MucA-Mediated Coordination of Type III Secretion and Alginate Synthesis in Pseudomonas aeruginosa

Abstract

The type III secretion system (T3SS) of Pseudomonas aeruginosa is an important virulence factor. The T3SS of P. aeruginosa can be induced by a low calcium signal or upon direct contact with the host cells. The exact pathway of signal sensing and T3SS activation is not clear. By screening a transposon insertion mutant library of the PAK strain, mutation in the mucA gene was found to cause repression of T3SS expression under both type III-inducing and -noninducing conditions. Mutation in the mucA gene is known to cause alginate overproduction, resulting in a mucoid phenotype. Alginate production responds to various environmental stresses and plays a protective role for P. aeruginosa. Comparison of global gene expression of mucA mutant and wild-type PAK under T3SS-inducing conditions confirmed the down regulation of T3SS genes and up regulation of genes involved in alginate biosynthesis. Further analysis indicated that the repression of T3SS in the mucA mutant was AlgU and AlgR dependent, as double mutants mucA/algU and mucA/algR showed normal type III expression. An algR::Gm mutant showed a higher level of type III expression, while overexpression of the algR gene inhibited type III gene expression; thus, it seems that the AlgR-regulated product inhibits the expression of the T3SS genes. It is likely that P. aeruginosa has evolved tight regulatory networks to turn off the energy-expensive T3SS when striving for survival under environmental stresses.

Pseudomonas aeruginosa is a versatile gram-negative bacterium that is present in soil, marshes, tap water, and coastal marine habitats. Patients with cystic fibrosis (CF), severe burns, or immunosuppression are at particularly high risk of P. aeruginosa infection (41, 45). Among CF patients, P. aeruginosa colonizes inside the thick mucus layer of the airway. In this anaerobic environment, P. aeruginosa overproduces the exopolysaccharide alginate and forms a biofilm which protects the bacterium from reactive oxygen intermediates and inhibits phagocytosis (15). More than 90% of P. aeruginosa strains isolated from CF patients show the mucoid phenotype, due to the overproduction of alginate (37). Clearly, alginate overproduction is a strategy to overcome environmental stresses and a number of stress signals trigger the overproduction of alginate, converting the bacterium to the mucoid phenotype (30).

The genes encoding enzymes for alginate synthesis form an operon (algD operon), and the expression of this operon is under the tight control of several regulators. The key regulatory gene of this operon is the algU gene (also called algT), encoded in an algU operon which consists of algU-mucA-mucB-mucC-mucD. The algU gene encodes a sigma factor, σ²², which autoregulates its own promoter and activates many other genes, including those for alginate biosynthesis (31). The second gene in the algU operon, the mucA gene, encodes a transmembrane protein with a cytoplasmic portion binding to and inactivating AlgU (σ²²). The third gene of the algU operon, the mucB gene, encodes a periplasmic protein, possibly sensing certain environmental signals. Upon sensing certain environmental signals, MucB transduces the signal to MucA, which in turn releases the bound form of AlgU (σ²²), resulting in activation of alginate production (31). The majority of P. aeruginosa isolates from the lungs of older CF patients carry mutations in the mucA or mucB gene and display a mucoid phenotype (29). In the AlgU regulon, two-component regulatory systems AlgB-FimS (27) and AlgR-AlgZ (46) and regulators AlgP (8) and AlgQ (25) are required for alginate synthesis. Among them, AlgR was also shown to be essential for P. aeruginosa pathogenesis (26). An algR mutant is less virulent than a wild-type strain in an acute septicemia infection mouse model (26), and AlgR is also required for twitching motility (42, 43). Proteomic analysis of the algR mutant suggested that AlgR is a global regulator, affecting expression of multiple genes (26).

P. aeruginosa also possesses a type III secretion system (T3SS), including a protein translocation apparatus and effector proteins. Upon establishing intimate contact with host cells, effector proteins are injected into host cells (22). T3SS mediates the killing of macrophages, polymorphonuclear phagocytes, and epithelial cells (4, 5). The T3SS regulon consists of a total of 32 genes in four operons encoding the protein secretion and translocation apparatus and four effector molecules, ExoS, ExoT, ExoU, and ExoY (22). However, so far no strain has been reported to encode all four effectors. Expression of the T3SS regulon can be stimulated by direct contact with the host cell or by growth under a low Ca²⁺ environment (22). ExsA, an AraC-type transcriptional regulator, is the master regulator of the T3SS regulon. ExsA binds to a consensus sequence present upstream of the operons belonging to the T3SS regulon (21). The mechanism by which ExsA senses low Ca²⁺ signal is not known. Recently, it was shown that under T3SS-inducing conditions (low Ca²⁺), the cyclic AMP level increased and a CRP homologue, Vfr, was also required for T3SS activation (44). Vfr is a global regulator which mediates activation of quorum sensing (35), twitching motility (1), type II secretion (44), and repression of flagellum synthesis (7).

In this study, a transposon (Tn) insertional mutant bank of the PAK strain was screened for mutants that are defective in T3SS expression. We found that mutation in the mucA gene suppresses the expression of T3SS genes, greatly reducing the response to low Ca²⁺ signal. Furthermore, the suppression is dependent on the AlgU and AlgR functions. Comparison of global gene expression of the mucA mutant and wild-type PAK under type III-inducing conditions confirmed the above observation. Several groups of genes have been found to be differently expressed in the mucA mutant and PAK, and their possible roles in T3SS expression are discussed.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Plasmids and bacterial strains used in this work are listed in Table 1. Bacteria were gown in Luria broth (LB) at 37°C. Antibiotics were used at the following concentrations: for Escherichia coli, ampicillin at 100 μg/ml, gentamicin at 10 μg/ml, tetracycline at 10 μg/ml, and kanamycin at 50 μg/ml; for P. aeruginosa, carbenicillin at 150 μg/ml, gentamicin at 100 μg/ml, tetracycline at 100 μg/ml, spectinomycin at 200 μg/ml, streptomycin at 200 μg/ml, and neomycin at 400 μg/ml. For β-galactosidase assays, an overnight culture was reinoculated into LB at 1% and supplemented with 5 mM EGTA. Bacteria were grown to an optical density at 600 nm (OD₆₀₀) between 1.0 and 2.0 before β-galactosidase assays (33).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Source or reference
E. coli strains
BW20767/pRL27	RP4-2-Tc::Mu-1 Kan::Tn7 integrant leu-63::IS10 recA1 zbf-5 creB510 hsdR17 endA1 thi uidA (ΔMluI::pir)/pRL27	24
DH5α/λpir	φ80dlacZΔM15 Δ(lacZYA-argF)U169 recA1 hsdR17 deoR thi-1 supE44 gyrA96 relA1/λpir	24
P. aeruginosa strains
PAK	Wild-type P. aeruginosa strain	David Bradley
PAK exsA::Ω	PAK with exsA disrupted by insertion of Ω cassette	12
PAK A44	PAK mucA::Tn5 mutant isolate	This study
PAK A61	PAK mucA::Tn5 mutant isolate	This study
PAK mucA22	Point mutation (ΔG440) in mucA gene of PAK	This study
mucA22 algU::Gm	mucA22 with algU disrupted by insertion of Gm cassette	This study
mucA22 algR::Gm	mucA22 with algR disrupted by insertion of Gm cassette	This study
PAK algU::Gm	PAK with algU disrupted by insertion of Gm cassette	This study
PAK algR::Gm	PAK with algR disrupted by insertion of Gm cassette	This study
Plasmids
pCR2.1-TOPO	Cloning vector for the PCR products	Invitrogen
pHW0005	exoS promoter of PAK fused to promoterless lacZ on pDN19lacZΩ	Unhwan Ha
pHW0006	exoT promoter of PAK fused to promoterless lacZ on pDN19lacZΩ	Unhwan Ha
pHW0024	pscN promoter of PAK fused to promoterless lacZ on pDN19lacZΩ	Unhwan Ha
pHW0032	exsA promoter of PAK fused to promoterless lacZ on pDN19lacZΩ	Unhwan Ha
pUCP19	Shuttle vector between E. coli and P. aeruginosa	38
pWW020	mucA gene on pUCP19 driven by algU promoter	This study
pWW021	mucA gene on pUCP19 driven by lac promoter	This study
pWW025	algU gene on pUCP19 driven by lac promoter	This study
pMMB67EH	Low-copy-number broad-host-range cloning vector	13
pWW022	algR gene on pMMB67EH driven by lac promoter	This study
pEX18Tc	Gene replacement vector, Tcr; oriT+sacB+	19
pEX18Ap	Gene replacement vector, Apr; oriT+sacB+	19
pPS856	Source of Gmr cassette; Apr Gmr	19
algU::Gm-pEX18Tc	algU disrupted by insertion of Gmr cassette on pEX18Tc	This study
algR::Gm-pEX18Ap	algR disrupted by insertion of Gmr cassette on pEX18Ap	This study

Construction of Tn insertional mutant bank.

The P. aeruginosa PAK strain containing the exoT-lacZ fusion plasmid (pHW0006) was grown overnight at 42°C, while E. coli donor strain BW20767/pRL27 was cultured to mid-log phase at 37°C. Cells of the two types of bacteria were washed with LB once to remove antibiotics in the culture medium. About 5 × 10⁸ PAK/pHW0006 cells were mixed with 10⁹ donor E. coli cells, and the mixture was filtered onto a sterile nitrocellulose membrane (pore size, 0.22 μm). The membrane was laid on top of nutrient agar and incubated at 37°C for 7 to 9 h before washing off the bacterial mixture from the membrane with LB. The bacterial suspension was serially diluted with LB and spread on L-agar plates containing spectinomycin at 100 μg/ml, streptomycin at 100 μg/ml, tetracycline at 50 μg/ml, neomycin at 400 μg/ml, and 20 μg of 5-bromo-4-chloro-3-indolyl-β-l-thiogalactopyranoside (X-Gal)/ml, and 2.5 mM EGTA for colony counting as well as mutant screening.

Determination of Tn insertion sites.

To locate the Tn insertion sites of the isolated mutants, the Tn with flanking DNA was rescued as a plasmid from the mutant chromosome. Plasmid rescue was carried out as described elsewhere (24). Briefly, genomic DNA of the Tn insertion mutants was isolated with the Wizard genomic DNA purification kit (Promega) and digested with PstI. The digested DNA was subjected to self-ligation with T4 DNA ligase and electroporated into DH5α/λpir. Plasmids were isolated from the transformants and sequenced with primers tpnRL17-1 (5′-AAC AAG CCA GGG ATG TAA CG-3′) and tpnRL13-2 (5′-CAG CAA CAC CTT CTT CAC GA-3′) for the DNA flanking the two ends of the Tn. The DNA sequences were then compared with the P. aeruginosa genomic sequence by using BLASTN (39).

Generation of knockout mutants.

Chromosome gene knockout mutants were generated as described previously (19). The target genes were amplified by PCR and cloned into pCR-TOPO2.1 (Invitrogen). After subcloning the PCR product into pEX18Tc or pEX18Ap, the target gene was disrupted by insertion of a gentamicin resistance cassette, leaving about 1 kb upstream and downstream of the insertion-mutation site. The plasmids were electroporated into wild-type PAK and single-crossover mutants were selected on LB plates containing gentamicin at 150 μg/ml and tetracycline at 100 μg/ml or carbenicillin at 150 μg/ml. Double-crossover mutants were selected by plating single-crossover mutants on LB plates containing 5% sucrose and gentamicin at 150 μg/ml. In the case of the mucA22 mutant, a 1.8-kb fragment of the mucA gene region was amplified from FRD1 (mucoid strain) (27) genomic DNA, and the fragment was cloned into the HindIII site of pEX18Gm. The plasmid was transformed into P. aeruginosa to select for single crossover mutants on LB agar plates containing gentamicin 150 μg/ml. Single-crossover mutants were plated on L-agar plates containing 5% sucrose to select for double-crossover mutants. The double-crossover mutants were mucoid, and the introduction of the mucA22 mutation was confirmed by sequencing of the mucA gene.

Plasmid constructs for complementation and overexpression.

Construction of reporter fusions between the exsA, exoT, exoS, and pscN genes and promoterless lacZ on pDN19lacZ was described elsewhere (14a). For mucA gene complementation, the mucA gene was amplified from PAK genomic DNA by PCR with primers MucA-1 (5′-CGG ATC CTC CGC GCT CGT GAA GCA ATC G-3′) and MucA-2 (5′-TAC TGC GGC GCA CGG TCT CGA CCC ATA C-3′). The PCR product was cloned into pCR-TOPO2.1 and transformed into E. coli TOP10F′. The obtained plasmid was digested with HindIII-XmnI and cloned into the HindIII-SmaI sites of pUCP19. The mucA gene in the resulting plasmid, pWW021, is driven by a lac promoter on the vector. To generate a mucA gene driven by the algU promoter, the mucA gene on the pCR-TOPO2.1 plasmid was subcloned into the BamHI and XmnI sites of pEX18Tc, resulting in mucA-pEX18Tc. To obtained the algU gene promoter, an 800-bp DNA fragment upstream of the algU gene open reading frame (ORF) was amplified by PCR with primers AlgT1 (5′-CCT TCG CGG GTC AGG TGG TAT TCG AAG C-3′) and AlgT2 (5′-TTG GAT CCG CGC TGT ACC CGT TCA ACC A-3′) and cloned into pCR-TOPO2.1. Then, this fragment was ligated into the EcoRI and BamHI sites upstream of the mucA gene on the plasmid mucA-pEX18Tc. The obtained plasmid was digested with EcoRI-XmnI, and the algU promoter and mucA gene ORF fragment were cloned into the EcoRI-SmaI sites of pUCP19. On the resulting plasmid (pWW020), the mucA gene is driven by the algU promoter and the transcriptional direction is opposite to that of the lac promoter on the vector.

For algR complementation, the algR gene was amplified from PAK genomic DNA by PCR with primers algR1 (5′-GGT CTA GAG GCC GAG CCC CTC GGG AAA G-3′) and algR2 (5′-GTG GAT CCT ACT GCT CTC GGC GGC GCT G-3′). The PCR product was initially cloned into pCR-TOPO2.1. The resulting plasmid was digested with ClaI, blunted ended with Klenow enzyme, and digested with XbaI. The algR gene-containing fragment was ligated into XbaI-SmaI sites of plasmid pMMB67EH, resulting in pWW022, on which the algR gene is driven by the tac promoter on the vector. For algU gene overexpression, the algU gene ORF was amplified from PAK genomic DNA by PCR with primers algU1 (5′-GGG AAA GCT TTT GCA AGA AGC CCG AGT C-3′) and algU2 (5′-GCT TCG TTA TCC ATC ACA GCG GAC AGA G-3′). The algU gene was cloned into HindIII-EcoRI sites of pUCP19, where the expression of the algU gene in the resulting plasmid pWW025 was driven by lac promoter on the vector.

Western blotting.

Overnight cultures of PAK and PAKmucA22 were subcultured (diluted to 1%) into LB or LB plus 5 mM EGTA. When the OD₆₀₀ of the culture reached 2.0 to 3.0, bacterial cultures were centrifuged at 10,000 × g for 3 min. The supernatants were directly mixed with loading buffer. The pellets were resuspended with phosphate-buffered saline and then mixed with loading buffer. Samples from equivalent numbers of bacterial cells were loaded and separated by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis. Proteins were transferred onto a nitrocellulose membrane, hybridized with anti-ExoS polyclonal antibody, and detected by the ECL method (Amersham).

RNA isolation and microarray analysis.

For RNA isolation, three single colonies of PAK and isogenic mutant PAKmucA22 were inoculated into 3 ml of LB and grown overnight. PAK and PAKmucA22 were subcultured into LB containing 5 mM EGTA. PAK started with an OD₆₀₀ of 0.03, and the mucA22 mutant started with an OD₆₀₀ of 0.06. After 3 to 4 h of culture, bacteria were harvested at an OD₆₀₀ of 1.0 to 1.2. Total RNA was isolated using an RNeasy mini kit (QIAGEN) according to the manufacturer's instructions. The purity and quantity were determined by spectrometry and electrophoresis. Fifteen micrograms of RNA of each sample was used for cDNA synthesis. cDNA fragmentation and biotin terminal labeling were carried out as instructed (Affymetrix). Microarray analysis was performed with the Affymetrix GeneChip P. aeruginosa genome array. The experimental procedure followed the manufacturer's instructions. Data were acquired and analyzed with Microarray Suite version 5.0 (Affymetrix). Significance analysis of microarrays (40) was used to detect differentially expressed ORFs. Then, a cutoff of 5% false discovery rate and a change greater than threefold were chosen to analyze the data.

RESULTS

Activation of T3SS requires a functional mucA gene.

To identify P. aeruginosa genes that affect the expression of T3SS, a Tn insertion mutant bank was constructed in PAK containing an exoT::lacZ (transcriptional fusion) reporter plasmid (pHW0006) (see Materials and Methods). On plates containing X-Gal and EGTA, the density of the blue color of each colony indicated the expression level of the exoT gene in that particular Tn insertion mutant. To identify optimal screening conditions, combinations of different concentrations of X-Gal and EGTA were tested. In the presence of 20 μg of X-Gal/ml and 2.5 mM EGTA, wild-type PAK and the type III-defective PAKexsA mutant harboring pHW0006 showed the greatest visual difference in colony color (blue) and thus these concentrations were adopted for the screening conditions. The mutant bank cells were grown on the screen plates, and we looked for colonies with lighter blue color. About 15,000 Tn insertion mutants were screened. Among four colonies with lighter blue color, two of them showed a mucoid phenotype. The mucoid mutants were picked to test their T3SS activity by β-galactosidase assay. As shown in Fig. 1A, the exoT gene promoter activity was three- to fourfold lower in the mutants than in the parent strain PAK/pHW0006. To confirm this observation, the exoT-lacZ reporter plasmid was cured from the Tn insertion mutants by passage in the absence of antibiotic selection and the pscN::lacZ reporter plasmid (pHW0024) was reintroduced. The resulting strain was subjected to a β-galactosidase assay. The assay results shown in Fig. 1B indicated that the expression of the pscN gene was also repressed in these mucoid mutants under both T3SS-inducing and -noninducing conditions. Similar results were also obtained by introducing exsA::lacZ (pHW0032) and exoS::lacZ (pHW0005) reporter plasmids and testing β-galactosidase activities (data not shown), confirming that the two Tn mutants were indeed defective in T3SS expression.

FIG. 1.

Expression of type III secretion genes in Tn insertional mutants of mucA. PAK, PAKexsA, and mucA mutants A44 and A61 harboring pHW0006 containing exoT::lacZ (A), pHW0032 containing exsA::lacZ (B), pHW0005 containing exoS::lacZ(C), or pHW0024 containing pscN::lacZ (D) were tested for β-galactosidase activities. Bacteria were grown in LB (white bars) or LB containing 5 mM EGTA (black bars) to an OD₆₀₀ of 1 to 2 before β-galactosidase assays. Each assay was done in triplicate, and the error bars indicate standard deviations.

The Tn and flanking DNA were rescued from the mutant strains and subjected to sequencing analysis (see Materials and Methods). Sequencing results showed that the Tn was inserted into two different positions on the mucA gene in these two mutants, explaining the mucoid phenotype of the isolates. Mutation in the mucA gene is commonly observed among P. aeruginosa isolates from CF patients, such as mucA22, where a nucleotide G was deleted within five G residues between positions 429 and 433 of the mucA coding region, causing protein truncation (3, 37). The identical mucA22 mutant was constructed in the background of PAK by allelic replacement with a mucA fragment amplified from FRD1 (27), which bears the mucA22 mutation (see Materials and Methods). Expression of the effector genes exoS and exoT in the resulting mutant strain PAKmucA22 was compared to that in PAK by Western blot analysis of the secreted and cell-associated proteins by using anti-ExoS antibody, which also cross-recognizes ExoT due to a high sequence homology between the ExoS and ExoT proteins. As shown in Fig. 2A, expression of ExoS and ExoT in the resulting PAKmucA22 was greatly reduced in comparison to that in wild-type PAK when grown under type III-inducing conditions. Reporter plasmids pHW0032 (exsA::lacZ) and pHW0005 (exoS::lacZ) were further introduced into PAKmucA22 and tested for β-galactosidase activity. Similar to the original isolates of the mucA Tn insertional mutants, expression of the exsA and exoS genes in PAKmucA22 was almost nonresponsive to low Ca²⁺ signal, compared to a three- to fourfold induction in the wild-type PAK background (Fig. 2B and C). Upon complementation of the PAKmucA22 mutant with the mucA gene in pUCP19, either driven by the algU promoter (pWW020) or lac promoter (pWW021), expression of the exsA and exoS genes in the resulting strains was restored to that of the wild type (Fig. 2C). These results clearly demonstrate that expression of the T3SS genes requires a functional mucA gene.

FIG. 2.

Expression and secretion of ExoS protein. (A) Comparison of cellular and secreted forms of ExoS in strains PAK and PAKmucA22 grown in LB or LB plus 5 mM EGTA. Supernatants and pellets from equivalent bacterial cell numbers were loaded onto sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels and immunoblotted with anti-ExoS antibody. Both ExoS and ExoT are indicated by the arrow. Anti-ExoS polyclonal antibody also recognizes ExoT due to high homology between these two proteins. (B) Expression of exsA::lacZ (pHW0032) in the backgrounds of PAK, PAKmucA22, and PAKmucA22algR::Gm with or without the mucA clone driven by an algU promoter (pWW020) or lac promoter (pWW021). (C) Expression of exoS::lacZ (pHW0005) in the same backgrounds as described above. Bacteria were grown to an OD₆₀₀ of 1 to 2 in LB with (black bars) or without (white bars) EGTA before β-galactosidase assays.

Microarray analysis of gene expression in the mucA mutant.

To further understand the mechanism of MucA-mediated regulation of T3SS genes, global gene expression profiles were compared between PAKmucA22 and its wild-type parent strain PAK grown under T3SS-inducing conditions. Previously, a microarray analysis compared global gene expression patterns between mucoid (mucA mutant) and wild-type P. aeruginosa under non-T3SS-inducing conditions (10). Under these conditions, the T3SS activity in both strains was low; thus, no obvious effect of the mucA gene on the T3SS was observed.

Results of our gene array analysis were consistent with the published data (15, 31); genes under the control of AlgU are up regulated in a PAKmucA22 mutant background compared to that in wild-type PAK, including genes for alginate biosynthesis (operon PA3540-3551) and regulation (Table 2). Also up regulated were an operon, PA4468-4471, consisting of the sodM gene (PA4468) encoding manganese superoxide dismutase, whose production is known to be higher in mucoid than nonmucoid P. aeruginosa (18), and the fumC gene (PA4470) encoding a tricarboxylic acid cycle enzyme fumarase C, which is essential for alginate production (17). Their results validated our gene array data.

TABLE 2.

Expression of AlgU regulon genes in PAKmucA22

Group and ID no.	Name	Function	Fold change in mucA22 vs wild type
Alginate biosynthesis genes
PA3540	algD	Alginate biosynthesis	64.2
PA3541	alg8	Alginate biosynthesis	29.9
PA3542	alg44	Alginate biosynthesis	28.9
PA3543	algK	Alginate biosynthesis	81.2
PA3544	algE	Alginate biosynthesis	47.9
PA3545	algG	Alginate biosynthesis	38.0
PA3546	algX	Alginate biosynthesis	86.0
PA3547	algL	Alginate biosynthesis	43.7
PA3548	algI	Alginate biosynthesis	55.2
PA3549	algJ	Alginate biosynthesis	27.2
PA3550	algF	Alginate biosynthesis	70.5
PA3551	algA	Phosphomannose isomerase	38.7
Alginate biosynthesis regulatory genes
PA0762	algU	Sigma factor	2.6
PA0763	mucA	Anti-sigma factor	2.4
PA0764	mucB	Negative regulator for alginate biosynthesis	1.3
PA5261	algR	Alginate biosynthesis; two-component system	1.5
PA5483	algB	Alginate biosynthesis; two-component system	2.0
PA5484	kinB	Two-component sensor	2.1
Genes known to be up regulated in mucA mutant
PA0059	osmC	Osmotically inducible protein	3.8
PA0376	rpoH	Sigma factor	1.3
PA4876	osmE	Osmotically inducible lipoprotein	3.0
PA5489	dsbA	Thiol:disulfide interchange protein	1.3

Meanwhile, the expression levels of exoS, exoT, exoY, and other T3SS-related genes were clearly down regulated in the mucA mutant background compared to those in wild-type PAK under T3SS-inducing conditions (Table 3), which confirmed our β-galactosidase assay and the Western blotting results. However, no significant changes in the expression of the exsA gene and a few other T3SS genes were observed. A previous gene array study also showed that expression of the exsA gene and the exsD-pscL operon is relatively nonresponsive to Ca²⁺ depletion (44), yet a clear difference in the β-galactosidase activities could be observed when PAK harboring exsA::lacZ (pHW0032) was grown in LB with or without EGTA. Similarly, we have seen differences in the β-galactosidase activities between PAK(pHW0032) and PAKmucA22(pHW0032) under type III-inducing conditions without observing such differences in gene array data, suggesting possible involvement of posttranscriptional control of the exsA gene.

TABLE 3.

Expression of T3SS-related genes in PAKmucA22

ID no.	Gene	Function	Fold change in mucA22 vs wild type
PA0044	exoT	Exoenzyme T	−2.0
PA2191	exoY	Adenylate cyclase	−1.3
PA3841	exoS	Exoenzyme S	−2.1
PA1707	pcrH	Regulatory protein	−1.4
PA1708	popB	Translocator protein	−1.6
PA1709	popD	Translocator outer membrane protein	−1.5
PA1718	pscE	Type III export protein	−1.4
PA1719	pscF	Type III export protein	−1.5

From the microarray analysis, genes that are differentially expressed more than threefold between PAKmucA22 and PAK are listed in Tables 4 and 5. A number of genes known to be inducible under iron deprivation were also elevated in the mucA22 mutant, including the sigma factor PvdS and genes regulated by PvdS for pyoverdine synthesis (17), the operon PA4468-4471 (17), and the probable two-component regulatory genes PA1300 and PA1301, encoding the extracytoplasmic function sigma-70 factor and a transmembrane sensor, respectively (34). Compared to the global gene expression profile of PAK grown under T3SS-inducing or -noninducing conditions, none of the above genes seem to be affected by Ca²⁺ depletion (44). The mechanism by which these genes are activated is not clear.

TABLE 4.

Genes up regulated in PAKmucA22

ID no.a	Gene	Function	Fold change in mucA22 vs wild type	TSBb (fold)	LBb (fold)
PA0059	osmC	Adaptation, protection	3.75	1.23	−1.29
PA2386	pvdA	Adaptation, protection	3.89	−1.50	1.07
#PA2397	pvdE	Adaptation, protection, membrane proteins, transport of small molecules	3.92	−3.00	−1.99
PA2401		Adaptation, protection	3.06	−1.00	−5.74
PA4468	sodM	Adaptation, protection	5.60	−1.30	1.02
PA2018		Antibiotic resistance and susceptibility, membrane proteins, transport of small molecules	3.88	−1.50	1.65
PA2019		Antibiotic resistance and susceptibility, transport of small molecules	4.16	−1.50	−1.36
PA1985	pqqA	Biosynthesis of cofactors, prosthetic groups, and carriers	2.99	−1.10	−1.24
PA1988	pqqD	Biosynthesis of cofactors, prosthetic groups, and carriers	3.18	−1.50	−1.28
PA1989	pqqE	Biosynthesis of cofactors, prosthetic groups, and carriers	3.00	1.44	−1.34
#PA2414		Carbon compound catabolism	5.17	−3.00	−1.29
PA3158	wbpB	Cell wall, LPS, and capsule; putative enzymes	5.76	−1.20	−1.08
PA0102		Central intermediary metabolism	3.38	−2.40	−1.15
PA2393		Central intermediary metabolism	3.27	−1.70	−4.76
PA2717	cpo	Central intermediary metabolism	4.75	1.21	−1.72
PA4470	fumC1	Energy metabolism	5.81	1.05	−1.62
PA5491		Energy metabolism	2.97	−1.30	1.08
#PA0320		Hypothetical	3.86	−7.80	−1.43
PA0586		Hypothetical	5.10	1.61	1.96
PA0587		Hypothetical	4.57	1.11	1.64
PA0588		Hypothetical	4.57	1.12	1.64
PA0613		Hypothetical	3.60	1.14	−1.55
#PA0737		Hypothetical	10.80	1.70	−1.00
PA0807		Hypothetical	3.97	1.21	−1.52
PA0990		Hypothetical	3.45	2.31	−1.48
PA1245		Hypothetical; membrane proteins	5.12	−1.10	−1.20
PA1323		Hypothetical	4.21	1.94	2.25
#PA1471		Hypothetical	12.4	1.46	1.07
#PA1784		Hypothetical	6.71	1.33	−1.31
PA1852		Hypothetical	3.34	1.90	−1.03
PA2159		Hypothetical	3.56	1.44	−2.46
PA2161		Hypothetical	2.95	1.50	−4.92
#PA2167		Hypothetical	9.54	1.16	−1.06
PA2168		Hypothetical	3.20	−1.20	−2.80
#PA2172		Hypothetical	3.86	−1.90	−1.99
PA2176		Hypothetical	7.30	1.79	2.10
PA2403		Hypothetical; membrane proteins	4.96	−1.30	−1.52
PA2404		Hypothetical; membrane proteins	5.83	1.27	−1.69
#PA2405		Hypothetical	10.00	1.10	−5.91
PA2406		Hypothetical	5.79	−1.10	−1.22
PA2412		Hypothetical	4.56	−1.30	−2.72
PA2485		Hypothetical	5.16	−1.20	5.51
PA2486		Hypothetical	6.30	1.17	1.41
PA2562		Hypothetical	3.30	1.27	2.38
PA3274		Hypothetical	5.30	1.61	−2.16
PA4154		Hypothetical	3.63	1.49	1.08
#PA4469		Hypothetical	8.78	−1.30	−1.64
PA4471		Hypothetical	3.00	−1.00	−3.49
PA5182		Hypothetical; membrane proteins	4.81	1.29	2.11
PA5183		Hypothetical; membrane proteins	3.88	1.20	2.02
PA5212		Hypothetical	3.24	1.32	1.83
PA2409		Membrane proteins, transport of small molecules	4.21	−1.20	−1.46
PA4876	osmE	Membrane proteins, adaptation, protection	3.00	1.36	1.27
PA2407		Motility and attachment	4.73	−1.20	−1.61
PA2385		Putative enzymes	3.19	1.05	−2.36
PA2394		Putative enzymes	3.56	1.00	−1.75
PA2402		Putative enzymes	3.18	−1.70	−2.00
PA2413		Putative enzymes	5.97	1.46	−2.32
PA4785		Putative enzymes	4.51	1.70	−1.98
PA0724		Related to phage, transposon, or plasmid	3.48	1.83	−2.39
PA1300		Transcriptional regulators	3.66	−1.70	−1.13
PA2426	pvdS	Transcriptional regulators	4.39	−1.40	−1.38
PA2408		Transport of small molecules	4.09	2.08	−3.14
PA3049	rmf	Translation, posttranslational modification, degradation	6.81	2.03	1.38
PA3188		Transport of small molecules	3.08	4.04	15.76
PA5470		Translation, posttranslational modification, degradation	3.31	−1.50	1.15
#PA2398	fpvA	Transport of small molecules	6.65	−2.10	−1.51

^a#, up regulated in mucA mutant compared to PAK, but down regulated in PAK under type III-inducing conditions versus noninducing conditions, and vice versa. Not included are those known to be affected by the growth medium, such as those varied in TSB versus LB (44).

^bChange in gene expression in PAK grown under T3SS-inducing conditions versus PAK grown under non-T3SS-inducing conditions (44). Bacteria were grown in TSB or LB.

TABLE 5.

Genes down regulated in PAKmucA22

ID no.a	Gene	Function	Fold change in mucA22 vs wild type	TSBb (fold)	LBb (fold)
PA3450		Adaptation, protection	−3.5	1.84	1.17
PA2138		DNA replication, recombination, modification, and repair	−3.2	−2.50	−3.19
PA0523	norC	Energy metabolism	−3.0	−1.70	−1.7
#PA3445		Hypothetical	−3.9	2.66	1.59
#PA3446		Hypothetical	−5.1	1.36	1.57
PA3931		Hypothetical	−3.8	1.91	1.22
PA0281	cysW	Membrane proteins, transport of small molecules	−3.5	1.18	1.13
PA0282	cysT	Membrane proteins, transport of small molecules	−3.0	−1.10	−1.07
PA1601		Putative enzymes	−3.5	1.35	−1.44
PA3444		Putative enzymes	−5.1	1.11	−1.51
PA1246	aprD	Secreted factors (toxins, enzymes, alginate); protein secretion-export apparatus	−3.4	2.05	−2.64
PA1312		Transcriptional regulators	−3.1	−1.10	−1.55
PA3927		Transcriptional regulators	−5.1	−1.70	−1.67
PA0198	exbB1	Transport of small molecules	−5.5	2.71	−3.68
PA0280	cysA	Transport of small molecules	−6.0	−1.00	−1.18
PA2204		Transport of small molecules	−4.7	1.16	−1.16

^a#, up regulated in mucA mutant compared to PAK but down regulated in PAK under type III-inducing conditions versus noninducing conditions and vice versa. Not included are those known to be affected by the growth medium, such as those varied in TSB versus LB (44).

^bChange in gene expression in PAK grown under T3SS-inducing conditions versus PAK grown under non-T3SS-inducing conditions (44). Bacteria were grown in TSB or LB.

T3SS repression in the mucA mutant is AlgU dependent.

MucA is an anti-sigma factor which represses the activity of AlgU (σ²²). In the mucA mutant, AlgU is derepressed and activates the expression of genes for alginate synthesis, resulting in a mucoid phenotype. AlgU can also activate the expression of itself and downstream genes (mucA-B-C-D) in the same operon. To determine the role of AlgU in the repression of T3SS in the mucA mutant, the algU gene was knocked out in the background of PAKmucA22, resulting in a PAKmucA22algU::Gm double mutant. Under T3SS-inducing conditions, expression of the exsA and exoS genes in this double mutant was similar to that in the wild type (Fig. 2B and C), indicating that AlgU is required for the T3SS repression in the mucA mutant. An algU::Gm mutant was further generated in the background of PAK, and T3SS activity in the resulting mutant was compared with that in PAK. As shown in Fig. 3, expression of the exsA and exoS genes was the same in the PAKalgU::Gm mutant and wild-type PAK under both T3SS-inducing and -noninducing conditions, suggesting that the basal level of AlgU in wild-type P. aeruginosa does not play a significant role in the regulation of T3SS genes. When the algU gene was overexpressed in wild-type PAK by introducing pWW025, T3SS activity was partially repressed under type III-inducing conditions (Fig. 3). Since AlgU mediates the activation of the algU-mucA operon, an extra copy of algU also increased the expression of its repressor MucA; thus, overexpression of the algU gene could not repress T3SS expression to the level seen in the mucA mutant.

FIG. 3.

Expression of exsA::lacZ (A) and exoS::lacZ (B) in strains PAK, PAKmucA22, PAKalgU::Gm, and PAK harboring algU overexpression plasmid pWW025. Bacteria were grown in LB (white bars) or LB plus 5 mM EGTA (black bars) to an OD₆₀₀ of 1 to 2 before β-galactosidase assays. *, P < 0.05.

AlgR has a negative regulatory function on T3SS.

algR is a regulatory gene required for alginate synthesis and is under the control of AlgU (27, 46). To investigate the role of AlgR in the regulation of T3SS, the algR gene was knocked out in the background of PAKmucA22. In the PAKmucA22algR::Gm double mutant, the expression of the exsA and exoS genes was restored to that of the wild type (Fig. 2A and B), suggesting that the repression of T3SS in the mucA mutant is also AlgR dependent. To test the function of AlgR on T3SS in wild-type P. aeruginosa, an algR::Gm mutant was generated in the PAK background. The expression of the exoS gene was consistently higher in the resulting PAKalgR::Gm mutant than in PAK under both type III-inducing and -noninducing conditions (Fig. 4B). However, the expression of the exsA gene was similar in the PAKalgR::Gm mutant and wild-type PAK. Complementation of the algR mutant with an algR-expressing clone (pWW022) decreased exsA and exoS expression under both inducing and noninducing conditions (Fig. 4). However, higher expression of algR induced by increasing the amount of isopropyl-β-d-thiogalactopyranoside (IPTG) could not further decrease exsA and exoS expression (Fig. 4). These results indicate that AlgR has a negative regulatory effect on the T3SS, but the up regulation of AlgR alone might not be sufficient to repress T3SS activity to the level seen in the mucA mutant. It is likely that in the mucA mutant, algR gene expression is activated by AlgU, which in turn represses T3SS activity.

FIG. 4.

Expression of exsA::lacZ (A) and exoS::lacZ (B) in the backgrounds of PAK, PAKmucA22, PAKalgR::Gm, and PAKalgR::Gm complemented with algR-expressing plasmid pWW022. For algR gene complementation, various concentrations of IPTG were added into the culture medium as indicated. Bacteria were grown in LB (white bars) or LB plus 5 mM EGTA (black bars) to an OD₆₀₀ of 1 to 2 before β-galactosidase assays. *, P < 0.05.

DISCUSSION

T3SS is an important virulence machinery for P. aeruginosa: it inhibits the host defense system by inducing apoptosis in macrophages, polymorphonuclear phagocytes, and epithelial cells. In our screen for mutants with lower T3SS activities, mucA mutants were found defective in exoT expression under type III-inducing conditions. Furthermore, the basal promoter activity of the type III master regulatory gene exsA was decreased two- to threefold in the mucA mutant compared to that in wild-type PAK, suggesting that the down regulation of T3SS genes occurs through repression of ExsA. Since ExsA is an autoactivator (21), the repression could be on the transcriptional or posttranscriptional level. Our microarray results showed that the transcript level of exsA in the mucA mutant was similar to that in wild-type PAK under type III-inducing conditions, which suggested that the activity of ExsA might be repressed at the posttranscriptional level. Similarly, a previous microarray study also showed that expression of the exsA gene was not affected by Ca²⁺ depletion (44). Further study is required to clarify the mechanism of exsA gene regulation.

MucA is a transmembrane protein, with its cytoplasmic domain binding to and repressing the sigma factor AlgU. Mutation in the mucA gene leads to derepression of AlgU, which in turn activates genes for alginate synthesis as well as others, such as dsbA, oprF, osmE, and rpoH (10, 28). In the mucA mutant, not only the sigma factor AlgU but also AlgQ, an anti-σ⁷⁰ factor, are activated (9), thus posing the possibility that sigma factor competition by AlgU and AlgQ effectively decreases the availability of σ⁷⁰-containing RNA polymerase for the expression of T3SS genes (23). However, the observation that AlgR, an AlgU-dependent transcriptional activator, is required for the T3SS suppression makes it unlikely that sigma factor competition leads to the type III gene suppression; instead, an AlgR-dependent repressor is likely involved. AlgR is a global regulator, affecting expression of multiple genes. Proteomics analysis of an algR::Gm mutant showed that more than 17 proteins were up regulated and 30 proteins were down regulated (26). In the present study, AlgR was also found to mediate the repression of type III secretion genes. In the PAKalgR::Gm mutant background, expression of the exoS gene was higher than in wild-type PAK and, when complemented by an algR gene clone, expression of exsA and exoS genes decreased to about 50% of that seen in wild-type PAK (Fig. 4). The inability to suppress T3SS genes to the level seen in the mucA mutant by pWW022 was possibly due to a lower level of expression of the algR gene from pWW022 than that in the PAKmucA background, in which algR is activated through the MucA-AlgU pathway. pMMB67HE is a low-copy-number plasmid (13), and the tac promoter is not as strong a promoter in P. aeruginosa as it is in E. coli. Alternatively, other regulatory genes might be involved in the repression of T3SS. Further study is needed to understand this observation.

We propose a model for T3SS repression in the mucA mutant (Fig. 5). With the activation of AlgU, the regulatory genes algP, algQ, algB, and algR are activated, which up regulates the expression of the algD operon. AlgR is required for T3SS repression in the mucA mutant, but whether the repression function is directly on ExsA or not is unclear. The involvement of other regulatory genes (algP, algQ, and algB) in T3SS regulation awaits further study.

FIG. 5.

Proposed model of MucA-mediated coordination of alginate production and T3SS expression. MucA is a transmembrane protein, with its cytoplasmic portion binding and inhibiting the sigma factor AlgU. Upon sensing of certain environmental stress signals by the periplasmic MucB, it signals MucA through the periplasmic domain to release the bound AlgU. Free AlgU is required for the expression of downstream transcriptional activators AlgP, AlgQ, AlgB, and AlgR, all of which contribute to the optimal expression of the algD operon, encoding enzymes for the synthesis of alginate. AlgR, on the other hand, also activates downstream genes which are responsible for the suppression of the type III secretion genes.

During chronic infection of CF patient airways, P. aeruginosa overproduces alginate and forms a biofilm (20). Alginate production is known to be activated by high osmolarity, nitrogen limitation, and membrane perturbation induced by ethanol (2); thus, the high salt concentration in the CF patient airway might be a signal for the overproduction of alginate. The biofilm mode of growth can help the bacterium survive in hostile environments and also render resistance against macrophages and polymorphonuclear cells (20). Our experimental data suggest that bacteria have evolved a mechanism to turn off T3SS when they need to synthesize alginate to overcome environmental stress. Such coordinated regulation of two energy-expensive processes is likely to render to the bacterium a survival advantage under environmental stress conditions. Indeed, the majority of P. aeruginosa isolates from CF patients at a late stage in the disease display the mucoid phenotype (11, 37) and are defective in type III gene expression (5). In a previous report, introduction of the wild-type exsA gene into type III secretion-defective clinical isolates restored type III secretion (5). However, our attempts to restore T3SS gene expression in 10 mucoid CF isolates by introducing a mucA gene clone failed, although all of the transformants were reverted back to the nonmucoid phenotype. It is possible that those mucoid clinical isolates may harbor additional mutations in the T3SS genes.

Known T3SS regulators include ExsA, Vfr, CyaA/B, and ExsD (21, 32, 44). Recently, DsbA and AceA and -B were also found to be necessary for the expression of T3SS. AceA and -B are subunits of pyruvate dehydrogenase, suggesting that metabolic imbalance influences the expression of T3SS (6, 36). DsbA is a periplasmic thiol-disulfide oxidoreductase and was shown to affect T3SS expression, twitching motility, and intracellular survival of P. aeruginosa upon infection of HeLa cells (14, 28). Interestingly, the dsbA gene is up regulated in the mucA mutant background, and its expression was shown to be affected by AlgU (28). However, the role of DsbA on the T3SS is believed to be through its general effect on protein disulfide bond formation in the periplasm, and up regulation of this gene may not be related to the MucA-AlgU-AglR-mediated suppression of the T3SS.

From the microarray analysis of the mucA mutant and wild-type strain under T3SS-inducing conditions, alginate synthesis genes and genes known to be under the control of AlgU were up regulated, while T3SS genes were down regulated in the mucA mutant (Tables 2 and 3). In addition, pyoverdine synthesis genes as well as an operon, PA4468-4471, which might be under the control of Fur (18), were up regulated in the mucA mutant under T3SS-inducing conditions (Table 4). These findings are consistent with published results, in which mucoid P. aeruginosa strains produced higher levels of pyoverdine, pyochelin, manganese superoxide dismutase (PA4468), and fumarase (PA4470) than wild-type strains (16, 17). However, pyochelin synthesis genes were not seen up regulated in our microarray data, suggesting again the possibility of translational-level control for these genes. The mechanism by which these genes are up regulated in the mucA mutant background is not known.

The mucA gene mutation-mediated suppression of the T3SS genes requires AlgR, which is a transcriptional activator; thus, it is likely that an AlgR-regulated repressor mediates the suppression of T3SS genes. To identify such candidate genes from the gene array data, we initially identified genes that were differentially expressed in the mucA mutant compared to wild-type PAK under type III-inducing conditions. Those genes that were up regulated in the mucA mutant compared to PAK under type III-inducing conditions but that were down regulated in PAK under type III-inducing conditions versus noninducing conditions, and vice versa, were identified. We further eliminated those known to be affected by the growth medium, such as those with varied responses in tryptic soy broth (TSB) versus LB (44). Based on the above criteria, 13 genes were identified (Tables 4 and 5). For example, expression of the PA2172 gene in mucA22 was up regulated about fourfold compared to wild-type PAK under T3SS-inducing conditions. From published data, the expression of this gene was down regulated twofold in wild-type PAK grown under type III-inducing conditions compared to that under noninducing conditions (44). Therefore, mutation in the mucA gene reversed the expression of PA2172 in response to the type III-inducing signal.

Among the 13 genes, pvdE and fpvA are involved in pyoverdine synthesis and absorption, respectively, and PA2414 is involved in carbon compound catabolism. The remaining 10 genes are all hypothetical genes. The expression of PA0737, PA2167, PA2176, and PA4785 seems to be ExsA dependent, since in the exsA mutant the expression of these genes was lower than in wild-type PAK under type III-inducing conditions and overexpression of exsA could activate expression of these genes under non-type III-inducing conditions (44). It is reasonable to hypothesize that one or more such differentially expressed genes mediate the repression of the T3SS in the mucA mutant. Attempts are under way to understand these possibilities by directed gene knockout of the identified candidate genes as well as screening of a random Tn insertional mutant bank generated in the background of PAKmucA22 for those mutants with restored wild-type T3SS activity.