ABSTRACT
Proteases play important roles in the virulence of Pseudomonas aeruginosa. Some are exported to act on host targets and facilitate tissue destruction and bacterial dissemination. Others work within the bacterial cell to process virulence factors and regulate virulence gene expression. Relatively little is known about the role of one class of bacterial serine proteases known as the carboxyl-terminal processing proteases (CTPs). The P. aeruginosa genome encodes two CTPs annotated as PA3257/Prc and PA5134/CtpA in strain PAO1. Prc degrades mutant forms of the anti-sigma factor MucA to promote mucoidy in some cystic fibrosis lung isolates. However, nothing is known about the role or importance of CtpA. We have now found that endogenous CtpA is a soluble periplasmic protein and that a ctpA null mutant has specific phenotypes consistent with an altered cell envelope. Although a ctpA null mutation has no major effect on bacterial growth in the laboratory, CtpA is essential for the normal function of the type 3 secretion system (T3SS), for cytotoxicity toward host cells, and for virulence in a mouse model of acute pneumonia. Conversely, increasing the amount of CtpA above its endogenous level induces an uncharacterized extracytoplasmic function sigma factor regulon, an event that has been reported to attenuate P. aeruginosa in a rat model of chronic lung infection. Therefore, a normal level of CtpA activity is critical for T3SS function and acute virulence, whereas too much activity can trigger an apparent stress response that is detrimental to chronic virulence.
INTRODUCTION
Pseudomonas aeruginosa is a ubiquitous Gram-negative bacterium and an opportunistic pathogen responsible for acute and chronic infections in both the community and health care settings. It is a prolific protein exporter with many virulence factors secreted by specialized machineries (1, 2). In fact, P. aeruginosa possesses all of the known secretion systems described in Gram-negative bacteria with the exception of type 4 secretion (2). Among these is the Psc type 3 secretion system (T3SS), which is critical for the virulence of P. aeruginosa in acute infections (3). There are only four known substrates exported by this T3SS, i.e., ExoS, -T, -Y, and -U, which play specific roles due to their different targets and mechanisms of action. However, most strains do not encode all four of these effectors (4–6). ExoS and ExoT are homologous dual-function proteins, each with GTPase-activating and ADP ribosyltransferase activities. They interfere with phagocytosis and host cell signaling and cause cytotoxicity (3). ExoU is a cytotoxic phospholipase (7), and ExoY is an adenylyl cyclase that upsets cyclic AMP (cAMP)-dependent signaling in host cells (8).
In addition to acute infections, P. aeruginosa is a notorious cause of chronic lung infections in people with cystic fibrosis (CF) (9). The lungs of individuals with CF are colonized by P. aeruginosa strains that often convert to a mucoid phenotype after prolonged infection. This mucoid conversion is caused by constitutive production of the polysaccharide alginate and is associated with a poor prognosis (10). The alginate biosynthesis genes are controlled by the AlgU/T extracytoplasmic function sigma factor (ECFσ), and the most common cause of mucoid conversion is a mutation that inactivates its inhibitory anti-sigma factor, MucA (11–13). P. aeruginosa also has 18 other putative ECFσ factors in addition to AlgU/T, most of which are not well characterized (14, 15).
Both the acute and chronic modes of P. aeruginosa virulence are influenced by proteases, including some that are exported and have destructive effects on host tissues (16). Proteases also control the wild-type AlgU/T system by regulated destruction of MucA, which can be triggered by d-cycloserine-induced cell envelope stress in the laboratory (17, 18). A protease named Prc has also been implicated in contributing to the mucoid conversion phenotype by degrading mutant forms of MucA that arise in CF lung isolates (19, 20). Prc is encoded by the gene annotated as PA3257 in strain PAO1 and is a periplasmic protease similar to Escherichia coli Prc/Tsp (tail-specific protease). Prc is a carboxyl-terminal protease (CTP), defined by a conserved serine/lysine catalytic dyad, cleavage within the C-terminal region of substrates, and the presence of a PDZ domain that is implicated in binding to nonpolar C termini of substrates (21, 22). E. coli Prc processes penicillin-binding protein 3 (23–25), degrades the phage λ repressor (26), and cleaves incorrectly synthesized proteins with a C-terminal Ssr tag (27). Additionally, in some pathogens CTPs affect virulence (28–30). However, our knowledge of bacterial CTPs is quite limited, and in most cases there has been no explanation for their effects on virulence.
Unlike E. coli K-12, sequenced P. aeruginosa genomes encode two putative CTPs, PA3257/Prc and PA5134/CtpA (31). Prc is in the CTP-1 subfamily and is approximately 30 kDa larger than CtpA, which is in the CTP-3 subfamily (31). As mentioned above, P. aeruginosa Prc has been implicated in mucoid conversion, but the only thing known about P. aeruginosa CtpA is that the overproduced protein is periplasmic (31). However, in this study we report that CtpA is essential for the normal function of the T3SS, for cytotoxicity toward CHO-K1 cells, and for virulence in a mouse model of acute pneumonia. Furthermore, elevating the level of CtpA induces an uncharacterized ECFσ regulon that is known to attenuate P. aeruginosa virulence in a rat model of chronic lung infection (32). These phenomena point to an important role for CtpA in the P. aeruginosa cell envelope with clear links to virulence.
MATERIALS AND METHODS
Bacterial strains and routine growth.Strains and plasmids are listed in Table 1. Bacteria were grown routinely in Luria-Bertani (LB) broth or on LB agar plates at 37°C. In some cases, P. aeruginosa was grown on Vogel-Bonner minimal (VBM) base agar or Pseudomonas isolation agar (Difco). All P. aeruginosa strains are derived from strain PAK (33). E. coli K-12 strain SM10 was used for conjugation of plasmids into P. aeruginosa (34). The following antibiotics were used: ampicillin (200 μg/ml for E. coli), tetracycline (15 μg/ml for E. coli and 75 μg/ml for P. aeruginosa), gentamicin (15 μg/ml for E. coli and 75 μg/ml for P. aeruginosa), carbenicillin (150 μg/ml for P. aeruginosa), spectinomycin (50 μg/ml for E. coli), streptomycin (50 μg/ml for E. coli and 250 μg/ml for P. aeruginosa), and irgasan (25 μg/ml for P. aeruginosa).
Strains and plasmids
Plasmid and strain constructions.All PCR-generated plasmid insert fragments were confirmed by DNA sequencing. In-frame deletion mutants were constructed using the sacB+ pEX18Ap suicide vector. Two ∼500-bp fragments from the regions immediately upstream and downstream of the area to be deleted were amplified by PCR and cloned into the pEX18Ap vector. The plasmids were then integrated into the P. aeruginosa chromosome following conjugation from E. coli, and sucrose-resistant, carbenicillin-sensitive segregants were isolated on agar containing 5% (wt/vol) sucrose. Deletions were verified by genomic PCR analysis using primers flanking the deleted region but outside the pEX18Ap clone insert.
The pscC null mutant was constructed by suicide plasmid insertion mutagenesis. An ∼500-bp fragment corresponding to the central region of pscC was amplified by PCR and cloned into suicide vector pAJD1. The resulting plasmid was integrated into the P. aeruginosa chromosome following conjugation from E. coli. The correct integration was confirmed by colony PCR analysis and the absence of Exo proteins in the culture supernatant after growth in LB medium containing 5 mM EGTA (data not shown).
Single-copy lacZ operon fusion strains were made by amplifying the noncoding regions upstream of each gene and cloning them into the pmini-CTX-lacZ vector. The plasmids were integrated into the attB site of the P. aeruginosa chromosome, and then the backbone vector DNA was removed by pFLP2-mediated excision and the integrations were confirmed by colony PCR analysis as described previously (35).
The araBp-ctpA-his6 expression plasmid pAJD2226 was constructed by amplifying ctpA from P. aeruginosa genomic DNA using a downstream primer that incorporated a region encoding His6 and cloning it into plasmid pHERD20T. A similar plasmid encoding the catalytically inactive mutant CtpA-S302A (pAJD2227) was generated by a splicing overlap extension (SOE) PCR approach as described previously (36). tacp-ctpA-his6 (pAJD2350) and tacp-ctpA-S302A-his6 (pAJD2351) expression plasmids were constructed by transferring the inserts from pAJD2226 or pAJD2227, respectively, into pVLT35 as SacI-XbaI fragments.
Polyclonal antiserum production and immunoblotting.The region encoding CtpA without its predicted N-terminal signal sequence was amplified by PCR as a BamHI-XhoI fragment and cloned into pET-24b(+) (Novagen) to encode a ′CtpA-His6 fusion protein (pAJD2290). E. coli strain ER2566 (NEB) containing pAJD2290 was grown to mid-log phase at 37°C with vigorous aeration, and fusion protein production was induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 2 h. The protein was purified under denaturing conditions by nickel-nitrilotriacetic acid affinity chromatography as described by the manufacturer (Qiagen). Polyclonal rabbit antiserum was raised against the purified protein by Covance Research Products Inc.
Samples separated by SDS-PAGE were transferred to nitrocellulose by semidry electroblotting. Enhanced chemiluminescent detection followed incubation with polyclonal antiserum or monoclonal antibody followed by goat anti-rabbit IgG (Bio-Rad), goat anti-mouse IgG (Bio-Rad), or rabbit anti-chicken IgY (IgG) (Sigma) horseradish peroxidase conjugates used at the manufacturer's recommended dilution. Primary antisera or antibodies were diluted 10,000-fold for anti-CtpA, anti-PA0943 (37), anti-PA4068 (37), anti-FLAG M2 (Sigma), and anti-ExoS (Agrisera) and 5,000-fold for anti-THE-His (GenScript).
PM and DB resistance assay.Wild-type PAK and its ΔctpA derivative AJDP730 were compared pairwise in two independent Biolog Phenotype MicroArray (PM) experiments (Biolog Inc.) (38). This technology monitors bacterial respiration by quantifying the accumulation of a purple color formed by reduction of a tetrazolium dye over time. Data are plotted as dye accumulation against time, and for each strain the “average height” for each well is calculated as the area under the curve divided by the number of reads. From this the “average height difference” between wells is derived. Any wells in which the difference in height between the two stains exceeded a chosen “average height difference” threshold in each of two independent experiments were deemed to be significant. For the analysis reported here, Biolog used the thresholds of 50 for metabolic tests and 60 for sensitivity tests. To confirm the domiphen bromide (DB) resistance phenotype, strains were grown at 37°C to saturation in M9 minimal medium containing 0.2% glucose, diluted 100-fold into the same medium supplemented with different concentrations of domiphen bromide, and incubated in a roller at 37°C for 24 h, and then the optical density at 600 nm (OD600)was measured. Results are the average from two independent experiments.
TEM.Sample fixation, slide preparation, and microscopy were performed by the Microscopy Core at the New York University Langone Medical Center (New York, NY). Bacterial cells were grown at 37°C to early log phase and collected by centrifugation. For chemical fixation, cells were mixed with fixative buffer (2% paraformaldehyde and 2% glutaraldehyde in 50 mM cacodylate), incubated for 2 h at room temperature, and washed with cacodylate buffer (50 mM, pH 7.2). Cells were then fixed with 2% osmium tetroxide in cacodylate buffer at room temperature for 2 h. Fixed cells were washed, embedded in 2% agar, and stained with 0.5% (wt/vol) uranyl acetate at room temperature for 1 h. The samples were then dehydrated with alcohol, sectioned thinly to adsorb onto electron microscope grids, and stained with 2% uranyl acetate and lead citrate. For high-pressure freezing, cells were mixed with fixative buffer, incubated for 2 h at room temperature, and washed with cacodylate buffer. The fixed cells were filled in planchette hats coated with hexadecane and then immediately frozen by being placed into a mixture of 2% osmium tetroxide in acetone at liquid nitrogen temperature for 96 h. Samples were slowly warmed by a Leica freeze substitution unit from −90°C to 0°C by raising the temperature 5°C/h. Samples were then infiltrated at room temperature in a 1:1 mixture of acetone-Epon for 1 h followed by a 1:2 mixture of acetone-Epon overnight. Samples were incubated in pure Epon for 4 h, embedded in embedding blocks, and sectioned at 60-nm thickness using a Leica UC6 ultramicrotome. The sections were placed on 200 Hex thinbar Cu grids and stained with uranyl acetate in 50% methanol and Reynold's lead citrate solution. Data were acquired on a Philips CM12 tungsten emission transmission electron microscope (TEM) and a Philips CM200 field emission gun TEM with a Gatan 4k by 2.7k side-mount camera and a Gatan 2k by 2k bottom-mount camera.
Subcellular fractionation.Cultures were grown at 37°C with shaking at 250 rpm in LB broth until the optical density at 600 nm was ∼1.5. Cells were collected by centrifugation at 8,000 × g for 10 min, and their density was adjusted to an OD600 of 10. Cells were washed in phosphate-buffered saline (PBS) and resuspended in Tris-HCl (pH 8.0) supplemented with 20% (wt/vol) sucrose, and then 5 mM EDTA and 20 μg/ml lysozyme were added and the sample was incubated on ice for 30 min without agitation. Spheroplasts were collected by centrifugation at 10,000 × g for 30 min, and the supernatant was collected as the periplasmic fraction.
Cell-free supernatant preparation.Strains were grown to saturation in tryptic soy broth (TSB) at 37°C with vigorous aeration. They were diluted to an OD600 of 0.05 in TSB supplemented with 5 mM EGTA and grown for 6 h at 37°C with vigorous aeration. The growth rates of the different test strains in this medium were not significantly different (data not shown). Bacterial cells were removed by centrifugation at 2,500 × g for 15 min, and the supernatants were passed through a 0.22-μm filter (Millipore). Proteins were precipitated from the filtered supernatant by adding 5% (vol/vol) trichloroacetic acid and incubating on ice overnight and then were collected by centrifugation at 13,000 × g for 30 min at 4°C. Pellets were washed with acetone and collected again by centrifugation at 13,000 × g. The protein pellet was resuspended in SDS-PAGE sample buffer containing β-mercaptoethanol.
Cytotoxicity assay.Chinese hamster ovary (CHO-K1) cells were propagated in Roswell Park Memorial Institute (RPMI) 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS) and 2 mM glutamine in a 5% CO2 atmosphere. For the assay, the CHO-K1 cells were seeded at 4 × 105 cells/well in a 24-well plate, incubated at 37°C for 16 h, washed once with PBS, and then covered with RPMI 1640 containing 1% FBS and 2 mM glutamine. P. aeruginosa was grown to mid-log phase at 37°C in TSB and added to the CHO-K1 cells at a multiplicity of infection of 10, followed by incubation at 37°C for 4 h. Supernatants were collected after centrifugation at 250 × g for 5 min, and the concentration of lactate dehydrogenase (LDH) in the supernatant was determined using the LDH Cytotox 96 nonradioactive cytotoxicity assay kit (Promega). For complementation of the ΔctpA mutation, 15 μM IPTG was added to bacterial growth media and the cell culture media to maintain tacp-ctpA expression.
Mouse model of acute pneumonia.Six-week-old female BALB/c mice were purchased from Charles River Laboratories Inc. and rested for 6 days. Mice were anesthetized deeply with isoflurane and inoculated intranasally with ∼1 × 108 CFU in 20 μl PBS. Twelve hours later, the mice were euthanized with a lethal dose of pentobarbital, and lungs and spleens were collected. The organs were homogenized in PBS, and bacteria were enumerated by spreading dilutions onto VBM agar and incubating at 37°C overnight. The bacterial load was calculated as CFU per organ. Statistical significance was assessed by one-way analysis of variance (ANOVA). We considered statistical significance to be a P value of <0.01. The New York University School of Medicine Animal Care and Use Committee approved all animal experiments.
Reverse transcription-PCR (RT-PCR).Bacteria were grown in LB broth in a test tube roller for 2 h at 37°C, and then in some cases 400 μg/ml d-cycloserine (freshly dissolved in 100 mM sodium phosphate buffer, pH 8.0) was added before growth was continued for another 2 h. Approximately 2 × 108 cells were used to isolate total RNA with a Qiagen RNeasy Mini-Kit, and residual genomic DNA was eliminated by treating with RNase-free DNase (NEB). Approximately 250 ng of total RNA was converted to cDNA using a Superscript RT III first-strand synthesis system (Invitrogen). Fourteen nanograms of cDNA was used as the template for PCR amplification with Taq DNA polymerase (Qiagen), with the exception of PA4495 experiments, in which 100 ng of cDNA was used. Each PCR mixture contained 0.4 μM each of two gene-specific primers (Table 2). The amplification protocol was as follows: 5 min at 98°C; 30 cycles of 30 s at 98°C, 45 s at 62°C, and 30 s at 72°C; and 5 min at 72°C. For each target gene, the PCR primers amplified an ∼200-bp fragment from the coding region. Five microliters of each PCR mixture was analyzed by 1.5% agarose gel electrophoresis and ethidium bromide staining.
Primers used for RT-PCR
RESULTS
Identification of a P. aeruginosa ctpA null mutant.Pore-forming outer membrane proteins called secretins are critical components of type 2 and 3 secretion systems and of type 4 pili (39). However, they can mislocalize in the envelope and kill the bacterial cell unless a specialized extracytoplasmic stress response called the phage shock protein (Psp) system prevents it (40). P. aeruginosa does not have a Psp system, but we reported the isolation of some transposon insertion mutants sensitive to increased production of the secretin XcpQ (37). After extending that screen, we have now isolated a mutant with a transposon insertion in the gene annotated as PA5134/ctpA in strain PAO1. This gene encodes a predicted carboxyl-terminal protease of 43.6 kDa after removal of a predicted N-terminal Sec-dependent signal sequence. As has been noted by others, P. aeruginosa CtpA is in the CTP-3 subfamily, with approximately 53% amino acid identity to Bartonella bacilliformis CtpA (see reference 31 for detailed sequence analysis). A global RNA-seq survey has indicated that ctpA is in the three-gene operon envC-ctpA-yibQ (41) (Fig. 1A). In E. coli K-12, envC and yibQ homologues are immediately adjacent, with no ctpA gene between them. EnvC activates amidases involved in peptidoglycan remodeling at the division septum (42), and YibQ is a predicted polysaccharide deacetylase. Therefore, the location of ctpA is this operon suggests that it might have a cell envelope-related function.
Phenotypes of a ctpA null mutant. (A) The ctpA chromosomal region. Approximate locations of promoters identified in global RNA-seq studies are shown as line arrows, one of which is extended to show the operon that contains ctpA. (B) Panel i, sensitivity to XcpQ overproduction. ctpA+ and ΔctpA strains contained either the tac promoter expression plasmid pVLT35 (−XcpQ) or the xcpQ+ derivative pAJD942 (+XcpQ). Serial 10-fold dilutions of saturated cultures were spotted onto VBM agar containing 1 mM IPTG and incubated at 37°C. Panel ii, complementation of XcpQ sensitivity. ctpA+ and ΔctpA strains contained the tacp-xcpQ+ derivative pAJD942 and either the araBp promoter expression plasmid pHERD20T (−) or a derivative encoding CtpA-His6 or the CtpA-S302A-His6 mutant. Serial 10-fold dilutions of saturated cultures were spotted onto VBM agar containing 1 mM IPTG to overexpress xcpQ and 0.005% arabinose to induce CtpA-His6 production to an approximately endogenous level and incubated at 37°C. (C) Resistance to domiphen bromide. Saturated cultures of ctpA+ and ΔctpA strains were diluted 100-fold into M9-glucose minimal medium supplemented with the indicated concentrations of domiphen bromide. Absorbance at 600 nm was measured after 24 h of incubation at 37°C with aeration. Data were averaged from two independent experiments, with the vertical bars showing the range of values from the two experiments. (D) Transmission electron microscopy. Images from samples prepared by either high-pressure freezing or conventional chemical fixation are shown. Arrows point to apparent separation of cell envelope layers. Scale bar, 2 μm.
To check if the XcpQ sensitivity of the ctpA transposon insertion mutant was due to loss of ctpA only, we constructed a ctpA in-frame deletion mutant. This mutant was also sensitive to XcpQ production, and the phenotype was complemented by a plasmid encoding CtpA (Fig. 1B). Sequence alignment with other CTPs identified the predicted CtpA catalytic serine as serine 302 (reference 43 and data not shown). A plasmid encoding CtpA with this serine changed to alanine (CtpA-S302A) did not complement the XcpQ sensitivity (Fig. 1B). Therefore, CtpA is required for resistance to XcpQ overproduction, and the proteolytic function of CtpA is probably involved.
A ctpA null mutant has other phenotypes consistent with an altered cell envelope.To learn more about the impact of a ctpA null mutation, we compared the wild type and ΔctpA strains in a Biolog Phenotype MicroArray analysis. This technique surveys almost 2,000 phenotypes, including the ability to use various nutrients and sensitivity to a wide variety of different antibiotics and stress conditions, several of which target the cell envelope (44). The results revealed that the ΔctpA mutant had increased resistance to domiphen bromide (DB) and dodine, two cationic surfactants that can damage the cytoplasmic membrane (45). We chose to validate the DB resistance phenotype by comparing the growth of ctpA+ and ΔctpA strains in M9 minimal medium supplemented with 0.2% glucose and different amounts of DB. The data confirmed that the ΔctpA mutant was more resistant to DB (Fig. 1C). This is consistent with the ΔctpA mutation altering the cell envelope of P. aeruginosa so that resistance to at least some cationic surfactants is increased. Nevertheless, resistance to DB and dodine were the only phenotypic differences between the ΔctpA and ctpA+ strains uncovered by the Biolog analysis, suggesting that the ctpA null mutation does not cause major pleotropic effects.
We also compared the ctpA+ and ΔctpA strains by transmission electron microscopy. When a traditional chemical fixation method was used, many of the ΔctpA cells appeared to have separation between layers of the cell envelope close to a cell pole (Fig. 1D). By surveying three independent fields, we determined that this occurred in approximately 34% of the ΔctpA cells that we observed (199 out of 590 cells), whereas it did not occur in any ctpA+ cells. We also saw this phenotype in ΔctpA cells when a less destructive high-pressure freezing preparation was used instead of traditional chemical fixation, but it occurred at a lower frequency (less than 5%). Therefore, it is possible that the ΔctpA mutation affects the bacterial cell envelope to alter its structure and/or make it more sensitive to disruption during preparative treatments for electron microscopy. This is also consistent with CtpA having a cell envelope-related function.
Endogenous CtpA is a soluble periplasmic protein.Previously, it was reported that CtpA was located in the periplasm of P. aeruginosa strain PAO1 (31). However, those authors could not detect endogenous CtpA by immunoblotting, causing them to speculate that it might have an extremely low abundance. Therefore, they had to overexpress ctpA from a plasmid to detect the protein and monitor its location. This overproduction could have influenced the behavior of CtpA, which motivated us to investigate the location of the endogenous protein. We raised an anti-CtpA polyclonal antiserum and found that it was able to detect endogenous CtpA in a total cell lysate of the wild-type PAK strain when used at 1-in-10,000 dilution (Fig. 2). We also investigated the location of the endogenous protein and found that the majority was in the soluble periplasmic fraction (Fig. 2), consistent with the conclusions from overproduced CtpA in strain PAO1 (31).
Endogenous CtpA is a soluble periplasmic protein. Total cell lysate (TCL), spheroplast (Sphero), and periplasmic (Peri) fractions of the wild-type PAK strain were separated by SDS-PAGE and transferred to nitrocellulose, and proteins were detected with polyclonal antisera against CtpA, PA0943 (periplasmic control) (37), and PA4068 (cytoplasmic control) (37).
CtpA is required for normal function of the T3SS and for cytotoxicity to CHO-K1 cells.As part of our characterization of a ctpA null mutant, we surveyed the extracellular (culture supernatant) protein profiles of ctpA+ and ΔctpA strains grown under various conditions (including type 2 and type 3 secretion-inducing conditions) (data not shown). This led to the discovery that the only condition where the ΔctpA mutation reduced the level of culture supernatant proteins was when the strains were grown under type 3 secretion-inducing conditions (with a low Ca2+ concentration, caused by adding EGTA to the growth medium). This suggested that the type 3 secretion system (T3SS) was defective in the ΔctpA mutant. To test this, we obtained a commercial antibody raised against the ExoS substrate of the T3SS. However, the homology between ExoS and ExoT results in the antibody recognizing both proteins. The ctpA+ and ΔctpA strains were grown under T3SS-inducing conditions, and culture supernatants were analyzed by SDS-PAGE and immunoblotting. The amounts of secreted ExoS and ExoT were significantly lower in the ΔctpA mutant (Fig. 3A). Furthermore, the ExoS and ExoT levels were restored to wild type by a plasmid encoding CtpA but not by a plasmid encoding the predicted catalytically inactive mutant CtpA-S302A (Fig. 3A). Therefore, CtpA is required for normal T3SS function, and this probably requires its proteolytic function. However, we do not yet know how CtpA affects the T3SS (e.g., regulation of expression, assembly, or rate of secretion) (see Discussion).
CtpA is required for normal function of the T3SS and for cytotoxicity. (A) A ΔctpA mutation reduces the levels of ExoS and ExoT in the culture supernatant. ctpA+ and ΔctpA strains contained the tac promoter expression plasmid pVLT35 (−) or derivatives encoding CtpA-His6 or CtpA-S302A-His6 were grown under T3SS-uninducing (−EGTA) or T3SS-inducing (+EGTA) conditions in medium with 15 μM IPTG to induce CtpA-His6 production to an approximately endogenous level. Cell-free supernatants derived from equivalent amounts of cells were separated by SDS-PAGE and transferred to nitrocellulose, and proteins were detected with an antibody that recognizes ExoS and ExoT. (B) Panel i, a ΔctpA mutation reduces cytotoxicity to CHO-K1 cells. Wild-type, ΔctpA, and pscC::pAJD1 strains were added to CHO-K1 cells at a multiplicity of infection of 10 and incubated for 4 h. Cell-free supernatants were then analyzed for lactate dehydrogenase (LDH) content. The amount of LDH in the supernatant following incubation with the wild-type strain was set to 100%, and the values for the mutants are shown as the relative percentage. Data were averaged from three independent experiments, with the error bars showing the positive standard deviation. Panel ii, complementation of the cytotoxicity defect. ctpA+ and ΔctpA strains contained the tac promoter expression plasmid pVLT35 (−) or derivatives encoding CtpA-His6 or CtpA-S302A-His6. The experiment was done exactly as for panel i except that the medium contained 15 μM IPTG to induce CtpA-His6 production to an approximately endogenous level. The amount of LDH in the supernatant following incubation with the wild-type strain containing pVLT35 was set to 100%.
P. aeruginosa exhibits T3SS-dependent cytotoxicity toward cultured mammalian cells such as CHO-K1 cells (see, e.g., reference 46). Therefore, if the ΔctpA mutation compromises the T3SS, the mutant should be less cytotoxic than the wild type. To test this hypothesis, we compared the cytotoxicities of ctpA+ and ΔctpA strains toward CHO-K1 cells by measuring the amount of lactate dehydrogenase (LDH) released into the cell culture supernatant, a method used by others to determine the cytotoxicity of the PAK strain (46). As a control, we constructed a strain with a null mutation in the pscC gene, which encodes the essential secretin component of the T3SS. Compared to the wild type, both the ctpA and pscC null mutants had similarly reduced cytotoxicity (Fig. 3B). Furthermore, the cytotoxicity of the ΔctpA mutant was fully restored by a plasmid encoding wild-type CtpA but not by a plasmid encoding CtpA-S302A.
CtpA is required for virulence in a mouse model of acute pneumonia.The preceding experiments showed that CtpA is required for T3SS function in the laboratory, but they did not address its importance during host infection. However, the T3SS is known to be important for P. aeruginosa virulence during acute infection (see, e.g., reference 5). Therefore, we tested the effect of a ctpA null mutation on the virulence of P. aeruginosa in a mouse model of acute pneumonia and dissemination. Once again, a pscC null mutant that is completely defective for type 3 secretion was used as a control. Mice were infected intranasally, and then at 12 h postinfection the numbers of bacteria in the lungs and spleens were determined. The results showed that the ctpA and pscC null mutations had indistinguishable effects on virulence. Both reduced the numbers of bacteria recovered from the lungs by ∼100-fold compared to the wild type (Fig. 4). Just as striking was the fact that neither mutant was detected in the spleen, whereas the wild type was. Therefore, CtpA is essential for virulence during an acute pneumonia infection in mice. This is probably because it is required for the normal function of the T3SS.
A ΔctpA mutant is attenuated in a mouse model of acute pneumonia. Mice were infected intranasally with 2 × 108 CFU of the wild-type, ΔctpA, or pscC::pAJD1 strain. The animals were euthanized at 12 h postinfection, and the numbers of bacteria in the lungs and spleens were determined. Each circular symbol represents the data from an individual mouse. Open circles indicate that no bacteria were detected from that organ and are shown at the approximate limit of detection (∼100 CFU for the spleen). The horizontal lines represent the geometric mean. One-way ANOVA statistical analysis was used to determine the statistical significance between the strains. For both mutants the P values were <0.0001 in the lung and <0.0005 in the spleen compared to the wild type.
An elevated CtpA level induces an uncharacterized extracytoplasmic function sigma factor regulon.Removing CtpA affects the cell envelope and compromises the T3SS. To complement these defects, we had controlled expression from inducible ctpA+ plasmids so that CtpA was produced at near-endogenous level (data not shown). However, early in the course of characterizing the ctpA+ plasmids, we noticed that ctpA overexpression led to two proteins becoming the most abundant in the cell (Fig. 5A). This did not happen when CtpA-S302A was overproduced, suggesting that it required CtpA proteolytic activity (data not shown). These two proteins were enriched in a periplasmic extract (data not shown). We separated them by SDS-PAGE, identified them by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry, and determined that they are encoded by the genes annotated as PA1494 and PA4495 in the strain PAO1 genome. PA1494 is a predicted periplasmic protein of 58.4 kDa after removal of an N-terminal signal sequence. PA1494 is annotated as muiA (mucoidy inhibitor A) because a random screen showed that its overexpression suppressed the mucoidy of a CF lung isolate (47). PA4495 is a predicted periplasmic protein of 22.8 kDa after removal of its signal sequence. We confirmed these identifications by modifying the chromosomal PA1494 and PA4495 genes to encode epitope-tagged derivatives and detected their induction by immunoblotting when CtpA was overproduced (data not shown).
CtpA overproduction activates a putative PA2896 ECFσ regulon. (A) PA1494 and PA4495 are the most abundant cellular proteins when CtpA is overproduced. PA2896+ and ΔPA2896 strains contained the araBp promoter expression plasmid pHERD20T (−) or a derivative encoding CtpA-His6 and were grown with gentle aeration in LB broth supplemented with 0.2% arabinose to overproduce CtpA-His6. Total cell lysates were separated by SDS-PAGE and stained with BioSafe Coomassie blue (Bio-Rad). Approximate locations of the PA1494 and PA4495 proteins are indicated. The sizes of molecular weight standards (in thousands) are indicated on the left. The panels at the bottom show anti-CtpA and anti-His6 immunoblots of the total cell lysates. (B) CtpA overproduction induces expression of PA2896, PA1494, and PA4495. Strains with the indicated single-copy lacZ operon fusions contained the araBp promoter expression plasmid pHERD20T (−) or a derivative encoding CtpA-His6 or CtpA-S302A-His6 and were grown in LB broth supplemented with 0.2% arabinose to overproduce CtpA-His6. β-Galactosidase activities were averaged from three independent experiments, with the error bars showing the positive standard deviation. The panels at the bottom show anti-CtpA and anti-His6 immunoblots of total cell lysates from one set of cultures. (C) d-Cycloserine induces the PA2896 regulon in a CtpA-independent manner. Total RNA was isolated from ctpA+ and ΔctpA strains grown in medium with or without 400 μg/ml d-cycloserine and used for semiquantitative reverse transcription-PCR. PCR of genomic DNA (gDNA) templates was used as a positive PCR control. Analysis of groEL expression served as a steady-state reference control. The groEL primers were also used for PCR of each RNA sample that had not been treated with reverse transcriptase (no RT) to check for DNA contamination. PCR products were separated on 1.5% agarose gels and photographed under UV illumination after ethidium bromide staining (negative images are shown).
A microarray study predicted PA1494 and PA4495 to be targets of an uncharacterized ECFσ factor encoded by PA2896 (32). PA2896 is in an operon with the downstream PA2895, and the latter is predicted to be the anti-sigma factor (32, 48). Interestingly, a PA2895 anti-sigma factor null mutant is attenuated in rat model of a chronic lung infection, suggesting that constitutive activation of the PA2896 regulon is detrimental to chronic infection (49). To test if CtpA-dependent induction of PA1494 and PA4495 was PA2896 ECFσ dependent, we generated a PA2896 in-frame deletion mutation. This mutation abolished the ability of CtpA overproduction to induce the PA1494 and PA4495 proteins (Fig. 5A). To address whether the effect of CtpA was transcriptional we constructed strains with single-copy ϕ(PA1494p-lacZ) and ϕ(PA4495p-lacZ) operon fusions. Their expression was induced by CtpA overproduction but not by CtpA-S302A overproduction (Fig. 5B). A ϕ(PA2896p-lacZ) operon fusion was also induced by CtpA overproduction, which is consistent with ECFσ proteins often being positive autoregulators (50). Together these data suggest that increasing the level of active CtpA induces PA2896 ECFσ activity, which leads to induction of its putative regulatory targets PA1494 and PA4495.
CtpA overproduction might induce the PA2896 ECFσ regulon by causing an extracytoplasmic stress signal.Many ECFσ regulons are induced by extracytoplasmic stress (50, 51). Among these is the AlgU/T regulon, which is activated by d-cycloserine, an antibiotic that inhibits peptidoglycan cross-linking (18). When the AlgU/T and d-cycloserine-dependent regulons were surveyed by microarray analysis, PA2895, PA2896, PA1494, and PA4495 were induced by d-cycloserine, but in an AlgU/T independent manner (52). Therefore, we tested the effect of d-cycloserine on PA2896, PA1494, and PA4495 expression by semiquantitative RT-PCR. Consistent with the microarray analysis, all were induced by d-cycloserine (Fig. 5C). This finding allowed us to design an experiment to distinguish between two hypotheses to explain why CtpA overproduction activated the regulon.
Regulated intramembrane proteolysis of an anti-sigma factor is a common mechanism for ECFσ activation (53). Therefore, one possibility is that CtpA is required for regulated intramembrane proteolysis of the anti-sigma factor PA2895 when an inducing signal is encountered, but overproducing CtpA can degrade the anti-sigma factor even without a normal inducing signal. If so, d-cycloserine-dependent induction of the PA2896 regulon should be prevented in a ΔctpA mutant. Alternatively, CtpA overproduction or d-cycloserine might be able to cause the same extracytoplasmic stress signal, such as a compromised cell wall, but do so independently from one another. If that is true, d-cycloserine should still be able to induce the PA2896 regulon in a ΔctpA mutant. Therefore, we tested the effect of a ΔctpA mutation on the ability of d-cycloserine to induce PA2896, PA1494 and PA4495 expression. The results showed that d-cycloserine-dependent induction of PA2896 and PA1494 occurred in both ctpA+ and ΔctpA strains (Fig. 5C). However, the involvement of CtpA in d-cycloserine-dependent induction of PA4495 was difficult to determine because deletion of ctpA increased basal PA4495 expression (Fig. 5C) [also confirmed by analysis of ϕ(PA4495p-lacZ) expression (data not shown)]. This suggests that PA4495 expression is sensitive to both reduced and increased CtpA levels.
These findings suggest that an increased CtpA level generates a stress signal, perhaps similar to that caused by d-cycloserine, which activates the PA2896 regulon. This is significant because PA2896 activation compromises chronic virulence (49). Therefore, reducing CtpA activity interferes with the T3SS and acute virulence, whereas increasing it might interfere with chronic virulence.
DISCUSSION
Proteases play important roles in bacteria, including destruction of damaged/misfolded proteins, processing preproteins to their mature forms, and controlling signal transduction mechanisms to regulate gene expression. The P. aeruginosa genome encodes two predicted members of a family of serine proteases that are relatively understudied in bacteria, the carboxyl-terminal processing proteases (CTPs). One (PA3257/Prc) is homologous to the only E. coli K-12 CTP and has been linked to promoting mucoidy in CF lung isolates by degrading truncated forms of MucA (19, 20). Here, we have discovered important phenotypes associated with the other P. aeruginosa CTP, PA5134/CtpA. CtpA is essential for normal type 3 secretion and virulence in a mouse model of acute pneumonia, whereas an increased CtpA level activates a putative ECFσ factor regulon that is known to compromise chronic virulence (49).
We first isolated a ctpA null mutant in an extension of our screen for mutants sensitive to XcpQ secretin overproduction (37). This phenotype is consistent with altered cell envelope properties, which is also suggested by increased resistance to domiphen bromide and by the altered appearance of cells prepared for electron microscopy (Fig. 1). In addition, a global screen of P. aeruginosa strain PA14 transposon mutants revealed that a ctpA mutant has reduced swarming motility over an agar surface (54), which we have confirmed in the PAK strain (data not shown). Burkholderia mallei and Brucella suis ctpA null mutants also have phenotypes that indicate changes in their cell envelope (28, 29). Similarly, an E. coli prc mutant leaks periplasmic proteins and has other phenotypes suggesting compromised envelope integrity (24). However, the P. aeruginosa ctpA null mutant does not appear to have the wide range of pleotropic phenotypes reported for CTP mutants in some other species (see, e.g., references 24, 29, and 55). For example, it is not more sensitive to antibiotics, low-salt media, or heat, it does not appear to leak periplasmic proteins, and it is competent for swimming motility and attachment to abiotic surfaces (data not shown). This suggests intact flagella and also type IV pili, because we did the surface attachment assays under conditions where type IV pili are required (39, 56). In fact, of the almost 2,000 tests in the Biolog analysis, increased resistance to domiphen bromide and increased resistance to dodine were the only phenotypes of the ΔctpA strain. Therefore, we speculate that CtpA has a restricted role in P. aeruginosa and probably few direct substrates.
The most significant phenotype of the ctpA null mutant is its defective T3SS, which probably explains its reduced cytotoxicity and acute virulence (Fig. 3 and 4) (although we cannot rule out the possibility that additional defects also contribute to these phenotypes). However, ExoS and ExoT do not accumulate inside ΔctpA cells, and our preliminary analysis suggests that exoT gene expression is actually decreased (J. Seo and A. Darwin, unpublished data). Reduced T3SS activity invokes feedback inhibition of T3SS gene expression, including that of the exo genes (57, 58). Therefore, reduced exo gene expression does not necessarily mean that the only role of CtpA is to positively regulate the exo genes. Instead, CtpA activity might be required for the proper assembly and/or function of the T3SS. The location of ctpA in an operon with envC, encoding a protein involved in peptidoglycan remodeling, might indicate that CtpA has a function related to the cell wall. This is notable because protein-peptidoglycan interactions are important for the correct assembly of secretion systems in bacteria, including T3SSs (54, 59, 60). Nevertheless, we cannot yet rule out any mechanism to explain the impact of CtpA on the T3SS, including gene regulation rather than T3SS assembly/function. Much more work will be needed to understand exactly how the ctpA null mutation compromises Exo protein export.
Another suggestion of a link between CtpA and the cell wall comes from the finding that CtpA overproduction induces PA1494 and PA4495 (Fig. 5). Data suggest that they are probably direct targets of the PA2896 ECFσ factor. First, in transcriptome analysis their expression was not detected in a PA2896 mutant but was massively upregulated in a PA2895 mutant, encoding the putative anti-sigma factor (32). Second, CtpA did not induce the PA1494 and PA4495 proteins in a ΔPA2896 strain (Fig. 5A). Third, when the regions upstream of the PA1494, PA4495, and PA2896 transcription start sites (+1) identified by global RNA-seq studies (41, 61) are aligned, they have identical −10 and −35 regions (TAACCCG-N16-CGTCTCA-N6-A[+1]). The putative PA2896 ECFσ regulon is also induced by d-cycloserine (52) (Fig. 5C), suggesting that it responds to cell wall stress. Furthermore, PA1494 is one of the genes most highly induced by GlmU depletion, which compromises the cell wall (62). Therefore, increased CtpA activity might also affect the cell wall adversely to generate an inducing signal for the PA2896 ECFσ regulon. However, it remains possible that CtpA overproduction does something different, such as degrading the anti-sigma factor directly. In fact, the other P. aeruginosa CTP (PA3257/Prc) degrades mutant forms of the MucA anti-sigma factor. Even so, we can rule out an essential role for CtpA in any regulated intramembrane proteolysis of the PA2895 anti-sigma factor that might be part of the normal activation mechanism. This is because d-cycloserine can still activate the regulon in a ΔctpA strain (Fig. 5C) and because ΔPA2896 and ΔctpA strains do not have any overlapping phenotypes in Biolog Phenotype MicroArray analysis (Seo and Darwin, unpublished data). Therefore, we favor the hypothesis that CtpA overproduction causes overcleavage of its substrate(s), generating cell envelope stress similar to that caused by d-cycloserine. Future work will be needed to rigorously test this hypothesis and to investigate the mechanism of PA2896 regulon induction in detail.
The role of the PA2896 ECFσ regulon is unknown, but it is probably related to the cell envelope because the two putative downstream targets, PA1494 and PA4495, encode periplasmic proteins. PA2895 anti-sigma factor mutants were identified in a signature-tagged mutagenesis screen for P. aeruginosa genes essential for survival in a rat model of chronic lung infection (49). This suggests that constitutive activation of the PA2896 ECFσ regulon attenuates chronic infection (a ΔPA2895 mutation induces PA1494 and PA4495 expression [Seo and Darwin, unpublished data]). This is intriguing because PA1494 was identified as an overexpression suppressor of mucoidy in a CF lung isolate, resulting in it being named mucoidy inhibitor A (MuiA) (47). Therefore, it is possible that constitutive production of PA1494/MuiA in the PA2895 mutant was responsible for attenuating the chronic infection by interfering with alginate production in the rat lung.
In summary, our work has provided insight into the previously uncharacterized P. aeruginosa CtpA and revealed it to be a highly significant protein. Removing CtpA compromises T3SS function, cytotoxicity, and acute virulence, whereas elevating it activates a putative regulon that can attenuate chronic virulence. This makes CtpA a worthy subject for continued study and raises the possibility that it might be an attractive target for drugs that either inactivate it or hyperactivate it. We speculate that there might be few direct substrates of CtpA. There is precedent for this in the cyanobacterium Synechocystis sp. strain PCC6803, where the only known target of its CtpA is the reaction center protein D1 of photosystem II (63, 64). CtpA processes D1 to its mature form by removing its C terminus, and the ctpA null mutant is deficient for photosynthesis but does not have other pleotropic phenotypes, which is in contrast to the case for an E. coli prc mutant. Similarly, the P. aeruginosa ctpA mutant does not have the same pleotropic phenotypes as an E. coli prc mutant. An intriguing possibility is that a single CtpA substrate could explain the phenotypes of both the ctpA null mutant and CtpA overproduction. For example, this substrate might be involved in some aspect of the cell wall. Failure to cleave it in the ctpA null mutant could alter cell wall properties so that T3SS assembly or function is compromised, whereas overcleaving it could cause a stress-inducing signal for the PA2896 ECFσ regulon. Therefore, a pressing goal for the future is to identify the target(s) of CtpA proteolysis and to explore the mechanism(s) underlying its impact on these different systems.
ACKNOWLEDGMENTS
We thank Terry-Ann Smith for helping with the early characterization of XcpQ-sensitive mutants and Steve Lory, Herbert Schweizer, and Hongwei Yu for providing plasmid vectors used in this work. We are grateful to Mary Ann Gawinowicz of the Protein Core Facility at Columbia University for protein identification by mass spectrometry and to the Microscopy Core at New York University Langone Medical Center for transmission electron microscopy. We also thank Josué Flores-Kim and Dana Harmon for critical review of a draft version of the manuscript. We are extremely grateful to Victor Torres for allowing us to share his laboratory space for the year following Superstorm Sandy.
A.J.D. holds an Investigators in Pathogenesis of Infectious Disease award from the Burroughs Wellcome Fund, which supported this work.
FOOTNOTES
- Received 19 August 2013.
- Returned for modification 5 September 2013.
- Accepted 23 September 2013.
- Accepted manuscript posted online 30 September 2013.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
