chromosomal DNA of CHA by PCR using the oligonucleotide primers psrA-R1-psrA-R2 and vfr-R1-vfr-R2, respectively. After NdeI-BamHI digestion, PCR fragments were cloned into pET15b (Novagen, Darmstadt, Germany) and digested with the same restriction enzymes to give psrAp-Rec and vfrp-Rec. The resulting recombinant constructs were verified by DNA sequence analysis. Recombinant proteins PsrA and Vfr (rPsrA and rVfr) were expressed in E. coli BL21(DE3) at 30°C for 3 h after IPTG (isopropyl-β-d-thiogalactopyranoside; 0.5 mM) induction at an OD₆₀₀ of 0.5 to 0.6. The bacterial pellet was resuspended in sonication buffer (25 mM HEPES [pH 7.3], 250 mM NaCl, 5 mM imidazole, 1 mM 4-[2-aminoethyl]-benzenesulfonylfluoride, and 10 μM pepstatin A), sonicated, centrifuged at 30,000 × g for 20 min, and then purified by using a nickel affinity column (Amersham Biosciences) according to the manufacturer's instructions. Recombinant proteins were eluted by imidazole, analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in combination with Coomassie staining, and confirmed by Western blotting using anti-His₆ antibody (QIAGEN, Courtaboeuf, France).

Gene reporting analysis.

Bacterial strains were grown in LB with aeration at 37°C for 14 to 16 h. Samples were precultured at 1:100 in LB for 3 h to an OD₆₀₀ of 1 to 1.5 and then cultured from an OD₆₀₀ of 0.05 in noninducing or inducing conditions. Relative luciferase units (RLU) of 0.2 ml of culture were measured at 1-h intervals starting at time zero by using a spectrophotoluminometer (Luminoskan Ascent; Labsystems, Helsinki, Finland); bacterial quantity at each time point was determined by measuring the OD₆₀₀ (we consider that there are 6 × 10⁸ bacteria in 1 ml of LB at an OD₆₀₀ of 1). Transcription level is expressed as the number of RLU per 6 × 10⁸ bacteria as noted.

To measure transcription in the presence of serum AB, precultured CHA(pC::lux), ΔpsrA(pC::lux), and ΔpsrA-psrAp(pC::lux) strains were washed twice with mHBS (3) and plated at 5 × 10⁵ cells/well in a 96-well culture plate with or without 10% human serum AB. RLU readings from the entire culture plate were taken immediately after addition of bacteria (time [t] = 0) and at 1-h intervals.

To measure transcription in response to host cell contact, differentiated PLB-985 cells (1) were washed twice with modified HEPES-buffered saline (mHBS), plated at 10⁵ cells/well in a 96-well culture plate, and then infected with precultured strain CHA(pC::lux) or ΔpsrA(pC::lux) at a multiplicity of infection of 5. Control infections were set up under identical conditions in the absence of PLB-985 cells. RLU readings from the entire culture plate were taken immediately after addition of bacteria (t = 0) and at 1-h intervals.

Functionality of the TTSS.

To examine the functionality of the TTSS, we analyzed the secreted proteins. Bacterial strains were cultured in LB containing appropriate antibiotics at 37°C overnight. The overnight bacterial cultures were washed twice with LB, reinoculated at 1:100 into inducing or noninducing LB, and then vigorously shaken at 37°C for 3 h. Supernatant of the bacterial culture (equivalent to 1 ml at an OD₆₀₀ of 1) was collected, and trichloroacetic acid was added to 13% to precipitate proteins at 4°C for 30 min. The precipitated proteins were collected by centrifugation at 15,000 × g for 15 min, washed twice with cold acetone, dried in air, resuspended in SDS-PAGE loading buffer and subjected to SDS-10% PAGE, and visualized by using Coomassie staining.

Resistance to phagocyte killing.

PLB-985 cells were maintained in RPMI 1640 (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum, 100 U of penicillin ml⁻¹, 100 μg of streptomycin ml⁻¹, and 2 mM l-glutamine at 37°C in a 5% CO₂ atmosphere. To obtain differentiated granulocytes, PLB-985 cells (5 × 10⁵ cells/ml) were exposed to 0.5% dimethylformamide (Sigma Chemical Co., St. Louis, Mo.) for 5 to 7 days as described previously (1). To measure the resistance to the bacterial killing activity of differentiated PLB-985 cells, the PLB-985 cells were washed twice with mHBS and 5 × 10⁶ cells were added into the culture tube and then infected with strain CHA(pC::lux), ΔpsrA(pC::lux), or ΔexsA at a multiplicity of infection of 5 in the presence of 10% human serum AB in a 1-ml reaction mixture volume. Two 1:10⁵ dilutions were prepared after different intervals and then plated (50 μl) onto a Pseudomonas isolation agar plate. CFU were counted after 16 h of incubation at 37°C.

RESULTS

PsrA was identified from purified ΔexsA strain extracts possessing pC binding activity. To isolate the protein(s) possibly implicated in the regulation of exsCEBA expression, an EMSA was performed. For this experiment, we used the 280-bp pC fragment containing the pC. It has been shown previously that this pC fragment fused to gfp can be used to monitor transcriptional regulation of the exsCEBA operon (4). The pC fragment was obtained from digestion of pIApC with BamHI and then was end labeled with biotin-16-dUTP as described in Materials and Methods. To exclude the pC binding activity due to ExsA, the strain ΔexsA (3) was used to produce extracts which were then purified by heparin-Sepharose affinity chromatography. Proteins from the fraction (eluted from 0.44 M NaCl) possessing the strongest pC binding activity and those from the fraction (eluted from 0.46 M NaCl) with no pC binding activity (Fig. 1A) were separated by SDS-10% PAGE and stained with Coomassie blue (Fig. 1B). The bands present in the active fraction and absent in the adjacent nonactive fraction were sliced, and a matrix-assisted laser desorption-time of flight (MALDI-TOF) analysis was performed. Results of the MALDI-TOF analysis were as follows, where PA refers to the P. aeruginosa assigned numbers in www.pseudomonas.com: band 1, mixture of PsrA (PA 3006) and Vfr (PA 0652); band 2, mixture of 50S ribosomal protein L14 (rplN; PA 4253) and a probable aminotransferase (PA 0299); and band 3, mixture of RplN and 50S ribosomal protein L19 (rplS; PA 3742). Because Vfr and PsrA are known transcriptional regulators of P. aeruginosa (25-27, 53), and moreover because Vfr is a proven activator of the TTSS by an unknown mechanism, we were interested to know whether Vfr and/or PsrA play a role in TTSS regulation by direct binding to the pC.

FIG. 1.

Isolation of protein(s) possessing pC binding activity. (A) Results of EMSA of ΔexsA crude extract or purified fractions on the pC. A 280-bp biotin-labeled pC fragment was incubated with ΔexsA extract or a fraction obtained from heparin-Sepharose affinity chromatography in the presence of 2 μg of poly(dI-dC) as described in Materials and Methods. Extract, total extract of ΔexsA; elution 1, preliminarily purified extract from 5-ml column; other lanes, fractions collected from second purification with elution at the indicated NaCl concentration. One microgram of protein was added in each lane. (B) SDS-PAGE analysis of fractions of ΔexsA extract purified using heparin-Sepharose affinity chromatography. The fraction (0.44 M) possessing the pC binding activity and the adjacent fraction (0.46 M) not having any pC binding activity (Fig. 1A) were separated by SDS-10% PAGE and stained with Coomassie blue. The bands present only in the 0.44 M fraction indicated by arrows were sliced, and a MALDI-TOF analysis was performed.

PsrA binds to the pC region.

To determine if PsrA and/or Vfr could bind to the pC, we produced rPsrA and rVfr in E. coli BL21(DE3). Recombinant proteins were purified by using a nickel affinity column according to the instructions of the manual, analyzed by SDS-PAGE, visualized with Coomassie staining, and confirmed by Western blotting using anti-His₆ antibody (data not shown). To verify if rPsrA and/or rVfr is able to bind the pC, an EMSA was performed. Purified rPsrA and rVfr were tested for the ability to bind the 280-bp DNA pC fragment containing the pC. As shown in Fig. 2, the mobility of the pC was retarded in the presence of 1.5 μM rPsrA. It has to be noted that the position at which this retardation occurred is not exactly the same as that due to the preliminarily purified extract from which PsrA was identified. In contrast, no retardation of the pC was observed in the presence of rVfr, although 20 μM cyclic AMP and the mixing of rPsrA and rVfr in an equimolar ratio did not modify the retardation profile of rPsrA alone (data not shown). This suggested that Vfr is not able to bind to the pC. The binding of rPsrA needed relatively high concentrations of recombinant protein, indicating either a low binding affinity for the pC or poor activity of the recombinant protein. In order to determine the binding capacity of our rPsrA on the pC, we also produced a 280-bp psrAp, a known promoter for PsrA (25), and a 280-bp fleQp, a proven promoter having binding sequences for Vfr (6). As shown in Fig. 2, no retardation was observed in the presence of fleQp and rPsrA even with a high concentration of rPsrA. In contrast, retardation was observed with fleQp and rVfr (Fig. 2), indicating (i) that the rVfr protein is correctly folded and active and (ii) that, for the first time, we showed that Vfr is not able to bind the pC. In the same manner, psrAp was retarded in the presence of rPsrA at a concentration 10 times lower than that needed for the pC, indicating that rPsrA is active and that, although the binding is specific, the in vitro binding affinity of PsrA for the pC is low.

FIG. 2.

rPsrA binds to the pC region. An EMSA experiment was performed to show that rPsrA interacts with the pC region. Biotin-labeled DNA fragments containing pC, fleQp, or psrAp were incubated in the presence of 2 μg of poly(dI-dC) with purified rPsrA of different concentrations as indicated in the figure. Elution 1, preliminarily purified extract (1 μg) fraction obtained from the first heparin-Sepharose affinity chromatography purification as described in Materials and Methods. *, rVFR was used for this lane to verify its activity.

PsrA is required for the full transcriptional activation of exsA and exoS.

As PsrA is able to bind the pC, we asked whether PsrA was indeed necessary for the transcription of exsA and logically the TTSS effectors, such as exoS, in response to various known stimuli of the TTSS. From our CHA parental strain we constructed, by allelic exchange, an unmarked in-frame deletion mutation of psrA as described in Materials and Methods, yielding the ΔpsrA strain which was confirmed by PCR (data not shown). In order to study precisely the level of transciption of exsCEBA and exoS promoters, we used the system developed by Laskowski and colleagues. In brief, the plasmids mini-CTX-exsCp-lux and mini-CTX-exsSp-lux (a kind gift of B. Kazmierczak) were used to obtain the CHA wild-type and ΔpsrA strains containing one single copy of transcriptional fusions of exsCEBA or exoS promoters with lux (29). The different strains were cultured from an OD₆₀₀ of 0.05 without or with either calcium depletion (by EGTA) or human serum AB stimuli. As seen in Fig. 3, the transcriptional level for exoS and exsA in the CHA wild type was two- to threefold higher than that in the ΔpsrA strain over 4 h of growth with EGTA. However, the difference in the fold activation of gene expression between the wild-type strain and the ΔpsrA strain was quite low (4.7/3.3 and 11.1/10.4). Note that under these culture conditions, no discernible growth difference was observed between the strains. Thus, it seemed that PsrA is implicated in the regulation of the level of expression of exsA and hence exoS but not in the regulation in response to calcium depletion. It is interesting that, in the presence of the serum stimulation for 4 h, the transcription of exsA was always about fourfold lower in the ΔpsrA strain than in the CHA wild type and the fold activation in ΔpsrA was two times lower than that in the wild type (Fig. 4). Complementation of ΔpsrA in trans with the plasmid pPsrA restored the level of transcription of exsA and exoS in response to both calcium depletion and serum AB stimuli (Fig. 3 and 4). Expression in trans of exsA only partially restored the level of transcription of exsCEBA under the calcium depletion condition; however, it completely restored the level of exoS expression (Fig. 3). Thus, PsrA appeared to be really required for the full expression of exsCEBA and hence the TTSS effectors, such as exoS.

FIG. 3.

Activation of TTSS transcription by calcium depletion requires PsrA. The CHA wild-type, ΔpsrA, ΔpsrA(pPsrA), ΔpsrA(pDD2), and ΔrpoS strains carrying a single copy of the pC::lux or pS::lux transcriptional reporter at its chromosome were grown at 37°C in LB under TTSS-noninducing (black bars) or -inducing (white bars) conditions. RLU were measured over 4 h of culture and calculated into normalized bacterial counts (6 × 10⁸ CFU). Bars represent the means plus the standard deviations of results for triplicate samples and show the averages of results from three independent experiments. *, fold activation of gene expression under inducing conditions compared with that under noninducing conditions. bac, bacteria.

FIG. 4.

Activation of TTSS transcription by serum requires PsrA. Cultured bacteria (5 × 10⁵) of the CHA wild type and strains ΔpsrA and ΔpsrA(pPsrA) carrying a single copy of the pC::lux transcriptional reporter at its chromosome were incubated with mHBS buffer (−) or 10% human serum AB (+) in a 96-well plate. RLU were measured over 4 h of incubation. RLU readings were taken immediately after addition of serum (t = 0) and at 1-h intervals. Bars represent the means plus the standard deviations of results for triplicate samples and show the averages of results from two independent experiments. *, fold activation of gene expression at hour 4 under inducing conditions compared with that under noninducing conditions. H0 to H4, hours 1 to 4.

PsrA is required for activation of exsCEBA transcription upon host cell contact.

Besides calcium depletion growth conditions, the expression of the P. aeruginosa TTSS and secretion of effector proteins could also be induced after contact with tissue culture cells (51). Therefore, we asked whether PsrA is required for such induction by measuring transcription of the pC::lux reporter construct in wild-type and ΔpsrA strains following infection of PLB-985 polymorphonuclear leukocyte-like cells. As seen in Fig. 5, exsCEBA transcription increased about fourfold at hour 4 in CHA in the presence of PLB-985 cells as did that in the ΔpsrA strain; on the contrary, the global transcription level of exsCEBA in the ΔpsrA mutant was twofold lower than that in the wild-type strain.

FIG. 5.

Transcriptional activation of exsCEBA during exposure of bacteria to differentiated PLB-985 cells requires PsrA. RLU produced by 5 × 10⁵ cultured bacteria of CHA and ΔpsrA strains carrying a single copy of the pC::lux transcriptional reporter at its chromosome were measured over 4 h following bacterial interaction with 10⁵ PLB-985 cells or mHBS buffer control in a 96-well plate. RLU readings were taken immediately after the addition of bacteria (t = 0) and at 1-h intervals. Bars represent the means plus the standard deviations of results for triplicate samples and show the averages of results from two independent experiments. *, fold activation of gene expression at hour 4 under inducing conditions compared with that under noninducing conditions. H0 to H4, hours 1 to 4.

Secretion of the type III effectors is decreased in the ΔpsrA strain.

Secretion of TTSS effectors is strongly correlated with P. aeruginosa virulence both in animal models and in studies of human disease (17, 41, 45). Therefore, we examined the secretion of TTSS effector molecules in wild-type CHA, the ΔpsrA strain, and the ΔpsrA strain carrying pPsrA [ΔpsrA(pPsrA)] under TTSS-noninducing or -inducing conditions. As seen in Fig. 6, the ΔpsrA mutant showed markedly decreased secretion of ExoS, ExoT, and PopB proteins compared to the parental CHA strain. Conversely, complementation of ΔpsrA restored the strain's ability to secrete ExoS, ExoT, and PopB proteins whose amounts seemed slightly higher than those of proteins from the wild-type CHA strain, confirming the positive regulatory function of PsrA for the TTSS. However, ectopic expression of psrA cannot overcome the requirement for a TTSS-inducing signal, such as calcium depletion, as demonstrated by the absence of ExoS, ExoT, and PopB secretion when ΔpsrA(pPsrA) was grown under noninducing conditions (Fig. 6).

FIG. 6.

Secretion of the type III effectors is decreased in the ΔpsrA strain. Bacteria were grown under TTSS-noninducing (lanes 1) or TTSS-inducing (lanes 2) conditions. Supernatants normalized to bacterial counts (6 × 10⁸ CFU lane⁻¹) were subjected to trichloroacetic acid precipitation. The extracellular proteins were separated by SDS-10% PAGE and stained with Coomassie blue. M, protein marker. Sizes in kilodaltons are indicated at the left.

Strain ΔpsrA is less resistant to the phagocyte-like PLB-985 cells.

To examine the effect of the psrA mutation on the ability of P. aeruginosa to resist the bacterial killing activity of polymorphonuclear leukocyte-like cells from differentiated PLB-985 cells, CHA, the ΔpsrA strain, and ΔexsA, a TTSS-defective strain, were compared. Figure 7 shows that the ΔpsrA strain is less resistant to the phagocyte-like PLB-985 cells than its parental strain, CHA, as the number of CFU after 80 min of coculture was two times lower for the ΔpsrA strain than for wild-type CHA. Nevertheless, the survival of ΔexsA in the presence of phagocyte-like PLB-985 cells reveals that, besides the TTSS, other factor(s) may also play a role in the resistance of P. aeruginosa to phagocytosis.

FIG. 7.

The ΔpsrA strain is less resistant to phagocyte-like PLB-985 cells. The numbers of viable bacteria were calculated from colony counts made from a 1:10⁵ dilution over 3 h following bacterial interaction with differentiated PLB-985 cells at different intervals. Points represent the means plus the standard deviations of results for duplicate samples and show the averages of results from two independent experiments.

DISCUSSION

The TTSS is known to be tightly controlled by the proteins encoded by the regulatory operon exsCEBA, particularly by the central positive regulator ExsA, though many other proteins, such as Vfr, RtsM, PtrA, and pyruvate dehydrogenase, are implicated in the regulation of the TTSS (5, 16, 29, 53) (Fig. 8). In this work, we describe a P. aeruginosa protein, PsrA, that is required for the full activation of the TTSS regulatory operon exsCEBA and hence the expression of effector genes, although the mutant still responds to depletion of calcium, exposure to serum, and contact with tissue culture cells. It is not surprising that P. aeruginosa has multiple regulatory networks, because 8% of its genome codes for regulatory genes, indicating that P. aeruginosa has dynamic and complicated regulatory mechanisms responding to various environmental signals (15). Because of the requirement for a large number of genes, the construction of the type III secretion apparatus is an energy-expensive process for the bacteria, rendering so important the ability to precisely regulate the TTSS in response to the environmental changes.

FIG. 8.

A model for the regulation of the TTSS. PsrA activates the TTSS through a positive effect (+) on the expression of ExsA and may also regulate another unknown protein(s) (X) which has a role in the TTSS. PsrA also has a positive effect on RpoS, which has a negative effect on the TTSS by an unknown mechanism. Other main molecules involved in the regulation of the TTSS that are mentioned in the text are also indicated.

PsrA was previously described as a regulator involved in positive regulation of RpoS and in negative autoregulation by direct binding to the promoters (25-27). RpoS, a central regulator during stationary phase (30), was identified also as an inhibitor of the TTSS (21) (Fig. 8). The PsrA-dependent induction of the synthesis of RpoS has been shown to take place in the late exponential and stationary phases (27). Here we measured the induction of the TTSS by different stimuli in the first part of the exponential growth phase when no RpoS induction of synthesis could occur. Furthermore, we confirmed the RpoS-dependent inhibition of TTSS synthesis in our strain CHA ΔrpoS(pC::lux) (Fig. 3). This indicated that the action of PsrA on TTSS gene transcription is independent of its role in RpoS regulation. A previous study showed that a ΔpsrA mutant had a 90% reduction in rpoS promoter activity (27); considering also the opposite functions of PsrA and RpoS in TTSS regulation, we could speculate that PsrA should have an even more important regulatory ability for the TTSS than that measured here. It seemed that PsrA is implicated rather in the regulation of the basic level of exsA and exoS expression than in the regulation in response to environmental signals, at least calcium depletion and host cell contact, since there was no or only a small difference in fold activation of gene expression under these inducing conditions compared with that under noninducing conditions for the mutant and the wild-type strain (Fig. 3 and 5). We also always observed that the growth curves of the CHA and ΔpsrA strains were identical, indicating that the deletion of psrA has no metabolic consequences for the CHA strain. Moreover, the complementation in trans by a high-copy-number plasmid harboring the psrA gene controlled by its own promoter showed an augmentation of the type III secretion. In addition, we showed that the decrease of exsCEBA expression due to psrA deletion could not be overcome by expression in trans of exsA, while exoS expression could be restored (Fig. 3). Taken altogether, these results suggest that PsrA is a positive regulator of the TTSS.

Using the PsrA binding motif (G/CAAAC N_{2 to 4} GTTTG/C, where N_{2 to 4} is any two to four nucleotides) to search against the P. aeruginosa genome sequence at http://www.pseudomonas.com revealed that at least 26 genes may be controlled by PsrA, and recently 15 genes were described as target genes of the PsrA regulator, indicating that the PsrA is a global regulator (26). It is reasonable to propose that PsrA could influence the TTSS by activating or repressing genes having a potent regulatory role in TTSS expression (Fig. 8). Identification of such unknown genes as well as the relationship with PsrA will help us to better understand the mechanism by which PsrA affects TTSS expression.

Though the rPsrA in vitro bound the pC with a relatively low affinity, it is still possible that PsrA may in vivo activate the TTSS by directly binding onto it. In fact, pC sequence analysis revealed no presence of a true consensus binding site for PsrA, but some sequences, such as GAAAC at the position −56 from the transcriptional start site, resemble a partial binding site. Considering the complicated regulatory network of the TTSS, it is possible that PsrA may bind the pC in vivo at a higher affinity in the presence of another partner(s) or it may also help another regulator(s) to bind the pC. The copurification of PsrA and Vfr, an activator of the TTSS (53), suggests possible interaction of PsrA and Vfr during the process of regulation of the expression of the TTSS. Exploring the interaction of PsrA with another molecule(s) and the pC would enrich understanding of the adaptation of bacteria to living conditions. The extract of the ΔpsrA strain possessed the same pC binding activity as ΔexsA (data not shown), indicating that the protein(s) having a strong pC binding activity has not been isolated in this work and is possibly the partner of PsrA in TTSS regulation. Work is under investigation to isolate this protein.

We do not know the signal(s) to which PsrA responds, but it may play an important role during infection, as ΔpsrA bacteria are less resistant to phagocyte-like PLB-985 cells. Futhermore, the presence of psrA homologues in other pathogenic bacteria that employ a TTSS, such as P. syringae, raises the possibility that these homologues also play roles in regulating the TTSS. If this proves to be the case, it will be interesting to ask whether the signal(s) transduced by PsrA and its homologues is also conserved among plant and human pathogens.