|Home | About | Journals | Submit | Contact Us | Français|
Pseudomonas aeruginosa is an opportunistic pathogen that causes both acute and chronic infections in immunocompromised individuals. This gram-negative bacterium produces a battery of virulence factors that allow it to infect and survive in many different hostile environments. The control of many of these virulence factors falls under the influence of one of three P. aeruginosa cell-to-cell signaling systems. The focus of this study, the quinolone signaling system, functions through the Pseudomonas quinolone signal (PQS), previously identified as 2-heptyl-3-hydroxy-4-quinolone. This signal binds to and activates the LysR-type transcriptional regulator PqsR (also known as MvfR), which in turn induces the expression of the pqsABCDE operon. The first four genes of this operon are required for PQS synthesis, but the fifth gene, pqsE, is not. The function of the pqsE gene is not known, but it is required for the production of multiple PQS-controlled virulence factors and for virulence in multiple models of infection. In this report, we show that PqsE can activate PQS-controlled genes in the absence of PqsR and PQS. Our data also suggest that the regulatory activity of PqsE requires RhlR and indicate that a pqsE mutant can be complemented for pyocyanin production by a large excess of exogenous N-butyryl homoserine lactone (C4-HSL). Finally, we show that PqsE enhances the ability of Escherichia coli expressing RhlR to respond to C4-HSL. Overall, our data lead us to conclude that PqsE functions as a regulator that is independent of PqsR and PQS but dependent on the rhl quorum-sensing system.
Pseudomonas aeruginosa is a serious opportunistic pathogen that can cause infections in a diverse group of organisms from the animal, plant, and insect kingdoms (9, 47, 53). In humans, P. aeruginosa can cause multiple types of infections that are generally difficult to treat due to the high level of antibiotic resistance exhibited by this microbe (15). These infections are complex and generally involve numerous virulence determinants, with some factors playing a larger role than others in different types of infections (16). One aspect of P. aeruginosa virulence that seems to be important in most infections is the ability of the bacteria to communicate via intercellular signals (11). P. aeruginosa uses at least three cell-to-cell signaling systems to control the expression of assorted virulence factors. There are two classical acyl-homoserine lactone (HSL)-based quorum-sensing systems, the las and rhl systems (see reference 43 for a review of P. aeruginosa quorum sensing). The las and rhl quorum-sensing systems function via the signals N-(3- oxododecanoyl) HSL (3-oxo-C12-HSL) and N-butyryl HSL (C4-HSL), respectively (40, 41). When at a threshold concentration, 3-oxo-C12-HSL and C4-HSL serve as coinducers for the transcriptional activators LasR and RhlR, respectively (21, 35, 36, 39). The other cell-to-cell signaling system of P. aeruginosa functions through the quinolone compound 2-heptyl-3-hydroxy-4-quinolone (the Pseudomonas quinolone signal [PQS]) (44). This signal serves as a coinducer for the transcriptional activator PqsR (also known as MvfR), which positively regulates PQS production to create an autoregulatory loop (5, 32, 54, 58). These three cell-to-cell signaling systems are connected by multiple regulatory pathways. The las system appears to be at the top of a hierarchy as it positively regulates both the rhl quorum-sensing system and the quinolone signaling system (27, 44, 45). The rhl and quinolone systems are also connected in that C4-HSL negatively affects PQS production, and PQS positively regulates rhlR (14, 26, 32, 33). With regard to genes that are controlled by cell-to-cell signaling, most are dominantly regulated by one system or the other (51, 55). Such is the case for rhlA, which is positively controlled by the rhl quorum-sensing system and is required for the production of rhamnolipids, surfactants that are important for virulence (35, 42, 49). However, there is also overlap between the systems with regard to some of the genes that are controlled. For example, lasB, which encodes the virulence factor LasB elastase, is directly regulated by both the las and rhl quorum-sensing systems and directly or indirectly regulated by quinolone signaling (14, 42, 44, 52). This complex interplay between the cell-to-cell signaling systems is one of the ways that P. aeruginosa finely tunes the expression of many genes in response to its environment.
With regard to PQS, the quinolone signaling system has been shown to be important for virulence, and PQS is produced in the lungs of cystic fibrosis patients infected by P. aeruginosa (7, 12, 19, 20, 28, 29, 48). PQS also can act as an iron chelator, and both the synthesis of PQS and the activity of PqsR-PQS are involved in iron homeostasis, another indication of the global importance of quinolone signaling (1, 37). In addition, PQS has a membrane-altering activity that is linked to the formation of membrane vesicles (30, 31). To date, quinolone signaling has been shown to directly or indirectly control 92 or 143 genes as determined by two separate transcriptome analysis studies (1, 12). To activate genes, PqsR primarily responds to PQS, but it can also be activated by the PQS precursor, 2-heptyl-4-quinolone, which is 100-fold less potent than PQS (58). PqsR can also interact with and be negatively affected by fungus-produced sesquiterpenes, which share some structural similarities to PQS (8).
It has been shown that PqsR-PQS directly activates at least the pqsABCDE and phnAB operons (5, 54). All of the members of these two operons, except for pqsE, appear to encode enzymes involved in the synthesis of 4-quinolone compounds, including PQS (10, 20). Curiously, mutation of the pqsE gene has no effect on the production of PQS or other 4-quinolone compounds, but the production of PQS-controlled virulence factors is negatively affected in pqsE mutants (13, 20). It has been reported that pqsE mutants do not produce pyocyanin or PA-IL lectin and that they produce much lower amounts of elastase and rhamnolipid (14, 20). In addition, like pqsA and pqsR mutants (which do not produce PQS), a pqsE mutant is avirulent in mice (12). Overall, the pqsE gene appears to play no role in the synthesis of PQS but seems to be absolutely required for the cells to respond to PQS. PqsE is predicted to be a 34.3-kDa protein that is most likely not secreted and is localized to the cytoplasm, according to in silico analysis with SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP-3.0/) and CELLO (version 2.5) subcellular localization predictor (http://cello.life.nctu.edu.tw/). A BLAST search of available genomes indicated that PqsE is related to the metallo-β-lactamase superfamily, members of which have a conserved motif able to bind up to two metal ions in their active sites (4). Unfortunately, this information provided no real indication of how PqsE allows the cell to respond to PQS. Therefore, we have begun a preliminary study of PqsE to try to understand how it can allow a cell to induce PQS-controlled virulence factors. We present data that suggest that PqsE serves a regulatory function and that this function is linked to the activation of RhlR.
Bacterial strains used in this study are listed in Table Table1.1. Strains of P. aeruginosa were maintained at −70°C in 10% skim milk (Becton Dickinson), and freshly plated bacteria were used to begin each experiment. Bacteria were cultured at 37°C in Luria-Bertani (LB) medium (50), peptone tryptic soy broth (PTSB) (38), or supplemented A medium (40) as noted below. When necessary to maintain plasmids, cultures were supplemented with 200 μg/ml carbenicillin for P. aeruginosa and 100 μg/ml ampicillin and/or 30 μg/ml chloramphenicol for Escherichia coli.
Plasmids used in this study are listed in Table Table1.1. In order to generate an expression plasmid for PqsE, a 953-bp DNA fragment, which began at the pqsE start codon (ATG) and ended 47 bp downstream from the stop codon, was amplified by PCR using cosmid pMTP58 as a template. The oligonucleotide primers used for this amplification were engineered to contain a single HindIII site downstream from the stop codon (all oligonucleotide primer sequences are available upon request). Vector plasmid pEX1.8, which contains a tac promoter (tacp) to control gene expression, was digested with EcoRI and then treated with Klenow fragment to generate blunt-ended fragments. The plasmid was then digested with HindIII. The prepared plasmid DNA was ligated with the pqsE-containing PCR fragment, which had also been digested with HindIII, to produce the tacp-pqsE expression plasmid pZOE01. This construct was designed to contain optimal spacing between the ribosome binding site and the pqsE start codon. To create a PqsE expression plasmid for use in a two-plasmid reporter system, plasmid pZOE01 was digested with BamHI and ScaI, yielding an approximately 2-kb fragment carrying the tacp-pqsE fusion. This fragment was purified from an agarose gel and then ligated into vector plasmid pACYC184, which had been digested with BamHI and EcoRV. The resulting plasmid, pROSE04, harbors the tacp-pqsE fusion and a gene that confers resistance to chloramphenicol.
To generate a PqsA expression plasmid for complementation experiments, a 1,553-bp fragment, which began at the pqsA start codon and ended at the pqsA stop codon, was amplified by PCR. The oligonucleotide primers used for this reaction were engineered to contain a PstI site upstream from the pqsA start codon and a HindIII site downstream from the stop codon. The PCR fragment was digested with PstI and HindIII and then ligated into pEX1.8, which had been digested with the same enzymes. This reaction produced pDSW7, which carries a tacp-pqsA fusion for expression of PqsA.
Mutant P. aeruginosa strains were derived using a modified version of our previous protocol (18). Mutant alleles for pqsE, pqsA, and pqsR were generated using PCR as described elsewhere (18). Alleles were constructed to contain in-frame deletions in the coding DNA sequence corresponding to amino acids 18 to 285 for pqsE (89% of protein sequence), 44 to 504 for pqsA (89% of protein sequence), and 7 to 267 for pqsR (78% of protein sequence). The mutated fragments were also designed to contain the following at each end: BamHI sites for pqsE, PstI sites for pqsA, and EcoRI sites for pqsR. Each fragment was digested with the appropriate restriction enzyme and then ligated into pEX18Ap, which had been previously digested with the same enzyme, to produce suicide plasmids pΔpqsE-suc, pΔpqsA-suc, and pΔpqsR-suc. To transfer the mutant alleles onto the P. aeruginosa strain PAO1 chromosome, each suicide plasmid was transferred via conjugation from E. coli strain SM10 into strain PAO1, and integrants were selected on LB medium containing carbenicillin and 25 μg/ml triclosan. Mutants were then selected by plating integrants on medium containing 6% sucrose to remove the vector sequence from the chromosome. Potential mutants were screened by PCR using appropriate flanking primers, and mutants were further confirmed by determining the DNA sequence of the PCR products.
To check for pyocyanin production, bacteria from frozen skim milk stocks were plated onto LB medium containing carbenicillin as necessary to maintain plasmids. After approximately 24 and 48 h of incubation at 37°C, the plates were inspected visually for the distinct blue-green color indicating the presence of pyocyanin. Alternatively, freshly plated cells were used to inoculate 10-ml cultures of LB medium, which were then incubated at 37°C for approximately 6 h with vigorous shaking. Cells from each culture were then washed in fresh LB medium and used to inoculate 10-ml cultures of fresh LB medium to an optical density at 660 nm (OD660) of 0.05. Subcultures were incubated at 37°C with vigorous shaking for 18 h, and then aliquots of each culture were transferred to glass test tubes and photographed. For experiments where cultures were supplemented with C4-HSL, aliquots of C4-HSL dissolved in acidified ethyl acetate (44) were dried in flasks under nitrogen before they were added to subcultures.
To quantitatively measure the amount of pyocyanin in culture supernatants, a modification of the method of Essar et al. (17) was used. Samples of cultures grown for 18 h were centrifuged to remove bacterial cells, and 500-μl aliquots of culture supernatants were extracted with 300 μl of chloroform. The organic phase was then extracted with 100 μl of 0.2 N HCl to give a pink solution containing pyocyanin. The absorbance of this solution at 520 nm was measured using a NanoDrop ND-1000 spectrophotometer, and the amount of pyocyanin present was calculated by comparison with results obtained using known quantities of pyocyanin (Cayman Chemical, Ann Arbor, MI). The range of detection for the assay was from 10 μM to 300 μM pyocyanin.
For rhamnolipid and elastase assays, bacteria were cultured in LB medium as described above for the detection of pyocyanin. After 18 h of incubation at 37°C, aliquots of each culture were centrifuged to remove bacterial cells. The cleared culture supernatant was filtered through a 0.45-μm-pore-size syringe filter, and samples were extracted immediately for the detection of rhamnolipids or stored at −80°C prior to being assayed for elastolytic activity.
The concentration of rhamnolipid in culture supernatants was determined using an orcinol assay as described elsewhere (35). Briefly, 300-μl samples of filtered culture supernatants were extracted twice with 600 μl of diethyl ether. The ether fractions were pooled in glass tubes and dried under a continuous stream of N2. Then, 100 μl of water, 100 μl of 1.6% orcinol (Sigma), and 800 μl of 60% H2SO4 were added to each tube, and the tubes were incubated at 80°C for 30 min. After tubes were cooled at room temperature for 10 min, the absorbance at 421 nm (A421) of each sample was measured. These data were compared with results obtained using known quantities of rhamnolipid (a mixture of mono- and dirhamnolipid; Jeneil Biosurfactant Co., Saukville, WI) to calculate the amount of rhamnolipid in each sample. The range of detection for the assay was from 2.5 μg to 25 μg of rhamnolipid.
The elastolytic activity of filtered culture supernatants was measured using a modification of the elastin-Congo red (ECR) assays described elsewhere (42). Duplicate 25-μl aliquots of culture supernatants were added to tubes containing 20 mg of ECR (Sigma) and 1 ml of ECR buffer (0.1 M Tris [ph 7.2], 1 mM CaCl2). Tubes were incubated at 37°C with rotation for 16 h, and then 100 μl of 0.12 M EDTA was added to each, and the tubes were placed on ice. Insoluble ECR was removed from samples by centrifugation, and the cleared samples were diluted 1:1 in ECR buffer. The A495 of each diluted sample was measured, and the values were corrected for the absorption due to P. aeruginosa pigments by subtracting the A495 of each sample that had been incubated in the absence of ECR.
For experiments involving the addition of both PQS and C4-HSL (see Fig. Fig.3),3), cells from overnight cultures of P. aeruginosa grown in PTSB medium were washed and resuspended in fresh medium at an OD660 of 0.05 and then incubated at 37°C with vigorous shaking until cells reached late-logarithmic phase. Then, cells from subcultures were washed and resuspended in fresh PTSB medium, and 1-ml aliquots of cell suspensions were transferred to tubes containing either PQS, C4-HSL, both, or neither. After 18 h of incubation with vigorous shaking at 37°C, cells were harvested from 1-ml cultures by centrifugation and suspended in 1 ml of supplemented A medium. The β-galactosidase (β-Gal) activity of each sample was assayed in duplicate. Data are reported in Miller units (34) as the mean ± standard deviation (SD) of at least three separate experiments.
For experiments involving the addition of C4-HSL only (see Fig. Fig.4),4), cells from cultures of P. aeruginosa grown in PTSB medium for 20 to 24 h were washed and resuspended in fresh medium at an OD660 of 0.05, and then 1-ml aliquots were transferred to tubes containing C4-HSL. After 18 h of incubation with vigorous shaking at 37°C, cells were diluted in supplemented A medium and assayed for β-Gal activity. Data are reported in Miller units as the mean ± SD of three separate experiments.
To measure C4-HSL production, culture supernatants from P. aeruginosa strains grown in LB medium for 18 h at 37°C with vigorous shaking were extracted with ethyl acetate as described elsewhere (44). The amount of C4-HSL in ethyl acetate extracts was determined using the C4-HSL bioassay (42).
To test the ability of pqsE to affect the function of RhlR-C4-HSL or LasR-3-oxo-C12-HSL in E. coli, a two-plasmid system was utilized in which one plasmid carried either tacp-rhlR or tacp-lasR and a reporter gene fusion, and the second plasmid was either the vector plasmid or a plasmid carrying tacp-pqsE. Cells from overnight cultures of E. coli carrying the appropriate plasmids and grown in supplemented A medium were diluted to an OD600 of 0.08 in fresh medium. Subcultures were incubated at 37°C with vigorous shaking for 3 h, and then 1-ml aliquots were transferred to tubes that contained dried signal compounds. After 90 min of incubation at 37°C with vigorous shaking, the β-Gal activity produced by each culture was measured. Data are reported in Miller units as the mean ± SD of at least three independent experiments.
Since the discovery of the PQS synthetic gene cluster (10, 20), the function of PqsE has been a mystery. pqsE is the final gene of a five-gene operon, of which the first four genes are required for PQS synthesis (10, 20, 32). However, mutation of pqsE has no effect on PQS synthesis (20). Instead, pqsE appears to be necessary for P. aeruginosa to respond to PQS and to produce PQS-controlled virulence factors such as pyocyanin (20). In order to begin studying the role of PqsE, we first constructed isogenic, in-frame deletion mutants for pqsE, pqsA, and pqsR. As observed previously (20), the pqsA and pqsR mutants did not make PQS, and PQS production was not affected in the pqsE mutant (data not shown). Furthermore, none of the three mutant strains produced pyocyanin (Fig. (Fig.11 and Table Table2),2), which was as expected from previous studies (20). Production of both pyocyanin and PQS was restored to the pqsA and pqsR mutant strains by genetic complementation with a plasmid carrying a copy of each respective gene (Fig. (Fig.11 and Table Table2;2; also data not shown). As a control, transfer of the vector plasmid into the strains being tested did not affect either pyocyanin or PQS production (data not shown). When we complemented the pqsE mutant with a plasmid carrying a tacp-pqsE fusion, pyocyanin production by the complemented strain greatly exceeded the level of the wild-type strain (Fig. (Fig.11 and Table Table2).2). Interested by this result, we then transferred the tacp-pqsE plasmid into the wild type and the pqsA and pqsR mutant strains. Much to our surprise, we found that the overexpression of PqsE in these mutants stimulated overproduction of pyocyanin in all three strains (Fig. (Fig.11 and Table Table2).2). This result was unexpected because all prior data showed that PqsR and PQS were required for pyocyanin production (5, 13, 14, 20). Our data indicated that PqsE was able to regulate pyocyanin production in the absence of both PQS and PqsR, implying that PqsE is an independent regulator.
Because pyocyanin production is also controlled by the other P. aeruginosa intercellular signaling systems (51), the las and rhl systems, we were interested in testing whether PqsE could act independently of these regulators as well. Therefore, we transferred the PqsE overexpression plasmid into various quorum-sensing mutant strains and observed the effects on pyocyanin production. In agreement with other reported observations (14, 22), the lasR and lasI mutants displayed delayed, but increased, production of pyocyanin (data not shown). The overexpression of PqsE in these strains did not alter this phenotype, nor was it able to restore pyocyanin to the lasR rhlR or lasI rhlI double mutants (data not shown), indicating that PqsE is unable to induce pyocyanin production in the absence of both the las and rhl quorum-sensing systems. Interestingly, the large positive effect of PqsE on pyocyanin production was abolished in an rhlR mutant containing the pqsE expression plasmid (Fig. (Fig.11 and Table Table2).2). However, a slight positive effect on pyocyanin production was seen in an rhlI mutant carrying this same plasmid (Fig. (Fig.11 and Table Table2).2). These data indicated that RhlR is necessary for PqsE to regulate pyocyanin production. This finding led us to speculate that PqsE might be functioning directly or indirectly through RhlR, which will be examined further below.
In addition to pyocyanin, numerous other virulence factors are known to be controlled by both PQS signaling and the rhl quorum-sensing system (1, 12, 14). To further assess the regulatory effects of PqsE, we examined the production of two other PQS- and rhl-controlled virulence factors, elastase and rhamnolipid. The results obtained were similar to those for pyocyanin production. Rhamnolipid production was decreased approximately 50% in the pqsE, pqsA, and pqsR mutants, and all three of these strains, along with the wild type, produced a much greater amount of rhamnolipid when they carried the pqsE expression plasmid (Fig. (Fig.2A).2A). This effect did not occur in an rhlR mutant, which produced no rhamnolipid with or without the pqsE expression plasmid (Fig. (Fig.2A).2A). Interestingly, rhamnolipid production was partly restored in an rhlI mutant that expressed PqsE (Fig. (Fig.2A),2A), which is similar to what occurred for pyocyanin production (Fig. (Fig.1).1). A similar result was seen with elastase production assays. The pqsE, pqsA, and pqsR mutants all doubled their elastase production when they harbored the pqsE expression plasmid (Fig. (Fig.2B).2B). It should be noted that under our growth conditions, elastase production in these three mutants was similar to that of the wild-type strain, which demonstrates that the las and rhl quorum-sensing systems are the dominant regulators of elastase production. In addition, as seen with pyocyanin and rhamnolipid production, the overexpression of PqsE in an rhlR mutant did not affect elastase production and caused only a slight increase in elastase activity from an rhlI mutant (Fig. (Fig.2B).2B). Taken together, the data shown in Fig. Fig.11 and and22 and in Table Table22 indicated that PqsE can regulate multiple virulence factors independently of PqsR and PQS, but it cannot act independently of RhlR.
Since PqsE is involved in the complex intercellular signaling regulatory scheme of P. aeruginosa, we wanted to learn more about the effect of PqsE on PQS activity in the absence of both C4-HSL and 3-oxo-C12-HSL. To test this, we examined rhlA′-lacZ induction in a P. aeruginosa lasI rhlI double mutant that expressed rhlR, which is a system previously used as a bioassay for C4-HSL activity (42). We also tested rhlA′-lacZ induction in strain PAO-JP2E, a derivative of strain PAO-JP2 (lasI rhlI) that also carries a mutation in pqsE. This system allowed us to assess the effects of PQS and C4-HSL on the rhl system independently. (Note that we previously showed that strain PAO-JP2 does not produce PQS when grown under similar conditions due to the lack of a functional las system [3, 44].) These experiments showed that, as expected, rhlA was not expressed above a background level in the absence of C4-HSL and 3-oxo-C12-HSL in either strain (Fig. (Fig.3).3). Interestingly, the addition of exogenous PQS to strain PAO-JP2 caused a large increase in rhlA induction (Fig. (Fig.3),3), indicating that PQS was active in the absence of C4-HSL and 3-oxo-C12-HSL. Furthermore, adding PQS to strain PAO-JP2E that was expressing RhlR caused only a slight induction of rhlA′-lacZ (Fig. (Fig.3).3). While the induction of rhlA caused by PQS in strain PAO-JP2 was considerable, we also observed a similar induction in response to 0.1 μM C4-HSL (Fig. (Fig.3),3), which is a concentration that is 200-fold lower than the tested concentration of PQS. When strain PAO-JP2 was supplemented with both PQS and C4-HSL, there was an increase in rhlA′-lacZ activity beyond that caused by either signal alone (Fig. (Fig.3).3). However, in strain PAO-JP2E the activity produced by the combination of the signals was equivalent to that of C4-HSL alone (Fig. (Fig.3).3). These data showed that PQS was able to induce rhlA expression in P. aeruginosa in the presence of RhlR and in the absence of acyl-HSL signals and that this effect required PqsE. Overall, our data suggested that a factor(s) that positively regulated rhlA was activated by both PQS and C4-HSL and that this unknown regulator was also dependent on PqsE.
While our results showed that PqsE required the presence of RhlR to regulate multiple virulence factors, it has been clearly shown that RhlR and C4-HSL are able to initiate gene expression in the absence of PqsE (42). Interestingly, we found that multiple rhl-controlled factors, such as rhamnolipid and pyocyanin, are downregulated in a pqsE mutant despite the fact that the mutant produced an amount of C4-HSL equivalent to that produced by the wild-type strain (data not shown), a finding also reported by others (14). This led us to investigate whether a P. aeruginosa pqsE mutant was still capable of responding to C4-HSL. In our first experiment, we cultured the pqsE mutant in the presence of excess C4-HSL and observed pyocyanin production. We found that exogenous C4-HSL restored pyocyanin production to a pqsE mutant but that relatively large amounts of C4-HSL were required to restore pyocyanin production to a level that was approximately 60% of that seen from the wild-type strain (Fig. (Fig.4A).4A). We continued this line of experiments and examined the expression of an rhlA′-lacZ fusion in the pqsE mutant. As expected from our earlier data (Fig. (Fig.2),2), rhlA expression was decreased approximately 60% in the pqsE mutant (Fig. (Fig.4B).4B). The addition of exogenous C4-HSL caused rhlA′-lacZ induction to increase in a dose-dependent manner, but restoration to a wild-type level of expression required greater than 100 μM C4-HSL (Fig. (Fig.4B)4B) (Note that this strain already produces a wild-type level of C4-HSL, which is approximately 10 to 20 μM [41; also data not shown].) To ensure that rhlR was still expressed, we analyzed rhlR transcription in strain PJF-QE1(pPCS1002) and found that only a minor decrease of 23% (from 158,118 ± 7,527 to 121,763 ± 10,450 Miller units of β-Gal activity) occurred in rhlR′-lacZ expression relative to that in the wild-type strain PAO1(pPCS1002). Taken together, these results demonstrated that a pqsE mutant is less sensitive to C4-HSL but that it can respond to C4-HSL, suggesting that RhlR is active in a pqsE mutant. The data also imply that PqsE may play a role in the ability of RhlR to respond to C4-HSL.
So far, our data have shown that factors controlled by the rhl system are overexpressed when pqsE is overexpressed (Fig. (Fig.11 and and2)2) and underexpressed when pqsE is absent (Fig. (Fig.4).4). This suggested that PqsE may have a role in the function of the rhl quorum-sensing system. Since the quorum-sensing circuitry of P. aeruginosa is complex and is modified by a wide array of factors, we wanted to try to establish a more direct link between PqsE and the rhl quorum-sensing system. To do this, we determined whether the effect of PqsE could be seen in E. coli cells by introducing two plasmids into E. coli. One contained tacp-pqsE, and the other contained tacp-rhlR and either rhlA′-lacZ or lasB′-lacZ. When these E. coli strains were grown in the presence of isopropyl-β-d-thiogalactopyranoside but with no exogenous signals, the strain containing the pqsE vector showed no increase in rhlA expression and was identical to that seen from a strain containing a control vector (Fig. (Fig.5A).5A). Most interestingly, as increasing amounts of C4-HSL were added to the cultures, the strain that contained pqsE exhibited a much more sensitive response to C4-HSL (Fig. (Fig.5A).5A). This effect also occurred with another PQS- and C4-HSL-controlled gene when lasB′-lacZ was substituted for rhlA′-lacZ. We found that RhlR was able to induce lasB′-lacZ at a much lower concentration of C4-HSL when PqsE was present (Fig. (Fig.5B).5B). In both strains that contained pqsE, the expression of the reporter gene (rhlA or lasB) in response to exogenous C4-HSL peaked and then leveled off at an induction level similar to that seen from the strain containing the control vector (Fig. 5A and B). This enhanced responsiveness to exogenous signal was not seen when we tested the ability of PqsE to affect lasB′-lacZ induction by LasR and 3-oxo-C12-HSL (Fig. (Fig.5C).5C). In this case, both the control vector and pqsE vector strains expressed lasB at the same levels as the concentration of 3-oxo-C12-HSL increased. Overall, these data showed that PqsE can enhance the ability of RhlR and C4-HSL to activate gene expression in the absence of other P. aeruginosa-encoded factors, and this effect appeared to be specific for the rhl quorum-sensing system.
In this study we have begun to clarify the role of PqsE in the P. aeruginosa quorum-sensing circuitry. Our previous findings showed that pqsE is not necessary for the synthesis of PQS but that pqsE is required for the expression of virulence factors controlled by PQS (20). This led us to speculate that pqsE was involved in the cellular response to PQS (20). In our current study, we used a pqsE overexpression plasmid to show that in the absence of the PQS-responsive transcriptional activator PqsR, PqsE can positively regulate expression of the virulence factors pyocyanin, rhamnolipid, and elastase (Fig. (Fig.11 and and22 and Table Table2).2). We also showed the same positive effects in a pqsA mutant, demonstrating that the regulatory activity of PqsE occurs independently of PQS or any of the other quinolone compounds produced through the action of the enzymes encoded by the pqsABCDE operon. These data suggested a model in which PQS and PqsR controlled the expression of pqsE and PqsE, in turn, regulated a second set of genes. This indirect regulation is different from that seen in the acyl-HSL-based signaling systems of P. aeruginosa, where LasR and RhlR each bind their cognate signals and directly induce the expression of multiple targets. Both PQS and PqsR have been shown to regulate a large number of genes (1, 12), but these data must now be reexamined to determine which genes are directly regulated by PqsR-PQS and which are controlled indirectly through the independent action of PqsE.
As our studies expanded, they led us away from the idea that pqsE is involved in the cell's response to PQS and instead pointed us toward investigating the role of pqsE in the rhl quorum-sensing system. Several previous studies have noted a significant overlap between PQS- and rhl-controlled genes (12, 14), but the relationship between these two systems has been unclear. Our experiments to study the effects of pqsE expression in various P. aeruginosa quorum-sensing mutants showed that the regulatory function of PqsE required the presence of RhlR (Fig. (Fig.11 and and2).2). In addition, we had previously shown that PQS and C4-HSL could cooperatively induce gene expression in the presence of RhlR (33). Therefore, we examined the involvement of PqsE in the interaction between the PQS and rhl signaling systems. Our results showed that PQS itself or in combination with C4-HSL could activate an rhlA′-lacZ fusion in the presence of RhlR (Fig. (Fig.3).3). However, these effects were observed only when a functional copy of pqsE was present, thereby providing further evidence that PQS indirectly regulates gene expression by controlling the expression of pqsE and establishing PqsE as a link between the quinolone and rhl signaling systems.
While these findings provided insight into the relationship between PQS and the rhl quorum-sensing system, it was still unclear as to why pqsE was required for the full production of numerous rhl-controlled virulence factors. This was especially puzzling since pqsE does not affect C4-HSL production (14; also data not shown). When we tested the ability of a pqsE mutant to respond to additional exogenous C4-HSL, we found that a great excess of C4-HSL restored the induction of rhl- controlled virulence factors (Fig. (Fig.4).4). This suggested that the rhl quorum-sensing system was still capable of functioning in a pqsE mutant but that the regulatory function of PqsE was necessary for the full induction of some genes. We then tested the effects of PqsE on the function of the rhl system in E. coli and found that the presence of PqsE enhanced the ability of RhlR and C4-HSL to induce gene expression (Fig. (Fig.5).5). This effect appeared to be specific for the rhl system since PqsE did not enhance the induction of lasB by LasR and 3-oxo-C12-HSL (Fig. (Fig.5).5). Taken together, our findings suggested that PqsE played a role in the rhl quorum-sensing system and that PQS was linked to this through its control of pqsE.
Overall, these findings begin to sort out some of the confusion regarding the interactions between the quinolone and rhl signaling systems. Previous data showed that PQS was important for the induction of numerous rhl-controlled factors and also suggested that PQS can upregulate rhlR and rhlI themselves (12, 14, 20, 26, 32, 33). Our current results show that PqsE affects the activity of the rhl signaling system but does not significantly affect the production of C4-HSL or the transcription of rhlR, implying that the role of PqsE is not to control the components of the rhl system. These findings suggest that the induction of rhlR and rhlI observed in response to PQS by ourselves and others was not through the induction of pqsE but may instead be the result of some other activity of PQS and that PQS may influence the rhl system in multiple ways. It has also been shown that RhlR and C4-HSL can negatively impact the production of PQS and other quinolones by influencing the transcription of the pqsABCDE operon (32, 59). Our findings showed that PqsE could enhance the ability of RhlR and C4-HSL to positively regulate gene expression but do not indicate whether it also enhances the negative regulatory function of the rhl system as well, although others have reported that a P. aeruginosa pqsE mutant and the wild-type strain produced similar amounts of 4-quinolone compounds (13).
The exact mechanism by which PqsE affects the rhl quorum-sensing system and achieves its regulatory function remains a mystery. However, our data suggested that PqsE may alter the function of RhlR rather than affecting C4-HSL. This is supported by our observation that PqsE had some regulatory activity in an rhlI mutant but none in an rhlR mutant (Fig. (Fig.11 and and2).2). We also saw pqsE-dependent regulatory activity in the absence of any acyl-HSL signals (Fig. (Fig.3),3), implying that PqsE may act on the rhl quorum-sensing system through a novel form of regulation. It is also interesting that Deziel et al. (12) found that a pqsE mutant, but not an rhlR mutant, had attenuated virulence in a burned mouse model of infection, suggesting that the influence of pqsE may extend beyond the rhl system. While the studies presented here do not determine the molecular mechanism through which PqsE acts, they provide a set of interesting data on which to build a model that will explain PqsE's role in the complex scheme that comprises P. aeruginosa intercellular signaling.
We thank H. Schweizer for kindly supplying pEX18Ap.
This work was supported by a research grant from the National Institute of Allergy and Infectious Disease (grant R01-AI076272).
Published ahead of print on 5 September 2008.