Roles of Pseudomonas aeruginosa las and rhl Quorum-Sensing ... · Pseudomonas aeruginosa is an opportunistic human patho-gen that infects immunocompromised individuals and people

JOURNAL OF BACTERIOLOGY,0021-9193/97/$04.0010

Sept. 1997, p. 5756–5767 Vol. 179, No. 18

Copyright © 1997, American Society for Microbiology

Roles of Pseudomonas aeruginosa las and rhl Quorum-Sensing Systemsin Control of Elastase and Rhamnolipid Biosynthesis Genes

JAMES P. PEARSON, EVERETT C. PESCI, AND BARBARA H. IGLEWSKI*

Department of Microbiology and Immunology, University of RochesterSchool of Medicine and Dentistry, Rochester, New York 14642

Received 31 January 1997/Accepted 14 July 1997

Two quorum-sensing systems (las and rhl) regulate virulence gene expression in Pseudomonas aeruginosa.The las system consists of a transcriptional activator, LasR, and LasI, which directs the synthesis of theautoinducer N-(3-oxododecanoyl) homoserine lactone (PAI-1). Induction of lasB (encoding elastase) and othervirulence genes requires LasR and PAI-1. The rhl system consists of a putative transcriptional activator, RhlR,and RhlI, which directs the synthesis of N-butyryl homoserine lactone (PAI-2). Rhamnolipid production inP. aeruginosa has been reported to require both the rhl system and rhlAB (encoding a rhamnosyltransferase).Here we report the generation of a DlasI mutant and both DlasI DrhlI and DlasR rhlR::Tn501 double mutantsof strain PAO1. Rhamnolipid production and elastolysis were reduced in the DlasI single mutant and abolishedin the double-mutant strains. rhlAB mRNA was not detected in these strains at mid-logarithmic phase but wasabundant in the parental strain. Further RNA analysis of the wild-type strain revealed that rhlAB is organizedas an operon. The rhlAB transcriptional start was mapped, and putative s54 and s70 promoters were identifiedupstream. To define components required for rhlAB expression, we developed a bioassay in Escherichia coli anddemonstrated that PAI-2 and RhlR are required and sufficient for expression of rhlA. To characterize theputative interaction between PAI-2 and RhlR, we demonstrated that [3H]PAI-2 binds to E. coli cells expressingRhlR and not to those expressing LasR. Finally, the specificity of the las and rhl systems was examined in E. colibioassays. The las system was capable of mildly activating rhlA, and similarly, the rhl system partly activatedlasB. However, these effects were much less than the activation of rhlA by the rhl system and lasB by the lassystem. The results presented here further characterize the roles of the rhl and las quorum-sensing systems invirulence gene expression.

Pseudomonas aeruginosa is an opportunistic human patho-gen that infects immunocompromised individuals and peoplewith cystic fibrosis (49, 60). Multiple cell-associated factorssuch as alginate, pili, and lipopolysaccharide are important inP. aeruginosa virulence. Additionally, many P. aeruginosa viru-lence factors, including toxins (exotoxin A and exoenzyme S),proteases (elastase, LasA protease, and alkaline protease), andhemolysins (phospholipase and rhamnolipid), are releasedfrom the bacterial cells. These factors have been found tocontribute to the virulence of P. aeruginosa in animal models(3, 20, 21, 32), in vitro studies (2, 9, 26, 36), and clinical studies(23, 69). Expression of many P. aeruginosa virulence productsis dependent on a particular environmental stimulus such asiron, nitrogen availability, temperature, or osmolarity, but ingeneral, virulence factor expression does not occur until highcell density is achieved (2, 26). A mechanism through which P.aeruginosa regulates virulence gene expression in a cell density-dependent manner is known as quorum sensing.

Quorum sensing involves emission of an N-acyl homoserinelactone signal (an autoinducer [AI]) generated by a LuxI-typeAI synthase. After reaching a threshold concentration, AI ac-tivates a LuxR-type transcriptional activator that induces spe-cific genes (for a review, see reference 13). P. aeruginosa con-tains two separate quorum-sensing systems, the las and rhlsystems. In las quorum sensing, the AI synthase, LasI, directsthe synthesis of N-(3-oxododecanoyl) homoserine lactone(PAI-1) (46), which triggers the lasR-encoded transcriptional

activator, LasR (15), to induce virulence genes such as lasB,lasA, apr, and toxA (14, 43, 65, 66). The las system also auto-regulates lasI, leading to the production of more PAI-1 (53).The importance of the las quorum-sensing system in P. aerugi-nosa virulence was recently demonstrated in vivo, where aDlasR mutant was shown to be avirulent in a mouse model ofpneumonia (64).

The second P. aeruginosa quorum-sensing system, the rhlsystem, regulates production of rhamnolipid which has bothhemolytic and biosurfactant properties (26, 29). Rhamnolipidis synthesized by the rhlAB-encoded rhamnosyltransferase(39). The four rhl genes required for rhamnolipid productionare organized as a regulon with all genes transcribed in thesame direction, 59-rhlABRI-39 (38–40). Components of the rhlquorum-sensing system include the rhlR-encoded putativetranscriptional activator, RhlR (38), and rhlI-encoded putativeAI synthase, RhlI (40). A second P. aeruginosa autoinducer(PAI-2), N-butyryl homoserine lactone (47), was shown to re-store rhamnolipid production in a P. aeruginosa rhlI mutant(40) and was later shown to require rhlI for its synthesis (68).Thus, it seemed likely that PAI-2 was the AI of rhamnolipidbiosynthesis genes. Recent results showed that PAI-2 andRhlR enhanced expression of rhlI and rpoS in Escherichia coli(31). Clearly, rhlR and rhlI are necessary for rhamnolipid bio-synthesis in P. aeruginosa (38–40). Multicopy plasmids carryingrhlR and rhlI have shown a positive regulatory effect on rhlAexpression in wild-type P. aeruginosa (40). This finding sug-gested rhlR and rhlI are components required for autoinduc-tion of the rhamnolipid biosynthesis genes rhlA and rhlB, butthis has not been directly demonstrated.

In this report, we describe the generation of a DlasI mutant,a DlasI DrhlI double mutant, and a DlasR rhlR::Tn501 double

* Corresponding author. Mailing address: Department of Microbi-ology and Immunology, University of Rochester, 601 Elmwood Ave.,Box 672, Rochester, NY 14642. Phone: (716) 275-3402. Fax: (716)473-9573. E-mail: [email protected].

5756

on August 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

mutant of P. aeruginosa PAO1. Northern blot analysis of thesemutants and the parent strain showed that the rhamnosyltrans-ferase genes (rhlAB) are transcribed as a bicistronic RNA andthat rhlR is transcribed independently of rhlAB. In addition, wemapped the transcriptional start site of rhlAB. By constructingplasmids containing rhlA9 or lasB9 fusions with lacZ and a rhlRor lasR under control of the tac promoter (tacp), we develop anrhl AI bioassay in E. coli and demonstrate the specificity of thelas and rhl systems in control of the elastase gene and rham-nolipid biosynthesis genes. Finally, using radiolabeled PAI-2,we provide evidence that PAI-2 binds to RhlR.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media. Bacterial strains and plasmids used inthis study are listed in Table 1. P. aeruginosa stock cultures were maintained asfollows. Late-log-phase cultures grown at 37°C with shaking in 50 ml of PTSBmedium (5% peptone, 0.25% tryptic soy broth [pH 7] [41]) plus appropriateantibiotics were centrifuged at 6,000 3 g for 10 min. Cells were resuspended in1.5 ml of 10% (wt/vol) skim milk (Becton Dickinson Corp., Cockeysville, Md.)and stored at 270°C. Freshly isolated colonies from stock cultures were used foreach P. aeruginosa experiment. P. aeruginosa was also grown in LB medium (51),Jensen’s defined medium (24) containing 0.4% glycerol instead of 70 mM glu-cose (47), or the defined, nitrogen-limited Guerra-Santos (GS) medium (18)modified by the presence of 2.0% glycerol instead of 0.1 M glucose (39). Pseudo-monas isolation agar (Difco Corp., Detroit, Mich.) was used when needed. E. coliwas stored by using standard procedures (51) and routinely grown in LB medium.E. coli was grown in supplemented A medium (46) for b-galactosidase determi-nations. Bacteria were grown at 37°C with shaking unless otherwise indicated.Antibiotics were used in the following concentrations when needed: for E. coli,ampicillin at 100 mg/ml, tetracycline at 20 mg/ml, kanamycin at 50 mg/ml, andchloramphenicol at 30 mg/ml; for P. aeruginosa, carbenicillin at 200 mg/ml, tet-racycline at 50 mg/ml (300 mg/ml in Pseudomonas isolation agar), and mercuricchloride at 15 mg/ml in solid medium or 7.5 mg/ml in liquid medium. Antibioticsand other chemicals were purchased from Sigma (St. Louis, Mo.) or AldrichChemical Corp. (Milwaukee, Wis.). Complex medium components were pur-chased from Difco. PAI-1 and PAI-2 were synthesized previously (45, 47).

DNA techniques. Plasmid DNA was purified and manipulated by using stan-dard techniques (51). E. coli DH5a was used as a host strain for most molecularcloning. Restriction endonucleases, DNA-modifying enzymes, and T4 DNA li-gase were purchased from Gibco/BRL (Gaithersburg, Md.) or New EnglandBiolabs (Beverly, Mass.). Plasmids were transformed (51) into E. coli and wereelectroporated (59) into P. aeruginosa as described previously. DNA fragmentswere excised from agarose gels and purified by using the GeneClean protocol(Bio101 Corp., La Jolla, Calif.). Oligonucleotides were synthesized at the CoreNucleic Acid Laboratory at the University of Rochester. DNA sequencing wasperformed with a Sequenase kit (U.S. Biochemical Corp., Cleveland, Ohio) and[a-35S]dATP (1,000 to 1,500 Ci/mmol; Dupont NEN, Boston, Mass.).

Construction of P. aeruginosa lasI and lasR mutant strains. To generate aP. aeruginosa lasI mutant, plasmid pJPP4 was constructed. A 1.8-kb VspI-EcoRIfragment from pKDT1.7 (carrying a 1.72-kb SacII-EcoRI fragment of strainPAO1 DNA which spans bp 21510 to 1212 relative to the lasI translationalstart) was ligated to VspI-EcoRI-digested pBR322, forming pJPP1.7. The EcoRIsite on pJPP1.7 was filled in with the Klenow fragment and religated to formpJPP1.7E. Next, the 3.0-kb KpnI-StyI fragment of pJPP1.7E, which contains the1.7-kb SacII-EcoRI fragment of strain PAO1 and 1.3-kb Tcr gene, was ligated toKpnI-StyI-digested pJMC31 to create pJPP3. This intermediate plasmid containsthe Tcr marker inserted in place of a 549-bp deletion from the EcoRI site withinlasI to the StyI site 116 bp downstream of the lasI translational stop, flanked by1.3-kb of strain PAO1 DNA (extending to the next EcoRI site downstream oflasI). Plasmid pJPP3 was digested with KpnI and EcoRI to release a 4.4-kbfragment (carrying the DlasI::Tcr cassette and the 1.7 kb upstream and 1.3 kbdownstream of P. aeruginosa DNA relative to the lasI deletion). This fragmentwas cloned into KpnI- and EcoRI-digested pGP704, yielding pJPP4. PlasmidpJPP4 is a suicide vector because it carries the lpir-dependent R6K ori, whichwill not replicate in bacteria that lack lpir (such as P. aeruginosa) (35). Next,pJPP4 was mobilized by conjugation from E. coli SY327 lpir into P. aeruginosaPAO1 (wild type) and PDO100 (DrhlI) with the help of E. coli HB101(pRK2013), using triparental matings as described by Deretic et al. (7). Tcr

exconjugants of strains PAO1 and PDO100 were selected on Pseudomonasisolation agar and screened by colony blot hybridization using nylon membranes(Hybond-N; Amersham Corp., Arlington Heights, Ill.) (67). Blots were hybrid-ized to a 32P-labeled pGP704 probe to screen for the absence of the vector DNA.After colonies that did not hybridize to the pGP704 probe were identified, theprobe was stripped from the membranes, which were probed with a 32P-labeled549-bp fragment of strain PAO1 DNA from the EcoRI site in lasI to the StyI sitedownstream. This probe was used to identify exconjugants which lacked thedeleted portion of lasI and thus were potentially the result of a double crossover.Approximately 20% (n 5 250 Tcr exconjugants of strains PAO1 and PDO100,

respectively) did not hybridize to either probe, indicating these exconjugantscontained neither vector DNA nor the deleted portion of lasI. DNA from four ofthese Tcr exconjugants from strains PAO1 and PDO100 was analyzed by South-ern blotting (described below).

In addition to the above-described mutants, we constructed a P. aeruginosalasR rhlR double mutant. Plasmid pMG319 has been used in the construction ofstrain PAO-R1, the P. aeruginosa DlasR mutant (15). Here, pMG319 was mobi-lized from E. coli HB101 into P. aeruginosa PDO111 (rhlR::Tn501) by conjuga-tion with the help of E. coli HB101(pRK2013). Tcr exconjugants were selectedand screened for the absence of the vector sequence and the deleted portion oflasR as described above. DNA from four of these exconjugants was analyzed bySouthern blotting.

Southern blot analysis. Chromosomal DNA was prepared from overnightcultures of P. aeruginosa exconjugants and parental strains grown in LB mediumplus appropriate antibiotics with shaking as described by Silhavy et al. (56). Afterdigestion with restriction endonucleases, DNA fragments were separated on0.7% agarose gels, transferred to nylon membranes, and hybridized to 32P-labeled DNA probes (as described by Amersham). Southern analyses of Tcr

exconjugants derived from strains PAO1, PDO100, and PDO111 are reported inResults.

Rhamnolipid assay. Rhamnolipid in P. aeruginosa culture fluids was detectedas previously described (29), with modifications. Briefly, cells from mid-log-phasecultures grown in PTSB were washed and resuspended in modified GS mediumto an optical density at 660 nm (OD660) of 0.15 and incubated at 37°C withshaking. Cultures reached stationary phase (OD660 of 3.7 6 0.7) after approxi-mately 40 h but were incubated for a total of 82 to 88 h. At this time, cultureswere centrifuged 16,000 3 g for 5 min, and supernatants were passed through0.22-mm-pore-size filters. Filtrates were extracted three times with 2 volumes ofdiethyl ether. The pooled ether extracts were extracted once with 20 mM HCl,and the ether phase was evaporated to dryness. The residue was dissolved inwater. Rhamnose content in each sample was determined by duplicate oricinolassays compared to rhamnose standards (29, 37). Rhamnolipid was determinedby the relation that 1.0 mg of rhamnose corresponds to 2.5 mg of rhamnolipid(37).

Elastolysis assay. Elastolytic activity in P. aeruginosa culture fluids was deter-mined by elastin Congo red (ECR) assays (2), with modifications. Briefly, cellsfrom mid-log-phase cultures in PTSB were washed and resuspended in PTSB toan OD660 of 0.05. After 21 h at 37°C with shaking, culture supernatants werefiltered (0.45-mm pore-size filter) and stored at 270°C. Triplicate 50-ml samplesof culture filtrates were added to tubes containing 20 mg of ECR (ElastinProducts Corp. Owensville, Mo.) and 1 ml of buffer (0.1 M Tris [pH 7.2], 1 mMCaCl2). Tubes were incubated 18 h at 37°C with rotation and then were placedon ice after 0.1 ml of 0.12 M EDTA was added. Insoluble ECR was removed bycentrifugation, and the OD495 was measured. Absorption due to pigments pro-duced by P. aeruginosa was corrected for by subtracting the OD495 of each samplethat had been incubated in the absence of ECR.

Complementation of mutations in P. aeruginosa. The lasI deletion in P. aerugi-nosa PAO-JP1 was complemented by lacp-lasI on pLASI-2. To complement boththe lasI and rhlI mutations in strain PAO-JP2, we constructed plasmid pJPP42,carrying both lacp-rhlI and lacp-lasI (each gene under control of a separate lacp).To do this, rhlI was obtained on a 950-bp fragment of strain PAO1 DNA frompMB41 (from the BglII site 150 bp upstream of the rhlI translational start to theSphI site approximately 200 bp downstream of the translational stop). Protrudingends were removed with the use of mung bean nuclease, and the fragment wascloned into the SmaI site on pBluescript II SK1, forming pJPP41 with rhlIdownstream of the lac promoter. Plasmid pJPP41 was partially digested withPvuII, releasing lacp-rhlI as a 1.5-kb fragment which was ligated into the mungbean nuclease-digested SphI site of pLASI-2, forming pJPP42. To complementthe rhlI mutation in strain PDO100, plasmid pJPP45 was constructed by firstligating the 1.5-kb PvuII fragment from pJPP41 to the EcoRV site ofpACYC184, forming pJPP44. A 1.7-kb fragment containing lacp-rhlI was re-leased from pJPP44 by HindIII and SphI and ligated to these sites in pSW200 toform pJPP45.

PAI-1 analysis. P. aeruginosa strains were grown in modified Jensen mediumto early stationary phase (OD660 of 5), and culture supernatants were extractedwith ethyl acetate (47). Extracts were tested in the PAI-1 bioassay in E. coli (46).

RNA preparation and Northern blot analysis. RNA was isolated by a previ-ously described method (8) from mid-log-phase cultures of P. aeruginosa grownin either LB or GS medium (see below). GS medium was developed for P. aerugi-nosa rhamnolipid production, and the expression of a rhlA9-lacZ fusion con-tained in P. aeruginosa grown in this medium is 20-fold higher than when cells arecultured in LB (39). Cells from primary cultures of P. aeruginosa that had beengrown to mid-log phase in PTSB medium were washed and diluted in GSmedium to an OD660 of 0.15, and these secondary cultures were grown tomid-logarithmic phase (OD660 of 0.9). RNA was obtained from cell lysates bycentrifugation through a 5.7 M CsCl cushion (8), extracted with chloroform, andstored in diethylpyrocarbonate-treated water at 270°C. Northern blot analysis ofRNA was performed as described previously (1). The following restriction frag-ments were used as probes in Northern blot analysis: a 560-bp BamHI internalfragment of rhlA obtained from pRL1500, a 610-bp EcoRV-XhoI internal frag-ment of rhlB obtained from pRL2000, a 470-bp BamHI-SacII internal fragmentof rhlR obtained from pSK825r, and a 280-bp Sau3A-BamHI internal fragment

VOL. 179, 1997 P. AERUGINOSA las AND rhl QUORUM SENSING 5757


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

TABLE 1. Bacterial strains and plasmids used

Strain orplasmid Relevant genotype and/or phenotype Source or

reference

P. aeruginosaPAO1 Wild-type prototroph This laboratoryPAO-R1 DlasR derivative of PAO1, Tcr 15PDO100 DrhlI::Tn501 derivative of PAO1, Hgr 4PDO111 rhlR::Tn501 derivative of PAO1, Hgr 4PAO-JP1 DlasI derivative of PAO1, Tcr This studyPAO-JP2 DlasI derivative of PDO100, Hgr Tcr This studyPAO-JP3 DlasR derivative of PDO111, Hgr Tcr This studyPAK Wild-type prototroph 22PAK-N1 DrpoN derivative of PAK, Tcr 22PG201 Wild-type prototroph 18

E. coliHB101 F2 mcrB mrr hsdS20 recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20 (Smr) supE44 GibcoDH5a F2 f80dlacZDM15 D(lacZYA-argF)U169 endA1 recA1 hsdR17 deoR gyrA96 thi-1 relA1 supE44 GibcoSY327 lpir (lpir) D(lac pro) argE(Am) rif nalA recA56 35

PlasmidsCloning vectors

pBR322 ori (ColE1) Tcr Apr GibcopBluescript II SK1 ori (ColE1) Apr StratagenepUC18, pUC19 ori (ColE1) Apr GibcopLG339 ori (pSC101) Kmr Tcr 63pACYC184 ori (P15A) Tcr Cmr New England

Biolabs

Plasmids used in construc-tion of suicide vector

pRO1614 pBR322 carrying ori (P. aeruginosa) on 1.8-kb PstI fragment 42pSW200 pUC18 with 1.8 kb PstI fragment of pRO1614 containing ori (P. aeruginosa) 15pMJG1.7 1.72-kb SacII-EcoRI fragment of strain PAO1 (lasR lasI9) on pSW200 15pKDT1.7 pUC19 with 1.74-kb BamHI-EcoRI fragment of pMJG1.7 This studypJPP1.7 pBR322 containing 1.72-kb SacII-EcoRI fragment of strain PAO1 from pKDT1.7, Tcr This studypJPP1.7E pJPP1.7 with filled-in EcoRI site This studypJMC31 1.8-kb EcoRI fragment of strain PAO1 with 9lasI on pBluescript II KS 43pJPP3 pJMC31 with 3.0-kb KpnI-StyI fragment of pJPP1.7E (DlasI), Tcr Apr This studypGP704 oriR6K mobRP4 Apr 35

Plasmids used in genereplacement

pJPP4 pGP704 with 4.2-kb KpnI-EcoRI fragment of pJPP3 DlasI, Tcr Apr This studypRK2013 ori (ColE1) tra1 (RK2) Kmr helper plasmid for conjugation 12pMG319 Vector for gene replacement, DlasR Tcr 15

Plasmids used as sourcesof P. aeruginosa DNA

pPAO-2 pUC19 with 0.28-kb Sau3A fragment of pilA from strain PAO1 52pRL1500 pUCP19 containing rhlA on 1.5-kb BglI-NcoI fragment from strain PG201 39pRL2000 pUCP19 containing rhlB on 2.0-kb BamHI fragment from strain PG201 39pSK825, pSK825r pBluescript II SK1 with rhlR on 825 bp BglI-BglII fragment from strain PG201 in both orientations 38

Plasmids used in comple-mentation

pMB41 2.2-kb BamHI fragment of PAO1 containing 9rhlR rhlI on pLG339, Kmr D. OhmanpJPP41 pBluescript II SK1 containing rhlI from pMB41 under lacp This studypLASI-2 lacp-lasI on pSW200 43pJPP42 pLASI-2 containing lacp-rhlI from pJPP41 This studypJPP44 pACYC184 with lacp-rhlI from pJPP41, Cmr This studypJPP45 pSW200 with 1.7-kb SphI-HindIII fragment from pJPP44 containing lacp-rhlI This study

Plasmids used in proteinexpression

pEX1 pKK223-3 derivative (Pharmacia Corp., Uppsala, Sweden) containing lacIq in PvuII site ori(ColE1), Apr

44

pEX1.8 pEX1 carrying ori (P. aeruginosa) as 1.8-kb PstI fragment from pRO1614 in StyI site This studypJPP8 pEX1.8 containing rhlR of strain PG201 under tacp This studypGroESL ori (P15A) lacp-groESL Cmr 16pKDT37 pEX1 containing lasR of strain PAO1 under tacp 45pECP59 pEX1.8 with tacp-lasR from pKDT37 This study

Continued on following page

5758 PEARSON ET AL. J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

of pilA obtained from pPAO-2. RNA molecular size standards were purchasedfrom Gibco/BRL.

Primer extension analysis. Primer extension analysis was performed as de-scribed previously (1). Two oligonucleotides, P1 (59-GAACGGCAGACAAGTAACTCAG-39) and P2 (59-AGACTTTCGCGCCGCATTTCAC-39), comple-mentary to nucleotides (nt) 289 to 2110 and 117 to 25 relative to the rhlA startcodon, respectively, were 59-end labeled with the use of [g-32P]dATP (3,000Ci/mmol; Dupont NEN) and T4 polynucleotide kinase (Gibco Corp.). 32P-la-beled primers were annealed to P. aeruginosa RNA, and avian myeloblastosisvirus reverse transcriptase (Promega Corp., Madison, Wis.) was used to extendeach primer. The resulting cDNAs were resolved on 8 M urea–8% polyacryl-amide gels adjacent to a DNA sequence obtained by using the same oligonucle-otide as used in each primer extension.

Overexpression of RhlR. We modified a tacp expression vector to allow ex-pression in E. coli or P. aeruginosa. Ends of a 1.8-kb PstI fragment of pRO1614(containing a P. aeruginosa origin of replication) and StyI-digested pEX1 (car-rying tacp and lacIq) were made blunt by T4 DNA polymerase and ligated toform pEX1.8. rhlR was fused to tacp on pEX1.8, and the spacing between therhlR translational start site and a Shine-Dalgarno region downstream of tacp wasoptimized. To do this, a DNA fragment containing rhlR was generated by PCRusing Vent polymerase (New England Biolabs) in a Hybaid thermal reactor(National Labnet Corp., Woodbridge, N.J.) with the following reaction condi-tions: melting temperature, 95°C (1 min); annealing temperature, 60°C (1 min);elongation temperature, 72°C (1 min); 25 cycles. pSK825 containing rhlR fromP. aeruginosa PG201 (which is 100% identical in amino acid sequence to rhlRfrom PAO1 [4]) was used as a template for the primer rhlR-ORF (59ATGAGGAATGACGGAGGC-39) and the SK primer of pBluescript II SK1 (59-GTAAAACGACGGCCAGT-39). Primer rhlR-ORF corresponds to nt 11 to 118 relativeto the rhlR start codon. The SK primer corresponds to a region of pBluescript IISK1 which is on the antisense strand of DNA relative to rhlR in pSK825. Theseprimers were used to generate a 790-bp DNA fragment which contained a singleHindIII site downstream of rhlR and a blunt-ended rhlR start. After digestionwith HindIII, the fragment was ligated to pEX1.8 that had been digested withEcoRI, treated with Klenow to produce a filled-in EcoRI site, and then digestedwith HindIII. This yielded pJPP8, which contains rhlR under control of tacp. TherhlR on pJPP8 was confirmed by sequencing, which also revealed the expected9-bp spacing between the Shine-Dalgarno region downstream of tacp and thetranslational start of rhlR. Protein expression was monitored by growing culturesof E. coli DH5a(pJPP8 or pEX1.8 [control vector]) to an OD600 of 0.5. Growthwas continued for 2 h in the presence of 1 mM isopropyl-b-D-thiogalactopyrano-side (IPTG). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of celllysates, followed with Coomassie brilliant blue staining (51), revealed that anIPTG-inducible 28-kDa polypeptide was overexpressed in cells carrying pJPP8but not in those with pEX1.8.

Bioassays for the rhl AI. To develop an rhl AI bioassay in E. coli, we con-structed a plasmid carrying tacp-rhlR and an rhlA9-lacZ reporter gene fusion. TherhlA9-lacZ fusion on pECP60, which contains 548 bp upstream of rhlA throughthe first 25 codons of rhlA in frame with lacZ, was released as a 4.70-kb XmnIfragment and ligated to the filled-in (by Klenow enzyme) SalI site upstream oftacp-rhlR on pJPP8, forming pECP61.5.

For rhl AI bioassays, overnight cultures of E. coli containing pECP61.5 werediluted to an OD600 of 0.08 and grown at 37°C with shaking to an OD600 of 0.3;1-ml aliquots of culture were then grown for 90 min with IPTG (1 mM) in thepresence or absence of AIs. b-Galactosidase activity was measured as describedpreviously (34).

AI bioassays in P. aeruginosa. Cells from overnight cultures of P. aeruginosaPAO-JP2 containing pECP61.5 were washed and resuspended in PTSB mediumto an OD660 of 0.05 and then grown at 37°C with shaking to an OD660 of 0.3. Atthis time, 1-ml aliquots of culture were transferred to tubes containing testsamples. Growth was continued for 60 min, and cultures were placed on ice. Cellswere washed in A medium, and b-galactosidase was assayed.

Specificity of rhl and las autoinduction systems. To examine the effects of thelas autoinduction system on rhlA expression, we constructed a plasmid withtacp-lasR and an rhlA9-lacZ fusion. First we ligated a BamHI-HindIII fragment of

pKDT37 containing tacp-lasR to pEX1.8 which had been digested with BamHIand HindIII. The resulting plasmid, pECP59, was digested with Tth111I, andends were filled in by Klenow enzyme and ligated to a 4.70-kb XmnI fragmentfrom pECP60 containing the rhlA9-lacZ fusion, resulting in pECP63. As a pos-itive control plasmid for the effect of the las autoinduction system on lasBexpression, a lasB9-lacZ fusion and tacp-lasR were cloned on the same pEX1.8vector as follows. Ends of Tth111I-digested pECP59 and the 3.75-kb Eco109Ifragment of pKDT37 (which carries the lasB9-lacZ translational fusion ofpTS400) were filled in by Klenow enzyme and ligated to form pECP64. To studythe effects of the rhl autoinduction system on lasB expression, plasmid pECP62.5was constructed. To do this, ends of the 3.75-kb Eco109I fragment from pKDT37(containing the lasB9-lacZ translational fusion) were filled in, and the fragmentwas ligated to the filled-in SalI site of pJPP8. The resulting plasmid, pECP62.5,contained tacp-rhlR and lasB9-lacZ.

Bioassays testing specificity of rhl and las autoinduction systems in E. coliDH5a containing either pECP62.5, pECP63, or pECP64 were performed asdescribed above for the rhl AI bioassay using E. coli DH5a(pECP61.5).

Synthesis of biologically active 3HPAI-2. Tritium-labeled PAI-2 (3HPAI-2)was synthesized by using procedures similar to those described for synthesis oftritium-labeled Vibrio fischeri AI-1, N-[3H](3-oxohexanoyl) homoserine lactone(3HVAI-1) (27), and 3HPAI-1 (45). First, N-crotanoyl-L-homoserine lactone, analkene analog of PAI-2, was synthesized by coupling sodium crotanate andL-homoserine lactone hydrochloride by the method of Eberhard et al. (11) andwas purified by reverse-phase high-performance liquid chromatography (HPLC)using water and methanol (C18 column, 1 by 25 cm; Beckman Corp., Fullerton,Calif.). Analysis by 1H nuclear magnetic resonance (NMR) spectroscopy (in[2H2]O and in C[2H]Cl3 on a 300-MHz Bruker instrument at the Department ofChemistry, University of Rochester) indicated the purified product was N-cro-tanoyl homoserine lactone. The N-crotanoyl homoserine lactone was radiola-beled by reduction using tritium gas in the presence of a Pd catalyst by DupontNEN. Tritium NMR spectroscopy of the synthesis product (3HPAI-2) was per-formed on a 300-MHz Bruker instrument by Dupont NEN (proton decoupled inC[2H]Cl3). The 3H NMR spectrum of the synthetic 3HPAI-2 contained onlythree peaks (2.20, 1.65, and 0.95 ppm), which correspond to tritium at positions2, 3, and 4, respectively, on the acyl side chain of PAI-2 in a ratio of 1:2:1. Thisresult indicated that half of the synthetic 3HPAI-2 contained tritium in themethylenes at positions 2 and 3 and half contained tritium in the methylenes atpositions 3 and 4, as could be expected for catalytic reduction of a,b unsaturatedgroups on acyl chains (57).

When 3HPAI-2 was applied to C18 reverse-phase HPLC, a single radioactivepeak that was the sole peak of PAI-2 activity [assayed in E. coli DH5a(pECP61.5)] was eluted with the same retention time as unlabeled PAI-2. Inaddition, when 3HPAI-2 was applied to the same HPLC together with excessunlabeled PAI-2 as the carrier, as described for the characterization of 3HVAI-1(27), a single peak of radioactivity coeluted with PAI-2 activity (data not shown).Specific activity of 3HPAI-2 was determined in the rhl AI bioassay in E. coli to be100 Ci/mmol, and bioactivity of 3HPAI-2 was indistinguishable from that ofunlabeled PAI-2 (data not shown). We stored 3HPAI-2 in ethyl acetate at 220°C.

3HPAI-2 binding assays. The binding assay was based on the method ofHanzelka and Greenberg (19) for binding of 3HVAI-1 to E. coli expressing theV. fischeri LuxR. Overnight cultures of E. coli DH5a containing pGroESL andeither pJPP8, pECP59, or pEX1.8 were diluted in LB medium to an OD600 of0.05 and incubated at 37°C with shaking until an OD600 of 0.5 was reached. Atthis point, IPTG was added to a final concentration of 1 mM and growth wascontinued for an additional 2 h. Next, 1 ml of culture was mixed with 60 ml of3HPAI-2 in water (solutions of 3HPAI-2 in ethyl acetate were evaporated intubes under a stream of N2 gas, and 3HPAI-2 was dissolved in water). Sampleswere incubated for 30 min at 25°C. After 5 min on ice, samples were centrifuged(16,000 3 g for 15 min at 4°C), and the cells were washed twice by resuspensionin 1 ml of ice-cold phosphate-buffered saline (pH 7) and by centrifugation asbefore. After this, cells were centrifuged for 10 min to remove all supernatantfluid. Finally, cells were resuspended in 30 ml of phosphate-buffered saline, andradioactivity in the cell suspension was determined by standard scintillationcounting procedures. All binding experiments were performed four times.

TABLE 1—Continued

Strain orplasmid Relevant genotype and/or phenotype Source or

reference

Plasmids used ingene expression

pSW205 ori (colE1) ori (P. aeruginosa) promoterless lacZ Apr 14pECP60 rhlA9-lacZ translational fusion on pSW205 48pECP61.5 rhlA9-lacZ from pECP60 on pJPP8, tacp-rhlR This studypTS400 lasB9-lacZ translational fusion on pSW205 43pECP62.5 lasB9-lacZ from pTS400 on pJPP8, tacp-rhlR This studypECP63 rhlA9-lacZ from pECP60 on pECP59, tacp-lasR This studypECP64 lasB9-lacZ from pTS400 on pECP59, tacp-lasR This study



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

RESULTS

Construction and characterization of quorum-sensing mu-tants of strain PAO1. Previous reports indicated that the twoP. aeruginosa quorum-sensing systems, las and rhl, regulateeach other to some extent (4, 31, 40, 47, 48). To determine thequorum-sensing systems’ effects on each other and how theycontrol virulence genes, a P. aeruginosa lasI mutant was de-sired. Previous attempts to obtain this mutant had been unsuc-cessful (47, 53). We speculated that these attempts failed fortechnical reasons and tried again. Figure 1A shows a schematicof pJPP4 which was used in allelic exchange with lasI of strainPAO1. Southern blot analysis of DNA from strain PAO1 andfour Tcr exconjugants by using a 32P-labeled fragment of lasI asa probe revealed that a chromosomal fragment carrying lasIwas replaced with the homologous DNA containing DlasI::Tcr

from pJPP4 (Fig. 1). Southern blot analysis using a 32P-labeledTcr cassette as a probe confirmed that the Tcr marker waspresent in the expected position on the chromosome of each ofthe four Tcr exconjugants (data not shown). One of these DlasImutants was named strain PAO-JP1. We also used the above-described allelic exchange procedure to generate Tcr exconju-gants of strain PDO100, a DrhlI mutant of PAO1 (4). Southernanalysis, similar to that performed on strain PAO-JP1, of thestrain PDO100 Tcr transconjugants indicated that lasI was re-

placed in the same manner as in strain PAO-JP1 (data notshown). This resulted in a DlasI DrhlI double mutant that wasnamed strain PAO-JP2.

Since lasI has been shown to direct the synthesis of PAI-1(46), we tested ethyl acetate extracts of culture fluids from theP. aeruginosa mutant strains PAO-JP1 and PAO-JP2 forPAI-1. As expected, culture fluid extracts from both strains didnot contain PAI-1, whereas culture fluid extracts of the paren-tal strains PAO1 and PDO100, respectively, contained PAI-1(data not shown). This finding demonstrated that lasI was re-quired in P. aeruginosa to direct the synthesis of PAI-1. Itfurther shows that rhlI is not required for synthesis of PAI-1and that the rhl quorum-sensing system does not regulate lasIexpression.

In addition to the above-described mutants, we also wantedto generate a lasR rhlR double mutant. To do this, we mobi-lized a previously constructed suicide plasmid containingDlasR::Tcr (15) into strain PDO111, a rhlR::Tn501 mutant ofPAO1 (4). Southern blot analysis of Tcr P. aeruginosa PDO111exconjugants resulted in an autoradiograph (data not shown)virtually identical to that reported for strain PAO-R1, theDlasR mutant (15). This result indicated that lasR had beendeleted from strain PDO111, and this lasR rhlR double-mutantstrain was named PAO-JP3.

We began our characterization of the mutants by examininghow major targets of quorum sensing (such as elastase andrhamnolipid) were affected. We analyzed cell-free culture flu-ids of P. aeruginosa PAO-JP1 (lasI), PDO100 (rhlI), PAO-JP2(lasI rhlI), and PAO1 (wild type) for elastolytic activity in ECRassays and for rhamnolipid in orcinol assays (Table 2). Bothelastolysis and rhamnolipid levels were significantly lower instrains PAO-JP1 and PDO100 than in the wild-type strainPAO1 (each strain contained the control vector pSW200 [Ta-ble 2]). Elastolytic activity and rhamnolipid production werecompletely restored when strain PDO100 contained pJPP45(rhlI1) or was grown in the presence of PAI-2 (Table 2). Bothphenotypes were also complemented when strain PAO-JP1contained pLASI-2 (lasI1) (Table 2). Interestingly, addition ofPAI-1 to cultures of strain PAO-JP1(pSW200) fully restoredelastolysis but only partially restored rhamnolipid (Table 2).This finding suggests that either (i) the PAI-1 concentrationtested was not optimal for rhamnolipid production or (ii) otherAIs known to be less abundantly synthesized by the lasI geneproduct (46) may contribute to rhamnolipid biosynthesis.

Neither rhamnolipid nor elastolysis was detectable in culturefluids of strain PAO-JP2 (lasI rhlI) (Table 2). Full complemen-tation of rhamnolipid and elastolysis occurred when PAI-1 andPAI-2 were added together to cultures of strain PAO-JP2(pSW200) and when strain PAO-JP2 contained pJPP42 (lasI1

rhlI1) (Table 2). Elastolysis and rhamnolipid were only par-tially restored when strain PAO-JP2 contained either pJPP45(rhlI1) or pLASI-2 (lasI1) (Table 2). These results confirmedour findings with the lasI and rhlI single mutants [strains PAO-JP1(pSW200) and PDO100(pSW200)], which indicated thatboth lasI and rhlI are required for wild-type levels of elastolysisand rhamnolipid.

Addition of either PAI-1 or PAI-2, at the concentrationstested, to strain PAO-JP2(pSW200) resulted in partial resto-ration of elastolysis (Table 2). Adding PAI-2 to this strain,however, essentially restored wild-type levels of rhamnolipid,whereas PAI-1 alone had no effect on rhamnolipid production(Table 2). Elsewhere we have shown PAI-1 alone had no effecton rhamnosyltransferase gene expression (rhlA9-lacZ) in strainPAO-JP2(pECP60) (48). There, we showed that PAI-2 onlypartially restored rhlA expression and that both PAI-1 andPAI-2 were needed for maximal expression (48). However, the

FIG. 1. Construction of a P. aeruginosa DlasI mutant strain. (A) Schematicdiagram showing lasR and lasI genes on the P. aeruginosa PAO1 chromosomeand relevant restriction sites for PstI (P) and SphI (Sp). Also shown is the regionof pJPP4 that contains P. aeruginosa DNA with a Tcr cartridge replacing part oflasI (described in Materials and Methods). The thick black line represents theprobe used for the Southern blot in panel B. (B) Autoradiograph of a Southernblot of chromosomal DNA isolated from four Tcr P. aeruginosa exconjugants(lanes A to D) and from wild-type P. aeruginosa (lane PAO1), digested with PstIand SphI as indicated. Uncut plasmid pJPP4 was run as a control. DNA waselectrophoresed on a 0.7% agarose gel, transferred to a nylon membrane, andhybridized to an 813-nt 32P-labeled DNA probe as indicated in panel A. Hybrid-ization of the probe to a 2.2-kb PstI fragment from the exconjugants indicatedthat the PstI site in the 39 end of lasI was lost. Hybridization to a 2.0-kb SphIfragment indicated that the SphI site in the Tcr cartridge was gained.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

levels of AIs that we tested in those studies were 10-fold lessthan those tested here. In addition, the experimental condi-tions in that study used early-stationary-phase cells grown inrich PTSB medium to measure rhlA9-lacZ gene expression(48), whereas here rhamnnolipid, which accumulates in culturefluids, was measured from late-stationary-phase cultures in GSminimal medium. Thus, different culture conditions and mea-surements likely reflect the observations that PAI-2 partiallycomplemented rhlA expression production in the lasI rhlI dou-ble mutant [strain PAO-JP2(pECP60) (48)] but almost fullycomplemented rhamnolipid in PAO-JP2(pSW200) (Table 2).

Northern analysis of rhlA, rhlB, and rhlR. Because lasI wasfound to be required for P. aeruginosa wild-type levels ofrhamnolipid production (Table 2), and elsewhere we showedPAI-1 is involved in rhlA expression (48), we examined theeffect of lasI on transcription of rhamnolipid biosynthesis genesby Northern blot analysis of P. aeruginosa RNA. Previous stud-ies indirectly suggested that rhlA and rhlB may be coexpressedfrom a putative promoter upstream of rhlA (39). However, thisobservation was based on complementation analysis of rham-nolipid production by P. aeruginosa rhlAB transposon mutantsand on expression of rhlA9, rhlB9, and rhlAB9 translationalfusions with lacZ in P. aeruginosa (39). We probed P. aerugi-nosa RNA from mid-log-phase cells (OD660 of 0.9) grown inGS medium with 32P-labeled internal fragments of rhlA, rhlB,rhlR (Fig. 2A), and pilA (positive control). The rhlA probeconsistently yielded a strong 1.15-kb band in addition to a lessintense 2.4-kb band when hybridized to RNA from eitherstrain PAO1 (Fig. 2B, lane 5) or strain PG201 (data notshown), but the rhlA probe did not hybridize to RNA fromstrains lacking lasI (PAO-JP2 and PAO-JP1 [Fig. 2B, lanes 3and 4, respectively]) or lasR (PAO-R1 and PAO-JP3 [Fig. 2B,lanes 1 and 2, respectively]). This finding indicated that bothlasI and lasR were required for transcription of rhlA in P.aeruginosa at least in mid-log growth phase.

We next probed P. aeruginosa RNA with an rhlB fragment,and the probe hybridized to a 2.4-kb transcript from eitherstrain PAO1 (Fig. 2C, lane 5) or strain PG201 (data not

shown), but the rhlB probe did not hybridize to RNA from anyof the mutant strains (Fig. 2C, lanes 1 to 4). This findingindicated that lasI and lasR are also required for expression ofrhlB. Interestingly, a 2.4-kb transcript was detected in Northernanalysis of strain PAO1 or PG201 RNA with either the rhlA orrhlB probe. However, unlike the rhlA probe, which also hybrid-ized to a 1.15-kb transcript (Fig. 2B, lane 5), the rhlB probehybridized only to a 2.4-kb transcript (Fig. 2C, lane 5, and datanot shown for strain PG201). Possibly a transcriptional termi-nator between rhlA and rhlB could account for the 1.15-kbtranscript detected by rhlA probes. The 2.4-kb transcript de-tected in P. aeruginosa wild-type RNA by using either the rhlAor rhlB probe is in agreement with the hypothesis that rhlA andrhlB are organized as an operon (39).

We next wanted to determine if the putative rhlAB transcriptwas independent from that of rhlR, which is found immediatelydownstream of rhlB. We probed for P. aeruginosa rhlR messageand observed a strong 0.85-kb band with RNA from strainsPAO-R1 and PAO1 (Fig. 2D, lanes 1 and 5, respectively) thatwas consistently less intense in strains PAO-JP2 and PAO-JP1(Fig. 2D, lanes 3 and 4, respectively) and totally absent asexpected in the rhlR lasR mutant, PAO-JP3 (Fig. 2D, lane 2).Thus, the 0.85-kb size of the rhlR transcript detected in ourNorthern blot hybridizations was consistent with that predictedby previous deletion analyses (4, 38). The fact that the rhlRprobe gave a 0.85-kb band and not a 2.4-kb band confirmedthat rhlR was not cotranscribed with the 2.4-kb rhlAB mRNA.Thus, the results in Fig. 2 prove that the 720-bp rhlR is tran-scribed as a monocistronic message from its own promoter andthat rhlA and rhlB are transcribed as a 2.4-kb bicistronic RNA.Our results also indicate that a second rhlA message (1.15 kb)is produced.

As a control to show that RNA from each strain was intactand loaded equally, we used an internal fragment of P. aerugi-nosa pilA, which encodes the structural subunit of pilin, as aprobe. The pilA probe hybridized to a 0.7-kb RNA from strainPAO1 and its derivatives as expected, indicating that RNA

TABLE 2. Effects of lasI and rhlI on elastase activity and rhamnolipid production of P. aeruginosa

Strain

Chromosomalgenotype Gene(s) on

plasmidAI

additionaElastolysisb

(A495)Rhamnolipidc

(mg/ml)rhlI lasI

PAO1(pSW200) 1 1 None 1.19 6 0.18 1.73 6 0.25PAI-1 1.22 6 0.06 NDPAI-2 1.33 6 0.05 NDboth 1.13 6 0.11 ND

PDO100(pSW200) 2 1 None 0.87 6 0.06 0.77 6 0.19PAI-2 1.30 6 0.25 2.08 6 0.09

PDO100(pJPP45) 2 1 rhlI1 None 1.24 6 0.15 2.60 6 0.39PAO-JP1(pSW200) 1 2 None 0.42 6 0.14 0.37 6 0.01

PAI-1 1.03 6 0.10 0.80 6 0.09PAO-JP1(pLASI-2) 1 2 lasI1 None 1.09 6 0.13 1.78 6 0.35PAO-JP2(pSW200) 2 2 None 0.01 6 0.00 0.01 6 0.01

PAI-1 0.66 6 0.03 0.02 6 0.01PAI-2 0.61 6 0.06 1.13 6 0.04both 1.34 6 0.19 1.65 6 0.49

PAO-JP2(pJPP45) 2 2 rhlI1 None 0.72 6 0.07 1.06 6 0.21PAO-JP2(pLASI-2) 2 2 lasI1 None 0.75 6 0.09 0.43 6 0.32PAO-JP2(pJPP42) 2 2 lasI1, rhlI1 None 1.31 6 0.23 2.57 6 0.44

a Where indicated, cultures were grown in the presence of each AI at 12 mM.b Elastase activities of culture fluids from 21-h cultures in PTSB medium were measured in ECR assays as described in Materials and Methods. Values are average

A495 of at least three experiments (each in triplicate) 6 standard deviation.c Rhamnolipid levels in culture fluids from 82- to 88-h cultures in modified GS medium were determined in orcinol assays as described in Materials and Methods.

Numbers are averages of two or three experiments (each in duplicate) 6 standard deviation. ND, not determined.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

from all strains was present in equal amounts (Fig. 2E andreference 14).

Transcriptional start site mapping of rhlAB mRNA. To fur-ther study rhlAB regulation, the transcriptional start site ofrhlA was mapped by primer extension with the use of twoseparate oligonucleotide primers (described in Materials andMethods). Primer extensions with both primers showed tworhlA transcriptional start sites. In Fig. 3, we present the resultsobtained with the use of primer P1 and total RNA from P.aeruginosa grown in GS or LB medium. Major and minorextension products obtained with RNA from cells grown in GSmedium are indicated by the upper and lower arrowheads,respectively, in lane GS (Fig. 3A). These products suggest thatrhlA transcription starts at nt 2228 (major start) or 2183(minor start) relative to the rhlA translational start codon.These results were identical to those obtained with primer P2(data not shown). No extended products were visible whenRNA from strains grown in LB medium was tested (Fig. 3A).However, longer exposure showed a very light band in lane LBat the same position as the upper arrowhead (data not shown).As shown in lane GS in Fig. 3A, a single major extensionproduct was produced and no products larger than this weredetected. A minor extension product mapped to nt 2183 up-

stream of the rhlA translational start may represent a secondtranscriptional start site of rhlA. Alternatively, this could bedue to a premature transcriptional stop caused by a compres-sion contained in the DNA sequence at this position upstreamof rhlA. Sequence analysis of the region immediately upstreamfrom the potential minor start site did not reveal any obviouspromoter or regulatory sequences. Therefore, we believe thatthe minor primer extension product is likely an artifact. In-spection of the DNA upstream from the major rhlA transcrip-tional start site revealed both a putative s54-type promoter atnt 227 to 211 (Fig. 3B) and a putative s70-type promoter(TTGACA-N17-TATAAT) at nt 232 to 26 (TTACCG-N17-TAAAAA). Site-directed mutagenesis will be required to con-firm the functional identity of the rhlA promoter.

Interestingly, some genes controlled by s54 are regulated byenvironmental signals such as nitrogen availability. We ob-served much less rhlA mRNA in cells grown in nitrogen-richLB medium than in those grown in the nitrogen-limited GSmedium (Fig. 3A), consistent with the predictions of earlierstudies that found both rhlA9-lacZ expression and rhamnolipidproduction to be dramatically increased when P. aeruginosa isgrown in GS medium (39). To begin to determine if rhlA is as54-controlled gene, we examined rhlA9-lacZ expression frompECP60 contained in a P. aeruginosa rpoN (s54) mutant, strainPAK-N1, and its wild-type parent, strain PAK. Under previ-ously described growth conditions (48), in the presence andabsence of rpoN, b-galactosidase activities were 224 6 15 and13.9 6 8.7 Miller units, respectively. Thus, rhlA9-lacZ expres-sion decreased 15-fold in the absence of s54. In addition, weprobed RNA from strains PAK and PAK-N1 with an internalBamHI fragment of rhlA and detected the same faint 2.4-kbband and more intense 1.15-kb band from strain PAK as weobserved from strains PAO1 and PG201, but no hybridizationoccurred with RNA from strain PAK-N1 (data not shown).Together, the Northern analysis and rhlA9-lacZ expression inthe presence or absence of rpoN indicated that s54 plays somerole in the expression of rhlA, but further study is needed todetermine if s54 directly controls rhlA.

Quorum-sensing control of rhlA. In P. aeruginosa, expressionof rhlA has been shown to require rhlR, and the production ofrhamnolipids was shown to depend on rhlR and either rhlI orPAI-2 (38, 40). To identify components of P. aeruginosa thatare both necessary and sufficient for expression of the rhlAgene, we developed a bioassay for quorum-sensing control ofrhlA in E. coli DH5a by cloning a plasmid containing an rhlA9-lacZ fusion and tacp-rhlR(pECP61.5). Expression of rhlA9-lacZin E. coli DH5a(pECP61.5) required both RhlR and at least100 nM PAI-2 (Fig. 4A). Half-maximal activation of the rhlApromoter occurred with approximately 2.5 mM PAI-2 (Fig.4A), which is fourfold less than the PAI-2 concentration foundin culture fluids of early-stationary-phase P. aeruginosa (47).Expression of rhlA increased more than 300-fold at a PAI-2concentration of 15 mM (Fig. 4A). In contrast, PAI-1 hadalmost no effect on rhlA expression at concentrations below 10mM and only a slight effect at 50 mM (data not shown). Fromthese data, we conclude that PAI-2 is the AI required forexpression of the P. aeruginosa rhlAB operon.

To confirm these results for E. coli, we developed a bioassayfor the rhl AI in P. aeruginosa. In the P. aeruginosa lasI rhlIdouble-mutant strain PAO-JP2(pECP61.5), half-maximal rhlAactivation occurred at a PAI-2 concentration 10-fold lowerthan that in E. coli (Fig. 4B). Half-maximal activation in strainPAO-JP2 occurred with approximately 250 nM PAI-2 (Fig.4B). Others have shown 250 nM PAI-2 restored rhamnolipidproduction in a P. aeruginosa rhlI mutant (40). The results inFig. 4 demonstrated that PAI-2 was the AI required for ex-

FIG. 2. Northern blot analysis of P. aeruginosa rhlA, rhlB, and rhlR mRNAs.(A) Schematic illustration showing restriction fragments that were labeled andused as probes for rhlA, rhlB, and rhlR mRNAs. Letters refer to restrictionenzyme sites that constitute the ends of each DNA fragment: B, BamHI; E,EcoRV; X, XhoI; S, SacII. (B to E) Autoradiographs of Northern blots of P.aeruginosa RNA hybridized to a 32P-labeled internal fragment of rhlA (B), rhlB(C), rhlR (D), and pilA (E). S, RNA molecular size standards, in kilobases. Lanes1 to 5 contain RNA (7 mg/lane [B, D, and E] or 12 mg/lane [C]) from P.aeruginosa PAO-R1 (lasR), PAO-JP3 (lasR rhlR), PAO-JP2 (lasI rhlI), PAO-JP1(lasI), and PAO1 (wild type), respectively. The top and bottom arrowheads inpanel B point to 2.4- and 1.15-kb bands found in lane 5. The arrowhead in panelC points to a 2.4-kb band detected only in lane 5; the arrowhead in panel Dpoints to an 850-bp band found in all lanes except lane 2; the arrowhead in panelE points to a 700-bp band found in all lanes. The 2.37- and 1.54-kb bands inpanels B to E appeared due to nonspecific hybridization of probes to the 23S and16S P. aeruginosa rRNAs.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

pression of the P. aeruginosa rhlAB operon and suggest that,like LuxR and LasR, which interact with VAI-1 (19) and PAI-1(45), respectively, PAI-2 most likely interacts with RhlR toinduce rhlA expression.

PAI-2 binds E. coli expressing rhlR. To characterize theinteractions between PAI-2 and its putative target, RhlR, wesynthesized 3HPAI-2 and used it to develop a binding assay tostudy the putative interaction between PAI-2 and RhlR. Pre-vious results of studies using E. coli have shown that PAI-2does not activate LasR to induce lasB expression (47) and doesnot block binding of 3HPAI-1 to E. coli expressing LasR (45).Figure 5A shows that 3HPAI-2 specifically binds E. coli ex-pressing RhlR but not to E. coli expressing LasR. Binding of3HVAI-1 to E. coli expressing LuxR has been shown to beenhanced about 10-fold when the cells contain pGroESL whichhas lacp-groESL for increased expression of the E. coli chap-eronins GroES and GroEL (19). Figure 5A demonstrates thatbinding of 3HPAI-2 to cells with RhlR was enhanced fourfoldin the presence of pGroESL. Nonspecific binding was 9% oftotal binding and was not affected by pGroESL. Half-maximalbinding of 3HPAI-2 to cells expressing both RhlR and GroESLoccurred with approximately 0.5 mM 3HPAI-2 (Fig. 5B). Thisis the same level of PAI-2 as needed for half-maximal induc-tion of rhamnolipid in a P. aeruginosa rhlI mutant (40). Theseresults strongly support the notion that PAI-2 binds RhlR andthat chaperonins encoded by groESL may facilitate properfolding of RhlR.

Specificities of the las and rhl quorum-sensing systems.Since rhlAB mRNA, rhamnolipid, and rhlA9-lacZ were notexpressed in P. aeruginosa PAO-JP2 (lasI rhlI) (Fig. 2, Table 2,and reference 48, respectively), we measured the effects of thelas and rhl quorum-sensing systems on our rhlA9-lacZ andlasB9-lacZ reporters in E. coli. To do this, we constructed aseries of plasmids carrying tacp-lasR or tacp-rhlR and either arhlA9-lacZ or lasB9-lacZ translational fusion. We monitoredb-galactosidase activities of E. coli containing the plasmidsto measure the expression of the rhlA9-lacZ and lacB9-lacZfusions. Addition of PAI-1 caused LasR (expressed frompECP63) to activate rhlA9-lacZ 10-fold, but this was much lessthan the 320-fold activation of rhlA9-lacZ by PAI-2 and RhlR(expressed from pECP61.5) (Table 3). As shown in Table 3,E. coli containing either pECP61.5 or pECP63 gave extremelylow basal expression (approximately 1.7 Miller units) of rhlA9-lacZ with no addition of AIs.

The rhl quorum-sensing system has been reported to affectlasB expression. In particular, both lasB9-cat activity and elas-tolysis were shown to be reduced in P. aeruginosa rhlR and rhlImutants (4). To determine if RhlR can activate lasB9-lacZ inE. coli, pECP62.5 was constructed. Expression of lasB9-lacZincreased sixfold when E. coli expressing RhlR (pECP62.5)was grown with 10 mM PAI-2 (Table 3). This result indicatedthe rhl quorum-sensing system is capable of partially activatinglasB. As a control for LasR-mediated lasB9-lacZ expression, weconstructed pECP64. PAI-1 (50 nM) and LasR (expressed

FIG. 3. Primer extension analysis of the rhlAB transcriptional start site. Shown is an autoradiograph of a urea-polyacrylamide gel. Lane GS contains the extensionproduct obtained with the use of primer P1 and RNA from P. aeruginosa PG201 grown in GS medium; lane LB contains the extension product as described above exceptthat RNA was from P. aeruginosa PG201 grown in LB medium; lanes A, C, G, and T contain DNA sequencing reactions generated by using the same oligonucleotideprimer. The upper arrowhead points to the major primer extension product in lane GS. A product of the same size was detected in lane LB after a longer exposure.The lower arrowhead points to a minor primer extension product in lane GS. ORF, open reading frame.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

from pECP64) fully activated lasB9-lacZ 20-fold (Table 3), asexpected (46). Interestingly, high concentrations of PAI-1 (10mM) caused both lasB expression and rhlA expression to de-crease in E. coli expressing LasR (pECP64 and pECP63, re-spectively [Table 3]). These negative effects appear similar toresults for E. coli expressing the V. fischeri LuxR (from tacp-luxR), where low levels of VAI-1 activated luxR expression(from a luxRp-luxAB reporter fusion) whereas high levels ofVAI-1 repressed luxRp expression (54). Basal levels of b-ga-lactosidase activity in E. coli containing lasB9-lacZ in the pres-ence of either RhlR (pECP62.5) or LasR (pECP64) were notidentical (33 and 87 Miller units, respectively). The results inTable 3 demonstrate that the rhl quorum-sensing system canrecognize and slightly activate lasB and that the las quorum-sensing system can recognize and slightly activate rhlA.These data also indicate the great specificity of each tran-scriptional activator protein and its cognate AI (RhlR plusPAI-1 and LasR plus PAI-2 do not activate either reporter).Clearly the rhl system has a much greater effect on rhlA thanon lasB, and similarly, the las system preferentially activateslasB over rhlA.

DISCUSSION

Here we report the generation of three mutants importantfor the study of P. aeruginosa quorum sensing. Despite failuresduring previous attempts to construct a lasI mutant strain (47,53), we were able to generate a lasI mutant and lasI rhlI doublemutant. In addition, we created a lasR rhlR double mutantof P. aeruginosa. Thus, with mutations in each of the two P.aeruginosa quorum-sensing systems, we were equipped to ex-amine the effects of the two systems on each other as well as tobetter define the regulation of the major target of the rhlsystem itself, the rhamnolipid biosynthesis operon rhlAB.

In this report, we showed that both P. aeruginosa rhamno-lipid production and rhlA expression are controlled by both therhl and las quorum-sensing systems. Particularly, we found thatproduction of elastolytic activity and rhamnolipid were re-duced in the P. aeruginosa lasI and rhlI mutant strains (PAO-JP1 and PDO100, respectively). Brint and Ohman have previ-ously reported a decrease in elastolysis and the total loss ofrhamnolipid activity (hemolysis) in strain PDO100 (4). Thisdiscrepancy between the studies most likely reflects differences

FIG. 4. Bioassays for the rhl AI in E. coli and P. aeruginosa. (A) Doseresponse of E. coli DH5a(pECP61.5) to increasing amounts of PAI-2. Expres-sion of rhlA in the presence of RhlR and increasing concentrations of PAI-2 wasmonitored with b-galactosidase activity assays. b-Galactosidase levels in culturesthat received no addition of PAI-2 were 1.7 6 0.3 Miller units. (B) rhl AI bioassayin P. aeruginosa PAO-JP2. The expression of the rhlA9-lacZ reporter in strainPAO-JP2(pECP61.5) in the presence of RhlR and increasing concentrations ofPAI-2 was monitored as in panel A. b-Galactosidase levels from cultures thatreceived no addition were 100 6 8 Miller units.

FIG. 5. Evidence for PAI-2 binding to RhlR. (A) Binding of 3HPAI-2 toE. coli overexpressing RhlR or LasR. Shown is radioactivity remaining with E.coli DH5a (in the presence and absence of pGroESL) carrying the vector control(pEX1.8) or expressing LasR (pECP59) or RhlR (pJPP8) after incubation with250 nM 3HPAI-2. (B) Dose response for binding of 3HPAI-2 to E. coli DH5aexpressing RhlR. Shown is radioactivity remaining with cells containingpGroESL and either pJPP8 (F) or pEX1.8 (E) after incubation with increasingconcentrations of 3HPAI-2. Assays were performed as described in Materials andMethods.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

in growth media (peptone-glycerol broth in reference 4; GSmedium here) and assay methods (they used an erythrocytehemolysis assay for rhamnolipid activity; we used the orcinolassay for the presence of rhamnolipid).

Rhamnolipid and elastolysis were totally absent in theP. aeruginosa lasI rhlI double mutant (strain PAO-JP2), whichindicated that production of these virulence factors requiresboth quorum-sensing systems. Both elastolysis and rhamno-lipid were fully complemented in strains PAO-JP1, PDO100,and PAO-JP2 by the addition of the appropriate wild-typegenes on plasmids (Table 2). Exogenously added PAI-1 andPAI-2 complemented elastolysis as expected (Table 2), butonly PAI-2 complemented rhamnolipid to the levels that weobserved with the wild-type genes on plasmids (Table 2). Ex-ogenously added PAI-1 partially complemented rhamnolipidproduction in the lasI mutant (strain PAO-JP1) and had noeffect on rhamnolipid in the lasI rhlI double mutant (strainPAO-JP2) (Table 2). We have shown that PAI-1 has no effecton rhlA expression in the P. aeruginosa lasI rhlI double mutant(48), and in Table 3 we observed only a minor effect of PAI-1on rhlA expression in E. coli expressing LasR (only 3% of therhlA expression compared to when PAI-2 was added to E. coliexpressing RhlR). Thus, LasR and PAI-1 had almost no directeffect on rhlA expression. In addition, PAI-1 does not signifi-cantly activate RhlR to induce either rhlA or lasB expression(Table 3), which indicates that RhlR has high specificity for itscognate AI, PAI-2. These results explain why PAI-1 does notcomplement rhamnolipid production in the P. aeruginosa lasIrhlI double mutant (strain PAO-JP2).

The fact that lasI on pLASI-2 partially restored rhamnolipidwhen carried by strain PAO-JP2, whereas PAI-1 did not, sug-gests that other AIs may be responsible. In addition to direct-ing PAI-1 synthesis, the lasI gene product makes smallamounts of other AIs (46) such as VAI-1 (10). These other AIsmay be responsible for the partial complementation of rham-nolipid that we observed by pLASI-2. VAI-1 has been shown toinduce rhamnolipid production when added to cultures of P.aeruginosa (40), and VAI-1 clearly does not activate LasR (17,45, 46). Thus, we speculate that the small amount of VAI-1that is likely to be made by the P. aeruginosa lasI product maybe capable of inducing rhamnolipid biosynthesis via RhlR.However, the fact that PAI-2 (i) is much more abundantlyproduced (.200-fold) than VAI-1 by P. aeruginosa (46, 47), (ii)

induces rhamnolipid biosynthesis at more than 20-fold-lowerconcentrations than VAI-1 (40), and (iii) alone almost fullyrestored rhamnolipid production by our lasI rhlI double mu-tant [strain PAO-JP2(pSW200) (Table 2)] emphasizes its roleas the major autoinducer of rhamnolipid production.

Northern blot analysis indicated the rhlA and rhlB geneswere expressed as a bicistronic 2.4-kb rhlAB transcript and as asmaller 1.15-kb rhlA transcript. The smaller rhlA RNA may bedue to an unknown transcriptional termination and antitermi-nation mechanism or may be a product of RNase processing ofthe rhlAB transcript. Our Northern blots also proved rhlR wastranscribed independently of rhlA and rhlB. Previous deletionanalyses of the rhlR upstream region (4, 38) and earlier obser-vations in P. aeruginosa, using both insertion mutants of rhlAand rhlB as well as rhlA9, rhlB9, and rhlAB9 fusions with lacZ(39), further support our evidence that rhlAB is transcribed asa bicistronic mRNA and that rhlR is transcribed from its ownpromoter. Elsewhere, we and others used rhlR9-lacZ transcrip-tional fusions to show that rhlR expression is increased fivefoldover its basal level by the las quorum-sensing system (31, 48).Based on our Northern analysis of rhlR mRNA presented here,however, it is clear that the las system is not required for basallevels of rhlR expression (Fig. 2D). As suggested before (31,48), activation of rhlR by the las system seems to be part of thequorum-sensing hierarchy where the las system exerts controlover the rhl system. Our Northern analysis supports this con-clusion but suggests that basal levels of rhlR expression aredependent on mechanisms other than the las quorum-sensingsystem.

A major rhlA transcriptional start site was identified at nt2228 upstream of the translational start of rhlA (upper arrow-head in Fig. 3A). The 2.4-kb rhlAB and 1.15-kb rhlA mRNAsdetected by Northern blotting (Fig. 2B and C) are likely pro-duced from the major rhlA transcriptional start site, but furthertranscript mapping will be needed to prove this.

By mapping the major transcriptional start of P. aeruginosarhlA, we were able to construct an appropriate reporter forrhlA activity and reconstitute the components necessary forexpression of rhlA in E. coli. Development of a bioassay al-lowed us to demonstrate that RhlR and PAI-2 are requiredand sufficient for induction of the rhamnolipid biosynthesisoperon in E. coli (Fig. 4A). Interestingly, 10-fold less PAI-2was necessary to induce the same reporter in our P. aeruginosalasI rhlI double mutant (Fig. 4B). This increased sensitivity ofthe reporter to PAI-2 in P. aeruginosa may be due to betteruptake of PAI-2 through P. aeruginosa membranes. To date,only one study has examined AI transport through bacterialcells. 3HVAI-1 was found to penetrate both V. fischeri andE. coli by diffusion (28). Future work using 3HPAI-2 will elu-cidate the mechanism of PAI-2 transport into P. aeruginosa.Another possible explanation for decreased AI sensitivity in E.coli could be the lack of other P. aeruginosa factors required formaximal expression of rhlA. These potential factors could beinvolved in AI binding (58) but more likely reflect differencesin transcriptional and translational machinery between P. aeru-ginosa and E. coli. Our rhl AI bioassays in both P. aeruginosaand E. coli clearly demonstrate that PAI-2 is the AI of therhamnosyltransferase operon, rhlAB. Elsewhere RhlR and PAI-2have recently been shown to enhance expression of rhlI andrpoS, which encodes a sigma factor expressed in stationaryphase (31). Together these results also suggest that RhlR in-teracts with PAI-2 to become a transcriptional activator.

To characterize the putative interaction between RhlR andPAI-2, we have demonstrated that 3HPAI-2 specifically bindsE. coli expressing RhlR but not LasR (Fig. 5). Elsewhere wehave shown that PAI-1 blocks binding of 3HPAI-2 to E. coli

TABLE 3. Specificities of P. aeruginosa las andrhl quorum-sensing components in E. colia

AI added

rhlA9-lacZ expression inthe presence of:

lasB9-lacZ expression inthe presence of:

RhlRb LasRc RhlRd LasRe

None 1.7 6 0.2 1.9 6 0.1 33 6 3.0 87 6 4.0PAI-2 (50 nM) 4.7 6 0.7 2.0 6 0.2 40 6 1.0 95 6 4.0PAI-2 (10 mM) 610 6 26 2.3 6 0.2 189 6 24 112 6 5.0PAI-1 (50 nM) 1.6 6 0.1 20 6 0.6 35 6 1.0 1,919 6 159PAI-1 (10 mM) 9.1 6 1.1 12.1 6 1.3 48 6 3.0 624 6 53

a Data are averages of four b-galactosidase assays in Miller units 6 1 standarddeviation for each. Assays were performed as described in Materials and Meth-ods.

b E. coli DH5a containing pECP61.5, carrying a rhlA9-lacZ reporter fusion andtacp-rhlR.

c E. coli DH5a containing pECP63, carrying a rhlA9-lacZ reporter fusion andtacp-lasR.

d E. coli DH5a containing pECP62.5, carrying a lasB9-lacZ reporter fusion andtacp-rhlR.

e E. coli DH5a containing pECP64, carrying a lasB9-lacZ reporter fusion andtacp-lasR.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

expressing RhlR and inhibits PAI-2 activity in the rhl AI bio-assay (48). Others have observed inhibition of LuxR andVAI-1 induction of lux genes by a second V. fischeri AI (N-octanoyl homoserine lactone) produced by the ainS gene prod-uct (30). Both are examples of posttranslational control by twoquorum-sensing systems of the same bacterium.

Our study supports the notion that PAI-2 binds specificallyto the RhlR protein. Previous studies by others (5, 19) providestrong evidence for an AI-binding region on LuxR, but theseresults were also obtained in E. coli expressing DLuxR proteinsand not with purified proteins. Future work with purified RhlRwill be essential to directly prove that PAI-2 binds RhlR and todetermine the position and number of PAI-2 binding sites.

Quorum-sensing and s factors. The results presented herehave demonstrated quorum-sensing control of the rhamnolipidbiosynthesis operon rhlAB. As expected for genes controlled byquorum sensing, we found a lux-type element (17) centered at240 bp upstream of the major rhlAB transcriptional start. Thisputative cis-acting promoter element shows identity to the luxand tra boxes and contains 13 of 20 nt identical to the OP1 lasbox upstream of the P. aeruginosa elastase gene, lasB (refer-ence 50 and Fig. 3B). Further inspection of the DNA upstreamof the rhlAB transcriptional start has revealed two possiblepromoter motifs (s54 and s70). Identification of the promoterand other cis-acting elements involved in rhlAB transcriptionwill require site-directed mutagenesis of this region. Proteinpurification and DNA footprinting studies will be necessary todetermine if either RhlR or LasR binds to the putative las box.

To date, quorum sensing has not been implicated in regu-lation of genes also controlled by s54-type promoters. Transcrip-tion from these promoters requires the binding of an activatorprotein to an upstream activation sequence (UAS), usually .100bp upstream of the transcriptional start site. This binding is fol-lowed by looping of the intervening DNA and direct activatorcontact with RNA polymerase-s54 holoenzyme (which canbind the promoter but is unable to form an open promotercomplex until contact is made with the activator protein boundto the UAS), resulting in transcription (33). Most of these invitro studies have focused on two such activators, NtrC andNifA, which are found in a number of gram-negative bacteria(6). In P. aeruginosa, PilR was shown to footprint a 40-bp UASconsisting of three repeating 59-N4(C/G)TGTC-39 sequencesin a tandem array from nt 120 to 80 upstream of the s54-controlled pilA transcriptional start (25). This type of UAS,however, is not present in the 300 bp upstream of the rhlAtranscriptional start.

In contrast to extensive in vitro studies of s54 and its acti-vators, much less is known about AI-LuxR-type transcriptionalactivation (13). Based on recent in vitro transcription assaysand DNA footprinting using RNA polymerase-s70 holoen-zyme and a purified AI-independent domain of LuxR to theV. fischeri lux promoter region (61, 62) as well as mutationalanalyses of the lux, tra, and las boxes (13, 50), quorum-sensingregulation of target genes has been shown to involve DNAelements that are overlapping or near the promoters of thesegenes. The 235 region of the putative s70 promoter at nt 232upstream of the rhlAB transcriptional start overlaps the las box(Fig. 3B), as would be expected based on those previous stud-ies. Determination of exactly how the las and rhl quorum-sensing systems directly interact with the rhlA and lasB pro-moters awaits future studies.

ACKNOWLEDGMENTS

We thank U. A. Ochsner for strains, plasmids, and advice; A. S.Kende and A. Eberhard for help with NMR; E. P. Greenberg forconstructive criticism; W. Kuhnert and K. Tucker for pEX1.8 and

pKDT1.7, respectively; D. E. Ohman and J. M. Brint for plasmids andstrains PDO100 and PDO111; J. S. Butler and R. P. Silver for theirexpertise; J. Nezezon for technical support; and L. Passador andC. Van Delden for help in manuscript preparation.

This work was supported by NIH grant AI 33713.

REFERENCES1. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. A. Smith, J. G.

Seidman, and K. Struhl (ed.). 1987. Current protocols in molecular biology.John Wiley & Sons, New York, N.Y.

2. Bjorn, M. J., P. A. Sokol, and B. H. Iglewski. 1979. Influence of iron on yieldsof extracellular products in Pseudomonas aeruginosa cultures. J. Bacteriol.138:193–200.

3. Blackwood, L. L., R. M. Stone, B. H. Iglewski, and J. E. Pennington. 1983.Evaluation of Pseudomonas aeruginosa exotoxin A and elastase as virulencefactors in acute lung infection. Infect. Immun. 39:198–201.

4. Brint, J. M., and D. E. Ohman. 1995. Synthesis of multiple exoproducts inPseudomonas aeruginosa is under the control of RhlR-RhlI, another set ofregulators in strain PAO1 with homology to the autoinducer-responsiveLuxR-LuxI family. J. Bacteriol. 177:7155–7163.

5. Choi, S. H., and E. P. Greenberg. 1992. Genetic dissection of DNA bindingand luminescence gene activation by the Vibrio fischeri LuxR protein. J.Bacteriol. 174:3856–3862.

6. Collado-Vides, J., B. Magasanik, and J. D. Gralla. 1991. Control site loca-tion and transcriptional regulation in Escherichia coli. Microbiol. Rev. 55:371–394.

7. Deretic, V., P. Tomasek, A. Darzins, and A. M. Chakrabarty. 1986. Geneamplification induces mucoid phenotype in rec-2 Pseudomonas aeruginosaexposed to kanamycin. J. Bacteriol. 165:510–516.

8. Deretic, V., J. F. Gill, and A. M. Chakrabarty. 1987. Gene algD coding forGDP mannose dehydrogenase is transcriptionally activated in mucoid Pseu-domonas aeruginosa. J. Bacteriol. 169:351–358.

9. Doring, G., J. Obernesser, and K. Botzenhart. 1981. Extracellular toxins ofPseudomonas aeruginosa II. Effects of two proteases on human immunoglob-ulin IgG, IgA, and secretory IgA. Zentralbl. Bacteriol. Mikrobiol. Hyg. Abt.1 Orig. Reihe A 249:89–98.

10. Eberhard, A., A. L. Burlingame, C. Eberhard, G. L. Kenyon, K. H. Nealson,and N. J. Oppenheimer. 1981. Structural identification of autoinducer ofPhotobacterium fischeri luciferase. Biochemistry 20:2444–2449.

11. Eberhard, A., C. Widrig, P. MacBath, and J. B. Schineller. 1986. Analogs ofthe autoinducer of bioluminescence in Vibrio fischeri. Arch. Microbiol. 146:35–40.

12. Figurski, D. H., and D. R. Helinski. 1979. Replication of an origin-containingderivative of plasmid RK2 dependent on a plasmid function provided intrans. Proc. Natl. Acad. Sci. USA 76:1648–1652.

13. Fuqua, W. C., S. C. Winans, and E. P. Greenberg. 1996. Census and con-sensus in bacterial ecosystems: the LuxR-LuxI family of quorum-sensingtranscriptional regulators. Annu. Rev. Microbiol. 50:727–751.

14. Gambello, M. J., S. Kaye, and B. H. Iglewski. 1993. LasR of Pseudomonasaeruginosa is a transcriptional activator of the alkaline protease gene (apr)and an enhancer of exotoxin A expression. Infect. Immun. 61:1180–1184.

15. Gambello, M. J., and B. H. Iglewski. 1991. Cloning and characterization ofthe Pseudomonas aeruginosa lasR, a transcriptional activator of elastase ex-pression. J. Bacteriol. 173:3000–3009.

16. Goloubinoff, P., A. A. Gatenby, and G. H. Lorimer. 1989. GroE heat shockproteins promote assembly of foreign prokaryotic ribulose bisphosphatecarboxylase oligomers in Escherichia coli. Nature 337:44–47.

17. Gray, K. M., L. Passador, B. H. Iglewski, and E. P. Greenberg. 1994. Inter-changeability and specificity of components from the quorum-sensing regu-latory systems of Vibrio fischeri and Pseudomonas aeruginosa. J. Bacteriol.176:3076–3080.

18. Guerra-Santos, L. H., O. Kappeli, and A. Fiechter. 1986. Dependence ofPseudomonas aeruginosa continuous culture biosurfactant production on nu-tritional and environmental factors. Appl. Microbiol. Biotechnol. 24:443–448.

19. Hanzelka, B. L., and E. P. Greenberg. 1995. Evidence that the N-terminalregion of the Vibrio fischeri LuxR protein constitutes an autoinducer bindingdomain. J. Bacteriol. 177:815–817.

20. Holder, I. A., and C. G. Haidaris. 1979. Experimental studies of the patho-genesis of infections due to Pseudomonas aeruginosa extracellular proteaseand elastase as in vivo virulence factors. Can. J. Microbiol. 25:593–599.

21. Howe, T., and B. H. Iglewski. 1984. Isolation and characterization of alkalineprotease-deficient mutants of Pseudomonas aeruginosa in vitro and in amouse eye model. Infect. Immun. 43:1058–1063.

22. Ishimito, K. S., and S. Lory. 1989. Formation of pilin in Pseudomonasaeruginosa requires the alternative s factor (RpoN) of RNA polymerase.Proc. Natl. Acad. Sci. USA 86:1954–1957.

23. Jaffar-Bandjee, M. C., A. Lazdunski, M. Bally, J. Carrere, J. P. Chazalette,and C. Galabert. 1995. Production of elastase, exotoxin A, and alkalineprotease in sputa during pulmonary exacerbation of cystic fibrosis in patientschronically infected by Pseudomonas aeruginosa. J. Clin. Microbiol. 33:924–929.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

24. Jensen, S. E., I. T. Fecycz, and J. N. Campbell. 1980. Nutritional factorscontrolling extracellular protease production by Pseudomonas aeruginosa. J.Bacteriol. 144:844–847.

25. Jin, S., K. S. Ishimoto, and S. Lory. 1994. PilR, a transcriptional regulator ofpiliation in Pseudomonas aeruginosa, binds to a cis-acting sequence upstreamof the pilin gene promoter. Mol. Microbiol. 14:1049–1057.

26. Johnson, M. K., and D. Boese-Marrazzo. 1980. Production and properties ofheat-stable extracellular hemolysin from Pseudomonas aeruginosa. Infect.Immun. 29:1028–1033.

27. Kaplan, H. B., A. Eberhard, C. Widrig, and E. P. Greenberg. 1985. Synthesisof N-[3-oxo-(4,5-3H2)-hexanoyl] homoserine lactone: biologically active tri-tium-labeled Vibrio fischeri autoinducer. J. Labelled Cmpd. Radiopharm.22:387–395.

28. Kaplan, H. B., and E. P. Greenberg. 1985. Diffusion of autoinducer isinvolved in regulation of the Vibrio fischeri luminescence system. J. Bacteriol.139:1210–1214.

29. Koch, A. K., O. Kappeli, A. Fiechter, and J. Reiser. 1989. Hydrocarbonassimilation and biosurfactant production in Pseudomonas aeruginosa mu-tants. J. Bacteriol. 173:4212–4219.

30. Kuo, A., S. M. Callahan, and P. V. Dunlap. 1996. Modulation of lumines-cence operon expression by N-octanoyl-L-homoserine lactone in ainS mu-tants of Vibrio fischeri. J. Bacteriol. 178:971–976.

31. Latifi, A., M. Foglino, K. Tanaka, P. Williams, and A. Lazdunski. 1996. Ahierarchical quorum-sensing cascade in Pseudomonas aeruginosa links thetranscriptional activators LasR and RhlR to expression of the stationary-phase sigma factor RpoS. Mol. Microbiol. 21:1137–1146.

32. Liu, P. V., Y. Abe, and J. L. Bates. 1961. The roles of various fractions ofPseudomonas aeruginosa in its pathogenesis. J. Infect. Dis. 108:218–228.

33. Merrick, M. J. 1993. In a class of its own—the RNA polymerase sigma factors54 (sN). Mol. Microbiol. 10:903–909.

34. Miller, J. H. 1972. Experiments in molecular genetics, p. 352–355. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

35. Miller, V. L., and J. J. Mekalonos. 1988. A novel suicide vector and its usein construction of insertion mutations: osmoregulation of outer membraneproteins and virulence determinants in Vibrio cholerae requires toxR. J.Bacteriol. 170:2575–2583.

36. Morihara, K., and J. Y. Homma. 1985. Pseudomonas protease, p. 41–79. InI. A. Holder (ed.), Bacterial enzymes and virulence. CRC Press, Boca Raton,Fla.

37. Ochsner, U. A. 1993. Genetics and biochemistry of Pseudomonas aeruginosarhamnolipid biosurfactant synthesis. Doctoral dissertation. Swiss FederalInstitute of Technology, Zurich, Switzerland.

38. Ochsner, U. A., A. K. Koch, A. Fiechter, and J. Reiser. 1994. Isolation andcharacterization of a regulatory gene affecting rhamnolipid biosurfactantsynthesis in Pseudomonas aeruginosa. J. Bacteriol. 176:2044–2054.

39. Ochsner, U. A., A. Fiechter, and J. Reiser. 1994. Isolation, characterization,and expression in Escherichia coli of the Pseudomonas aeruginosa rhlABgenes encoding a rhamnosyltransferase involved in rhamnolipid biosurfac-tant synthesis. J. Biol. Chem. 269:19787–19795.

40. Ochsner, U. A., and J. Reiser. 1995. Autoinducer-mediated regulation ofrhamnolipid biosurfactant synthesis in Pseudomonas aeruginosa. Proc. Natl.Acad. Sci. USA 92:6424–6428.

41. Ohman, D. E., S. J. Cryz, and B. H. Iglewski. 1980. Isolation and character-ization of a Pseudomonas aeruginosa PAO mutant that produces alteredelastase. J. Bacteriol. 142:836–842.

42. Olsen, R. H., G. DeBusscher, and W. R. McCombie. 1982. Development ofbroad-host-range vectors and gene banks: self-cloning of the Pseudomonasaeruginosa PAO chromosome. J. Bacteriol. 150:60–69.

43. Passador, L., J. Cook, M. J. Gambello, L. Rust, and B. H. Iglewski. 1993.Expression of Pseudomonas aeruginosa virulence genes requires cell-to-cellcommunication. Science 260:1127–1130.

44. Passador, L., and T. Linn. 1989. Autogenous regulation of the RNA poly-merase b subunit of Escherichia coli occurs at the translational level in vivo.J. Bacteriol. 171:6234–6242.

45. Passador, L., K. D. Tucker, K. R. Guertin, M. P. Journet, A. S. Kende, andB. H. Iglewski. 1996. Functional analysis of the Pseudomonas aeruginosaautoinducer PAI. J. Bacteriol. 178:5995–6000.

46. Pearson, J. P., K. M. Gray, L. Passador, K. D. Tucker, A. Eberhard, B. H.Iglewski, and E. P. Greenberg. 1994. Structure of the autoinducer requiredfor expression of Pseudomonas aeruginosa virulence genes. Proc. Natl. Acad.Sci. USA 91:197–201.

47. Pearson, J. P., L. Passador, B. H. Iglewski, and E. P. Greenberg. 1995. Asecond N-acyl homoserine lactone signal produced by Pseudomonas aerugi-nosa. Proc. Natl. Acad. Sci. USA 92:1490–1494.

48. Pesci, E. C., J. P. Pearson, P. C. Seed, and B. H. Iglewski. 1997. Regulationof las and rhl quorum sensing in Pseudomonas aeruginosa. J. Bacteriol.179:3127–3123.

49. Pollack, M. 1990. Pseudomonas aeruginosa, p. 1673–1691. In G. L. Mandell,R. G. Douglas, and J. E. Bennett (ed.), Principles and practices of infectiousdiseases. Churchill Livingstone, Edinburgh, Scotland.

50. Rust, L., E. C. Pesci, and B. H. Iglewski. 1996. Analysis of the Pseudomonasaeruginosa elastase (lasB) regulatory region. J. Bacteriol. 178:1134–1140.

51. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

52. Sastry, P. A., B. P. Finlay, B. L. Pasloske, W. Paranchych, J. R. Pearlstone,and L. B. Smillie. 1985. Comparative studies of the amino acid sequences ofpilin derived from Pseudomonas aeruginosa PAK and PAO. J. Bacteriol.164:571–577.

53. Seed, P. C., L. Passador, and B. H. Iglewski. 1995. Activation of the Pseudo-monas aeruginosa lasI gene by LasR and the Pseudomonas autoinducer PAI:an autoinduction regulatory hierarchy. J. Bacteriol. 177:654–659.

54. Shadel, G. S., and T. O. Baldwin. 1991. The Vibrio fischeri LuxR protein iscapable of bidirectional stimulation of transcription and both positive andnegative regulation of the luxR gene. J. Bacteriol. 173:568–574.

55. Siegmund, I., and F. Wagner. 1991. New method for detecting rhamnolipidsexcreted by Pseudomonas aeruginosa during growth on mineral agar. Bio-technol. Tech. 5:265–268.

56. Silhavy, T. J., M. L. Berman, and L. Enquist. 1984. Experiments with genefusions, p. 217. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

57. Simon, H., and O. Berngruber. 1968. Mechanistic studies on the catalytichydrogenation of a, b unsaturated carbonyl compounds with platinum orpalladium, based on intramolecular tritium distribution. Tetrahedron Lett.45:4711–4714.

58. Sitnikov, D., J. B. Schineller, and T. O. Baldwin. 1995. Transcriptionalregulation of bioluminescence genes from Vibrio fischeri. Mol. Microbiol.17:801–812.

59. Smith, A. W., and B. H. Iglewski. 1989. Transformation of Pseudomonasaeruginosa by electoporation. Nucleic Acids Res. 17:10509.

60. Smith, J. J., S. Travis, E. P. Greenberg, and M. J. Welsh. 1996. Cystic fibrosisairway epithelia fail to kill bacteria because of abnormal airway surface fluid.Cell 85:229–236.

61. Stevens, A. M., K. M. Dolan, and E. P. Greenberg. 1994. Synergistic bindingof the LuxR transcriptional activator domain and RNA polymerase to the luxpromoter region. Proc. Natl. Acad. Sci. USA 91:12619–12623.

62. Stevens, A. M., and E. P. Greenberg. 1997. Quorum-sensing in Vibrio fischeri:essential elements for activation of the luminescence genes. J. Bacteriol.179:557–562.

63. Stoker, N. G., N. F. Fairweather, and B. G. Spratt. 1982. Versatile low-copynumber plasmid vectors for cloning in Escherichia coli. Gene 18:335–341.

64. Tang, H. B., E. Dimango, R. Bryan, M. J. Gambello, B. H. Iglewski, J. B.Goldberg, and A. Prince. 1996. Contribution of specific Pseudomonas aerugi-nosa virulence factors to pathogenesis of pneumonia in a neonatal mousemodel of infection. Infect. Immun. 64:37–43.

65. Toder, D. S., M. J. Gambello, and B. H. Iglewski. 1991. Pseudomonas aerugi-nosa LasA: a second elastase under the transcriptional control of lasR. Mol.Microbiol. 5:2003–2010.

66. Toder, D. S., S. J. Ferrell, J. L. Nezezon, L. Rust, and B. H. Iglewski. 1994.lasA and lasB genes of Pseudomonas aeruginosa: analysis of transcription andgene product activity. Infect. Immun. 62:1320–1327.

67. Vogeli, G., and P. S. Kaytes. 1987. Amplification, storage, and replication oflibraries. Methods Enzymol. 152:407–415.

68. Winson, M. K., M. Camara, A. Latifi, M. Foglino, S. R. Chhabra, M. Daykin,M. Bally, V. Chapon, G. P. C. Salmond, B. W. Bycroft, A. Lazdunski,G. S. A. B. Stewart, and P. Williams. 1995. Multiple N-acyl-L-homoserinelactone signal molecules regulate production of virulence determinants andsecondary metabolites in Pseudomonas aeruginosa. Proc. Natl. Acad. Sci.USA 92:9427–9431.

69. Woods, D. E., P. A. Sokol, L. E. Bryan, D. G. Storey, M. J. Mattingly, H. J.Vogel, and H. Ceri. 1991. In vivo regulation of virulence in Pseudomonasaeruginosa associated with genetic rearrangement. J. Infect. Dis. 163:143–149.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Documents

Roles of Pseudomonas aeruginosa las and rhl Quorum-Sensing ... · Pseudomonas aeruginosa is an opportunistic human patho-gen that infects immunocompromised individuals and people