Secretion of Pseudomonas aeruginosa Type III Cytotoxins is Dependent on Pseudomonas Quinolone Signal Concentration

Abstract

Pseudomonas aeruginosa is an opportunistic pathogen that can, like other bacterial species, exist in antimicrobial resistant sessile biofilms and as free-swimming, planktonic cells. Specific virulence factors are typically associated with each lifestyle and several two-component response regulators have been shown to reciprocally regulate transition between biofilm-associated chronic, and free-swimming acute infections. Quorum sensing (QS) signal molecules belonging to the las and rhl systems are known to regulate virulence gene expression by P. aeruginosa. However the impact of a recently described family of novel quorum sensing signals produced by the Pseudomonas Quinolone Signal (PQS) biosynthetic pathway, on the transition between these modes of infection is less clear. Using clonal isolates from a patient developing ventilator-associated pneumonia, we demonstrated that clinical observations were mirrored by an in vitro temporal shift in isolate phenotype from a non-secreting, to a Type III cytotoxin secreting (TTSS) phenotype and further, that this phenotypic change was PQS-dependent. While intracellular type III cytotoxin levels were unaffected by PQS concentration, cytotoxin secretion was dependent on this signal molecule. Elevated PQS concentrations were associated with inhibition of cytotoxin secretion coincident with expression of virulence factors such as elastase and pyoverdin. In contrast, low concentrations or the inability to biosynthesize PQS resulted in a reversal of this phenotype. These data suggest that expression of specific P. aeruginosa virulence factors appears to be reciprocally regulated and that an additional level of PQS-dependent posttranslational control, specifically governing type III cytotoxin secretion, exists in this species.

Introduction

Pseudomonas aeruginosa is a ubiquitous environmental gram negative bacterium that can cause a variety of acute and chronic infections (1). This pathogenic plasticity is attributed to the large, complex genome encoding a vast arsenal of virulence factors and regulatory switches which permit P. aeruginosa to thrive in a multitude of niches and rapidly adapt to fluctuations in environmental conditions (46). Like many other bacterial species, P. aeruginosa can form biofilms, communities of bacterial cells encased in an exopolysaccharide matrix that protects the cells within from antimicrobial activity and host immune responses (27, 33, 39). In vitro, biofilms have been associated with production of specific virulence factors such as elastase and pyocyanin (14) which are regulated by quorum sensing molecules, diffusible signaling compounds produced by bacterial species (27, 33, 39). Chronic infections, including pulmonary disease of cystic fibrosis patients, have been associated with P. aeruginosa biofilms (5, 30, 44), elevated elastase secretion (17) and high concentrations of quorum sensing signals in respiratory samples (43).

Outside of the biofilm, bacteria may also exist as free-swimming or planktonic cells. Motile free-swimming P. aeruginosa cells are associated with acute infections and secrete a number of cytotoxins through a needle-like complex, the type III secretion system [TTSS;(2, 41)]. Pseudomonas aeruginosa communities are in constant state of flux with planktonic cells capable of detaching from the biofilm, suggesting that both modes of lifestyle can exist in a given niche in parallel. This is supported by recent findings demonstrating the presence of type III cytotoxins in the flow through from biofilm cultures of P. aeruginosa (32). Transition between biofilm and planktonic forms appears to involve reprogramming of a number of physiological aspects, including lipopolysaccharide structure, putatively to facilitate the virulence factors expressed during each mode of lifestyle (2). Transition between P. aeruginosa biofilm and free-swimming cells is known to be controlled by a number of two component regulators e.g. the Sad, Lad, Ret and Gac sensor-response systems, which clearly respond to specific environmental signals (25, 47). Regulatory interplay between these molecular switches is postulated to fine tune reciprocal expression of a number of genes, including those typically associated with chronic or acute infection (19, 47).

Regulation of type III secretion genes has previously been associated with the las and rhl quorum sensing system (6), a cell-to-cell signaling mechanism that permits rapid response to environmental changes through production of diffusible molecules that, when sensed by their cognate cell surface associated receptor, regulate gene expression. This sophisticated communication system, enables coordinated gene expression in bacterial populations facilitating cooperative behavior by bacterial communities. P. aeruginosa coordinates population gene expression primarily through three main quorum sensing systems: the las and rhl systems which employ acylated homoserine lactone signal molecules (3O-C₁₂-HSL and C₄-HSL, respectively), and the recently described 2-alkyl-4(1-H)-quinolone (AHQ) family (3, 48) The AHQ’s include more than 50 quinolone compounds, two of which, Pseudomonas quinolone signal (PQS; 2-heptyl-3-hydroxy-4-quinolone) and its immediate precursor, 2-heptyl-4-quinolone (HHQ) have been shown to act as quorum sensing signals (16).

Here we describe the basis for altered virulence profiles in a set of clonal P. aeruginosa clinical strains isolated from a single patient developing ventilator-associated pneumonia. In parallel with clinical observations made over a period of 3 weeks, in vitro observations suggested that these serial clonal isolates transitioned from a non-secreting to a an acute, type III-secreting phenotype. This transition coincided with a shift in the concentration of extracellular secreted organic molecules (including PQS) produced by the strains. The effect of PQS concentration on type III secretion was confirmed by a series of experiments exposing strains to cell-free supernatant from PQS-sufficient or -deficient strains and by inhibiting quinolone production. These data suggest that P. aeruginosa strains have the ability to transition over relatively short periods of time to acutely infectious cells and that type III cytotoxin secretion, a hallmark of acute infection by this species, is regulated by the relatively recently described quinolone signal, PQS. That this phenotype was maintained in vitro suggests that it is “hard-wired” into the genetic blueprint of the organism. These observations open up the possibilities for novel therapeutic approaches involving physiological transition of this species towards the more antimicrobial susceptible free-swimming lifestyle by blocking quinolone production in combination with antimicrobial and anti-type III secretion virulence therapies in cases of refractory chronic infection.

Materials and Methods

P. aeruginosa clinical isolates used in this study

P. aeruginosa isolates were obtained from endotracheal aspirates (ETA), collected sequentially from a single patient who was culture-positive for P. aeruginosa and subsequently met the criteria for ventilator associated pneumonia (VAP) following three weeks of mechanical ventilation. Five clonal isolates from this patient collected pre- and post-VAP diagnosis were used for this study (Table 1). For all experiments, fresh cultures generated from glycerol stocks of these isolates were used to ensure that the strains maintained their phenotype during the in vitro experiments performed.

Table 1.

Clinical isolates used in this study and patient status at time of sample collection.

Strain number	Date of isolation	Colony Appearance	VAP development
			Sputum quantity and appearance	X-ray findings
1	12/10/’04	Glossy	Little; non-purulent	Patchy consolidation
2	12/14/’04	Glossy	Unknown	Increased consolidation
3	12/25/’04	Dry	Large; purulent	Diffuse infiltrate
4	12/28/’04	Dry	Moderate; purulent	Diffuse infiltrate
5	01/02/’05	Dry	Moderate; purulent	Diffuse infiltrate

Pulsed field gel electrophoresis (PFGE) analysis of isolates

P. aeruginosa isolates were cultured overnight in Luria broth at 37°C with shaking. Agarose plugs containing 5 × 10⁸ cells/ml of overnight culture were prepared according to the Bio-Rad PFGE protocol (Bio-rad Laboratories, CA) prior to overnight digestion using 40 units of SpeI at 37°C. PFGE was carried out using a Lambda ladder size standard (Bio-Rad Laboratories, CA) and a CHEF-DR III apparatus (Bio-Rad Laboratories, CA). PFGE conditions were as follows: initial switch time 1.0 s, final switch time 25 s, total run time 24 h, included angle 120°C, electric field 6 V/cm, buffer 0.5x TBE, temperature 14°C, and agarose 1% pulsed field certified (Bio-Rad Laboratories, CA).

Pyoverdin and Elastase assays

Isolates were cultured in LB broth overnight prior to dilution (OD₆₀₀, 0.1) in fresh medium and reincubation for 24 h at 37 °C with shaking. Spectrophotometric measurements (405 nm) of triplicate cell-free supernatants from each strain were performed to determine pyoverdin production. Parallel elastase activity was determined in these same cultures using elastin congo red and spectrophotometric measurements at 490 nm, according to methods previously described by Nicas et al (36).

Type III secretion assay

Overnight cultures of each strain in MIN-S medium (37) were diluted to an OD₆₀₀ of 0.1 in fresh MIN-S and incubated with shaking for a further 4 h. Cultures were harvested by centrifugation, the cell-free supernatant was removed and concentrated using centricon tubes (MWCO = 10,000; Millipore, MA). The cell pellet was lysed using SDS buffer (2% SDS in Tris-HCl, pH 7.0). Protein concentration for both fractions was determined using the Bio-Rad DC protein assay (Bio-Rad Laboratories, CA) according to the manufacturer’s instructions. Standardized protein concentrations (20 ng ug⁻¹) of cell lysate and cell-free supernatant (secreted) fractions from each isolate were electrophoresed on a 12% Criterion SDS polyacrylamide gel (Bio-rad Laboratories, CA), prior to transfer onto PVDF membrane (Amersham, NJ) and immunoblotted using antibodies against ExoU, ExoT as previously described (2).

Cytotoxicity assay using a BEAS2-B lung epithelial cell line

Human immortalized BEAS-2B lung epithelial cells (200 μl) were seeded at a density of 2.5 × 10⁵ ml⁻¹ in Dulbeccos Modified Eagle Medium supplemented with 10% BCS in 96-well cultures and incubated overnight to reach confluency. Each P. aeruginosa isolate was grown overnight in LB medium as described above. Strains were harvested, washed three times in sterile PBS, and finally resuspended in 100 μl Ringers: PBS solution (2:1). Each well received 5 × 10⁶ bacteria. Samples were taken at 30 min time intervals and cytotoxicity assayed by lactate dehydrogenase release using the Cyto Tox96® kit (Promega, CA) according to the manufacturer’s instructions.

Quinolone Signal assays

Thin layer chromatography analysis of organic compound production

Each strain was cultured overnight in LB as described above, prior to dilution to OD₆₀₀ of 0.1 and reincubation for 24 h. Cell-free supernatant was then removed from harvested cultures, using methods as described above. Pooled extracts of 1 volume of acidified ethyl acetate (0.1% glacial acetic acid [v/v]) performed twice, which has been shown to extract PQS and other organic compounds (8), was performed on each cultured supernatant prior to quantification by TLC.

Real-time quantitative PCR for expression of pqsA

Each strain was cultured overnight in LB medium as described above, diluted to an OD₆₀₀ of 0.1 and reincubated for 4 h. Total RNA was purified using Trizol (Invitrogen, CA) according to the manufacturer’s instructions. After incubation with DNase I (Invitrogen, CA), total RNA was checked for integrity by gel electrophoresis, and absence of DNA by PCR, prior to cDNA synthesis. RNA from each strain was quantified and cDNA generated using 1 μg total RNA and Omniscript reverse transcriptase (Qiagen, CA). One μl of this reverse transcriptase reaction was used for each Q-PCR reaction. Q-PCR primers for pqsA (pqsAF, 5′-GAGTGCCTTCCATCGCCAGG-3′ and pqsAR, 5′-AGCGAACGCCCGGCAGAAAC-3′), the first gene in the PQS biosynthetic operon, were used to quantify expression in all 5 isolates. mRNA transcript copy number was measured using the QuantiTect SYBR Green PCR Kit (Qiagen, CA) according to the manufacturer’s instructions. Reactions were carried out using a Mx3000P Real-Time PCR System (Stratagene, CA) as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 58°C for 1 min, and 72°C for 30 s. The data acquisition step was set at 58°C with a final melting-curve analysis to ensure amplification of a single product. Quantification of pqsA copy number was determined using a standard curve of known pqsA copy number, as described in the manufacturer’s instructions (Qiagen, CA).

Exposure of clinical isolates to exogenous PQS-sufficient or deficient supernatant extracts

Large scale (2 × 200 ml LB) preparations of cell-free quinolone signal extracts were carried out on overnight cultures of clinical strain 1 (PQS-sufficient), the laboratory strain PAO1 (PQS-sufficient), or PAO1ΔpqsH (generously provided by Stephan Heeb and Paul Williams, Unversity of Nottingham, UK), which produces the PQS precursor, 4-hydroxy-2-heptylquinoline (HHQ) and related AHQs, but not PQS (16). Extracts were prepared using 1 volume of acidified ethyl acetate as described above. Extractions were pooled and 40 ml of this pooled solution added to each 50 ml falcon tube prior to drying under compressed air. Each P. aeruginosa isolate was cultured overnight in MIN-S and diluted to an OD₆₀₀ of 0.1 in fresh MIN-S. Twenty ml of each diluted culture was placed in a 50 ml Falcon tube (screw caps were replaced with cotton bungs) that contained dried quinolone signal extract (PQS-sufficient or deficient extracts), a set of tubes containing 20 ml of culture but no quinolone extract served as controls. All tubes were incubated for 24 h with shaking at 37°C. Following incubation, 5 ml of culture from each of the quinolone-containing and control tubes was harvested for immunoblot analysis using anti-ExoU antibody to assess cytotoxin secretion profiles as previously described (2).

Exposure of clinical isolates to farnesol

Overnight MIN-S cultures of each strain were diluted to an OD₆₀₀ of 0.1 in fresh MIN-S. To 5 ml of diluted culture, a final concentration of 10 mM farnesol, a compound reported to inhibit PQS (11), was added. Cultures were incubated with shaking for a further 4 h. Following incubation, each culture was harvested for immunoblot analysis with ExoU to assess cytotoxin secretion, as described previously (2).

Results

Clonal isolates from a patient developing VAP exhibit a temporal shift in virulence phenotype

Visual inspection of the 5 sequential clinical samples collected from a single patient who developed VAP demonstrated a notable temporal change in appearance. Initial ETA samples (1 and 2) appeared green, while the latter three (3, 4 and 5) became progressively bloodier (Fig. 1A), suggesting temporal progression of the lower respiratory tract infection from a chronic (pyocyanin-producing) to an acute (type III secretion) pathoglogy. Examination of the P. aeruginosa colonies isolated from these ETA’s on Pseudomonas isolation agar revealed a corresponding change in colony morphology. The dominant morphology (~ 95% of all colonies) of colonies cultured from these first two ETA’s (strain 1 and 2) was glossy and representative strains produced a fluorescent compound when visualized under UV light. In contrast colonies cultured from the later three respiratory samples (3, 4 and 5) appeared dry and their representative strains produced little or no fluorescence (Fig. 1B). To determine if all 5 strains isolated over time from this single patient were isogenic, SpeI digested genomic DNA from each of the 5 isolates were subjected to PFGE. Under standard PFGE conditions, all 5 clinical isolates appeared clonal (Fig. 1C), suggesting that the decline in patient’s clinical status was not due to acquisition of a novel, more virulent strain of P. aeruginosa, but to a change in virulence gene expression and/or secretion by the original infecting strain.

Fig. 1.

Fig. 1A. Serial ETA’s collected over a 3 week period from a patient developing VAP exhibit evidence of a transition from chronic (green tinged appearance) to acute (bloody appearance) mode of infection. B. P. aeruginosa isolates demonstrate a shift in fluorescent metabolite production upon exposure to UV light. C. Serial isolates are clonal by standard PFGE analysis.

We hypothesized that the 5 strains isolated from this patient may represent a set of clonal isolates transitioning from a chronic (strains 1 and 2) to an acute (strains 3, 4 and 5) mode of infection that contributed to the development of VAP in this patient. As previously mentioned, both chronic and acute P. aeruginosa infections are typically associated with expression of specific virulence systems, therefore, analysis of these traits was performed for each isolate. Elastase production, typically associated with chronic infections (14) was measured for all 5 isolates. Strains 1 and 2 produced approximately 2-fold higher concentrations of elastase, while strains 3, 4 and 5 did not (Fig. 2A). Measurement of pyoverdin production, the fluorescent siderophore produced by P. aeruginosa and co-expressed with other biofilm-associated virulence factors such as alginate and elastase (26), also demonstrated significantly (p < 0.0001) increased production by the first two isolates compared to the latter three (Fig. 2B). Examination of biofilm formation by all five strains also demonstrated that the first two strains produced substantially more biofilm than the latter three (Fig. 2C).

Fig. 2.

Fig. 2A. Elastase production is approximately 2-fold higher in strains 1 and 2 compared to strains 3, 4 and 5. B. Pyoverdin production is also greater in the initial two isolates compared to the latter 3. C. Biofilm formation by all five strains. D. Analysis of intracellular (IC) and secreted total protein fractions (S) of each strain demonstrates that while all strains produce ExoU and ExoT, only the latter 3 isolates secrete these cytotoxins. E. Cytotoxicity assays profiling epithelial cell toxicity of all 5 strains demonstrate that the latter 3 isolates are significantly more cytotoxic compared to strains isolated earlier in the infectious process.

To further examine the hypothesis that transition from chronic to acute modes of infection was occurring in these clonal isolates, type III cytotoxin secretion, a hallmark of acute infection (41), was analyzed for all 5 isolates. Both intracellular and secreted fractions were analyzed and demonstrated that while all strains produced the cytotoxins ExoT and ExoU, only the latter 3 isolates secreted them (Fig. 2D). Type III cytotoxins mediate acute lung injury, therefore these observations appeared to correlate with the appearance of bloody ETA samples from this patient as a diagnosis of VAP was made. To confirm that the latter 3 strains were more cytotoxic compared to the first two strains, in vitro cell culture cytotoxicity assays were performed for each strain. Cytotoxicity towards human epithelial cells following a 90 min co-incubation was significantly higher for strains 3, 4 and 5 compared to strains 1 and 2 (Fig. 2E). Collectively, these results supported the hypothesis that during the course of pulmonary infection, a transition from a chronic biofilm mode of virulence to one of planktonic acute, type III mediated pathogenicity occurred in these clonal isolates and moreover that these phenotypes were maintained in vitro.

PQS signal concentration regulates secretion of type III cytotoxins

A clue as to how this physiological shift was regulated came from the observation that pyoverdin production decreased dramatically in the latter 3 strains. Recently, it has been demonstrated that pyoverdin production is regulated by the quorum sensing molecule PQS (7, 13), suggesting that altered concentrations of this signal molecule may be responsible for the physiological changes observed. To examine this hypothesis, an ethyl acetate extract of secreted organic compounds (typically used to purify quinolone signals) was quantified for each strain by TLC, as described in material and methods. As hypothesized, strains 1 and 2 which expressed virulence factors associated with a biofilm mode of lifestyle, produced large concentrations of organic compounds, while the latter 3 type III secreting strains, did not (Fig. 3A). To determine if quinolone signal concentration was specifically affected, Q-PCR analysis of pqsA (the first gene in the PQS biosynthetic operon) copy number was performed for all 5 clinical isolates. pqsA copy number was on average more than 12-fold higher in the initial two strains compared to the average copy number detected in the latter 3 strains (Fig. 3B), confirming that quinolone signal production decreased substantially in the latter 3 isolates.

Fig. 3.

Fig. 3A. Substantially higher concentrations of extracellular organic compounds (including quinolone signals) in cell-free extract are produced by strains 1 and 2, compared to strains 3, 4 and 5. B. Q-PCR analysis confirms that strains 1 and 2 express approximately 12-fold higher pqsA copy number compared to strains 3, 4 and 5.

To confirm involvement of the PQS biosynthetic cluster in regulation of type III secretion, all isolates were exposed to farnesol, a sesquiterpene produced by many organisms, including Candida albicans, and recently identified as an inhibitor of PQS production via promotion of a non-productive binding between PqsR (mfvR) and the pqsA promoter (11). Following 4 h of growth in the presence of farnesol (which did not impact culture denisty), all 5 of the clinical isolates secreted ExoU (Fig. 4A). Notably, the first two isolates, which consistently produced high concentrations of PQS and had not previously demonstrated cytotoxin secretion, exhibited ExoU secretion in the presence of this inhibitor of PQS biosynthesis. To determine specifically which of the two known quinolone signals, HHQ or PQS was responsible for the differential type III secretion observed, purified extracts from PQS-sufficient (clinical strain 1 and PAO1) and deficient (PAO1ΔpqsH) strains were added exogenously to cultures of the 5 clinical isolates. As a control, a set of cultures without exogenous extract was incubated under identical conditions. Western analysis of the secreted fraction of each culture demonstrated that in the absence of exogenously added extract (control), strains exhibited the same profile as previously observed: no cytotoxin secretion by strains 1 and 2, while strains 3, 4 and 5 secreted ExoU (Fig. 4B). Strains exposed to exogenous supernatant extract from PQS-sufficient strains: clinical strain 1 (this study), or PAO1 (to control for molecules that may be secreted specifically by the clinical strain), exhibited no detectible ExoU secretion by any isolate (Fig. 4B), suggesting that high concentrations of a secreted molecule in these supernatants negatively regulated type III cytotoxin secretion. The identity of this signal was confirmed as PQS by the observation that the same strains cultured in the presence of a PQS-deficient extract (purified from PAO1ΔpqsH that produces all other quinolone compounds except PQS), exhibited the same secretion profile as the unexposed control cultures (Fig. 4B). It is pertinent to note that the culture conditions employed for all of these experiments likely results in reduced aeration (see materials and methods), suggesting that PQS regulation of type III cytotoxin secretion may be specific to low oxygen conditions e.g. hypoxic pockets in chronically infected airways in the host.

Fig. 4.

Fig. 4A. Inhibition of AHQ biosynthesis by farnesol addition induces a type III secretion phenotype in all five isolates. B. Strains cultured without exogenous extract exhibit cytotoxin secretion by the latter 3 strains (Control), those exposed to PQS-containing extract (PQS⁺) obtained from clinical strain 1 (this study) or the laboratory strain PAO1 exhibit no evidence of cytotoxin secretion, while strains cultured in the presence of extract from PAO1ΔPqsH exhibit a secretion profile identical to the control cultures, confirming the role of PQS in type III cytotoxin secretion.

Discussion

To date, the bacterial signals that regulate transition between reciprocally regulated chronic and acute modes of infection in P. aeruginosa are not well understood. While high concentrations of specific quorum sensing signals have previously been shown to induce expression of virulence factors associated with the chronic mode of infection (8), the role of these molecules on the type III secretion system which is active during acute infection is less clear. In this study, a series of ETA samples from a patient developing VAP over a 3-week period provided a collection of clonal P. aeruginosa isolates that appeared to transition from a chronic to an acute infection during the interval of observation. Analysis of virulence factors associated with both modes of infection, confirmed that strains of the dominant morphotype isolated initially from this patient exhibited a proclivity for chronic infection, even under culture conditions that induce the TTSS. Conversely, the latter 3 isolates exhibited type III secretion phenotypes consistent with acute infection. Persistence of phenotypic characteristics of chronic or acute virulence gene expression associated with each strain in vitro suggests that subtle genomic changes, not identified by standard PFGE e.g. mutations, small-scale chromosomal rearrangements or epigenetic modifications, may have “hard-wired” these strains into the respective mode of infection. This is particularly interesting, given the recent finding that antibiotic administration (which occurred during the course of this patients illness), can induce hypermutability in bacterial strains through reactive oxygen species generation (23), providing a potential explanation for evolution of more virulent sub-populations in otherwise “clonal” strains.

A clue to the nature of the signal that governed this phenotypic switch came from the observation that the first two strains produced substantially more pyoverdin compared to the latter 3 isolates. Pyoverdin is a fluorescent secondary metabolite produced by P. aeruginosa which functions as a siderophore and facilitates iron sequestration (10). Recently it has been demonstrated that the quorum sensing signal molecule PQS regulates pyoverdin production (7, 13). Examination of crude supernatant extracts from each isolate supported the hypothesis that the concentration of an extracellular product was substantially increased in the initial isolates compared to the latter strains. In addition, inhibition of PQS production (and all other quinolones) by the sesquiterpene farnesol, restored type III secretion phenotype to all strains, suggesting a role for quinolone signaling in controlling TTSS by P. aeruginosa under reduced aeration culture conditions. Early isolates that exhibited expression of virulence factors associated with chronic infection produced substantially higher pqsA mRNA, compared to the latter isolates that exhibited type III cytotoxin secretion (and produced little or no pqsA transcript). Addition of PQS-sufficient or -deficient supernatant extracts to cultures of these clinical isolates, or inhibition of PQS biosynthesis by using farnesol, demonstrated that PQS concentration, either directly or indirectly specifically regulates secretion of type III cytotoxins. Collectively, these observations support the hypothesis that elevated concentrations of PQS promote expression of virulence factors associated with chronic infection while low concentrations or inability to synthesize PQS appear to trigger the reciprocal, acute mode of infection through physiological reprogramming of P. aeruginosa that induces motility (1, 2) and secretion of type III cytotoxins. This data is supported by a recent study (1) demonstrating that inactivation of the PA2663 gene of P. aeruginosa repressed PQS production, leading to a substantial (seven-fold) decrease in elastase activity and induction of genes for motility. Previously we have demonstrated that type III secretion is associated with motile cells that possess a relatively uncomplex lipopolysaccharide (2), which we hypothesized enhanced motility and contact with mammalian cells. These motile cells would, presumably, also be exposed to lower concentrations of PQS, which may explain their proclivity to secrete type III cytotoxins.

Bleves et al demonstrated down-regulation of type III secretion using P. aeruginosa las and rhl mutants, as well as increased secretion of ExoS in a rhl mutant strain (6), quorum sensing systems that are known to impact PQS production (12, 31). More recently, addition of acylated homoserine lactones (AHL’s) has been shown to down-regulate expression, production and supernatant (secreted) LcrV concentrations. LcrV is a key protein in the type III secretion system of Yersinia pestis and homolog of the P. aeruginosa PcrV protein both of which are located at the tip of the needle complex (18). Conflicting data exists in the literature regarding the role of PQS in regulating P. aeruginosa type III cytotoxin gene expression. Shen et al, demonstrated that addition of exogenous PQS to cultures of P. aeruginosa did not impact expression of the exsCEBA operon that encodes the major transcriptional regulators of type III secretion genes (42). However, more recently, Kong et al, demonstrated that a psqR mutant exhibited significant increases in expression of ExoS and ExoT using a luminescent reporter fused to individual promoters of each of the these cytotoxins (24). Shen and colleagues equated expression with secretion, which, based on our data may not be true of P. aeruginosa type III secretion. We demonstrate that, in agreement with the Shen study, intracellular concentrations of ExoT and U are relatively unaffected by PQS concentration under our specific culture conditions; all strains demonstrated relatively equivalent intracellular concentrations of these cytotoxins. However, our data quite clearly demonstrates that secretion of these cytotoxins is dependent on the concentration of PQS, suggesting that a further level of PQS-mediated post-translational regulation governing cytotoxin translocation exists in P. aeruginosa.

The mechanism by which PQS plays into the regulatory circuitry associated with cytotoxin secretion remains to be determined. Given that several two component regulators are known to reciprocally control acute and chronic virulence gene expression by P. aeruginosa (47) and are known to respond to environmental signals, it is likely that PQS exerts its effect through at least one of these molecular switches. This hypothesis is bolstered by recent reports that mutation in retS, the sensor component of one of these reciprocal switches, attenuates type III virulence (51), coincident with increased exopolysaccharide production (19). Our data generated under conditions of potentially reduced oxygen availability, suggest that PQS may regulate type III secretion under these circumstances. Nonetheless, low oxygen conditions are clinically relevant in lower respiratory tract infections where bronchial obstruction from mucus plugging leads to hypoventilation and diffusion of alveolar oxygen into pulmonary arterioles. In addition, there are several reports of hypoxic pockets in the airways of patients with chronic pulmonary infection, and, moreover that P. aeruginosa exists in these mucopurulent niches (49) suggesting that PQS may play a key role in regulating virulence gene expression under these conditions.

The natural history of clinical P. aeruginosa isolates in chronic airway infections is reported to involve a loss of type III secretion over time in sequential isolates (22). Our data suggests that strains may also transition in the opposite direction from a “chronic” to an “acute” phenotype under specific conditions. While loss of type III secretion ability appears true of the specific populations studied (22) and may be due to the selection of a dominant biofilm-forming genotype under conditions of repeated antimicrobial administration, it is also possible that very small sub-populations exist within these communities that retain the ability to secrete type III cytotoxins. Certainly there is evidence for phenotypically distinct sub-populations in single-species bacterial biofilms [reviewed in (45)] and the recent demonstration of antimicrobial-induced hypermutability leading to generation of “libraries” of mutants (23) make it highly plausible that small subsets of cells with subtle genotypic distinctions may coexist in these communities. Under the appropriate conditions e.g. when microbial competition is low due to antimicrobial administration, this small subset could potentially become dominant, a strategy that would increase the overall fitness of the population.

That QS molecules play such a key role in regulating virulence factors in P. aeruginosa, makes them an attractive target for development of therapeutics. Since the identification of the first furanone compounds produced by the marine algae Delisia pulchra that exhibit quorum sensing inhibiting (QSI) activity, many groups have reported natural and synthetic compounds that exhibit this ability (4, 20, 29, 50) and protect against bacterial infection in a murine model (28, 34, 50). Human epithelial cells also produce enzymes that interfere with QS (9, 38), suggesting that this strategy may be employed as a form of innate immunity against bacterial infection. As a result, quorum sensing inhibitors are being developed as promising novel antimicrobial therapeutics (40). While anti-Pseudomonal QSI compounds may block PQS production and decrease expression of virulence factors associated with a chronic, biofilm lifestyle, there is the possibility that preventing biofilm formation could potentially promote motility and type III cytotoxin secretion. Given our findings, QS-blocking therapies may be most efficacious when used in combination with additional therapeutics e.g. the use of QS inhibitor in combination with a type III blocking therapy and appropriate antimicrobials. Such a combination could force P. aeruginosa out of the protective environment of the biofilm and into a free-swimming mode of lifestyle, thus increasing antimicrobial efficacy. Since we have previously demonstrated that these free-swimming cells are also associated with type III secretion, use of an effective anti-type III secretion therapeutic (15, 21, 35) to prevent acute injury in combination with a QSI and antimicrobial administration may provide an excellent means of resolving recalcitrant chronic infection while preserving epithelial integrity in a number of patient groups. The effects of these types of therapeutic combinations to combat P. aeruginosa infections have yet to be determined. What remains clear however, is that the full impact of these emerging species-specific therapeutics must be extensively evaluated in the context of a clinical strain physiology to truly determine their efficacy.