The Pseudomonas aeruginosa quinolone signal molecule overcomes the cell density-dependency of the quorum sensing hierarchy, regulates rhl-dependent genes at the onset of stationary phase and can be produced in the absence of LasR

Authors

  • Stephen P. Diggle,

    1. Institute of Infection, Immunity and Inflammation, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2RD, UK.
    2. School of Pharmaceutical Sciences University Park, University of Nottingham, Nottingham NG7 2RD, UK.
    Search for more papers by this author
  • Klaus Winzer,

    1. Institute of Infection, Immunity and Inflammation, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2RD, UK.
    Search for more papers by this author
  • Siri Ram Chhabra,

    1. School of Pharmaceutical Sciences University Park, University of Nottingham, Nottingham NG7 2RD, UK.
    Search for more papers by this author
  • Kathryn E. Worrall,

    1. Institute of Infection, Immunity and Inflammation, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2RD, UK.
    2. School of Pharmaceutical Sciences University Park, University of Nottingham, Nottingham NG7 2RD, UK.
    Search for more papers by this author
  • Miguel Cámara,

    1. Institute of Infection, Immunity and Inflammation, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2RD, UK.
    2. School of Pharmaceutical Sciences University Park, University of Nottingham, Nottingham NG7 2RD, UK.
    Search for more papers by this author
  • Paul Williams

    Corresponding author
    1. Institute of Infection, Immunity and Inflammation, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2RD, UK.
    2. School of Pharmaceutical Sciences University Park, University of Nottingham, Nottingham NG7 2RD, UK.
      E-mail paul.williams@nottingham.ac.uk; Tel. (+44) (0)115 9515047; Fax (+44) (0)115 8466296.
    Search for more papers by this author

E-mail paul.williams@nottingham.ac.uk; Tel. (+44) (0)115 9515047; Fax (+44) (0)115 8466296.

Summary

In Pseudomonas aeruginosa, diverse exoproduct virulence determinants are regulated via N-acylhomoserine lactone-dependent quorum sensing. Here we show that 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS) is also an integral component of the quorum sensing circuitry and is required for the production of rhl-dependent exoproducts at the onset of stationary phase. Analysis of spent P. aeruginosa culture supernatants revealed that PQS is produced at the end of exponential phase in the parent strain and in the late stationary phase of a lasR mutant. Mutants defective in both PQS production (pqsR-) and response (pqsE-) produced substantially reduced levels of exoproducts but retained wild-type N-butanoyl homoserine lactone (C4-HSL) levels. In the wild type, provision of exogenous PQS at the time of inoculation significantly increased PA-IL lectin, pyocyanin and elastase production during early stationary phase and promoted biofilm formation. Exogenous PQS but not PQS derivatives lacking the 3-hydroxy group overcame the cell density but not growth phase-dependent production of exoproducts. PQS also overcame the transcriptional and post-transcriptional repression of lecA (which codes for the PA-IL lectin) mediated via the negative regulators MvaT and RsmA respectively. Increased expression of lecA in the presence of exogenous PQS can be explained partially by increases in RhlR, RpoS and C4-HSL levels. A refined model for quorum sensing in P. aeruginosa is presented.

Introduction

Pseudomonas aeruginosa is a Gram-negative bacterium, capable of causing disease in plants, animals and humans (Rahme et al., 1995). It is a major source of nosocomial infections, and is a leading cause of mortality in cystic fibrosis patients (Govan and Deretic, 1996). As an opportunistic human pathogen, P. aeruginosa can colonize a wide variety of anatomical sites. This is because the organism produces an arsenal of extracellular virulence factors which are capable of causing extensive tissue damage, bloodstream invasion and consequently promoting systemic dissemination. Many of these exoproducts are regulated in a cell density-dependent manner via cell-to-cell communication or ‘quorum sensing’ (reviewed in Swift et al., 2001). Pseudomonas aeruginosa possesses two N-acylhomoserine lactone (AHL)-dependent quorum sensing systems. These are termed the las and rhl systems, comprising of the LuxRI homologues, LasRI (Gambello and Iglewski, 1991; Passador et al., 1993) and RhlRI (Ochsner et al., 1994; Latifi et al., 1995) respectively. LasI directs the synthesis of primarily N-(3-oxododecanoyl)-l-homoserine lactone (3-oxo-C12-HSL) and together with the transcriptional regulator LasR regulates the production of virulence factors including elastase, the LasA protease, alkaline protease and exotoxin A (Gambello and Iglewski, 1991; Toder et al., 1991). In addition, the las system controls expression of the xcpP and xcpR genes which are involved in the regulation of the Type II general secretion pathway (Chapon-Herve et al., 1997) and has been implicated in the maturation of P. aeruginosa biofilms (Davies et al., 1998). RhlI directs the synthesis of N-butanoyl-l-homoserine lactone (C4-HSL) (Winson et al., 1995) which activates RhlR and in turn RhlR/C4-HSL induces the production of rhamnolipid, elastase, LasA protease, hydrogen cyanide, pyocyanin, siderophores and the cytotoxic lectins PA-I and PA-II (Ochsner et al., 1994; Brint and Ohman, 1995; Latifi et al., 1995; Pearson et al., 1995; Winson et al., 1995; Latifi et al., 1996; Winzer et al., 2000; Diggle et al., 2002). The las and the rhl systems are considered to be organized in a hierarchical manner such that the las system exerts transcriptional control over both rhlR and rhlI (Latifi et al., 1996).

In addition to 3-oxo-C12-HSL and C4-HSL, P. aeruginosa releases a 4-quinolone signal molecule into the extracellular milieu, the synthesis and bioactivity of which has been reported to be mediated via the las and rhl systems respectively. This molecule has been chemically identified as 2-heptyl-3-hydroxy-4(1H)-quinolone and termed the Pseudomonas Quinolone Signal (PQS) (Pesci et al., 1999) and culture supernatants were found to contain approximately 6 µM, although the authors considered this to be an underestimate (Pesci et al., 1999). LasR has been shown to regulate PQS production and the provision of exogenous PQS induces expression of lasB (coding for elastase), rhlI and rhlR (Pesci et al., 1999; McKnight et al., 2000) suggesting that PQS activity constitutes a regulatory link between the las and rhl quorum sensing systems. However, McKnight et al. (2000) suggested that PQS is not involved in sensing cell density as it was identified much later in the growth cycle than is typical for a quorum sensing signal molecule. It was speculated that the purpose of PQS in regulating rhlI expression, may be to further upregulate the rhl quorum sensing system in late stationary phase cultures. Interestingly, PQS has recently been isolated from the lungs of CF patients infected with P. aeruginosa (Collier et al., 2002; Guina et al., 2003) and the presence of the molecule in vivo may be a factor in allowing P. aeruginosa to develop or maintain a chronic state (Collier et al., 2002). The regulation of PQS synthesis is becoming better understood. Calfee et al. (2001) demonstrated that anthranilate is a precursor of PQS and that inhibition of anthranilate results in a loss of PQS production. More recently, the structural genes required for PQS have been identified (pqsABCDH) along with a transcriptional regulator (pqsR) and a response effector (pqsE) (Gallagher et al., 2002). The transcription of pqsH is regulated by the las quorum sensing system, linking quorum sensing and PQS regulation. Mutations in the PQS genes results in a loss of PQS synthesis and a corresponding loss in pyocyanin production. A mutation in the pqsE gene also results in a loss of pyocyanin even though PQS synthesis remains intact. This suggests that pqsE is not required for PQS biosynthesis and may have a role in the cellular response to PQS (Gallagher et al., 2002).

The quorum sensing-dependent production of exoproducts in P. aeruginosa is tightly regulated with respect to growth phase and growth environment. In contrast to the AHL-dependent induction of bioluminescence in V. fischeri (Eberhard et al., 1981) and carbapenem antibiotic production in Erwinia carotovora (Williams et al., 1992), the provision of exogenous AHLs does not advance the expression of several quorum sensing dependent genes in wild type P. aeruginosa PAO1 (Diggle et al., 2002; Pearson, 2002). For example when supplied to the wild-type strain at the start of growth, exogenous C4-HSL and/or 3-oxo-C12-HSL does not advance lecA, lasB or rhlR expression (Diggle et al., 2002). This is likely to be due to the contribution of additional regulatory factors in addition to LasR and RhlR. These include the stationary phase sigma factor RpoS (Winzer et al., 2000) and the transcriptional regulators GacA (Reimmann et al., 1997) and MvaT (Diggle et al., 2002). Furthermore, it has been shown that in PAO1, the rhlAB operon cannot be induced in the logarithmic phase of growth in the presence of both RhlR and C4-HSL (Medina et al., 2003). Mutation of the third P. aeruginosa LuxR homologue termed QscR (for quorum sensing control repressor) (Chugani et al., 2001) results in the premature transcription of lasI and rhlI which in turn results in the earlier production of both 3-oxo-C12-HSL and C4-HSL, respectively (Chugani et al., 2001). The expression of lasI has also been reported to be delayed by the product of rsaL, a gene located in the intergenic region between lasR and lasI and which itself is regulated by LasR and 3-oxo-C12-HSL (DeKievit et al., 1999). Furthermore, the post-transcriptional regulator RsmA has also been shown to negatively regulate quorum sensing dependent genes in P. aeruginosa (Pessi et al., 2001). In an rsmA mutant, lasI but not rhlI expression is advanced and enhanced as is the AHL-dependent exoproduct, hydrogen cyanide. Overexpression of the guanosine 3′,5′-bisphosphate (ppGpp) synthetase, RelA in P. aeruginosa, has also been reported to prematurely induce expression of lasR, rhlR and production of 3-oxo-C12-HSL (Van Delden et al., 2001). Hence, the quorum sensing circuitry in P. aeruginosa is just one component of an intricate regulatory network which integrates multiple sensory inputs including cell population density and growth phase. Furthermore, some P. aeruginosa quorum sensing target genes are dependent on both las and rhl (e.g. lasB; rhlAB) whereas others can be considered primarily las- or primarily rhl-dependent (e.g. lecA; Latifi et al., 1995; Whiteley et al., 1999; Winzer et al., 2000).

In this paper we show that PQS is capable of overcoming the cell density-dependent but not growth phase-dependent expression of P. aeruginosa virulence determinants and promoting biofilm production. In contrast to previous reports that PQS is produced primarily in late stationary phase cultures (McKnight et al., 2000), we demonstrate that PQS is substantially produced in late logarithmic phase and is also produced in the absence of lasR. The loss of PQS biosynthesis is shown to result in the abolition of primarily rhl-dependent quorum sensing phenotypes despite continued C4-HSL synthesis. Furthermore although exogenously provided PQS does not overcome the requirement for RhlR and RpoS with respect to lecA expression, PQS stimulates the LasR-independent activation of the rhl system and overcomes both transcriptional (via MvaT) and post-transcriptional (via RsmA)-mediated repression of lecA expression. A refined model for quorum sensing in P. aeruginosa integrating the action of all three signal molecules is presented.

Results

PQS is detectable at the onset of stationary phase and is produced in the absence of LasR

As PQS plays a role in the regulation of virulence determinants such as elastase in P. aeruginosa, we sought to determine when in growth PQS was produced. Previous work has suggested that PQS is produced maximally in late stationary phase and hence the primary function of PQS was suggested to be the upregulation of the rhl system late in the stationary phase of growth (McKnight et al., 2000). To test this hypothesis, we sampled P. aeruginosa PAO1 culture supernatants at intervals throughout growth and assayed for the presence of PQS using thin layer chromatography. Figure 1A demonstrates that PQS can be detected in the logarithmic phase of growth after only 4 h. Semi-quantitative TLC analysis indicated that after 6 h (approximate onset of stationary phase; Fig. 1C), the culture contained 5–10 µM PQS which increased up to 25 µM after 24 h (data not shown). The presence of PQS at 6 h coincided with the production of both pyocyanin and PA-IL lectin (data not shown). Previously it has been shown that LasR is required for PQS production (Pesci et al., 1999). It was therefore surprising to find that under our growth conditions, PQS was strongly detected in a lasR mutant 24 h post inoculation (Fig. 1B), which corresponds with high levels of pyocyanin, the RhlR protein and C4-HSL (data not shown) as well as PA-IL lectin (Winzer et al., 2000). This data demonstrates that under certain growth conditions, there is LasR-independent activation of the rhl quorum sensing system, which is accompanied by LasR-independent production of PQS.

Figure 1.

Production of PQS occurs at the onset of stationary phase. Cultures were grown in 100 ml LB broth in a 1 L flask at 37°C with shaking 200 r.p.m. Ten millilitres of culture supernatant was taken for PQS analysis by thin layer chromatography as described in the Experimental procedures after 2, 4, 6, 8 and 24 h. Strains analysed were PAO1 (A) and PAO-P4 (lasR) (B). The first lane in each panel contains a synthetic PQS standard. A growth curve for PAO1 is provided showing the onset of stationary phase is at ∼6 h post inoculation (C).

rhl-dependent phenotypes are abolished in the absence of PQS signalling

Recently, Gallagher et al. (2002) identified a number of genes which are required for PQS production. A mutation in the LysR-type transcriptional regulator pqsR results in a loss of PQS biosynthesis and pyocyanin production. A mutation in the pqsE gene results in a loss of pyocyanin but PQS production remains like the wild type. To analyse the role of PQS in regulating rhl-dependent phenotypes, we investigated the effects of pqsR and pqsE mutations on the production of PA-IL, pyocyanin, elastase and C4-HSL by growing MP551 (pqsR), MP605 (pqsE) and the corresponding wild type (MPAO1) in the presence/absence 60 µM PQS for 8 h (to early stationary phase). In addition, rhamnolipid production was examined after growth on SW agar plates. Pyocyanin (Fig. 2A) and PA-IL lectin (Fig. 2C) were virtually abolished in both MP551 and MP605 which also produced substantially reduced levels of elastase (Fig. 2B) and rhamnolipids (Fig. 2D) when compared with the wild type. This was despite the fact that both MP551 and MP605 produce wild-type levels of C4-HSL (data not shown), indicating that the rhl system was functioning normally. In addition, RpoS levels were also unchanged (data not shown). This suggests that PQS is required for the production of primarily rhl-dependent phenotypes and exerts its effect via a novel regulatory pathway. However, whereas the addition of 60 µM PQS to the culture medium of MPAO1 resulted in enhancement of pyocyanin, PA-IL lectin and elastase levels, it failed to restore these phenotypes in MP551 and MP605. To confirm that PQS is required for the expression of primarily rhl-dependent genes we grew P. aeruginosa PAO1 containing a chromosomal lecA::lux reporter gene fusion (Winzer et al., 2000) in the presence of methyl anthranilate, a compound known to efficiently inhibit the synthesis of PQS (Calfee et al., 2001). When PAO1 lecA::lux was grown in the presence of a range of concentrations of methyl anthranilate, lecA expression was inhibited in a concentration dependent manner (Fig. 2E). Concentrations of 500 µM methyl anthranilate and above, resulted in complete inhibition of lecA expression without affecting growth. Interestingly, the inhibition of lecA expression observed in the presence of 1 mM methyl anthranilate could be partially restored by the addition of 60 µM PQS (Fig. 2F). Furthermore, addition of 1 mM methyl anthranilate also completely inhibited pyocyanin production and this also could be partially restored by PQS (data not shown). These data suggest that methyl anthranilate blocks the action as well as the production of PQS.

Figure 2.

The presence of PQS is required for the production of rhl-dependent virulence determinants. MPAO1 (wild type), MP551 (pqsR -) and MP605 (pqsE) were grown in 100 ml LB broth in a 1 L flask at 37°C with shaking 200 r.p.m. Production of pyocyanin (A) and elastase (B) was assayed 8 h post inoculation in the absence (clear bars) and presence (grey bars) of 60 µM PQS. PA-IL lectin (C) was analysed 8 h post inoculation by SDS-PAGE and Western blotting (see Experimental procedures) after growth in the absence (–) and presence (+) of 60 µM PQS. Rhamnolipid production (D) was assayed using SW plates (see Experimental procedures). (E) Effect of methyl anthranilate on lecA::lux expression. PAO1 lecA::lux was grown in LB medium in the absence of exogenously added methyl anthranilate (◆) or in the presence of 100 µM methyl anthranilate (□); 500 µM (▵); 1 mM (X). (F) PQS/methyl anthranilate competition assay. PAO1 lecA::lux was grown in the absence of any molecule (◆); PQS 60 µM and methyl anthranilate 500 µM (□) to determine whether enhanced and advanced lecA expression in the presence of PQS could be inhibited by methyl anthranilate.

PQS overcomes the cell density-dependence of lecA expression but not the growth phase-dependency

The influence of PQS and two PQS analogues (to examine PQS structural specificity) on growth phase-dependent gene regulation was subjected to a more thorough analysis. First, we first assessed the general effect of PQS on the growth of P. aeruginosa. LB broth was supplemented with PQS at concentrations from 0 to 200 µM, inoculated, and cultures were grown overnight at 37°C with shaking in either microtitre plates or in culture flasks. Figure 3A shows that addition of increasing concentrations of PQS in microtitre plates resulted in an extended lag phase (at high concentrations) and lower stationary phase optical densities compared to growth in LB medium alone. No effects on growth were seen with the PQS analogue 2-heptyl-4(1H)-quinolone even at concentrations as high as 100 µM and only a slight reduction in growth was seen with 3-formyl-2-heptyl-4(1H)-quinolone at 100 µM concentration (data not shown). Similar results were obtained for cultures grown with shaking in Erlenmeyer flasks (data not shown).

Figure 3.

A. Growth curves of the P. aeruginosa lecA::lux in the absence/presence of PQS. Strains were grown at 37°C for 22 h in LUCYI (Anthos, Labtech). No PQS (◆); 10 µM PQS (□); 50 µM PQS (▵); 100 µM PQS (X); 150 µM PQS (O); 200 µM PQS (•).
B. Effect of PQS on lecA::lux expression in P. aeruginosa. PAO1 lecA::lux was grown in LB medium in the absence of exogenously added PQS (◆) or in the presence of 10 µM PQS (□); 50 µM PQS (▵); 100 µM PQS (X); 150 µM PQS (O); 200 µM PQS (•).
C. Light output (lecA expression) of the culture with time in the absence of exogenously added PQS (◆) or in the presence of 10 µM PQS (□); 50 µM PQS (▵); 100 µM PQS (X); 150 µM PQS (O); 200 µM PQS (•). In all experiments, RLU (relative light units) and OD495 were determined as described in Experimental procedures. The results represent single experiments although each experiment was repeated three times with similar results.

Previously we have shown that expression of the lecA gene is quorum sensing-regulated (Winzer et al., 2000), growth phase- and RpoS- dependent (Diggle et al., 2002). To determine whether PQS could override this type of regulation, the P. aeruginosa PAO1 lecA::lux reporter strain (Winzer et al., 2000) was grown in the absence or presence of a range of PQS concentrations (Fig. 3B). Addition of increasing concentrations of PQS enhanced transcription of lecA in a concentration-dependent manner, leading to expression at significantly lower cell densities (due to its effect on growth). However, a comparison of the growth profiles (Fig. 3A) with the time-point of lecA induction (Fig. 3C) shows that PQS could not advance lecA expression to the logarithmic phase of growth: lecA expression always occurred at the transition to stationary phase. This suggests that although exogenous PQS is capable of overcoming the cell density-dependent regulation of lecA, it fails to override the growth phase-dependency. When cultures of the P. aeruginosa lecA::lux reporter were supplemented with one of two PQS analogues either lacking the 3-hydroxy substituent or carrying a 3-formyl substituent, i.e. 2-heptyl-4(1H)-quinolone and 3-formyl-2-heptyl-4(1H)-quinolone, respectively, lecA expression was neither enhanced nor advanced (data not shown). This finding indicates the importance of the 3-hydroxy group for PQS functionality.

PQS advances elastase and pyocyanin production and enhances biofilm development in P. aeruginosa

To determine whether PQS could overcome the growth phase-dependent super regulation of other quorum sensing-controlled genes, we examined the effect of PQS on both elastase and pyocyanin as a function of cell density and growth phase. PAO1 was grown in the presence/absence of 60 µM PQS and assayed throughout the growth curve for both quorum sensing-dependent exoproducts. Both pyocyanin (Fig. 4A) and elastase (Fig. 4B) were clearly enhanced and advanced in the presence of PQS with respect to cell density. However, a comparison of the growth profiles revealed that production of neither virulence determinant could be advanced into the logarithmic phase (data not shown). Finally, biofilm formation in the presence of PQS was examined. PQS concentrations of 60 µM and above, significantly enhanced surface coverage of PAO1 on stainless steel coupons (Fig. 4C and D).

Figure 4.

Influence of PQS on multiple phenotypes in P. aeruginosa. PAO1 was grown in 100 ml LB broth, 200 r.p.m. at 37°C in the absence or presence of 60 µM PQS and samples were taken throughout growth and analysed for (A) pyocyanin production in the absence of PQS (▪) or presence of 60 µM PQS (▵) and (B) elastolytic activity in the absence of PQS (▪) or presence of 60 µM PQS (▵).
C. The effect of PQS on biofilm development. The percentage surface coverage of P. aeruginosa PAO1 cells on stainless steel coupons in the presence of increasing (0–100 µM) PQS concentrations.
D. A representation of surface coverage of acridine orange stained P. aeruginosa biofilms grown on stainless steel coupons in the absence or presence of 60 µM PQS.

Effect of PQS on RpoS, RhlR and C4-HSL production

The enhancement of lecA expression as well as elastase and pyocyanin production by PQS suggested that this quinolone signal molecule regulates rhl-dependent quorum sensing phenotypes. As the rhlRI locus is essential for production of the PA-IL and PA-IIL lectins (Winzer et al., 2000; Diggle et al., 2002) as well as rhamnolipid and pyocyanin production, and the rpoS gene product is either required or has a major impact on these phenotypes, we examined the effect of PQS on RhlR and RpoS protein levels. Figure 5 shows that both RhlR (Fig. 5A) and RpoS (Fig. 5B) protein levels were enhanced in the presence of 100 µM PQS after 12 h of growth in shaking cultures. Similar results were obtained for earlier and later time-points, although production of neither of these proteins was advanced with respect to growth phase (data not shown). In the presence of 60 µM PQS, it is clear that C4-HSL levels are substantially elevated (Fig. 5C). At the transition into stationary phase, after 5 h of growth, there is approximately 3.5 times more C4-HSL in cultures supplemented with PQS compared with the control and an elevated level of C4-HSL is maintained for several hours. After 24 h, however, little C4-HSL is detected after the growth of P. aeruginosa in LB medium as a consequence of pH-dependent lactonolysis (Yates et al., 2002).

Figure 5.

Effect of PQS on RhlR (A), RpoS (B) and C4-HSL production (C). PAO1 was grown in LB medium in the absence or presence of 100 µM PQS. Protein samples for (A) and (B) were taken throughout growth and analysed by SDS-PAGE and Western blotting as described in Experimental procedures. For (C), supernatant samples were taken from a culture of PAO-P4 grown in LB broth at 37°C in the absence (clear bars) or presence (grey bars) of 60 µM PQS as previously described (Diggle et al., 2002) and quantified using an E. coli strain harbouring the reporter plasmid pSB536 (Swift et al., 1997) as described in Experimental procedures.

PQS cannot overcome the RhlR and RpoS dependence of lecA expression, but stimulates LasR-independent activation of the rhl quorum sensing system

Because the expression of lecA is completely dependent on both rhlR and rpoS and only partially dependent on lasR (Winzer et al., 2000; Diggle et al., 2002), we sought to determine whether exogenous PQS could overcome the loss of lecA expression in the corresponding mutants. PAO1 lecA::lux, PAO-P4 (lasR), PAO-P34 (rhlR) and PAO-P9 (rpoS) were grown in the presence of 10 µM C4-HSL, 10 µM 3-oxo-C12-HSL and 60 µM PQS or in different combinations of these molecules (Fig. 6). Whereas PQS alone both enhanced and advanced (with respect to cell density) lecA expression in PAO1 lecA::lux, a combination of PQS and C4-HSL and/or 3-oxo-C12-HSL enhanced (but not advanced) expression even further (Fig. 6A), suggesting that lecA is synergistically regulated by AHLs and PQS. Neither PQS nor a combination of PQS and AHLs could substantially override the inhibition of lecA caused by mutation of rhlR or rpoS although PQS either singularly or in combination with AHLs, slightly increased expression of the fusion (Fig. 6C and D). The addition of PQS alone had little effect on lecA in a lasR mutant (PAO-P4), however, when PQS was added in combination with C4-HSL, partial restoration of lecA expression was observed (Fig. 6B). Thus, the addition of C4-HSL and PQS at the time of inoculation are sufficient to induce rhlRI-dependent phenotypes during the transition to stationary phase in a LasR-independent manner. We conclude that LasR-independent activation of the rhlRI quorum sensing system, as for instance observed during mid- to late stationary phase with shaking cultures of P. aeruginosa lasR mutants (see previous text), is mediated via LasR-independent activation of PQS and C4-HSL production.

Figure 6.

Effects of quorum sensing mutations on lecA expression in the presence/absence PQS and/or AHLs. A, PAO1 lecA::lux;B, PAO-P4 (lasR; C, PAO-P9 (rpoS), and D, PAO-P34 (rhlR) were grown in LB medium in the presence of no signal molecules (◆); C4-HSL (□); 3-oxo-C12-HSL (▵); PQS (X); C4-HSL and 3-oxo-C12-HSL ( inline image); PQS and C4-HSL (•); PQS and 3-oxo-C12-HSL (O); PQS, C4-HSL and 3-oxo-C12-HSL (▴). Relative light units (RLU) and OD495 were determined as described in Experimental procedures. The results represent single experiments although each experiment was repeated three times with similar results.

PQS overcomes MvaT and RsmA-mediated inhibition of lecA

Previously we have shown that MvaT and RsmA negatively regulate the PA-IL lectin at the transcriptional and post-transcriptional levels, respectively (Pessi et al., 2001; Diggle et al., 2002). To determine whether PQS could overcome the action of these regulators, PQS was added at a concentration of 60 µM to strains overexpressing plasmid-based copies of mvaT or rsmA. Figure 7A shows that PQS can override the MvaT-mediated inhibition of lecA expression. Furthermore, PQS advances lecA transcription, even under the conditions of mvaT overexpression, in a manner similar to that seen in the PAO1 lecA::lux parent strain (Fig. 3B) and to a much greater extent in the mvaT mutant (Fig. 7A). This suggests that the effect of PQS on lecA expression in the wild type is not simply mediated by the release of MvaT repression. Similarly, PQS also overcomes the RsmA-mediated inhibition of lecA expression (Fig. 7B), which has been attributed, at least in part, to a reduction in the levels of C4-HSL and 3-oxo-C12-HSL. (Pessi et al., 2001).

Figure 7.

PQS can override MvaT (A) and RsmA (B)-mediated inhibition of lecA. PAO-P10 (mvaT) containing either pUCP18 (◆); pUCPmvaT (□) or pUCPmvaT and 60 µM PQS (▵) was grown in LB medium in LUCYI (Anthos) (A). PAO1 lecA::lux containing pME3849 was grown in LUCYI (Anthos) in the absence (◆) or presence (□) of 60 µM PQS (B). Relative light units (RLU) and OD495 were determined as described in Experimental procedures. The results represent single experiments although each experiment was repeated three times with similar results.

Discussion

Pseudomonas aeruginosa produces three chemically distinct signal molecules (C4-HSL, 3-oxo-C12-HSL and PQS) which are involved in the control of numerous exoproducts including the lecA gene product, PA-IL. In this paper we demonstrate that PQS: (i) is present at the onset of stationary phase; (ii) is required for the production of primarily rhl-regulated virulence determinants, and (iii) influences growth and overcomes the cell density but not growth phase-dependency of certain quorum sensing-controlled genes. Furthermore, we have established that the effect of PQS is partially mediated via upregulation of the rhlRI quorum sensing system and RpoS. How-ever, a rhlRI- and rpoS-independent mechanism must also exist, as RhlR, C4-HSL, and RpoS levels during entry into stationary phase are not significantly altered in mutants defective in PQS production or response, and yet these mutants fail to induce RhlR/C4-HSL-dependent virulence determinants such as PA-IL and pyocyanin. PQS also plays a role in LasR/3O-C12–3HSL-independent activation of the rhlRI quorum sensing circuit. We therefore propose a refined model for the quorum sensing hierarchy in P. aeruginosa, where PQS plays a central role as a third signalling molecule essential for gene expression during and after entry into stationary phase (Fig. 8). Details of this model, which accommodates the findings of this study and previous observations by other investigators, will be discussed in the following paragraphs.

Figure 8.

PQS-dependent regulation of RhlR/C4-HSL-controlled genes. As the cells enter stationary phase, a critical concentration of LasR/3-oxo-C12-HSL induces expression of rhlR and the switch on of the rhl quorum sensing system (1). LasR/3-oxo-C12-HSL also positively regulates the PQS biosynthetic genes (in particular pqsH) and hence PQS levels (2). Thus at the onset of stationary phase, a combination of RhlR/C4-HSL and PQS regulates production of rhl-dependent virulence determinants. Under certain growth conditions, induction of the RhlR/C4-HSL system can occur in late stationary phase in the absence of LasR (3). This induction is probably mediated via the LasR-independent production of PQS. Expression of some RhlR/C4-HSL controlled genes is also dependent on RpoS (e.g. lecA; rhlAB) whereas others are negatively regulated (e.g. genes encoding pyocyanin production) by this sigma factor. We suggest that this negative regulation occurs through an increase in PQS levels (4).* Pyocyanin production and lecA expression are completely dependent on the presence of PQS whereas rhlAB appears to be partially dependent. (→) indicates + ve regulation whilst (—|) indicates –ve regulation.

PQS is an integral part of the quorum sensing hierarchy

PQS influences the production of quorum sensing-dependent exoproducts such as pyocyanin (Gallagher et al., 2002) and elastase (Pesci et al., 1999), and also upregulates the C4-HSL synthase gene, rhlI (McKnight et al., 2000). By analysing the PQS production profile as a function of growth using a lasB-lacZ reporter-based bioassay, McKnight et al. (2000) suggested that PQS is not maximally produced until late stationary phase (well after the AHLs) and that very little PQS is present at the time when the bacterial cells enter stationary phase and induce many quorum sensing-dependent genes. They therefore proposed that PQS is not involved in sensing cell density but that its primary function is to upregulate the rhl system in late stationary phase, possibly to counter stress conditions (McKnight et al., 2000). However, direct detection of the molecule by TLC in the present study (Fig. 1) demonstrates that although PQS levels are indeed maximal in late stationary phase, they are already detectable in the logarithmic phase of growth, and that substantial levels of the molecule are present at the onset of stationary phase. Indeed, by TLC analysis, we estimate that at the onset of stationary phase, the PQS concentration is between 5 and 10 µM and increases up to 25 µM in late stationary phase. Early production of PQS has recently also been noted for P. aeruginosa isolates from CF patients and for strains grown under magnesium limitation (Guina et al., 2003). These findings indicate that PQS, apart from being important in late stationary phase, may also have a function at a much earlier stage of growth. Indeed, we have shown that the PQS signal is required for the induction of RhlR/C4-HSL-dependent phenotypes during the transition to stationary phase, and the data presented in other studies are in agreement with this function. This is because: (i) PQS accumulation in the wild type or in a lasR mutant always corresponds with the production of PA-IL lectin, elastase and pyocyanin; (ii) addition of synthetic PQS enhances PA-IL lectin, elastase and pyocyanin immediately after their induction; (iii) in mutants where PQS biosynthesis is abolished or the cellular response to PQS is disrupted, a loss of PA-IL lectin and pyocyanin (which are both primarily rhl-dependent) as well as a reduction in rhamnolipid and elastase production (which both depend on the las and the rhl systems) is observed; (iv) inhibition of PQS production by methyl anthranilate results in the loss of RhlR/C4-HSL-dependent phenotypes such as lectin and pyocyanin production (this study) and reduces lasB expression (Calfee et al., 2001) and (v) the methyl anthranilate-mediated loss of RhlR-dependent phenotypes could be overcome by the addition of synthetic PQS, demonstrating the central role of this signal molecule within the quorum sensing hierarchy of P. aeruginosa.

Factors controlling PQS production and routes of PQS action

The genes required for the generation of PQS from chorismate are under the control of various regulators. These include LasR/3-oxoC12-HSL, a LysR-type regulator, PqsR, and the np20 protein (Gallagher et al., 2002), which is highly induced by mucus from cystic fibrosis patients (Wang et al., 1996). PqsR controls the phnAB locus, the neighbouring pqsABCDE operon and pqsH (which is also regulated by LasR/3-oxo-C12-HSL and np20) (Gallagher et al., 2002). Previous studies have demonstrated that the expression of several quorum sensing-dependent genes (e.g. lecA, rhlR, lasB) is induced during the transition to stationary phase and in the wild type cannot be advanced through the addition of exogenous 3-oxo-C12-HSL and C4-HSL (Whiteley et al., 1999; Winzer et al., 2000; Diggle et al., 2002). We therefore sought to determine whether PQS was one of the signals required for the growth phase-dependent expression of these genes. However, addition of exogenous PQS, even in combination with 3-oxo-C12-HSL and C4-HSL, could not advance the expression of lecA into the logarithmic phase of growth. Because PQS affected growth at concentrations greater than the physiological level, cells entered stationary phase at a lower optical density, and consequently induced quorum sensing-dependent genes at lower densities than observed with the parent strain. This observation further supports the hypothesis of Diggle et al. (2002) that P. aeruginosa cultures reach a ‘quorum’ at comparatively low cell densities, but only respond upon entering stationary phase.

The mechanism by which PQS controls gene expression is not known. No receptor protein or signal transduction system has yet been identified, although it is clear from the present study that a PQS-controlled regulatory pathway must act at several different levels. We have shown that PQS upregulates the production of RhlR and C4-HSL, which is in agreement with the observed upregulation of rhlR and rhlI reporter gene fusions described by McKnight et al. (2000). In addition, we have shown that PQS enhances RpoS production. RpoS is required for the production of PA-IL and PA-IIL lectins (Winzer et al., 2000), rhamnolipid (Medina et al., 2003), and influences elastase production (Diggle et al., 2002). However, the fact that RpoS, RhlR and C4-HSL levels are not advanced with respect to growth phase shows that PQS, although involved in controlling levels of these regulatory elements does not determine the time point of their induction. Furthermore, a negative feedback mechanism between RpoS and PQS production appears to exist, because PQS levels are elevated in an rpoS mutant (unpublished data). Such a regulatory link would explain earlier reports that rpoS inactivation significantly increases pyocyanin production (Suh et al., 1999; Whiteley et al., 2000; Diggle et al., 2002). In addition, PQS controls target genes independently of RpoS and RhlR/C4-HSL, because these were present at wild-type levels in pqsR and pqsE mutants, at least during early stationary phase, even though no lectin or pyocyanin was detectable. These findings concur with previous work by Gallagher et al. (2002) who showed that although loss of PQS signalling resulted in a loss of pyocyanin production, there was only a slight reduction in the expression of the rhlI gene. However, this regulatory pathway cannot activate target genes independently of RpoS and RhlR, as PQS addition did not restore lecA expression in either an rpoS or a rhlR mutant. Thus, PQS constitutes an essential regulatory input. In addition, the mechanism by which PQS is able to overcome the negative effects of the transcriptional regulator MvaT and the post-transcriptional regulator RsmA on lecA expression is not yet apparent.

Although the present study suggests that PQS is not required for the activation of rhlRI during entry into stationary phase, our data indicate that PQS is important for the lasR-independent activation of this system later in stationary phase. Winzer et al. (2000) described the production of PA-IL lectin in mid to late stationary phase cultures of a lasR mutant. Similar observations have also been made for pyocyanin production in this study and by Beatson et al. (2002) using Pseudomonas isolation media. Significant amounts of RhlR, C4-HSL, and PQS were also present in late stationary phase cultures of a lasR mutant (this study). Taken together, these findings indicate that under certain conditions, LasR-independent activation of the rhl system and PQS production, respectively, can occur. Furthermore, in the presence of exogenous PQS and C4-HSL, expression of the primarily RhlR/C4-HSL-dependent lecA gene was advanced to the onset of stationary phase in a lasR mutant, suggesting that activation of the rhl system in the absence of LasR is driven by a PQS controlled mechanism.

PQS signalling influences biofilm formation

Pseudomonas aeruginosa biofilm development on stainless steel coupons was substantially enhanced in the presence of PQS. Whereas the molecular basis for this PQS-dependent phenomenon has not yet been established, AHL-mediated quorum sensing has previously been implicated in biofilm maturation. For example, Davies et al. (1998) reported that a lasI mutant forms thin biofilms lacking the differentiated biofilm architecture characteristic of the parent, although a rhlI mutant biofilm resembled that of the parent strain. Both lasI and rhlI are expressed during the biofilm mode of growth although lasI expression was found to decline over time whereas rhlI expression remained constant albeit in a lower percentage of cells (DeKievit et al., 2001). Furthermore, Heydorn et al. (2002) have shown that a flow chamber grown P. aeruginosa rpoS mutant formed densely packed biofilms that were significantly thicker than those of the wild type. Thus, the stimulating influence of PQS on both RhlR/C4-HSL and RpoS may be responsible for the biofilm phenotype observed in the presence of synthetic PQS. We have also obtained preliminary evidence to suggest that the PA-IL lectin, which is controlled by PQS, RhlR/C4-HSL, and RpoS, contributes directly to biofilm development as lecA mutants form poor biofilms (R. E. Stacey, S. P. Diggle, K. Winzer, M. Cámara and P. Williams, unpubl. obs.).

Structural requirements for PQS signalling

Pesci et al. (1999) have demonstrated that in contrast to PQS, 3-heptyl-2-hydroxy-4(1H)-quinolone, in which the positions of the hydroxy- and heptyl- substituents of PQS have been exchanged, was unable to activate lasB expression in a lasR mutant. The authors also demonstrated that an isomer of PQS (2-heptyl-4-hydroxyquinoline N-oxide) which lacks the 3-hydroxy group was inactive (Pesci et al., 1999). This study has provided further evidence that the activity of PQS depends on the presence of the 3-hydroxy group as neither 2-heptyl-4(1H)-quinolone nor 3-formyl-2-heptyl-4(1H)-quinolone exhibited any biological activity comparable with PQS. Thus, it seems likely that the PQS signal transduction pathway contains a sensor protein which recognizes PQS with a high degree of structural specificity. In this context it is of interest that methyl anthranilate at high concentrations (0.5 mM) not only completely prevented lecA expression, but also expression of this gene was only partially restored in the presence of 60 µM PQS supplied exogenously. This inhibitory activity of methyl anthranilate on PQS action may be at the level of transport or signal transduction and requires further investigation.

Taken together the data presented here demonstrate that the PQS molecule is a central component of the P. aeruginosa quorum sensing hierarchy. Further work is required to elucidate the regulation of PQS synthesis, release and mode of action.

Experimental procedures

Bacteria, growth conditions and plasmids

All bacterial strains and plasmids used in this work are listed in Table 1. Pseudomonas aeruginosa strains were routinely grown in Luria–Bertani broth (LB) or on Pseudomonas Isolation Agar (PIA) (Difco). All strains were grown at 37°C in 100 ml of broth and 1000 ml baffled Erlenmeyer flasks with shaking at 200 r.p.m. Where required, synthetic PQS was added at concentrations ranging from 10 to 200 µM. The PQS analogues 2-heptyl-4(1H)-quinolone and 3-formyl-2-heptyl-4(1H)-quinolone were added at concentrations ranging from 5 to 100 µM.

Table 1. . Bacterial strains and plasmids used in this work.
Strain/plasmidRelevant characteristicsSource/reference
  1. AmpR, ampicillin resistant; GmR, gentamycin resistant; CarR, carbenicillin resistant

Strain
P. aeruginosa
 PAO1PAO1 Nottingham wild typeHolloway collection
 MPAO1PAO1 wild typeGallagher et al. (2002)
 PAO1 lecA::luxlecA::luxCDABE genomic chromosomal reporter fusion in PAO1Winzer et al. (2000)
 PAO-P4Tn5-B21 lasR mutant derived from PAO1 lecA::luxDiggle et al. (2002)
 PAO-P9Tn5-B21 rpoS mutant derived from PAO1 lecA::luxDiggle et al. (2002)
 PAO-P34Tn5-B21 rhlR mutant derived from PAO1 lecA::luxDiggle et al. (2002)
 PAO-P10Tn5-B21 mvaT mutant derived from PAO1 lecA::luxDiggle et al. (2002)
 MP605pqsE mutant of MPAO1Gallagher et al. (2002)
 MP551pqsR mutant of MPAO1Gallagher and Manoil (2001)
Plasmid
 pME3849PME6001 with rsmA gene on a 311-bp BamHI/SmaI fragment (GmR)Pessi et al. (2001)
 pUCP18pUC18 containing an additional 1.8 kb stabilising fragment for maintenance in Pseudomonas spp. (CarR)Schweizer (1991)
 pSB536AHL biosensor; ahyR′′::luxCDABE in pAHP13 (AmpR)Swift et al. (1997)
 pUCPmvaT1.6 kb BamHI/PstI fragment PAO1 chromosomal DNA fragment containing mvaT gene in pUCP18 (CarR)Diggle et al. (2002)

Time and cell-density-dependent measurement of bioluminescence

Bioluminescence was determined as a function of cell density using a combined, automated luminometer-spectrometer (the Anthos Labtech LUCYI). Overnight cultures of P. aeruginosa were diluted 1:1000 in fresh LB medium, and 0.2 ml cultures were grown in microtitre plates. Luminescence and turbidity were automatically determined every 30 min. Luminescence is given in relative light units (RLU) divided by OD495. Where required, synthetic PQS or the PQS analogues 2-heptyl-4(1H)-quinolone and 3-formyl-2-heptyl-4(1H)-quinolone were added at concentrations ranging from 10 to 200 µM. AHLs were added as previously described (Diggle et al., 2002). Methyl anthranilate (Sigma-Aldrich) was added in concentrations ranging from 10 µM to 200 µM.

Assays for elastolytic and rhamnolipid activity

Elastolytic activity of bacterial supernatants was determined using the elastin Congo red (ECR, Sigma) assay (Ohman et al., 1980). A 100 µl aliquot of bacterial supernatant was added to 900 µl ECR buffer (100 mM Tris, 1 mM CaCl2, pH 7.5) containing 20 mg ECR and incubated with shaking at 37°C for 3 h. Insoluble ECR was removed by centrifugation and the absorption of the supernatant was measured at 495 nm. LB medium was used as a negative control. The production of rhamnolipids was analysed using SW agar plates previously described by Siegmund and Wagner (1991).

Assay for pyocyanin production

Pyocyanin was extracted from culture supernatants and measured using the method of (Essar et al., 1990). Briefly, 3 ml of chloroform was mixed with 5 ml culture supernatant. The chloroform layer was transferred to a fresh tube and mixed with 1 ml 0.2 M HCl. After centrifugation, the top layer (0.2 M HCl) was removed and its absorption measured at 520 nm.

Analysis of C4-HSL production

Supernatant samples were taken from a culture of PAO-P4 (lasR-) grown in LB broth at 37°C as previously described (Diggle et al., 2002) and quantified using an E. coli strain harbouring the reporter plasmid pSB536 (Swift et al., 1997). Supernatants were prepared at intervals (0–8 h and at 24 h) in the presence or absence of 60 µM PQS. Supernatants were diluted 1:10 in LB broth and 100 µl was added to a microtitre plate containing 100 µl 1:1000 diluted E. coli[pSB536]. Plates were incubated in LUCYI (Anthos). Luminescence and turbidity were automatically determined every 30 min Luminescence is given in relative light units (RLU) divided by OD495.

Extraction of PQS from P. aeruginosa cultures

Aliquots of 10 ml P. aeruginosa were extracted with 10 ml acidified ethyl acetate (Pesci et al., 1999), vortexed vigorously and centrifuged at 10 000 r.p.m. for 5 min. The organic phase was transferred to a fresh tube and dried to completion under a stream of nitrogen gas. The solute was resuspended in 50 µl methanol for future analysis.

Thin layer chromatography (TLC) analysis of PQS

A 10 µl sample of ethyl acetate extracted culture supernatant was spotted onto a normal phase silica 60F254 (Merck) TLC plate which had been previously soaked for 30 min in 5% KH2PO4 and activated at 100°C for 1 h. Extracts were separated using a dichloromethane:methanol (95:5) system until the solvent front reached the top of the plate. The plate was visualized using a UV transilluminator and photographed. Synthetic PQS (2 µl of a 10-mM stock concentration) was used as a positive control. 2-heptyl-3-hydroxy-4(1H)-quinolone levels in culture supernatants were estimated by comparing a series of dilutions of the synthetic standard against diluted ethyl acetate extracts using the TLC method described above.

Production of polyclonal antibodies

Polyclonal antibodies to the recombinant PA-IL, RhlR and RpoS proteins were raised in New Zealand White rabbits as described before (Winzer et al., 2000). Rabbits were immunized with 100 µg of antigen, administered subcutaneously, three times at two weekly intervals. Primary immunization used Freund's Complete adjuvant and the two subsequent immunizations used Freund's Incomplete adjuvant. Serum was obtained from the rabbits two weeks after the third immunization.

SDS-PAGE and immunoblotting

Pseudomonas aeruginosa cells were grown in LB broth and 1 ml samples taken at points throughout the growth curve. All strains demonstrated a similar growth pattern and the OD600 of each sample time was similar between strains. Cells were resuspended in SDS-PAGE sample buffer and lysed by sonication. Samples were then boiled before loading onto 15% SDS-polyacrylamide gels. Proteins were then electrophoretically transferred onto nitro-cellulose membranes (Hybond C) then probed with polyclonal antibody (1:500 dilution for PA-I lectin and 1:10 000 dilution for RhlR and RpoS). Detection was achieved using a secondary anti-rabbit IgG-horse radish peroxidase conjugate (1:3000 dilution) (Amersham Life Sciences) and developed using the ECL chemiluminescence system (Amersham Life Sciences).

Biofilm assays

After overnight growth of P. aeruginosa PAO1 in LB at 37°C, the optical density (OD600) of the culture was adjusted to 1.0 and 100 µl aliquots were used to inoculate sterile Petri dishes each containing three stainless steel coupons (10 mm by 20 mm cut from a batch of stainless steel 316 l) suspended in 10 ml of 1/10 strength LB broth containing a range of PQS concentrations from 0 to 100 µM. The plates were then incubated at 37°C on a rotary shaker (60 r.p.m.) for 72 h.

Visualization and analysis of biofilms

Stainless steel coupons were rinsed twice in sterile phosphate-buffered saline (PBS; pH 7.4) and air-dried. They were then heat-fixed and stained in 0.1% wt/vol acridine orange for 2.5 min, and rinsed in PBS. The coupons were then examined for bacterial attachment using an inverted fluorescent microscope (Nikon Eclipse TE200) Pictures were taken using a JVC KY-F58 video camera. Surface coverage was calculated with Lucia G/Comet for Nikon UK.

Synthesis of AHLs and PQS analogues

C4-HSL and 3-oxo-C12-HSL were synthesized as described by (Chhabra et al., 1993) and (Chhabra et al., 2003) respectively. AHLs were dissolved in acetonitrile before adding to growth media at the indicated concentrations. 2-Heptyl-4(1H)-quinolone, m.p. 145–146°C, was synthesized by the acid catalysed cyclocondensation of ethyl 3-oxodecanoate with aniline in 50% yield (Somanathan and Kevin, 1981). Starting from 2-heptyl-4(1H)-quinolone, 3-formyl-2-heptyl-4(1H)-quinolone, m.p. 245–248°C (dec.), and 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS), m.p. 195–197°C were synthesized in 40% and 70% yields, respectively, by the procedures described by Pesci et al. (1999). PQS analogues were dissolved in methanol before adding to growth media at the indicated concentrations.

Acknowledgements

This work was supported in part by grant QLK3-2000–01759 from the European Union (Vth Framework Programme), by a grant and studentship from the Biotechnology and Biological Sciences Research Council, UK and via an MRC Co-operative Group. We would like to thank Chris Harty for synthesising PQS and Colin Manoil for supplying the pqsR and pqsE mutants. Also we thank Stephan Heeb for technical advice and assistance.

Ancillary

Advertisement