|Home | About | Journals | Submit | Contact Us | Français|
Cyclic AMP (cAMP) is an important second messenger signaling molecule that controls a wide variety of eukaryotic and prokaryotic responses to extracellular cues. For cAMP-dependent signaling pathways to be effective, the intracellular cAMP concentration is tightly controlled at the level of synthesis and degradation. In the opportunistic human pathogen Pseudomonas aeruginosa, cAMP is a key regulator of virulence gene expression. To better understand the role of cAMP homeostasis in this organism, we identified and characterized the enzyme CpdA, a putative cAMP phosphodiesterase. We demonstrate that CpdA possesses 3′,5′-cAMP phosphodiesterase activity in vitro and that it utilizes an iron-dependent catalytic mechanism. Deletion of cpdA results in the accumulation of intracellular cAMP and altered regulation of P. aeruginosa virulence traits. Further, we demonstrate that the cAMP-dependent transcription factor Vfr directly regulates cpdA expression in response to intracellular cAMP accumulation, thus providing a feedback mechanism for controlling cAMP levels and fine-tuning virulence factor expression.
The second messenger signaling molecule 3′,5′-cyclic AMP (cAMP) plays diverse and vital roles in multiple cellular processes in both eukaryotes and prokaryotes (4, 7). Intracellular cAMP levels are tightly controlled, and homeostasis is achieved by balancing cyclic nucleotide synthesis with degradation. Cyclic AMP is synthesized by the enzyme adenylate cyclase (AC) in response to intracellular and extracellular cues. ACs are ubiquitous in nature, and six structurally distinct classes have been identified (47, 49). Genome sequencing has revealed that these enzymes are widely distributed among bacteria and that many species possess multiple ACs. For most bacterial species, however, the cellular processes that are affected by AC activity, and thus the biological role(s) of cAMP, have yet to be defined.
Cyclic AMP is a stable, membrane-impermeable molecule; in order to maintain homeostasis and to reset cAMP signaling cascades following AC activation, signal attenuation is necessary. This function can be achieved through degradation of cAMP to 5′-AMP by the enzyme cyclic 3′,5′-AMP phosphodiesterase. Like ACs, cyclic nucleotide phosphodiesterases are found in all branches of life and encompass a family of enzymes that are grouped into three classes based on their primary amino acid sequence: class I (eukaryotes), class II (Saccharomyces cerevisiae, Schizosaccharomyces pombe, Candida albicans, Dictyostelium discoideum, and Vibrio fischeri), and class III (most bacteria). Class III enzymes contain highly conserved amino acid residues and secondary structural elements that are homologous to those of purple acid phosphatases and other dimetallophosphoesterases (41). Although the exact catalytic mechanism of purple acid phosphatases is unknown, their activity requires the presence of two active-site metal ions (Fe3+ and Fe2+, Zn2+, or Mn2+). Biochemical analyses of cAMP phosphodiesterases from Escherichia coli, Mycobacterium tuberculosis, and Anabaena sp. strain PCC 7120 indicate a similar requirement, suggesting that the structural and functional organization of the catalytic site of class III cyclic nucleotide phosphodiesterases is similar to that of the purple acid phosphatases (15, 21, 46).
E. coli possesses the best-characterized bacterial cAMP-dependent signaling system, which primarily controls the metabolic response to environmental carbon availability. Cyclic AMP acts as a positive allosteric regulator of the cAMP-dependent transcriptional regulator cyclic AMP receptor protein (CRP) (7). The cAMP-CRP complex regulates the expression of more than 100 genes and operons, many of which are involved in the transport and catabolism of carbon compounds (17). E. coli possesses a single class III cyclic 3′,5′-AMP phosphodiesterase (CpdA). The activity of this enzyme has been characterized in vitro (21, 33, 34); however, its global role in cAMP homeostasis and signaling has not been examined.
In contrast to E. coli, where cAMP is involved in metabolite repression, cAMP plays a critical role in controlling virulence in the opportunistic human pathogen Pseudomonas aeruginosa (51, 59). P. aeruginosa is an environmental bacterium capable of causing a variety of life-threatening infections in immunocompromised individuals and those receiving critical care (10). In addition, P. aeruginosa is the primary cause of chronic debilitating lung infection in individuals with cystic fibrosis. Cyclic AMP serves as an allosteric regulator of the transcription factor Vfr, a member of the CRP family (56). Together, cAMP and Vfr directly or indirectly regulate the expression of numerous genes encoding virulence factors such as ToxA (exotoxin A or ETA), a type III secretion system (T3SS) and its effectors (ExoS, ExoT, and ExoY), protease IV, and type IV pili (5, 9, 56, 59). Whole-genome expression profiling revealed that P. aeruginosa mutants defective in cAMP synthesis or lacking vfr produced nearly identical transcriptomes, affecting the expression of approximately 200 genes, many of which are known to be involved in virulence (59). These studies suggest that Vfr activity is ultimately dependent on the availability of cAMP. Thus, characterization of the mechanisms that regulate cAMP homeostasis in P. aeruginosa will provide a clearer understanding of how this bacterial pathogen establishes and maintains infection.
In P. aeruginosa, intracellular cAMP is synthesized by two adenylate cyclases, designated CyaA and CyaB (59). CyaA shares homology with the class I AC of E. coli (27), and CyaB is a member of the class III family, widely distributed in both eukaryotes and prokaryotes (49). P. aeruginosa mutants lacking either cyaA or cyaB have reduced intracellular levels of cAMP; however, the contribution of CyaB to the cAMP pool is substantially greater than that of CyaA (59). Elimination of CyaB alone results in attenuation of virulence in an adult mouse model of acute pneumonia (51).
Although cAMP phosphodiesterase activity was previously detected in the cytoplasmic fraction of crude cell lysates of P. aeruginosa (48), the enzyme(s) involved has not been identified. In the present study, we identify a single putative cAMP phosphodiesterase-encoding gene (PA4969) within the P. aeruginosa genome. We demonstrate that the purified gene product, which we have designated CpdA, has cyclic 3′,5′-AMP phosphodiesterase activity in vitro. In addition, we show that P. aeruginosa CpdA plays a significant role in cAMP homeostasis in vivo and is specifically required for virulence factor regulation. Finally, we found that Vfr directly activates expression of cpdA in response to elevated intracellular cAMP, thus providing a complex multilayered mechanism for controlling cAMP homeostasis in P. aeruginosa.
For routine passage, E. coli and P. aeruginosa strains were grown at 37°C in LB medium (Difco). pMMB-based expression plasmids were maintained in P. aeruginosa with 150 μg/ml carbenicillin (Cb) except where noted. Bacterial growth in broth culture was assessed by optical density at 600 nm (OD600), and doubling times during exponential phase growth were calculated based on time required for 2-fold increases in the OD600.
The strains and plasmids used in this study are listed in Table Table1.1. Chromosomal deletion of P. aeruginosa cpdA (codons 2 to 272) was carried out as previously described (59). Transcriptional reporters were constructed by PCR amplification of vfr, exoS, or cpdA promoter fragments from P. aeruginosa (strain PAK) genomic DNA using primers containing EcoRI and BamHI restriction sites. The fragments correspond to the following promoter regions in bp relative to the translational start codon: −310 to +60 for vfr, −553 to +78 for exoS, and −500 to + 100 cpdA; fragment were digested with EcoRI and BamHI and cloned into the same sites of plasmid mini-CTX-lacZ (19, 44). The resulting plasmids were used to integrate the promoter-lacZ fusions onto the chromosome at the CTX phage attachment site of wild-type and mutant P. aeruginosa strains as described previously (19).
For CpdA expression in E. coli strain M15(pREP4) (Qiagen), cpdA (+4 to +819 bp relative to translational start) was PCR amplified from P. aeruginosa strain PAK genomic DNA using primers containing SphI and HindIII restriction sites. The resulting product was cloned into the SphI-HindIII sites of pQE30 (Qiagen) to create pQE30-cpdA such that expression results in the production of an amino-terminal His6-tagged version of full-length CpdA (His6-CpdA). Site-directed mutagenesis of pQE30-cpdA was performed to generate His6-CpdA amino acid substitution mutations [H23A, D63A, and N93A, yielding His6-CpdA(H23A), His6-CpdA(D63A), and His6-CpdA(N93A), respectively] according to the manufacturer's protocol (Stratagene QuikChange II XL site-directed mutagenesis kit).
To create P. aeruginosa gene expression plasmids, cpdA was PCR amplified from P. aeruginosa strain PAK and E. coli strain DH1 genomic DNA and cloned into pMMBV1GW (encoding pCpdA) or pMMBGW (encoding pEcCpdA), respectively, using Gateway cloning (Invitrogen) as described previously (59). For expression of His6-CpdA or mutant variants in P. aeruginosa, pQE30-cpdA and its derivatives were digested with EcoRI and HindIII, and the cpdA-containing fragments were cloned into the EcoRI and HindIII sites of pMMBV2GW. The P. aeruginosa expression plasmid pMMBV1GW is a version of the previously described vector pMMBGW (59) in which the sequence of the −35 region of the tac promoter was changed from TTGACA to TTTACA. pMMBV2GW contains the alteration to the −35 region described above and an additional modification of the −10 region, changing TATAAT to CATTAT (S. Lory, unpublished data). These promoter region modifications reduce the efficiency of transcriptional initiation at the tac promoter, allowing for reduced levels of expression (data not shown).
For plasmid expression of vfr, the corresponding coding sequence was PCR amplified from genomic DNA of P. aeruginosa strain PA103 using primers incorporating NdeI and SacI restriction sites, and the resulting product was cloned into the NdeI and SacI sites of the arabinose-inducible expression vector pUY30 (54), resulting in plasmid p2UY22.
For His6-CpdA purification, an overnight culture of E. coli strain M15(pREP4) (Qiagen) expressing wild-type P. aeruginosa cpdA or mutant derivatives was diluted 1:100 in 500 ml of LB medium containing Cb (30 μg/ml) and kanamycin (Kn; 25 μg/ml). Cultures were grown with shaking at 37°C. When the OD600 reached 1.0, 500 μM isopropyl β-d-1-thiogalactopyranoside (IPTG) was added, and cultures were grown for 1 additional hour. Bacteria were collected by centrifugation (10,000 × g for 15 min at 4°C), and pellets were stored at −80°C. Cell lysates were prepared as described previously (38) with the exception that a protease inhibitor cocktail (Sigma) and 20 mM imidazole were included in the lysis buffer. His6-tagged CpdA was isolated under native conditions using Ni-nitrilotriacetic acid (NTA) agarose (Qiagen) according to the manufacturer's protocol (38) except that the buffer consisted of 50 mM NaH2PO4 (pH 7.0), 300 mM NaCl, 80 mM imidazole, 2 mM MgCl2, 20 mM β-mercaptoethanol, 30% glycerol, and 0.1% Tween. Protein was desalted into 50 mM NaH2PO4 (pH 7.0), 300 mM NaCl, and 10% glycerol using a HiTrap desalting column (GE Healthcare) and stored at −20°C.
For Vfr purification, P. aeruginosa strain PA103 exsA::Ω carrying p2UY22 was grown overnight at 37°C on Vogel-Bonner minimal (VBM) medium (55) containing gentamicin (Gm; 100 μg/ml). Bacteria were inoculated into 100 ml of LB medium containing 100 μg/ml Gm and grown at 37°C with shaking. When an OD600 of 0.3 was reached, arabinose was added to 0.4% to induce Vfr expression. Following an additional 4.5-h incubation, bacteria were harvested by centrifugation (5,000 × g for 5 min at 4°C), suspended in 2 ml of Vfr buffer (50 mM Tris-HCl [pH 7.0], 100 mM KCl, 50 mM NaCl, 2 mM dithiothreitol [DTT], 2 mM EDTA, 10% glycerol, 0.5% Tween-20) and lysed by sonication in an ethanol-ice bath. Lysates were centrifuged (16,000 × g for 5 min at 4°C) to remove unbroken cells and then again at 100,000 × g for 60 min to remove insoluble material. The soluble material was applied to a 1-ml cAMP-agarose (Sigma) column and washed with 15 ml of Vfr buffer, and protein was eluted with Vfr buffer containing 5 mM cAMP. The eluted protein was dialyzed against Vfr buffer and stored at −80°C.
Size exclusion chromatography of purified CpdA was performed on a Sephacryl S-200 high-resolution column (GE Healthcare) attached to an ÄKTAxpress system (GE Healthcare) using 50 mM NaH2PO4 (pH 6.8), 300 mM NaCl. A 1-ml aliquot of 10 μM CpdA was loaded onto the column at a flow rate of 1.2 ml/min. The size exclusion elution profile of CpdA was compared to profiles of the following protein standards to determine the oligomeric state: thyroglobulin (670 kDa), gamma-globulin (158 kDa), ovalbumin (44 kDa), myoglobin (17 kDa), and vitamin B12 (1.35 kDa).
Using a PDELight HTS cAMP phosphodiesterase kit (Lonza Rockland, Inc.), reaction mixtures (40 μl total volume) containing various concentrations of cAMP (0 to 20 μM), purified CpdA (2 ng), and reaction buffer (50 mM Tris-HCl [pH 7.6], 0.1 mM DTT) were incubated for 30 min at 25°C and then terminated. As indicated below and in the figure legends, reaction mixtures contained ferrous chloride (10 μM or 1 mM) and/or CpdA that was previously incubated with 100 μM α-α′-dipyridyl (Sigma) for 2 h at 4°C. In this assay, the 5′-AMP product formed during the reaction was quantified by a bioluminescent readout that was measured using a Veritas microplate luminometer. A standard curve in which the amount of bioluminescence produced from concentrations (0 to 0.2 μM) of pure 5′-AMP (Sigma) was used to calculate CpdA activity (nM 5′-AMP produced/min/ng of CpdA). Cyclic GMP (cGMP) phosphodiesterase activity for CpdA was measured using a PDEGlo phosphodiesterase assay (Promega) as described in the legend of Fig. S2 in the supplemental material.
Overnight cultures of wild-type and mutant P. aeruginosa strains were diluted 1:100 into LB medium. For plasmid-carrying strains, LB medium was supplemented with Cb (30 μg/ml) and IPTG (50 μM). When cultures reached an OD600 of 3.0, 2-ml aliquots of cells were collected by centrifugation (13,200 × g for 3 min) and washed twice with charcoal-treated phosphate-buffered saline (PBS; pH 7.0) containing 1% Bacto Protease Peptone (BD). Half of each sample was resuspended in charcoal-treated 0.1 N hydrochloric acid (100 μl), vortexed five times during a 15-min incubation on ice, and centrifuged (13,200 × g for 5 min). The supernatant was used for cAMP (or cGMP) determination using an immunoassay (Cyclic AMP EIA kit or Cyclic GMP EIA kit; Cayman Chemical Co.) following the acetylation protocol. The remaining sample was resuspended in PBS (100 μl), subjected to three freeze-thaw cycles, and centrifuged (13,200 × g for 3 min), and the supernatant was used for total protein determination (bicinchoninic acid [BCA] protein assay; Thermo Scientific). Assay results were used to calculate the intracellular cAMP (or cGMP) concentration based on the estimated cellular volume for P. aeruginosa per mg of protein (4.92 μl/mg of total cellular protein) (11). As indicated above, reagents were mixed with charcoal (10 g per liter) for 30 min to remove cyclic nucleotide contaminants and then filter sterilized.
Overnight cultures of strains containing vfr, exoS, and cpdA promoter-lacZ fusions were diluted 1:100 to 1:200 into LB medium and grown to mid-log phase (OD600 of 1.0). Cb (30 μg/ml), IPTG (50 μM), EGTA (5 mM), and MgCl2 (5 mM) were added to the medium as indicated in the figure legends. Levels of β-galactosidase were measured as described previously (31). Each assay was repeated at least three times.
For detection of Vfr protein, overnight cultures were diluted 1:200 into LB medium and grown to mid-log phase (OD600 of 1.0). Culture aliquots (1 ml) were centrifuged (13,200 × g for 3 min), and cell pellets were resuspended in 100 μl of PBS (pH 7.0) and subjected to three freeze-thaw cycles. After centrifugation (13,200 × g for 3 min), the protein concentration of the supernatant was determined (BCA protein assay). Samples were normalized for total protein, diluted in SDS loading buffer, boiled for 1 min, and loaded onto a 12% SDS-polyacrylamide gel for electrophoresis. Proteins were transferred to a nitrocellulose membrane, and Vfr was detected using anti-Vfr rabbit serum (1:25,000 dilution) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1:25,000 dilution; Rockland). Blots were developed with Immobilon Western chemiluminescence reagents (Millipore) and visualized by autoradiography. Detection of His6-CpdA and mutant derivatives was carried out as described above except that overnight cultures were diluted 1:100 into LB medium containing Cb (30 μg/ml) and IPTG (50 μM) and were grown to an OD600 of 3.0. Anti-His tag monoclonal antibody (1:10,000 dilution; Millipore) and HRP-conjugated anti-mouse IgG (1:10,000 dilution; Jackson) were used as primary and secondary antibodies, respectively.
For detection of secreted ExoS protein, overnight cultures were diluted 1:100 into LB medium containing 5 mM EGTA and 5 mM MgCl2 and grown to an OD600 of ~1.0. For detection of secreted ToxA, overnight cultures were diluted 1:100 into deferrated Bacto tryptic soy broth (BD) (36) and grown to an OD600 of ~2.0. Aliquots of culture supernatants (1 ml) were subjected to precipitation by trichloroacetic acid (75 μl) for 1 h on ice, centrifuged (13,200 × g for 15 min at 4°C), and washed with acetone (1 ml) overnight at 4°C. Protein pellets were dried and resuspended in 50 μl of PBS. Samples were normalized and analyzed by Western blotting as described above using anti-ExoS rabbit serum (1:3,000 dilution), anti-Pseudomonas exotoxin A (ToxA) rabbit serum (1:100,000 dilution; Sigma), and HRP-conjugated goat anti-rabbit IgG (1:25,000 dilution for ExoS; 1:100,000 dilution for ToxA; Rockland). All Western blotting experiments were repeated a minimum of three times with independently derived protein samples, and representative blots are shown.
Overnight cultures were diluted 1:100 into LB medium and grown to early stationary phase (OD600 of 4.0). Bacteria were pelleted by centrifugation (13,200 × g for 3 min), and supernatants were concentrated 40× using Vivaspin 2 spin filters (molecular-weight-cutoff [MWCO] of 10 kDa; Sartorius Stedium Biotech) and assayed for total protein content (BCA protein assay). Protease IV activity was determined as described previously (35) with the exception that 50 mM Tris-HCl (pH 7.6)-0.1 M NaCl was used as the reaction buffer, and the absorbance at 410 nm was measured at 10-min intervals for 1 h upon addition of the substrate N-(p-tosyl)-Gly-Pro-Lys-4-nitroanilide (Chromozym PL; Roche) to the reaction mixture. Protease IV activity was calculated as the change in absorbance at 410 nm over time, divided by the protein concentration of concentrated culture supernatant.
DNA probes were generated by PCR and end labeled using 10 μCi of [γ-32P]ATP (GE Healthcare) and 10 U of T4 polynucleotide kinase (New England Biolabs). Electrophoretic mobility shift assays (EMSAs) were performed as previously described (8). Briefly, probes (0.25 nM each) were incubated in binding buffer (10 mM Tris [pH 7.5], 50 mM KCl, 1 mM EDTA, 1 mM DTT, 5% glycerol and 100 μg/ml bovine serum albumin) containing 5 μg/ml poly(2′-deoxyinosinic-2′-deoxycytidylic acid) (Sigma) for 5 min at 25°C. Purified Vfr protein was added to the concentration indicated in the figure legends for a final reaction volume of 20 μl, and the mixture was incubated for an additional 15 min at 25°C. Samples were subjected to electrophoresis on a 5% polyacrylamide glycine gel (10 mM Tris [pH 7.5], 380 mM glycine, 1 mM EDTA) at 4°C. Imaging and data analyses were performed using an FLA-7000 phosphorimager (Fujifilm) and MultiGauge, version 3.0, software (Fujifilm).
Single end-labeled γ-32P double-stranded DNA probes were generated by PCR in which one oligonucleotide primer was modified (5 Amino-MC6; Integrated DNA Technologies) at the 5′ end to prevent phosphorylation. Probes were subsequently labeled as described above. Footprinting reaction mixtures consisted of end-labeled probe (10 fmol) with 5 μg/ml poly(2′-deoxyinosinic-2′-deoxycytidylic acid) (Sigma) in DNase I reaction buffer (10 mM Tris [pH 8.0], 50 mM KCl, 2 mM MgCl2, 0.5 mM DTT, 100 μg/ml bovine serum albumin, 10% glycerol). Purified Vfr was added for a final reaction volume of 25 μl, and the mixture was incubated for 15 min at 25°C. DNase I footprinting and DNA sequencing reactions were performed as previously described (30, 43, 50).
A two-tailed unpaired t test was used for data comparison where appropriate. Statistical calculations and estimation of enzyme kinetic parameters were preformed using Prism software, version 5.0b (GraphPad Software).
The cpdA gene (locus PA4969) was identified upon searching the Pseudomonas Genome Database (58) for genes homologous to E. coli cpdA (21, 33). Sequence corresponding to the cpdA open reading frame was PCR amplified from genomic DNA of P. aeruginosa strain PAK using primers designed from the P. aeruginosa strain PAO1 genomic sequence (58). The DNA fragment was cloned into pMMBV1GW as described previously (59), and both DNA strands were sequenced (Eurofins MWG Operon); specific primer sequences are available upon request. The PAK cpdA nucleotide sequence (which contains nucleotide differences compared to the PAO1 cpdA sequence) was assembled with Sequencher, version 4.8, software.
The P. aeruginosa PAK cpdA (PA4969) nucleotide sequence was deposited in the GenBank sequence data library under the accession number GU551724.
To identify and characterize the enzyme(s) responsible for cAMP degradation in P. aeruginosa, we searched the Pseudomonas Genome Database (58) for genes whose products were homologous to members of the class I, II, and III families of cAMP phosphodiesterases. We found a single gene (PA4969) whose predicted product shares 41% identity with E. coli CpdA (21) and 21% identity with a previously characterized cyclic nucleotide phosphodiesterase of M. tuberculosis (Rv0805) (46) (Fig. (Fig.1A).1A). Like E. coli CpdA and Rv0805, the P. aeruginosa protein contains the highly conserved sequence motif D-(X)n-GD-(X)n-GNH(E/D)-(X)n-H-(X)n-GHXH (where X is any residue, and n is the number of repeats) that is characteristic of known class III cyclic nucleotide phosphodiesterases, purple acid phosphatases, and other dimetallophosphoesterases (41). Thus, based on primary sequence analysis, we have designated the PA4969 gene product CpdA.
To determine whether P. aeruginosa CpdA possessed phosphodiesterase activity, a full-length amino-terminal hexahistidine-tagged version of CpdA was purified under nondenaturing conditions (see Fig. S1 in the supplemental material). Size exclusion chromatography revealed that CpdA had an elution profile representative of an ~31-kDa protein, indicating that CpdA (which has a predicted molecular mass of 30 kDa) is monomeric in solution (data not shown). Cyclic AMP phosphodiesterase activity was determined in vitro by measuring the amount of 5′-AMP produced at various concentrations of cAMP substrate (0 to 20 μM) with a constant amount of CpdA (2 ng) protein. In the presence of CpdA, 5′-AMP was produced in a linear fashion with increasing cAMP concentrations up to ~8 μM (Fig. (Fig.1B).1B). In contrast, there was no detectable increase in 5′-AMP when CpdA was not included in the reaction mixture (data not shown). These results demonstrate that P. aeruginosa CpdA is, in fact, a cAMP phosphodiesterase with an estimated Km of 7.2 ± 1.4 μM cAMP and Vmax of 3.4 ± 0.3 nM 5′-AMP produced/min/ng of enzyme. While a proven biological role for 3′,5′-cyclic GMP (cGMP) in bacteria remains to be determined (26), some class III cyclic nucleotide phosphodiesterases can use both cAMP and cGMP as substrates in vitro (6, 15, 46). We observed that CpdA of P. aeruginosa also exhibited cGMP phosphodiesterase activity in vitro (see Fig. S2); however, enzyme kinetics could not be determined by the assay method employed, and this activity was not pursued further due to our inability to detect the presence of cGMP in P. aeruginosa strains (see below).
In purple acid phosphatases and other dimetallophosphoesterases, residues within the conserved sequence motif D-(X)n-GD-(X)n-GNH(E/D)-(X)n-H-(X)n-GHXH coordinate a Fe3+-Me2+ center (where Me2+ can be Fe2+, Zn2+, or Mn2+) that is required for enzyme activity (41). Since CpdA shares this sequence motif, we examined the metal ion requirement for CpdA activity. Similar to previous in vitro studies of E. coli CpdA (21, 33), addition of 10 μM ferrous chloride, as a source of Fe2+, resulted in an ~2-fold stimulation of P. aeruginosa CpdA activity (Fig. (Fig.1B).1B). The addition of ferrous chloride did not significantly influence substrate affinity (Km of 6.7 ± 1.2 μM cAMP) but did increase the rate of the 5′-AMP production (Vmax of 5.6 ± 0.4 nM 5′-AMP produced/min/ng of enzyme). The addition of other divalent metals (Mg2+, Mn2+, Zn2+, and Ca2+) had no effect (data not shown). Furthermore, treatment of CpdA with the Fe2+-specific chelator α-α′-dipyridyl resulted in a nearly complete loss of activity (Fig. (Fig.1B);1B); addition of ferrous chloride (1 mM) to the α-α′-dipyridyl-treated CpdA fully restored cAMP phosphodiesterase activity (data not shown). These results suggest that the catalytic mechanism for P. aeruginosa CpdA utilizes a Fe3+-Fe2+ center.
To further investigate which residues within the D-(X)n-GD-(X)n-GNH(E/D)-(X)n-H-(X)n-GHXH motif likely participate in the binding of iron, we constructed a homology model of CpdA based on the solved crystal structure of Rv0805 (45) (Fig. (Fig.1C).1C). This model reveals that the highly conserved amino acids D21, H23, D63, N93, H162, H200, and H202 (Fig. (Fig.1A)1A) cluster together in the three-dimensional protein structure (Fig. (Fig.1C),1C), suggesting that they may form the Fe3+-Fe2+ binding site. To address the significance of these conserved residues in P. aeruginosa CpdA function, alanine substitutions of three different CpdA amino acids (H23, D63, and N93) were constructed. The resulting mutant proteins were purified and tested for cAMP phosphodiesterase activity in vitro. Unlike the wild-type CpdA, there was no detectable activity (≤0.1% of wild-type CpdA) for any of the mutant proteins [data for CpdA(H23A) are shown in Fig. Fig.1B].1B]. In addition, the mutant proteins had no detectable activity when assayed in the presence of 10 μM ferrous chloride (data not shown). Taken together, these results indicate that the highly conserved residues H23, D63, and N93 are essential for CpdA activity and are consistent with their proposed role in coordinating iron ions within the active site.
To determine if CpdA plays a role in regulating cAMP levels in vivo, the intracellular cAMP concentration was determined for wild-type P. aeruginosa strain PAK and an isogenic mutant in which the cpdA gene was deleted. The cpdA mutant contained nearly 30-fold higher cellular cAMP levels than the wild-type strain (Fig. (Fig.2A),2A), suggesting that the cpdA gene product participates in decreasing the cAMP concentration in vivo. Introduction of the cpdA deletion into other commonly studied P. aeruginosa strains (PA14, PA103, and PAO1) also resulted in a dramatic increase in intracellular cAMP (see Fig. S3A in the supplemental material), indicating that CpdA function is conserved among P. aeruginosa strains. The cellular cAMP concentration in the cpdA mutant could be reduced upon plasmid-based expression of either P. aeruginosa cpdA or E. coli cpdA (Fig. (Fig.2B),2B), demonstrating that the mutant phenotype was specifically due to deletion of cpdA. In addition, plasmid-encoded His6-CpdA also complemented the cpdA mutant, indicating that His6-CpdA is functional in vivo (Fig. (Fig.2B).2B). In contrast, when expressed to an equivalent level, the putative active-site mutants of His6-CpdA were unable to complement the cpdA mutant (Fig. (Fig.2B).2B). Thus, these results correlate with those observed in the in vitro assay and further demonstrate that the conserved residues H23, D63, and N93 are required for CpdA activity.
To evaluate the balance of cAMP synthesis and degradation, we also determined the effect of the cpdA deletion on cAMP levels in mutants lacking the endogenous adenylate cyclases (ACs) CyaA and CyaB. As previously observed (59), cellular cAMP levels were slightly lower in a cyaA mutant and reduced by ~50% in a cyaB mutant relative to the wild type (Fig. (Fig.2A).2A). These results are consistent with the notion that in P. aeruginosa, the contribution of CyaB to the cAMP pool is substantially greater than that of CyaA. Upon deletion of cpdA, cAMP levels increased 25-fold in the cyaA mutant. Deletion of cpdA in the cyaB mutant strain resulted in a 9.5-fold increase in intracellular cAMP. These results indicate that in wild-type P. aeruginosa, CpdA and CyaB are the major enzymes responsible for controlling cellular cAMP concentration. Cyclic AMP levels were ~3-fold higher in the cyaA cyaB cpdA triple mutant than in the cyaA cyaB double mutant even though both strains lack the two known P. aeruginosa ACs responsible for cAMP synthesis. P. aeruginosa, like other Gram-negative bacteria, is capable of acquiring extracellular cAMP by an unknown transport mechanism (7, 15a, 42). Since contaminating cAMP was not removed from the bacterial growth medium (LB), its uptake may account for the elevated level detected in the triple mutant. However, low activity associated with a previously unidentified AC cannot be ruled out.
To determine whether CpdA also affects intracellular cGMP levels in P. aeruginosa, we measured the cGMP content in wild-type and cpdA mutant versions of strains PAK, PA14, PA103, and PAO1. In all cases, the intracellular level of cGMP was below the limit of detection (≤0.01 μM) (data not shown), suggesting that cGMP is unlikely to be synthesized by P. aeruginosa and that cAMP is the biologically relevant substrate for CpdA in vivo.
We have recently shown that the cAMP-Vfr complex, in addition to regulating hundreds of genes involved in virulence (59), directly controls vfr promoter activation (E. L. Fuchs et al., submitted for publication). Thus, vfr gene expression is autoregulated and cAMP dependent. To first address how virulence gene expression is affected by elevated intracellular cAMP levels, we determined if vfr promoter activity was altered in the cpdA mutant by measuring β-galactosidase activity in strains containing a chromosomal copy of the vfr promoter-lacZ fusion. In the cpdA mutant, vfr promoter activity was ~12-fold higher than in the parental wild-type strain (Fig. (Fig.3A).3A). Similar increases were observed in cpdA mutants that lacked cyaA, cyaB, or both AC genes. Plasmid-based expression of either P. aeruginosa cpdA or E. coli cpdA restored vfr promoter activity in the cpdA mutant to the level observed for the wild-type strain (Fig. (Fig.3C).3C). Thus, increased concentrations of cAMP resulted in elevated vfr gene expression. Consistent with this result, Western blot analysis revealed that Vfr protein levels were substantially higher in cpdA mutant strains than in the respective parent strains (Fig. (Fig.3B).3B). Taken together, these data suggest that elimination of cpdA results in elevated levels of the cAMP-Vfr complex.
Given that CpdA plays a role in controlling intracellular cAMP-Vfr levels, we examined the effect of the cpdA deletion on virulence factor production. ExoS is a toxic effector protein secreted by the T3SS directly into host cells (12). Expression of the T3SS and its effector-encoding genes is regulated by cAMP-Vfr; however, the mechanism of regulation is currently unknown (59, 60). To measure expression of exoS, we created a transcriptional reporter by fusing the exoS promoter region to the lacZ gene and introduced the fusion in single copy onto the chromosome of wild-type and mutant strains. β-Galactosidase assays revealed that with exception of the vfr mutant, exoS promoter activity was significantly higher in all cpdA mutants than in the respective cpdA+ parent strains (Fig. (Fig.4A).4A). In addition, Western blot analysis showed that the amount of secreted ExoS protein was higher for the cpdA mutants (Fig. (Fig.4B).4B). Thus, transcription of exoS and secretion of its gene product are increased in response to elevated levels of cAMP-Vfr.
Similar trends were observed for two additional secreted virulence factors: ToxA (an ADP-ribosyltransferase) and protease IV. Unlike ExoS, ToxA and protease IV are secreted by a type II secretion system. Previous studies have shown that Vfr is required for the production and secretion of ToxA and protease IV and that cAMP-Vfr directly binds the promoter region of their respective genes, toxA and prpL (piv) (22, 56). Western blot analysis revealed that the level of ToxA in bacterial culture supernatants was higher in mutant strains lacking cpdA (Fig. (Fig.4C).4C). Protease IV secretion was examined indirectly by measuring protease IV-dependent proteolysis of the chromogenic substrate N-(p-tosyl)-Gly-Pro-Lys-4-nitroanilide. Protease IV activity was elevated ~ 2-fold in culture supernatants derived from the cpdA and cyaA cpdA mutant strains compared to wild-type and cyaA strains, respectively (Fig. (Fig.4D).4D). Consistent with this observation, the cpdA mutant produced a larger zone of proteolytic clearing than the wild type on low-salt LB agar containing 1.5% milk (Fig. (Fig.4D).4D). Taken together, these results indicate that CpdA plays a role in maintaining an intracellular cAMP level that is most likely optimized for appropriate virulence factor expression.
In other bacterial species, deletion of genes encoding CpdA homologs results in pleiotropic effects on cell growth. For instance, cpdA mutants of Salmonella enterica serovar Typhimurium have increased antibiotic sensitivities and faster growth in the presence of certain carbon sources than their parent strains (2). In addition, it was recently reported that in Vibrio vulnificus, deletion of cpdA resulted in an elongated cell morphology compared to a wild-type strain (24). During the course of this study, we observed a growth phenotype associated with the loss of cpdA in P. aeruginosa. When grown in rich liquid medium (LB medium) (Fig. (Fig.5A),5A), the cpdA mutant exhibited a significantly reduced growth rate compared to that of the parental wild-type strain; the doubling time for the cpdA mutant (51 ± 6 min) was approximately twice that of the wild type (28 ± 5 min). The reduced growth rate was specific for rich medium since the cpdA mutant and wild-type strains had similar doubling times when they were grown in a minimal defined medium containing either 10 mM succinate (see Fig. S4A in the supplemental material) or 10 mM glucose (data not shown) as the sole carbon source. In addition to the slow-growth phenotype in LB broth, the cpdA mutant also produced small colonies on solid medium (LB agar) (Fig. (Fig.5B).5B). The growth rate and colony morphology could be restored in the cpdA mutant when it was complemented with plasmid-expressed P. aeruginosa cpdA or E. coli cpdA (Fig. (Fig.55).
To determine whether elevated intracellular cAMP levels associated with the loss of cpdA accounted for the defect in cell growth, we examined the contribution of the endogenous ACs CyaA and CyaB. The doubling times for cpdA+ AC mutants were similar to the doubling time of the wild type. However, deletion of cpdA in the AC mutant strains resulted in significantly slower growth in rich liquid medium (Fig. (Fig.5A)5A) and the formation of small colonies (data not shown). These results indicate that increased cAMP is not responsible for the altered growth phenotype and suggest that P. aeruginosa CpdA is likely to have an additional function beyond its role in cAMP degradation. Interestingly, the cpdA-specific small-colony phenotype could not be complemented with plasmids expressing the active-site mutants CpdA(H23A), CpdA(D63A), and CpdA(N93A) (see Fig. S4B in the supplemental material), indicating that these conserved residues are required for wild-type growth.
As shown in Fig. Fig.2A,2A, introduction of the cpdA deletion into wild-type strain PAK and the isogenic cyaA and cyaB mutant strains resulted in increased intracellular cAMP. However, the contribution of CpdA in each background was disproportionate. For example, cAMP levels in the cpdA mutant were nearly 30-fold higher than the level in the wild-type strain but only 9.5-fold higher in the cpdA cyaB double mutant than in the cyaB mutant. This finding suggests that CpdA activity or expression may be downregulated in strains that synthesize less cAMP. To further explore this possibility, we tested whether cAMP-Vfr regulates cpdA expression. Specifically, we constructed a single-copy cpdA transcriptional reporter gene by fusing the putative cpdA promoter region to a promoterless lacZ gene at a vacant CTX phage integration site on the P. aeruginosa chromosome. Expression of the cpdA promoter-lacZ fusion in strains lacking the ACs, vfr, or cpdA was assessed using β-galactosidase assays. Compared to the wild-type strain, β-galactosidase activity decreased ~3- and 5-fold in the cyaA cyaB double mutant and the vfr mutant, respectively. In contrast, activity increased ~2-fold in the cpdA mutant (Fig. (Fig.6A).6A). These results suggest that cpdA expression is regulated in response to the cellular concentration of cAMP via the cAMP-Vfr complex.
To determine if cAMP-Vfr directly regulates cpdA, an electrophoretic mobility shift assay (EMSA) was performed with purified cAMP-Vfr and three DNA probes, encompassing different regions of the cpdA promoter (Fig. (Fig.6B).6B). The mobility of probe 3, corresponding to −231 bp to −1 bp relative to the cpdA translational start codon, was specifically retarded by the cAMP-Vfr complex in a concentration-dependent manner. Furthermore, DNase I footprinting analysis revealed that cAMP-Vfr altered the cleavage pattern of a region within the cpdA promoter spanning from −112 bp to −76 bp relative to the cpdA translational start codon (Fig. (Fig.6C).6C). Specifically, cAMP-Vfr protected a 25-bp sequence (−104 bp to −80 bp) from DNase I cleavage with the exception of strong hypersensitive sites at positions −86, −96, and −97. There was also enhanced DNase I cleavage at positions −76 and −112 bp. A similar pattern of DNase I hypersensitivity was observed in footprints of the lasR, ptxR, toxA, and regA promoters in the presence of cAMP-Vfr (1, 13, 22). Furthermore, 14 of the bases (underlined) within the protected sequence of the cpdA promoter (5′-CAACTGTGATCT·GTTCCGCTT-3′) matched the reported consensus Vfr binding sequence (5′-ANWWTGNGAWNY·AGWTCACAT-3′, where W is A or T, Y is T or C, and N is any nucleotide) (22). Taken together, these data reveal that cAMP-Vfr binds to the cpdA promoter region, suggesting that in vivo, cpdA transcription is directly activated by cAMP-Vfr.
We have provided definitive in vitro evidence that purified CpdA of P. aeruginosa hydrolyzes the second messenger signaling molecule cAMP to form biologically inactive 5′-AMP (Fig. (Fig.1B).1B). Furthermore, the dramatic increase in intracellular cAMP in cpdA mutant strains indicates that CpdA exhibits cAMP phosphodiesterase activity in vivo (Fig. (Fig.2;2; see also Fig. S3A in the supplemental material). While CpdA has features in common with other class III cyclic nucleotide phosphodiesterases, such as the metal ion requirement and conserved amino acid residues, several substantial differences in the properties of these bacterial enzymes exist. For instance, the Km calculated for P. aeruginosa CpdA (7.2 ± 1.4 μM for cAMP) is ~6-fold lower than that of E. coli CpdA (~45 uM) (21, 33), indicating that P. aeruginosa CpdA may exhibit a stronger affinity for cAMP. In addition, while both of these enzymes function as monomers, a previous study has shown that Rv0805 of M. tuberculosis functions as a dimer (46). Differences in oligomeric state may account for alterations in binding of cAMP and/or the metal ion cofactors and may explain why alanine substitution of specific conserved active-site residues had variable effects on Rv0805 activity (46) but completely inactivated P. aeruginosa CpdA. Finally, in contrast to the finding that E. coli CpdA does not degrade cGMP (21), we observed that P. aeruginosa CpdA exhibited cGMP phosphodiesterase activity in vitro (see Fig. S2), similar to Rv0805 (46) and the CpdA enzyme from the cyanobacterium Anabaena sp. strain PCC 7120 (15). While these enzymes appear to have dual specificity like many class I mammalian cyclic nucleotide phosphodiesterases (6), the relevance of cGMP degradation in P. aeruginosa is unclear, given the lack of measurable cGMP synthesis in vivo.
Although cAMP phosphodiesterases are ubiquitous in nature, the global role these enzymes play in bacteria is not well understood. In this study, we investigated how the cAMP phosphodiesterase activity of CpdA contributes to cAMP signaling in P. aeruginosa. Since the cAMP-Vfr complex affects the expression of ~200 genes, many of which are involved in virulence (59), it is not surprising that cAMP levels are tightly controlled. Our data reveal that CpdA plays a major role in this process by reducing intracellular cAMP and, in turn, affecting cAMP-Vfr-dependent virulence factor expression. We propose that CpdA activity would be advantageous in the absence of a host in order to maintain basal cAMP levels. For example, in the environment, CpdA activity would limit cAMP accumulation, thus preventing the unnecessary expression of virulence factors. Furthermore, in the context of infection, CpdA activity may allow P. aeruginosa to reset the cAMP signaling cascade following AC activation in response to spatiotemporal signals present in the host. Studies have demonstrated that inhibition of virulence factor production occurs when P. aeruginosa has established a chronic infection, perhaps providing a means to escape host immune detection (32). Thus, the ability to regulate cAMP levels through synthesis (CyaB) and degradation (CpdA) may enable P. aeruginosa to readily adapt to its surroundings.
Currently, the reason for the slow-growth, small-colony phenotype associated with the cpdA mutant is not clear as we have ruled out the possibility that elevated cAMP levels or Vfr-dependent transcriptional activity are responsible (Fig. (Fig.5).5). In addition, the lack of detectable cGMP in wild-type and cpdA mutant strains suggests that it is unlikely that altered cGMP levels cause the growth phenotype. Interestingly, a recent study has shown that the M. tuberculosis cAMP phosphodiesterase Rv0805 can hydrolyze multiple phosphoester substrates, suggesting that the growth defect in P. aeruginosa cpdA mutants may reflect the involvement of CpdA in other metabolic pathways (37). In addition, Rv0805 was shown to localize to the M. tuberculosis membrane and cell wall, suggesting a possible role in membrane function (37). Increased antibiotic sensitivity and elongated cell morphology of S. typhimurium and V. vulnificus cpdA mutants, respectively, are consistent with the idea that bacterial cAMP phosphodiesterases play an ancillary role in cell membrane function (2, 24) and could account for the slow-growth, small-colony phenotype of P. aeruginosa cpdA mutants.
In this study, we used a combination of in vitro and in vivo experiments to show that cpdA is a novel target of the cAMP-Vfr regulon (Fig. (Fig.6).6). Although whole-genome transcriptional profiling did not previously identify cpdA as being regulated by cAMP or Vfr (59), different bacterial growth conditions used in our study may explain this inconsistency. Our data clearly show that cpdA transcription is activated by the cAMP-Vfr complex in vivo. In turn, CpdA protein reduces cAMP levels and, hence, cAMP-Vfr-dependent gene expression. This feedback mode of regulation would be advantageous in optimizing the intracellular cAMP concentration and raises the question as to whether there is a hierarchy of gene expression within the cAMP-Vfr regulon. We propose that activation of cAMP synthesis (in response to some as yet unknown host cue) would, in turn, increase cAMP-Vfr levels and subsequent Vfr-dependent virulence gene expression. The simultaneous upregulation of cpdA by cAMP-Vfr would lead to subsequent degradation of cAMP (signal attenuation) and would facilitate the finite expression of cAMP-Vfr dependent virulence genes.
Transcriptional regulation of cpdA also appears to occur in other bacterial systems. For instance, within the P2 promoter region of icc, a cpdA homolog in Haemophilus influenzae, a putative CRP binding site has been identified, suggesting that cAMP-CRP may regulate icc expression (29). In addition, it was recently shown that in V. vulnificus, the cAMP-CRP complex directly activates transcription of the cpdA gene (24). Interestingly, in V. vulnificus, cpdA is part of the mutT-yqiB-cpdA-yqiA operon whose expression requires a CRP binding site located upstream of the mutT transcription start site (24). In contrast, the region immediately upstream of P. aeruginosa cpdA (−500 to +100 bp relative to the cpdA start codon) was sufficient for regulation of cpdA by cAMP-Vfr in vivo (Fig. (Fig.6A).6A). Furthermore, DNase I footprinting results indicate that cAMP-Vfr directly binds to a region of the cpdA promoter which corresponds to −104 bp to −80 bp relative to the cpdA translational start codon (Fig. (Fig.6C).6C). Thus, it appears that differences in the locations of Vfr and CRP binding sites relative to the cpdA coding sequence may exist among bacterial species.
The finding that cAMP-Vfr controls cpdA transcription raises the question as to whether there are additional modes of CpdA regulation. Interestingly, cpdA translation from both P. aeruginosa and E. coli initiates at a rare start codon, UUG (21). It has been proposed that expression of E. coli cpdA may be limited at the translational level based on the finding that expression of E. coli cyaA, which also possesses the UUG start codon, increases upon changing the start codon to GUG or the more common AUG (40). Furthermore, there is evidence that in E. coli, inorganic phosphate reduces cAMP phosphodiesterase activity in vivo (3) and that CpdA activity is inhibited by both phosphate and succinate in vitro (34). However, we observed that wild-type and cpdA mutant strains grew at similar rates in minimal defined medium containing either succinate or glucose, suggesting that succinate does not inhibit P. aeruginosa CpdA activity in vivo. Nevertheless, further experiments are needed to establish whether translation or enzyme activity of P. aeruginosa CpdA is subject to regulation.
In summary, the results of this study establish the function of P. aeruginosa CpdA as a cAMP phosphodiesterase and provide evidence that cAMP degradation plays a central role in maintaining the intracellular levels of cAMP in P. aeruginosa. Our findings provide valuable, new insight as to how this bacterial pathogen maintains control of virulence gene expression. Furthermore, the finding that CpdA is regulated at the transcriptional level has shed light on an additional mechanism that contributes to cAMP homeostasis. These results not only demonstrate the significance of cAMP phosphodiesterases in cAMP signaling but also open new areas of inquiry regarding the contribution of cAMP phosphodiesterases to other cellular processes.
This work was supported by grants from the Cystic Fibrosis Foundation (to M.C.W.) and the National Institutes of Health (AI069116 to M.C.W. and AI055042 to T.L.Y.). E.L.F. was supported by a Pathogenesis Training Grant from the University of North Carolina Center for Infectious Diseases.
We thank members of the Wolfgang and Yahr laboratories for their constructive suggestions and critical review of the manuscript. We thank Mark Urbanowski for purified Vfr protein, Katrina Forest for anti-Vfr serum, Åke Forsberg for anti-ExoS serum, Matthew Redinbo for assistance in generating a structural model of CpdA and Michael Johnson for his help in performing size exclusion chromatography.
Published ahead of print on 26 March 2010.
†Supplemental material for this article may be found at http://jb.asm.org/.