Negative Control of Quorum Sensing by RpoN ({sigma}54) in Pseudomonas aeruginosa PAO1

In Pseudomonas aeruginosa PAO1, the expression of several virulencefactors such as elastase, rhamnolipids, and hydrogen cyanidedepends on quorum-sensing regulation, which involves the lasRIand rhlRI systems controlled by N-(3-oxododecanoyl)-L-homoserinelactone and N-butyryl-L-homoserine lactone, respectively, assignal molecules. In rpoN mutants lacking the transcriptionfactor ${sigma}$ ⁵⁴, the expression of the lasR and lasI genes was elevatedat low cell densities, whereas expression of the rhlR and rhlIgenes was markedly enhanced throughout growth by comparisonwith the wild type and the complemented mutant strains. As aconsequence, the rpoN mutants had elevated levels of both signalmolecules and overexpressed the biosynthetic genes for elastase,rhamnolipids, and hydrogen cyanide. The quorum-sensing regulatoryprotein QscR was not involved in the negative control exertedby RpoN. By contrast, in an rpoN mutant, the expression of thegacA global regulatory gene was significantly increased duringthe entire growth cycle, whereas another global regulatory gene,vfr, was downregulated at high cell densities. In conclusion,it appears that GacA levels play an important role, probablyindirectly, in the RpoN-dependent modulation of the quorum-sensingmachinery of P. aeruginosa.

INTRODUCTION

Top

Introduction

Intercellular communication systems allow bacteria to monitorenvironmental conditions and to coordinate the expression ofseveral genes, particularly those specifying extracellular productsand virulence factors, in a cell density-dependent manner (44).This communication, termed quorum sensing, is based on the productionof diffusible signal molecules (autoinducers) which accumulatein the environment during bacterial growth. Upon reaching athreshold concentration at high cellular densities, they canbind to and activate specific transcriptional regulators. Inmany gram-negative bacteria, quorum-sensing systems consistof pairs of genes in which a luxI-type gene encodes an N-acylhomoserinelactone (AHL; autoinducer) synthase and a luxR-type gene encodesa transcriptional regulator activated by the cognate autoinducer(15, 60, 67).

In the pathogen Pseudomonas aeruginosa, which is responsiblefor nosocomial infections (63) as well as for serious infectionsin patients suffering from cystic fibrosis, cancer, or burnwounds (6, 62), pathogenicity is due to the production of bothcell-associated and extracellular virulence factors, most ofwhich are regulated by quorum sensing (15, 60, 67). P. aeruginosacontains two interdependent quorum-sensing systems (26, 32,45, 70). In the lasRI system, the LasI synthase catalyzes thebiosynthesis of N-(3-oxododecanoyl)-L-homoserine lactone (OdDHL).The LasR-OdDHL complex positively regulates the expression ofvirulence factors such as elastases (LasB and LasA), exotoxinA, alkaline protease, hydrogen cyanide (HCN), and pyocyaninand, moreover, induces lasI itself, forming an autoinductionloop. LasR activated by OdDHL also positively regulates theexpression of the rhlR gene. RhlR is the transcriptional regulatorof the second autoinducer system and functions with N-butyryl-L-homoserinelactone (BHL), whose biosynthesis requires the RhlI synthase.The rhlRI system enhances the expression of multiple exoproductssuch as LasB elastase, HCN, pyocyanin, and rhamnolipids (14).Thus, there is a quorum-sensing hierarchy in which the las systemis dominant.

Both quorum-sensing systems are positively regulated by theglobal regulator GacA; in particular, the response regulatorGacA has a marked enhancing effect on BHL formation and BHL-dependentvirulence factor production (48, 52). In addition, the CRP (cyclicAMP receptor protein) homologue Vfr specifically activates thelasR promoter (1, 66). Recently, the QscR protein, a homologof LasR and RhlR, has been shown to act as a repressor of lasIand rhlI, leading to a downregulation of quorum-sensing-dependentvirulence factors (8). Thus, LasR- and RhlR-controlled quorum-sensingmechanisms are embedded in global regulatory networks. Recentwork on sigma factor ${sigma}$ ⁵⁴ (RpoN) in P. aeruginosa revealed thatthis alternative sigma factor is important for virulence inseveral models, e.g., infection of the respiratory epitheliumin cystic fibrosis xenografts and of burned mice (9, 21). Reducedvirulence of P. aeruginosa rpoN mutants may be explained, inpart, by the fact that rpoN function is necessary for the expressionof pili and flagella, two important adherence factors (25, 42,61). However, considering the pleiotropic effects of rpoN mutationson metabolic functions in P. aeruginosa (5, 27, 40, 61), ithas been postulated that the role of RpoN in the regulationof virulence factors could be quite complex, extending beyondpositive control of pilin and flagellin synthesis (9). Herewe show that RpoN exerts global negative control on the quorum-sensingmachinery of P. aeruginosa. At least part of this effect appearsto be mediated by GacA.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and growth conditions. The strains and plasmids used in this study are listed in Table1. Both Escherichia coli and P. aeruginosa strains were routinelygrown in nutrient yeast broth (NYB) or on nutrient agar platesat 37°C (58). L-Glutamine was added at a concentration of1 mM for growing the rpoN mutant strains in NYB or in minimalmedium E (64). For ß-galactosidase and autoinducerextraction experiments, P. aeruginosa strains were grown withaeration in NYB in 50-ml Erlenmeyer flasks. When required, antibioticswere added to media at the following concentrations: tetracycline(Tc), 25 µg ml^-1 (E. coli) or 125 µg ml^-1 (P. aeruginosa);ampicillin (Ap), 100 µg ml^-1 (E. coli); gentamicin (Gm),10 µg ml^-1; kanamycin (Km), 400 µg ml^-1; and spectinomycin(Sp), 1,500 µg ml^-1 (P. aeruginosa).

View this table:
[in this window]
[in a new window]
TABLE 1. Bacterial strains and plasmids

To counterselect E. coli S17-1 donor cells in matings with P.aeruginosa for gene replacement, chloramphenicol (Cm) was usedat a concentration of 10 µg ml^-1; mutant enrichment experimentswere performed with tetracycline at a final concentration of20 µg ml^-1 and carbenicillin at a final concentrationof 2,000 µg ml^-1. Flagellar swimming was tested as describedby Rashid and Kornberg (51) on NYB containing 1 mM L-glutaminesolidified with 0.3% (wt/vol) agar. Swarming was tested on platescontaining 0.5% (wt/vol) agar, 8 g of nutrient broth per liter(Oxoid), 5 g of glucose per liter, and 1 mM L-glutamine (51).Twitching motility was assayed on 1% (wt/vol) agar supplementedwith Luria broth and 1 mM L-glutamine (29).

DNA manipulation and cloning procedures. Small- and large-scale preparations of plasmid DNA were madeby the cetyltrimethylammonium bromide method (13) and with JetStarcolumns (Genomed, Basel, Switzerland), respectively. ChromosomalDNA was purified from P. aeruginosa as described elsewhere (16).Restriction enzyme digestions, PCRs, ligation, electrophoresis,and electroporation were performed with standard procedures(52) and as described elsewhere (17, 46). Nucleotide sequencesof PCR-derived constructs were determined on both strands witha dye terminator kit (Perkin-Elmer product 402080) and an ABIPrism 373 sequencer. Comparison of nucleotide and deduced aminoacid sequences was performed with the Genetics Computer Groupprogram GAP.

Plasmid constructions. Plasmid pME3829 was obtained by subcloning a 3.1-kb BamHI-HindIIIfragment from pRP1-rpoN (kindly provided by Y. Itoh), containinggenes PA4461 and rpoN, into pBBR1MCS (Fig. 1). Plasmid pME3836was constructed by PCR-amplifying a 744-bp fragment (3 min at95°C; 25 cycles of 1 min at 95°C, 30 s at 58°C,1 min at 72°C; and 3 min at 72°C) with PAO1 chromosomalDNA as the template, with the primers qscR-3 (5'-AGGCCAGGATCCTGTTTATTGTCT-3')and qscR-4 (5'-GACAAAATCTGCAGATATCCCTCT-3'). Artificial restrictionsites (italic) for BamHI and PstI, respectively, were incorporatedinto these primers. The resulting 730-bp BamHI-PstI fragment,including a potential Lux box located 486 bp upstream of theATG and the first eight codons of qscR, was fused in-frame withthe 'lacZ reporter gene from pNM482 in vector pME6010.

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. Chromosomal region of P. aeruginosa PAO1 surrounding the rpoN gene (59). The construction of strains PAO6358 and PAO6359 and of plasmids pME3338 and pME3829 is described in Materials and Methods. KH4, KH5, and KH10 indicate primers used in PCR. The assignment of pstN to open reading frame PA4464 is based on sequence comparison with the rpoN gene cluster in E. coli (28).

To obtain plasmid pME3338, a 2.1-kb EcoRI-HindIII fragment containingthe rpoN gene (Fig. 1) was excised from pME3829 and insertedinto the SpeI site of pME3280b (S. Zuber and D. Haas, unpublisheddata). To construct plasmid pME3851, the rhlR promoter regionand first nine rhlR codons were amplified by PCR as above withthe primers RR1 (5'-CGCTTGCTCGAGACCCGGC-3') and RR2 (5'-AAAACTGCAGCAGCAAAAAGCCTCCGTC-3'),which are XhoI and PstI tagged, respectively (italic); plasmidpME3840 was used as the template. The 0.3-kb PCR product wascut with XhoI, blunted, cut with PstI, fused in frame with the'lacZ reporter gene of pNM482, and cloned into pME6010. To generateplasmid pME3858, a 0.4-kb SmaI-PstI fragment from pME3827, containingthe lasR promoter region and the first 23 codons of lasR, wasfused in frame with the 'lacZ reporter gene from pNM482 in vectorpME6010.

A putative rpoN recognition site overlapping the 3' end of thelasR (lux) box in the rhlI promoter was mutated as follows.Primer rhlR-1 (5'-CCGTGGATCCGGCGATCCTC-3'), which anneals aroundthe BamHI site (italic) in the rhlR coding sequence, and primerAcl-2 (5'-GATGAACGTTTGGCAACCTGCCAGATCTGGT-3'), which is complementaryto part of the lasR box in the rhlI promoter and carries anartificial AclI site (italic), were used to PCR-amplify a 551-bpfragment from pME3846. A second PCR fragment of 183 bp was amplifiedfrom pME3846 with primer Acl-3 (5'-GCCAAACGTTCATCCTCCTTTAGTCT-3'),which is tagged with an artificial AclI site (italic) and annealsimmediately downstream of the putative rpoN box, and primerrhlI-4 (5'-AAAACTGCAGCGGAAAGCCCTTCCAGCG-3'), which is complementaryto codons 8 to 13 of rhlI and carries an artificial PstI site(italic). The two PCR fragments were ligated to each other attheir AclI sites, fused as a 0.71-kb BamHI-PstI fragment tothe 'lacZ gene of pNM481, and cloned into pME6010. Sequenceanalysis of the resulting construct, pME6186, confirmed thatin the putative rpoN recognition sequence GGCAGGTTGCCTGC, thelast three bases were replaced by AAA, destroying the secondhalf-site of the GG-N₁₀-GC motif.

P. aeruginosa mutant constructions. Care was taken to construct all mutants and chromosomal reporterfusions in the same PAO1 background. In PAO6358, a 0.9-kb fragmentwas deleted in-frame in the rpoN gene. The deletion includesthe helix-turn-helix motif and the rpoN box (Fig. 1) and wasobtained as follows. A 604-bp KpnI-SacI fragment from pME3829,including the first 190 codons of rpoN, was linked to a 620-bpSacI-BamHI fragment containing the last six codons of rpoN,gene PA4463, and the beginning of pstN (Fig. 1), which had beenamplified by PCR as above with primers KH4 (5'-AAAAGAGCTCCGCAAGCGACTGGTGTGA-3')and KH5 (5'-AAAAGGATCCGGCGATGCCATTGCCGAA-3') and cut at theartificial restriction sites for SacI and BamHI (italic). Theresulting 1.22-kb fragment was cloned into the suicide plasmidpME3087 digested with KpnI and BamHI, giving plasmid pME3834.In conjugation with PAO1 as the recipient and S17-1/pME3834as the donor, tetracycline-resistant transconjugants havinga chromosomally integrated pME3834 plasmid were selected. Aftercarbenicillin enrichment, glutamine-auxotrophic colonies wereobtained and verified by PCR for their 0.9-kb deletion in rpoNwith primers KH5 and KH10 (5'-TCCAGCAGGAAATCCAGGAAG-3') (Fig.1).

In strain PAO6359, the rpoN gene was interrupted by the insertionof a kanamycin resistance cassette ( ${Omega}$ -Km). We obtained this mutantby transducing the mutation from PAO4460 (40) into PAO1 withthe temperate phage E79tv-2 (39) as described before (17). Kanamycin-resistanttransductants were auxotrophic for glutamine. The double rpoNgacA mutant PAO6363 was obtained by transduction of the gacAmutation of PAO6281 into PAO6358 with phage E79tv-2.

In strain PAO6366, the qscR gene was interrupted by the insertionof a spectinomycin-streptomycin resistance cassette ( ${Omega}$ -Sp/Sm).The qscR gene was amplified by PCR with PAO1 chromosomal DNAas the template and primers QSCR1 (5'-AAAAGAGCTCATGGAGCGTGCGAGAAGAAC-3')and QSCR2 (5'-TAAAGGATCCTATCCGGCCATTCGGTGAAT-3'), which containartificial restriction sites for SacI and BamHI (italics). The2-kb ${Omega}$ -Sp/Sm resistance cassette from pHP45 ${Omega}$ was inserted as anEcoRI fragment into the EcoRI site located 30 codons downstreamof the qscR ATG start codon. The resulting 3.15-kb fragmentwas cloned into the suicide plasmid pME3087, giving pME3828.This plasmid, carrying qscR:: ${Omega}$ -Sp/Sm, was mobilized into PAO1and chromosomally integrated, with selection for tetracyclineresistance. The chromosomal insertion in a tetracycline-sensitive,spectinomycin-resistant clone was verified by Southern blotting.

Construction of chromosomal insertions in the Tn7 attachment site. Complementation by a single copy of the rpoN⁺ gene was carriedout with a Tn7-based system developed for gram-negative bacteria(3, 22). Chromosomal insertion of the mini-Tn7 construct pME3338carrying rpoN⁺ (Fig. 1) was obtained via a triparental matingbetween the recipient PAO6358 (grown overnight at 43°C),E. coli SM10/ ${lambda}$ pir carrying the pUX-BF13 helper plasmid, and E.coli S17-1/pME3338, with selection for gentamicin and chloramphenicolresistance. The rpoN⁺ insertion in the resulting strain, PAO6360,was verified by the loss of the auxotrophy for glutamine andby Southern blotting analysis.

Chromosomal vfr'-'lacZ strains were constructed as follows.First, the vfr gene was amplified by PCR from chromosomal DNAof PAO1 with the XhoI (italic)-tagged primer pVFR1 (5'-CATCCTCGAGGAAGGCTTCGC-3')and the EcoRI (italic)-tagged primer pVFR2 (5'-GGAATTCATGGGTGCTGTTCA-3').The resulting 1.15-kb PCR fragment was cleaved with EcoRI andXhoI and inserted into pBluescriptII-SK to give pME6157. Theblunted 0.6-kb XhoI-PvuI fragment of pME6157 was fused to 'lacZin the SmaI site of pNM482, and the resulting vfr'-'lacZ fusionon a 3.7-kb EcoRI-DraI fragment was ligated to a 0.35-kb SphI(T4 DNA polymerase treated)-EcoRI fragment carrying the transcriptionstop signal of the ${Omega}$ -Sp/Sm cassette, and cloned into the bluntedHindIII site of the Tn7 delivery vector pME6313. The transcriptionstop signal upstream of the vfr'-'lacZ fusion in this construct,named pME6165, prevents potential readthrough from the gentamicinresistance gene. The vfr'-'lacZ fusion was delivered to thechromosome of PAO1 and PAO6358 by triparental mating with E.coli S17-1/pME6165 and E. coli SM10 ${lambda}$ pir/pUX-BF13 as donors. Gentamicin-resistanttransconjugants were checked by Southern analysis.

A translational gacA'-'lacZ fusion excised from plasmid pME6118as a 4.5-kb EcoRI-DraI fragment was ligated to the SphI (T4DNA polymerase treated)-EcoRI fragment carrying the transcriptionstop signal of ${Omega}$ -Sp/Sm and inserted into the blunted HindIIIsite of pME6313. The mini-Tn7 gacA'-'lacZ construct of the resultingplasmid, pME6166, was delivered to the chromosome of PAO1 andPAO6358 as described above. The resulting strains, PAO6320 andPAO6361, were checked by Southern analysis.

Semiquantitative determination of autoinducer concentrations by thin-layer chromatography. P. aeruginosa strains were cultivated with shaking in 20 mlof NYB amended with 1 mM glutamine in 50-ml Erlenmeyer flasksat 37°C to obtain an optical density at 600 nm (OD₆₀₀) of0.6 or 1.5. Cells were removed by centrifugation, and the pHof supernatants was adjusted to 5.0 prior to extraction with3 volumes of dichloromethane in a separating funnel. Water wasremoved from the solvent phase with anhydrous Na₂SO₄, and dichloromethanewas evaporated with a rotary evaporator. The extracts were concentrated200-fold by dissolving them in aqueous 50% (vol/vol) acetonitrile.The presence of AHLs was tested by C₁₈ reverse-phase (Merck)thin-layer chromatography, developed by elution in methanol-water(60:40, vol/vol), and revealed by overlaying either Chromobacteriumviolaceum CV026 (36), for BHL, or Agrobacterium tumefaciensNTL4/pZLR4 (7), for OdDHL. The amounts of BHL and OdDHL wereestimated by comparison with standards, i.e., 4, 6, or 8 nmolof BHL and 50, 100, or 150 pmol of OdDHL.

ß-Galactosidase assay. P. aeruginosa strains were cultivated with shaking in 20 mlof NYB with 1 mM glutamine in 50-ml Erlenmeyer flasks at 37°C.ß-Galactosidase specific activities were determinedby the method of Miller (55).

RESULTS

Top

Results

Growth characteristics of P. aeruginosa rpoN mutants. As the rpoN gene is the first gene in a cluster of five (27,59), we constructed an in-frame deletion mutation in rpoN toavoid potential polar effects on the expression of the downstreamgenes. The resulting rpoN mutant, PAO6358, was complementedby a single rpoN⁺ copy inserted into the unique chromosomalTn7 attachment site in strain PAO6360 (Fig. 1). An rpoN:: ${Omega}$ -Kminsertion mutant, PAO6359, was also constructed (Fig. 1). BothrpoN-negative strains were auxotrophic for glutamine and unableto swim, swarm, and twitch, in agreement with previous studiesshowing that P. aeruginosa rpoN mutants are defective for flagellaand type IV pili (21, 25, 29, 40, 61). In rich medium (NYB)amended with 1 mM L-glutamine, both rpoN mutants had a longerdoubling time (about 70 min) than the wild-type PAO1 and thecomplemented mutant PAO6360 (about 40 min), indicating thatsubstrate utilization was somewhat impeded by the loss of RpoNfunction. It is known that the utilization of several aminoacids as C and N sources depends on RpoN (40, 61). However,both the wild type and the rpoN mutant reached stationary phaseat similar levels (2.5 x 10⁹ to 3 x 10⁹ cells/ml). Under anaerobicconditions in a GasPak jar, the rpoN mutant PAO6358 and thewild-type PAO1 grew similarly on nutrient agar amended with1 mM L-glutamine and either 100 mM KNO₃ or 5 mM KNO₂, suggestingthat RpoN is not essential for denitrification.

Effects of rpoN null mutations on AHL production in P. aeruginosa. To investigate whether RpoN influences the production of thequorum-sensing signals OdDHL and BHL, we quantified AHL levelsat cell densities corresponding to early exponential phase (OD₆₀₀= 0.6, i.e., about 6 x 10⁸ cells/ml) and to late exponentialphase (OD₆₀₀ = 1.5) (Table 2). The rpoN mutants PAO6358 andPAO6359 produced about two times more OdDHL and about five timesmore BHL than did the parent strain during both early and lateexponential phases of growth. As the rpoN deletion mutant wasphenotypically similar to the insertion mutant, only the formerwas complemented. In the complemented rpoN mutant, PAO6360,autoinducer levels were close to those of the wild type (Table2). A similar regulation phenomenon was also detected in overnightcultures of an rpoN mutant of strain PAK-N1 (PAK-SR rpoN:: ${Omega}$ -Tc[25]), which produced six to seven times more AHL than did theparental wild type, PAK-SR (data not shown). By comparison withstrain PAO1, strain PAK-SR yielded fivefold-lower autoinducerlevels (data not shown).

View this table:
[in this window]
[in a new window]
TABLE 2. Negative control of autoinducer production by RpoN in P. aeruginosa^a

Effects of rpoN null mutations on expression of lasRI and rhlRI quorum-sensing genes. To confirm the regulatory effects of RpoN on AHL synthesis,we studied the expression of the lasR, lasI, rhlR, and rhlIgenes with translational lacZ fusions in the wild-type PAO1,in both rpoN mutants PAO6358 and PAO6359, and in the complementedmutant PAO6360. Null mutation of the rpoN gene resulted in anapproximately threefold derepression of lasR on pME3858 andlasI on pME3853 at cell densities below an OD₆₀₀ of 1.0, butthese effects were reversed at high cell densities (Fig. 2Aand B). The complemented mutant PAO6360 showed lasR and lasIexpression similar to that in the parent (Fig. 2A and B). RpoNwas also found to control the second quorum-sensing circuitin that translational rhlR'-'lacZ and rhlI'-'lacZ fusions (onpME3851 and pME3846, respectively) were induced four- to fivefoldat low and high cell densities (Fig. 2C and D). We additionallytested transcriptional lasR-lacZ and rhlR-lacZ fusions in anrpoN and in a wild-type background. The derepressing effectof the rpoN mutation was similar to that found with the analogoustranslational fusions (data not shown). In conclusion, the ß-galactosidaseactivities of the lasRI and rhlRI fusion constructs closelyparalleled AHL levels (Table 2).

View larger version (18K):
[in this window]
[in a new window]
FIG. 2. RpoN control of lasR, lasI, rhlR, and rhlI expression. ß-Galactosidase activities resulting from the translational fusions lasR'-'lacZ carried by pME3858 (A), lasI'-'lacZ on pME3853 (B), rhlR'-'lacZ on pME3851 (C), and rhlI'-'lacZ on pME3846 (D) were determined in P. aeruginosa wild-type PAO1 ( ${square}$ ), in the rpoN deletion mutant PAO6358 ( ${diamond}$ ), in the rpoN insertion mutant PAO6359 ( ${circ}$ ), and in the rpoN deletion mutant complemented with monocopy rpoN⁺ PAO6360 ( ${triangleup}$ ), grown in NYB with 1 mM glutamine. Each point is the mean ± standard deviation for three cultures.

Mutation in rpoN leads to increased expression of exoproduct genes. Quorum sensing controls the production of numerous extracellularenzymes and metabolites in P. aeruginosa; we chose to measurethree of them, elastase, rhamnolipids, and HCN (67), by followingthe expression of some representative structural genes: lasB(for elastase), rhlA (the first of two genes for rhamnolipidsynthesis), and hcnA (the first of three genes encoding HCNsynthase). All three genes are most strongly induced by a synergisticaction of both the las and rhl quorum-sensing systems (47, 68).Translational lacZ fusions of the lasB and rhlA genes on plasmidspTS400 and pECP60, respectively, showed a two- to fourfold increaseof expression in the rpoN mutant PAO6358 compared to the levelsin the wild type and in the complemented rpoN mutant. An hcnA-lacZtranscriptional fusion on pME3850.1 gave a similar result (datanot shown). These experiments as well as the autoinducer determinations(Table 2) and the quorum-sensing expression studies (Fig. 2)were all conducted in the same growth medium, NYB amended with1 mM glutamine. Growth reached a plateau at an OD₆₀₀ of 2.5to 3.0 under these conditions. The results obtained with thelasB, rhlA, and hcnA fusions are consistent with an overallnegative effect of RpoN acting transiently on lasRI and constantlyon rhlRI expression.

Mechanisms of quorum-sensing control by RpoN. We considered four hypotheses, which are not mutually exclusive,to explain the effects of RpoN on quorum-sensing in strain PAO1.(i) RpoN could directly repress the lasR, lasI, rhlR, and rhlIgenes at their promoters, similar to the negative effect that ${sigma}$ ⁵⁴ can exert on alginate synthesis by binding directly to thealgD promoter in P. aeruginosa (5). (ii) RpoN could repressVfr, a positive regulator of lasR gene expression (1). (iii)RpoN could activate QscR, a negative effector of quorum sensing(8, 33). (iv) RpoN could repress GacA, a positive regulatorof the expression of the lasR, rhlR, and rhlI genes (48, 52).

(i) RpoN recognizes a TGGCAC-N₅-TTGCA consensus sequence inwhich the TGGC-N₉-GC motif is most strongly conserved (4, 53).We searched for the presence of such a motif in the lasR, lasI,rhlR, and rhlI promoter regions; note that the transcriptionstart sites have been determined experimentally for the lasR,lasI, and rhlI genes (1, 41, 48, 56). We did not find evidencefor a conserved RpoN motif in the lasR, lasI, and rhlR promoterregions at positions that would be compatible with a repressiveeffect of RpoN, nor did we detect such a motif in the promoterof the rsaL gene, which exerts negative control on the lasIgene (11). By contrast, in the rhlI promoter, a potential RpoNrecognition sequence, TGGCAG-N₅-CTGCC, was found at positions-43 to -31 relative to the +1 site. This sequence overlaps alasR (lux) box placed at -57 to -38. A 3-bp mutation replacingTGC at -33 to -31 with AAA was constructed in the rhlI promoterregion. This mutation leaves the lasR box intact but would beexpected to interfere with recognition of RpoN. However, theexpression of an rhlI'-'lacZ fusion was not influenced by thismutation in the wild type or the rpoN mutant PAO6358 (data notshown).

(ii) In order to determine if the influence of RpoN on quorumsensing was mediated by Vfr, we measured the expression of achromosomal translational vfr'-'lacZ fusion in a wild-type (PAO6304)and an rpoN mutant background (PAO6362). The expression of vfrwas low throughout growth and significantly reduced in the rpoNbackground at high cell densities (Fig. 3). Inspection of thevfr promoter (54), however, did not reveal the presence of anRpoN recognition sequence, suggesting that the positive effectof RpoN on vfr expression may be indirect. The consequence ofthis regulation will be considered in the Discussion.

View larger version (21K):
[in this window]
[in a new window]
FIG. 3. Influence of RpoN on vfr expression. ß-Galactosidase activities from the chromosomally located translational fusion vfr'-'lacZ were determined in the wild-type PAO6304 ( ${square}$ ) and in the rpoN mutant background strain PAO6362 ( ${diamond}$ ). Each result is the mean ± standard deviation for three cultures. Bacteria were grown in NYB with 1 mM glutamine.

(iii) A qscR:: ${Omega}$ -Sp/Sm insertion mutant, PAO6366, was constructedand tested for expression of the lasR, lasI, rhlR, and rhlIgenes with the lacZ fusion constructs pME3858, pME3853, pME3851,and pME3846, respectively. However, each fusion gave a similarexpression profile in the wild-type PAO1 and in the qscR mutantPAO6366 (data not shown). A translational qscR'-'lacZ fusion(constructed as detailed in Materials and Methods) was expressedat a low level of 3 to 4 Miller units in the wild-type PAO1as well as in the rpoN mutant PAO6358, at an OD₆₀₀ of 2.0. Theseresults suggest that RpoN does not modulate the quorum-sensingmachinery via QscR.

(iv) In order to check if the influence of RpoN on quorum sensingwas a result of RpoN-mediated control of gacA expression, weassayed a chromosomal translational gacA'-'lacZ fusion in thewild type (PAO6320) and the rpoN deletion mutant (PAO6361).A marked negative effect of RpoN on gacA expression occurredthroughout growth (Fig. 4). In the wild type as well as in thecomplemented rpoN mutant, gacA expression was about three timeslower than in the rpoN mutant (Fig. 4). In agreement with previousdata (52), the expression levels of the lasB'-'lacZ, rhlA'-'lacZ,and hcnA-lacZ fusions were very low in a gacA mutant comparedwith those in the wild type and the rpoN mutant (Fig. 5A, B,and C).

View larger version (25K):
[in this window]
[in a new window]
FIG. 4. Influence of RpoN on gacA expression. ß-Galactosidase activities from the chromosomally located translational fusion gacA'-'lacZ were determined in the wild-type PAO6320 ( ${square}$ ), in the rpoN mutant PAO6361 ( ${diamond}$ ), and in the rpoN mutant PAO6361 complemented with pME3829 ( ${triangleup}$ ). Each result is the mean ± standard deviation for three cultures. Bacteria were grown in NYB with 1 mM glutamine.

View larger version (18K):
[in this window]
[in a new window]
FIG. 5. Influence of RpoN and GacA on lasB, rhlA, and hcnA expression. ß-Galactosidase activities from the translational fusions lasB'-'lacZ on pTS400 (A) and rhlA'-'lacZ on pECP60 (B) and from the transcriptional fusion hcnA-lacZ on pME3850.1 (C) were determined in the wild-type PAO1 ( ${square}$ ), the rpoN mutant PAO6358 ( ${diamond}$ ), the rpoN gacA double mutant PAO6363 ( ${circ}$ ), and the gacA mutant PAO6281 ( ${triangleup}$ ). Each result is the mean ± standard deviation for three cultures. Bacteria were grown in NYB with 1 mM glutamine.

If the repressive effects of RpoN on the quorum-sensing machinerywere essentially a consequence of repression of GacA, then wemight expect that a gacA rpoN double mutant would not differsubstantially from a gacA mutant in terms of quorum-sensing-dependentexpression. In the case of the lasB'-'lacZ fusion, this wasindeed observed (Fig. 5A). However, in the case of the rhlA'-'lacZand hcnA-lacZ fusions, the gacA rpoN mutant gave intermediateexpression levels (Fig. 5B and C), indicating that a simplelinear model (RpoN —| GacA $->$ LasRI/RhlRI $->$ RhlA/HcnA) wouldincompletely describe the situation and that RpoN may act onthe expression of the rhlA and hcnA genes via GacA-independentpathways. The hcnA-lacZ fusion of pME3850.1 used is controlledtightly and probably exclusively by LasR and RhlR (47). It islikely that RpoN also acts on quorum sensing via regulatorsother than GacA, but these remain to be identified.

DISCUSSION

Top

Discussion

In P. aeruginosa, the alternative sigma factor RpoN has severalroles. As in enteric bacteria, it is a vital component of nitrogenassimilation. rpoN mutants of strains PAO1 and PAK but not strainPA14 are auxotrophic for glutamine (21, 40, 61). In strain PAO,RpoN cooperates with the two-component systems NtrB-NtrC andCbrA-CbrB in the utilization of various carbon and nitrogensources, e.g., arginine, histidine, and polyamines (40, 61).Moreover, PAO and PAK mutants that are defective for rpoN donot produce flagella and pili and therefore do not show swimmingand twitching mobility (25, 61; this study). For the transcriptionof flagellar and pilus genes, ${sigma}$ ⁵⁴ needs the transcriptional regulatorsFleQ, FleR, and PilR (2, 35). Remarkably, the genomic sequenceof strain PAO predicts >20 transcriptional regulators activating ${sigma}$ ⁵⁴; however, the function of most of these has not yet beendiscovered (59). Whether any of these is involved in the modulationof quorum-sensing activity reported here is also unknown. Althougha direct repressive effect of RpoN on some quorum-sensing genepromoter has not been rigorously excluded, it seems more likelythat RpoN, together with some transcriptional regulator(s),would transcribe one or several regulatory elements that controlquorum sensing.

Certain transcriptional regulators that interact with ${sigma}$ ⁵⁴, e.g.,NtrC and CbrB, are response regulators of two-component systemsin P. aeruginosa. Since the cognate sensor kinases NtrB andCbrA are involved in the utilization of various N sources (40),we considered the possibility that the production of AHLs byP. aeruginosa might be influenced by the N source. However,we did not observe any significant differences in AHL levelsproduced aerobically by P. aeruginosa PAO1 cells when the growthmedium contained either ammonium (a good N source) or nitrate(a poor N source) (data not shown).

One role of the postulated RpoN-dependent control element(s)is to downregulate the expression of the global regulator GacA.This downregulation occurs at both low and high cell densities(Fig. 4). By contrast, RpoN has a positive effect on the expressionof another quorum-sensing regulator, Vfr, especially duringlate growth phases (Fig. 3). Vfr positively controls the expressionof the lasR regulator (1). It therefore appears that the derepressingeffect of an rpoN mutation on lasR expression (Fig. 2A and B)could be the result of the enhanced expression of GacA duringearly growth phases and at low cell densities. During laterstages of growth and at higher cell densities, the downregulationof vfr may compensate for the upregulation of gacA in an rpoNmutant. This would explain the observation that lasR and, indirectly,lasI are not overexpressed in an rpoN mutant background at anOD₆₀₀ of $>=$ 1.5 (Fig. 2A and B).

The signal transduction pathways by which GacA acts on the expressionof target genes, many of which are involved in the synthesisof exoproducts and biofilm formation (43), are incompletelyunderstood at present (19). In P. aeruginosa, a detailed analysisof hcnABC expression revealed the existence of two GacA-dependentpathways (47, 48, 52). In the first pathway, GacA exerts a positiveeffect on the expression of the lasR and rhlRI genes and, asa consequence, also on the transcription of LasR- and RhlR-dependentgenes, including hcnABC. In the second, AHL-independent pathway,GacA exerts a positive effect on target gene expression at aposttranscriptional level; in the case of the hcnABC genes,this effect requires a sequence surrounding the hcnA ribosome-bindingsite (48) and the RNA-binding protein RsmA (49). In both pathways,the DNA sequences directly recognized by the GacA protein, presumablyin its phosphorylated form, have remained elusive. We consideredthe possibility that RpoN might regulate the expression of RsmA.However, we found no evidence for such an effect in experimentswith an rsmA'-'lacZ fusion and Western blotting (our unpublisheddata).

It is interesting that another alternative sigma factor, thestress and stationary-phase sigma factor RpoS ( ${sigma}$ ³⁸), is alsoinvolved in quorum-sensing regulation in P. aeruginosa. In anrpoS null mutant, BHL levels are elevated throughout growth,essentially due to derepression of rhlI expression, by comparisonwith the wild-type strain PAO1. As a consequence, expressionof exoproduct genes, e.g., hcnB, is increased in an rpoS mutant(69). There is also evidence that the rhlRI system can positivelycontrol the RpoS level (31). Environmental conditions greatlyinfluence the relative amounts of sigma factors in bacterialcells (24). These considerations led us to propose an empiricalmodel in which GacA and some sigma factors (such as RpoN andRpoS) globally exert opposite effects on the quorum-sensingmachinery in P. aeruginosa.

Interactions between quorum sensing and RpoN have also beenproposed in E. coli. Addition of a signal molecule termed autoinducer2 (a furanone compound) to E. coli cultures induces, among alarge number of genes, the ybhH gene (corresponding to the genePA4463 lying downstream of rpoN; Fig. 1) as well as the ygeVgene, encoding a ${sigma}$ ⁵⁴-dependent regulator homologous to LuxO,which is a component of the quorum-sensing cascade regulatingbioluminescence in Vibrio harveyi (12, 34).

Considering the fact that inactivation of rpoN leads to theloss of two important adherence factors (pili and flagella)in P. aeruginosa on the one hand (21, 25, 61) and to an overexpressionof several quorum-sensing-regulated virulence genes (lasB, rhlAB,and hcnABC) on the other hand, it would have been difficultto predict the virulence properties of a P. aeruginosa rpoNmutant in an animal model. The work of Hendrickson et al. (21)showed that an rpoN mutant of strain PA14 can manifest pathogenicitydifferently depending on the host. For nematodes and burnt mice,the rpoN mutant was less virulent than the wild type, whereasboth strains were equally able to kill wax moth larvae (21).

RpoN is required in P. aeruginosa strain CHA for type III secretionof exotoxin S and exotoxin T, and thus RpoN makes a contributionto cytotoxicity in this strain (10). Moreover, since many pathogenicitymodels use relatively high infectious doses, the adherence propertiesof P. aeruginosa in these models may be less important thanin most clinical situations. To some extent, the decreased virulenceof P. aeruginosa rpoN mutants in some models might then be aconsequence of a decreased ability to utilize a large numberof organic substrates, not just N sources (40), and to producesecreted toxins (10).

ACKNOWLEDGMENTS

We thank Christoph Keel for discussion, Yoshifumi Itoh for supplyingplasmid pRP1-rpoN and strain PAO4460, and Nazife Beqa for technicalhelp. We also thank Barbara Iglewski for providing plasmidspTS400 and pECP60, Dan Hassett for informing us of his workon RpoN prior to publication, and Glaxo-SmithKline for generouslysupplying carbenicillin.

This study was supported by the Swiss National Foundation forScientific Research (project 31-56608.99) and the program GénieBiomédical.

FOOTNOTES

* Corresponding author. Mailing address: Institut de Microbiologie Fondamentale, Université de Lausanne, CH-1015 Lausanne, Switzerland. Phone: 41 21 692 56 31. Fax: 41 21 692 56 35. E-mail: Dieter.Haas{at}imf.unil.ch.

Present address: Department of Genetics and Developmental Biology,Center for Microbial Pathogenesis, University of ConnecticutHealth Center, Farmington, CT 06030-3710.

REFERENCES

Top