Pseudomonas aeruginosa AmpR Is a Global Transcriptional Factor That Regulates Expression of AmpC and PoxB {beta}-Lactamases, Proteases, Quorum Sensing, and Other Virulence Factors

In members of the family Enterobacteriaceae, ampC, which encodesa ß-lactamase, is regulated by an upstream, divergentlytranscribed gene, ampR. However, in Pseudomonas aeruginosa,the regulation of ampC is not understood. In this study, wecompared the characteristics of a P. aeruginosa ampR mutant,PAOampR, with that of an isogenic ampR⁺ parent. The ampR mutationgreatly altered AmpC production. In the absence of antibiotic,PAOampR expressed increased basal ß-lactamase levels.However, this increase was not followed by a concomitant increasein the P_ampC promoter activity. The discrepancy in protein andtranscription analyses led us to discover the presence of anotherchromosomal AmpR-regulated ß-lactamase, PoxB. We foundthat the expression of P. aeruginosa ampR greatly altered theß-lactamase production from ampC and poxB in Escherichiacoli: it up-regulated AmpC but down-regulated PoxB activities.In addition, the constitutive P_ampR promoter activity in PAOampRindicated that AmpR did not autoregulate in the absence or presenceof inducers. We further demonstrated that AmpR is a global regulatorbecause the strain carrying the ampR mutation produced higherlevels of pyocyanin and LasA protease and lower levels of LasBelastase than the wild-type strain. The increase in LasA levelswas positively correlated with the P_lasA, P_lasI, and P_lasR expression.The reduction in the LasB activity was positively correlatedwith the P_rhlR expression. Thus, AmpR plays a dual role, positivelyregulating the ampC, lasB, and rhlR expression levels and negativelyregulating the poxB, lasA, lasI, and lasR expression levels.

INTRODUCTION

Top

Introduction

Pseudomonasaeruginosa is an important opportunistic pathogencausing infections in patients suffering from cystic fibrosis,chronic obstructive pulmonary disease, or severe burns and inpatients in intensive care units. It poses an overwhelming threatto these individuals, leading to a high risk of morbidity andmortality. The current treatment against P. aeruginosa usuallyincludes a combination of ß-lactam antibiotics andaminoglycosides (20). However, resistance to these antibioticsis not uncommon, resulting in the failure of chemotherapy. P.aeruginosa also possesses various mechanisms to counteract antimicrobialagents. These include inducible AmpC ß-lactamase (33),plasmid- and transposon-encoded acquired ß-lactamases(12), multidrug efflux pumps (17), low outer membrane permeability(53), the ability to grow in biofilms (7), and hypermutability(41). The rapid emergence of resistance to ß-lactamantibiotics in P. aeruginosa isolates from cystic fibrosis patientsin the early 1990s has been attributed to the in vivo selectionof constitutive ß-lactamase-producing strains (6,15, 16). Overproduction of ß-lactamase in some ofthe P. aeruginosa isolates has been ascribed to mutations inampD encoding a negative regulator of AmpC (2, 28), but geneticevents underlying overproduction of AmpC ß-lactamasestill remain to be fully elucidated.

Many members of the Enterobacteriaceae family have genes thatcode for a chromosomal Ambler class C ß-lactamase(38). The analysis of chromosomally mediated ß-lactamasesrevealed that, in species possessing an ampR gene, synthesisof ß-lactamase is inducible, whereas in strains withoutampR, synthesis is constitutive (44, 50). The ampR genes inall the species studied thus far are located immediately upstreamof ampC and are transcribed divergently (18). The short intercistronicregion of Citrobacter freundii ampR-ampC accommodates two overlappingpromoters (3, 31). The P. aeruginosa ampC and ampR genes havesimilar gene organization (32, 33). However, the mechanism ofampR regulation in P. aeruginosa is not yet clear.

To address the role of P. aeruginosa ampR, we generated an ampRmutant in the prototypic, wild-type PAO1 strain. Our analysisof PAO1 and PAOampR showed that a mutation in ampR results ina high constitutive expression of ß-lactamases andthat AmpR is required for AmpC induction. In addition, AmpRmodulates expression of yet another chromosomally encoded ß-lactamase,extracellular proteases, and pyocyanin. Subsequent analysessuggest that AmpR affects the virulence factors through quorumsensing.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains, plasmids, and media. Bacterial strains, plasmids, and primers used in this studyare shown in Table 1. Escherichia coli and P. aeruginosa wereroutinely cultured in Luria-Bertani medium (10 g tryptone, 5g yeast extract, 5 g NaCl, per liter). Pseudomonas isolationagar (Difco) was used in triparental mating experiments. Antibioticswere used at the following concentrations (per milliliter) unlessotherwise indicated: ampicillin at 50 µg, tetracyclineat 20 µg, and gentamicin at 15 µg for E. coli andcarbenicillin at 300 µg, gentamicin at 300 µg, andtetracycline at 60 µg for P. aeruginosa. For induction,500 µg/ml of benzylpenicillin was used.

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

DNA manipulations. All molecular techniques were performed according to standardprotocols (1).

Cloning of ampC and poxB genes. Primers were designed to PCR amplify ampC and poxB. SBJ05ampCForand SBJ06ampCRev were used to generate a 1,220-bp fragment withampC, while KKF28poxBFor and KKF30poxBRev generated a 2,170-bpfragment containing poxB (Table 1). These PCR amplicons wereelectrophoresed and purified according to manufacturer's instructions(QIAGEN, California). The cleaned products were ligated intopCRII-TOPO and transformed into TOP10F' (Invitrogen, California).Because we were concerned that the presence of the bla genein the pCRII-TOPO vector might interfere with data interpretation,an internal 712-bp DraI fragment of the bla gene was deleted,and the resulting vector was transformed into E. coli DH5 ${alpha}$ . Selectionwas carried out using kanamycin, as this resistance marker wasalso present in pCRII-TOPO. The plasmids pSJ02 and pKKF675 harboredthe ampC and poxB genes, respectively. The negative control,plasmid pKKF0679, contained a self-ligated pCRII-TOPO in whichthe bla gene was deleted (Table 1).

Insertional inactivation of the ampR gene. A 1,220-bp ampR fragment was PCR amplified using SBJ01ampRForand SBJ02ampRRev with flanking EcoRI and BamHI sites, respectively(Table 1). The PCR product was ligated to an EcoRI-BamHI-digestedpGEMEX-1 (Promega, California), generating pSJ01 (Fig. 1). Agentamicin cassette, aacCI, was retrieved from pUCGm (47) andinserted into the PstI site of ampR (Fig. 1). This disruptedthe reading frame of ampR in pSJ01. Subsequently, ampR::Gm wassubcloned as an EcoRI-HindIII-cut, blunt-ended fragment intothe SmaI site of pEX100T (48), which is a mobilizable suicideplasmid, creating pSJ07. This plasmid was conjugated into P.aeruginosa PAO1, with a helper strain harboring pRK2013 (11).The merodiploids, resulting from homologous recombination, wereselected with pseudomonas isolation agar containing gentamicin.Gm^r colonies were then screened for gentamicin resistance andcarbenicillin sensitivity by replica plating. The insertionwas confirmed by PCR and restriction analysis on the PCR product(data not shown). The PAO1 isogenic strain with defective ampR,henceforth referred to as PAOampR, was named PKM300.

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. Physical map of ampR-ampC loci (A) and plasmids (B). (A) The restriction map of the region is based on the PAO1 genome sequence with relevant restriction sites. The 5' end of ampC (labeled ampC') and the entire ampR open reading frames are indicated. The 1.2-kb fragment has the PAO1 genome coordinates 4594088 to 4592869. SBJ03 and SBJ04 show the position where the primers annealed for the promoter amplification. The striped box at positions 197 to 207 is where ligand-bound AmpR is postulated to bind. (B) The horizontal lines below the map show the amount of PAO1 genome DNA present in each plasmid. The restriction sites, EcoRI and BamHI, flanking the ends were introduced in the primers used for PCR (Table 1). Plasmids pSJ01 and pSJ06 are derivatives of pGEMEX (Promega) and pME6030 (19), respectively. The plasmid pSJ06 is used for complementation analysis and is referred to as pAmpR. The gentamicin cassette ( ${blacktriangledown}$ ) was inserted into the SalI site of pSJ01, creating pSJ05. The ampR::aacCI (ampR::Gm) fragment was subcloned into the suicide vector pEX100T (48) to generate pSJ07, which was introduced into PAO1 to generate PAOampR by homologous recombination. A 342-bp fragment containing the promoters was cloned in both directions, upstream of promoterless lacZ (indicated by the striped arrows) in mini-CTX-lacZ (4) to generate the transcriptional fusions for ampC (pSJ10) and ampR (pSJ11) promoters. These fusions were introduced into PAO1 and PAOampR by site-specific recombination using the attP site (4).

Construction of a plasmid with intact ampR for complementation. Plasmid pSJ06 was generated by subcloning an EcoRI-BamHI-cutfragment from pSJ01 containing ampR into an EcoRI-BamHI-cut,broad-host-range, low-copy-number plasmid, pME6030 (19). Thisplasmid, pSJ06, also referred to as pAmpR, was used in all complementationanalyses by conjugation into PAOampR.

Gene fusions with the lacZ gene. A 330-bp fragment was PCR amplified with the flanking primersSBJ03ampCRFor and SBJ04ampCRRev (Table 1) and then cloned intopGEMEX-1, generating pSJ09 (Fig. 1). This fragment containedthe ampC-ampR intergenic region with the putative promoters(33). To avoid PCR artifacts, pSJ09 was sequenced by using M13primers. The fragment containing the promoters was subclonedas an EcoRI-BamHI fragment into an EcoRI-BamHI-cut, promoterlesslacZ, mini-CTX-lacZ reporter plasmid (4), creating pSJ10 (P_ampC-lacZ)and pSJ11 (P_ampR-lacZ). This suicide vector contained the integration-proficientattP site, which recombines into the chromosomal attB site togenerate a single-copy reporter fusion (4). The resulting cloneswere mobilized into PAO1 and PAOampR (Table 1). The presenceof the chromosomal insertions was confirmed by PCR and restrictionanalysis of the product.

Pyocyanin quantitation assay. The pyocyanin assay is based upon the absorbance of excretedpyocyanin at 520 nm in acidic solution (9). A 5-ml supernatantfrom a stationary-phase culture ( $~$ 16 h) in LB broth was mixedwith 3 ml of chloroform. The pyocyanin from the chloroform phasewas then extracted into 1 ml of 0.2 N HCl, giving it a pinkto deep red color, indicating the presence of pyocyanin. Theabsorbance was measured at 520 nm. Concentrations, expressedas micrograms of pyocyanin produced per milliliter of culturesupernatant, were determined by multiplying the optical densityat 520 nm (OD₅₂₀) by 17.072 (9).

LasA staphylolytic assay. LasA protease activity was measured by determining the abilityof stationary-phase P. aeruginosa culture supernatants to lyseboiled Staphylococcus aureus (25). A 30-ml volume of an overnightS. aureus culture was boiled for 10 min and then centrifugedfor 10 min at 10,000 x g. The resulting pellet was resuspendedin 10 mM Na₂PO₄ (pH 4.5) to an OD₆₀₀ of approximately 0.8. A100-µl aliquot of bacterial supernatant was added to 900µl of S. aureus suspension, and the OD₆₀₀ was determinedafter 0, 5, 10, 15, 20, 30, 45, and 60 min (8, 25).

LasB elastolytic assay. The elastolytic activity of bacterial suspension was determinedwith the elastin Congo red (ECR; Sigma, St. Louis, MO) assay(40). A 100-µl aliquot of bacterial supernatant of 16-hculture was added to 900 µl of ECR buffer (100 mM Tris,1 mM CaCl₂, pH 7.5) containing 20 mg of ECR and then incubatedwith shaking at 37°C for 3 h. Insoluble ECR was removedby centrifugation, and the absorption of the supernatant wasmeasured at 495 nm. LB medium was used as a negative control.Activity was expressed as change in OD₄₉₅ per µg protein.

ß-Lactamase assay. The assay of the P. aeruginosa chromosomal ß-lactamasewas modified based on previously published protocols (39). Briefly,stationary-phase cultures were diluted 1:100 into 10 ml of LBbroth and incubated with shaking at 37°C until an OD₆₀₀of 0.6 to 0.8 was attained. For induction (500 µg/ml benzylpenicillin),the cultures were incubated for an additional 3 hours beforeharvesting. The cells were washed once with 50 mM sodium phosphatebuffer and were then resuspended to a final volume of 2 ml ofthe same buffer. Following disruption of cells on ice with sonication,the cell lysate was centrifuged at 10,000 x g for 30 min at4°C and the supernatant was retained. A 2-µl portionof the ß-lactamase-containing supernatant was addedto 1 ml (final volume) of assay buffer containing nitrocefin(final concentration, 100 µM) at room temperature. Thereaction was allowed to take place for 20 min before nitrocefinhydrolysis was measured spectrophotometrically at 482 nm. Concurrently,the total protein was determined with the Bradford assay usingthe same supernatant. One milliunit of ß-lactamaseis defined as 1 nanomole of nitrocefin hydrolyzed per minuteper microgram of protein.

ß-Galactosidase assay. Assays for ß-galactosidase in P. aeruginosa were performedas previously described (34).

Statistical analysis. All data were analyzed with one-way analysis of variance usingthe statistical software package SPSS (Chicago, IL).

RESULTS AND DISCUSSION

Top

Results and Discussion

Typically the presence of ß-lactam antibiotics ina medium induces the synthesis of chromosome-encoded AmpC ß-lactamase.The regulator of ampC, the divergently transcribed ampR geneproduct, has been identified in more than 10 species (18). P.aeruginosa AmpR has a high sequence identity with the AmpR fromC. freundii and Enterobacter cloacae (58.3 and 62.5% identity,respectively) (33). Because the three species have the sameampR and ampC gene organization in the chromosome, it has beenassumed that they also use the same regulatory mechanism. Todate, the role of AmpR is restricted to the induction of ampC.The goal of this study was to examine the role of P. aeruginosaAmpR by generating an isogenic ampR mutant in the prototypicP. aeruginosa strain PAO1 (PAOampR).

P. aeruginosa AmpR does not autoregulate its own promoter. AmpR is a DNA-binding protein that belongs to the LysR familyof regulatory proteins (22, 29, 31). Many LysR-type transcriptionalregulators act as transcriptional activators of the target genesbut repress their own expression (45). In addition, the P. aeruginosaampR-ampC intergenic region is only 159 bp long, leading tothe suggestion that the AmpR expression may be autoregulated(32). To address this, we constructed strains with single copiesof the ampR promoter upstream of a promoterless reporter gene,lacZ. The P_ampR-lacZ fusion was integrated into the PAO1 andthe PAOampR chromosomes via attB-attP site-specific recombination,thus mimicking the chromosomal regulation (4). As a negativecontrol, the promoterless lacZ gene was also integrated in bothPAO1 and PAOampR. In the absence of the inducer, PAO1 showeda basal level of expression from the ampR promoter, which wasnot significantly affected by the presence of the inducer (Table2). This finding concurs with previous C. freundii ampR studiesshowing that the amount of C. freundii AmpR expressed in E.coli minicells remains constant in the absence and presenceof inducer (30), that the constitutive ampR expression is lowwith a short half-life of mRNA (22), and that the ampR transcriptionand the level of AmpR remain unaffected in the absence and presenceof antibiotics (31). Expression of P_ampR-lacZ in the PAOampRmutant was not significantly different, suggesting that ampRdoes not regulate its own transcription (Table 2).

View this table:
[in this window]
[in a new window]
TABLE 2. ß-Galactosidase activities of ampC and ampR promoters

P. aeruginosa AmpR is a positive regulator of ampC expression. Class C ß-lactamase has long been implicated as theculprit in the cephalosporin resistance of P. aeruginosa (5).Based on previously published work in other systems (22, 30),we postulated that the mutation in ampR would increase the activityof the ampC promoter. To test this hypothesis, we generateda P_ampC-lacZ reporter and integrated it as a single-copy fusioninto the PAO1 and the PAOampR chromosomes via attB-attP site-specificrecombination, thus mimicking the chromosomal regulation (4).As a negative control, the promoterless lacZ gene was also integratedin both PAO1 and PAOampR. In PAO1, little activity was synthesizedfrom the ampC promoter (Table 2). As expected, the presenceof benzylpenicillin resulted in a significant elevation in theampC promoter activity in PAO1 (Table 2). However, no inductionof the ampC promoter was detectable in PAOampR in the absenceor presence of inducer. The significant loss of transcriptionalinduction of the ampC promoter (Table 2) was likely due to theabsence of AmpR, suggesting that AmpR is a positive regulatorof ampC expression in the presence of an inducer. This is inagreement with previous studies of C. freundii and E. cloacaedemonstrating that AmpR modulates ampC expression. The clonedampC genes from C. freundii and E. cloacae are activated inE. coli only if accompanied by their ampR gene (22, 30).

AmpR exhibits differential regulation on AmpC and PoxB. To ascertain the role of AmpR in the expression of ß-lactamase,the activities in both PAO1 and the isogenic ampR mutant werecompared. The parental PAO1 strain showed a low noninduced levelof activity, which was increased 11-fold by addition of theß-lactam agent (Table 3). However, PAOampR expresseda high ß-lactamase level, 12-fold higher than thatof the parental strain in the absence of inducer (Table 3).No further induction was exhibited in the presence of a ß-lactamagent. The inducible phenotype was restored in the PAOampR mutantby complementation with the plasmid pSJ06, harboring ampR.

View this table:
[in this window]
[in a new window]
TABLE 3. Expression of ß-lactamases in P. aeruginosa and E. coli

However, based on the P_ampC-lacZ expression assay (Table 2),we expected the AmpC ß-lactamase levels to remainthe same in the absence of ampR (Table 3). This discrepancyin the biochemical and transcriptional analyses of ampC andthe characterization of a P. aeruginosa ampC mutant led to thediscovery of a novel chromosomally encoded class D ß-lactamase,encoded by poxB (27). This class D ß-lactamase wasindependently reported by Girlich et al., and the enzyme wasnamed OXA-50 (13). However, our analysis of the sequence arguesthat PoxB is distinct from OXA enzymes (27). This discoveryled us to propose that AmpR might also regulate poxB. The presenceof more than one copy of ß-lactamases in P. aeruginosacomplicated the analysis of the effects of AmpR on each of thesegenes in a homologous host model. Therefore, we cloned ampCand poxB genes into E. coli and analyzed their activity in theabsence and presence of the ampR gene.

In strain E. coli DH5 ${alpha}$ , AmpC expression was very low (Table 3).This is in contrast to what has been observed with C. freundiiAmpC activity (30). In the absence of ß-lactam antibioticand C. freundii ampR, an increase in C. freundii AmpC ß-lactamaseactivity in E. coli was observed, suggesting that C. freundiiAmpR is a negative regulator (30). However, in our analysis,the presence of ampR is required for ampC expression.

Introduction of P. aeruginosa ampR into E. coli strain DH5 ${alpha}$ failedto induce the endogenous ß-lactamase from the chromosomalampC (Table 3). However, when P. aeruginosa ampR was introducedalong with its cognate ampC into E. coli strain DH5 ${alpha}$ , the ß-lactamaseactivity was significantly elevated both in the absence (fourfold)and in the presence (65-fold) of inducer (Table 3), furthersupporting the notion that ampR is a positive regulator requiredfor ampC expression as well as induction.

In contrast to ampC, the activity of poxB was constitutivelyhigh in E. coli (Table 3). When AmpR was supplied in trans,poxB ß-lactamase activity significantly dropped 65%and 96% in the absence and presence, respectively, of inducer(Table 3). These data suggest that AmpR suppresses the expressionof PoxB.

It has been demonstrated that the increased resistance in clinicalisolates of P. aeruginosa is also due to the derepression ofß-lactamase (5, 14). Since the existence of PoxB wasnot known, and the biochemical assays used were not sensitiveenough to distinguish these two ß-lactamases, it waspresumed that AmpC was overexpressed in the hyperproducing clinicalisolates. A high constitutive expression of ß-lactamaseseen in PAOampR suggests that the clinical P. aeruginosa isolatesthat are hyperproducing ß-lactamase may harbor poxBand a mutation in ampR. A Southern blot analysis of a few ofthese clinical P. aeruginosa isolates with ampR- and ampC-specificprobes showed no gross changes in the ampR-ampC intergenic DNA(5). This led the authors to rule out any changes in ampR andampC genes, although this technique is not sensitive enoughto detect minor deletions or point mutations (5). In fact, oneof these P. aeruginosa ß-lactamase hyperproducerswas shown to have a point mutation in ampR (2). It is not clearif this clinical isolate harbors the poxB gene, though we havedemonstrated that poxB is fairly ubiquitous and is found inchromosomes of both clinical and environmental P. aeruginosaisolates (27).

Mutation in ampR affects pyocyanin production. A change in virulence factor levels was first macroscopicallyobserved when PAOampR produced a highly pigmented culture inmid-log phase, compared to the parent PAO1. Pyocyanin is a typicalgreen pigment produced by P. aeruginosa that is involved inmicrobial competitiveness, the suppression of soilborne plantpathogens, and virulence in human and animal hosts (35). Twoseven-gene phenazine biosynthetic operons, designated phzA1B1C1D1E1F1G1and phzA2B2C2D2E2F2G2, are involved in pyocyanin production(35). Besides the presence of a LasR binding site upstream ofthe phzA1 operon (35), little is known about the regulationof these operons.

In the parental strain PAO1, pyocyanin production was relativelylow and inducible on antibiotic challenge, albeit insignificantly(Table 4). However, pyocyanin production in the ampR mutantwas significantly higher in the absence and presence of theß-lactam agent (Table 4). The high constitutive pyocyaninlevel was reduced by ampR in trans from the low-copy-numberplasmid (Table 4). This suggests that AmpR negatively regulatespyocyanin production. It remains to be clarified if AmpR regulatesone or both of these phz operons.

View this table:
[in this window]
[in a new window]
TABLE 4. Effect of ampR mutation on pyocyanin production and LasA staphylolytic and LasB elastase activity

AmpR regulates staphylolytic protease LasA and LasB elastase. P. aeruginosa strains that overproduce pyocyanin are also knownto produce other extracellular virulence factors. P. aeruginosasecretes a number of proteases, which are believed to play amajor role in pathogenesis, especially during acute infections(26, 37). We determined that the mutation in ampR also affectsthe expression of two proteases, the LasB elastase and the LasAstaphylolytic protease.

LasA staphylolytic protease is a 20-kDa zinc metalloendopeptidasebelonging to the ß-lytic endopeptidase family of proteases(24). The PAOampR supernatant lysed the Staphylococcus aureuscells at a significantly higher rate than PAO1 (Table 4), indicatingan increased production of LasA protease in the absence of ampR.This phenotype can also be rescued by ampR expressed in trans.This suggests that AmpR negatively regulates LasA protease production.Since AmpR is a homolog of LysR, the differential expressionof LasA in PAOampR could be a result of the effect of AmpR onthe transcription of lasA. To address this, we introduced plasmidpLPA containing P_lasA-lacZ (43) into PAO1 and PAOampR (Table5). In PAOampR, the P_lasA expression was significantly increasedcompared to PAO1, suggesting that the increased transcriptionmay contribute to a higher production of LasA staphylolyticprotease. As expected, the induction was not influenced by thepresence of inducer. It is interesting that the lasA promoteris significantly up-regulated in the presence of inducer.

View this table:
[in this window]
[in a new window]
TABLE 5. Transcriptional activity of lasA, lasB, lasI, lasR, rhlI, and rhlR promoters

Elastase, a zinc metalloprotease encoded by lasB, is capableof inactivating a wide range of biological tissues and immunologicalagents (43). In the absence of inducer, there is a twofold inductionof elastase activity in PAO1. However, in the presence of inducer,the ampR mutation failed to induce the production of LasB elastase.This induction could be restored with ampR expressed in trans(Table 4). This suggests that AmpR positively regulates LasBactivity in the presence of inducer. To determine if the effectof AmpR is on the transcription of the lasB gene, we introducedplasmid pLPB, containing the P_lasB-lacZ fusion (43), into PAO1and PAOampR. The P_lasB expression did not differ significantlybetween PAO1 and PAOampR (Table 5). This suggests that AmpRdoes not affect P_lasB directly and that the noninducibilityof LasB protease activity is likely due to an indirect effectof AmpR.

Loss of ampR affects the expression of quorum sensing genes. The regulation of P. aeruginosa lasA and lasB genes has beenexplored extensively. Both of these genes are regulated by thelas and rhl quorum sensing systems (49). We wanted to determineif AmpR affects the production of protease indirectly via thequorum sensing systems. To address this, plasmids containingP_lasI-lacZ, P_lasR-lacZ, P_rhlI-lacZ, and P_rhlR-lacZ transcriptionalfusions were introduced into PAO1 and PAOampR (42). The analysesof these fusions in PAOampR(pAmpR) were not investigated. However,based on all the analyses above we predict that pAmpR will restorethe wild-type phenotype in the mutants.

In P. aeruginosa, there is a direct association between theproduction of LasA protease and the LasI-LasR quoromone system(43, 51). Since LasR positively regulates LasA expression, wepostulated that the higher production of LasA protease in thePAOampR strain may also be a direct result of the higher transcriptionof lasI and lasR genes (Table 5). In fact, there was a significantincrease in the expression of the las genes, suggesting thatAmpR is also a negative regulator of lasI and lasR, adding yetanother layer of complexity to the regulation of these genes(49). Similar elevation was seen in PAOampR with P_lasI-lacZand P_lasR-lacZ fusions in the presence of the inducer (Table5). This suggests that AmpR negatively regulates lasI and lasRtranscription. In addition, the presence of inducer in PAO1resulted in a significant (P < 0.05) decrease in P_lasR-lacZactivity. This loss of activity was recovered in the absenceof ampR, which further supports the idea that AmpR is a negativeregulator of lasR expression.

The rhl quorum sensing is also known to regulate the transcriptionof lasA and lasB genes. There was no significant change in theP_rhlI-lacZ fusion between the PAO1 and PAOampR strains (Table5). However, in PAOampR, we saw a significant reduction of rhlRtranscription compared to that in PAO1 either in the absenceor in the presence of inducer (Table 5), suggesting that AmpRis a positive regulator of rhlR expression.

Though the LasB elastase was slightly reduced in the ampR mutant(Table 4), there was no concomitant decrease in the expressionof the lasB promoter in PAOampR (Table 5). The expression ofLasB is more complicated. It involves the coordinate regulationof three quorum sensing systems, including lasR-lasI, rhlR-rhlI,and Pseudomonas quinolone signal (PQS) (36). Thus, AmpR mayact indirectly via the rhlR-rhlI and Pseudomonas quinolone signal(PQS) quorum sensing systems. It appears that the reductionof LasB activity seen in PAOampR is directly correlated withthe decrease of rhlR expression. These results suggest thatAmpR positively regulates rhlR expression and LasB activity.

Concluding remarks. (i) P. aeruginosa AmpR is a global regulator. E. coli naturally lacks ampR, and its native ampC gene is notinducible by ß-lactam antibiotics (22). Thus, mostof the studies address the role of ampR in E. coli using heterologoussystems in a multicopy plasmid background. The role of AmpRas a transcriptional activator of ß-lactamase production,and also as a global regulator of other virulence factors, wasa serendipitous conjecture drawn from the direct observationof PAOampR on LB plates and liquid culture that appeared tosynthesize higher levels of pyocyanin. In fact, our studiescomparing PAO1 with its isogenic ampR mutant revealed that AmpRpositively regulates the expression of ampC and rhlR and theactivities of AmpC ß-lactamase and LasB. In additionAmpR negatively regulates the expression of poxB, lasA, lasI,and lasR and the production of pyocyanin, PoxB ß-lactamase,and LasA. In fact this is typical of LysR-type transcriptionalregulators that are known for activating divergent transcriptionof linked target genes or unlinked regulons encoding extremelydiverse functions (45). However, this is the first report linkingAmpR to many diverse functions.

(ii) Mechanism of AmpR action. As in other members of the LysR family of regulatory proteins,the DNA-binding motif is found in the N terminus of the AmpRprotein (22, 29, 31). The binding of P. aeruginosa AmpR to theP. aeruginosa ampC promoter was suggested by the ability ofP. aeruginosa AmpR in the E. coli extract to immobilize theampC promoter-containing DNA fragment in a gel shift assay (33).The LysR-type proteins generally bind to inverted repeat motifs,including the T-N(11)-A motif (45). However, the binding motifsof C. freundii and E. cloacae AmpR have been determined in theampR-ampC intergenic regions (31). Using this consensus sequence(5'-TCTGCTGCTAAATTT), we found a strong conserved sequence (5'-TCTGCTCCAAATTT-3')within the ampR-ampC intergenic region of PAO1.

Since AmpR appears to regulate also other genes, we analyzedthe intergenic regions of all the AmpR-regulated genes for thepresence of its binding site. Since the transcription startsites for many of these genes have not been determined, we searchedfor a potential AmpR binding sequence 200 bp immediately upstreamfrom their open reading frames, both manually and using computersoftware such as AlignACE (23). No AmpR binding site was foundupstream of phzA1, rhlI, and rhlR. However, the putative sitesin other genes are relatively weak. A detailed promoter analysis,followed by footprinting with purified protein, is necessaryto further elucidate the significance of these sequences andthe mechanism of regulation.

Although much is known about the ubiquity of ampC-ampR, littleis understood about the regulation mechanism of ß-lactamase.The study presented here provides new insights into the roleof P. aeruginosa AmpR as a global regulator involved in antibioticresistance and the expression of many virulence factors. Thecombination of increased resistance to antibiotics and hyperproductionof pyocyanin and LasA protease would provide a mechanism ofprotection and survival in the harsh lung environment.

ACKNOWLEDGMENTS

This work has been supported by the NIH-MBRS SCORE (S06 GM08205;K.M.), the MACEE Fulbright Malaysian Scholar Program Award (C.L.K.),and Florida International University Teaching Assistantships(S.R.J., K.F.K., and S.D.I.).

We thank Barbara Iglewski for generously sharing all the promoter-fusionplasmids. We are grateful to Oana Ciofu and Joan Campbell forthe helpful discussion at the beginning of the project. We areindebted to Elaine Newman, Giri Narasimhan, and Erliang Zengfor their scientific consultation and the Mathee lab crew fortheir critical reading. We also thank Brooke Crandall for hereditorial assistance. We immensely appreciate the reviewersfor their critical review of the manuscript.

FOOTNOTES

* Corresponding author. Mailing address: Department of Biological Sciences, Florida International University, University Park, Miami, FL 33199. Phone: (305) 348-1261. Fax: (305) 348-2913. E-mail: matheek{at}fiu.edu.

Present address: National Institute of Education, Nanyang TechnologicalUniversity, Singapore 637616.

REFERENCES

Top