Co-regulation of {beta}-lactam resistance, alginate production and quorum sensing in Pseudomonas aeruginosa.

Development of β-lactam resistance, production of alginate and modulation of virulence factor expression that alters host immune responses are the hallmarks of chronic Pseudomonas aeruginosa infection in cystic fibrosis patients. In this study, we propose that a co-regulatory network exists between these mechanisms. We compared the promoter activities of ampR, algT/U, lasR, lasI, rhlR, rhlI and lasA genes, representing the β-lactam antibiotic resistance master regulatory gene, the alginate switch operon, the las and rhl quorum-sensing (QS) genes, and the LasA staphylolytic protease, respectively. Four isogenic P. aeruginosa strains, the prototypic Alg(-) PAO1, Alg(-) PAOampR, the mucoid Alg(+) PAOmucA22 (Alg(+) PDO300) and Alg(+) PAOmucA22ampR (Alg(+) PDOampR) were used. We found that in the presence of AmpR regulator and β-lactam antibiotic, the extracytoplasmic function sigma factor AlgT/U positively regulated P(ampR), whereas AmpR negatively regulated P(algT/U). On the basis of this finding we suggest the presence of a negative feedback loop to limit algT/U expression. In addition, the functional AlgT/U caused a significant decrease in the expression of QS genes, whereas loss of ampR only resulted in increased P(lasI) and P(lasR) transcription. The upregulation of the las QS system is likely to be responsible for the increased lasA promoter and the LasA protease activities in Alg(-) PAOampR and Alg(+) PDOampR. The enhanced expression of virulence factors in the ampR strains correlated with a higher rate of Caenorhabditis elegans paralysis. Hence, this study shows that the loss of ampR results in increased virulence, and is indicative of the existence of a co-regulatory network between β-lactam resistance, alginate production, QS and virulence factor production, with AmpR playing a central role.

INTRODUCTION

Pseudomonas aeruginosa, a ubiquitous, versatile saprophytic bacterium, is a major aetiological agent of nosocomial infections and the leading cause of mortality among patients with cystic fibrosis (CF) (Greenberg, 2000; Rahme ). This Gram-negative bacillus is equipped with an impressive arsenal of virulence factors to resist host defence mechanisms, counteract antibacterial agents, circumvent nutrient deprivation and withstand harsh environmental changes (Govan & Harris, 1986; Lyczak ; Pedersen, 1992). One distinctive feature of P. aeruginosa lung isolates of patients with advanced CF is that a higher proportion of them are mucoid as compared to those from other sites of infection (Doggett, 1969; Fick ). This mucoid phenotype is the result of an overproduction of the exopolysaccharide alginate (Evans & Linker, 1973). The activation of genes for alginate overproduction occurs primarily through the deregulation of algT/U or its product, σ22, a member of the extracytoplasmic function (ECF) sigma factors (DeVries & Ohman, 1994; Hershberger ; Martin ). Genomic, proteomic and microarray analyses have shown that AlgT/U regulates a diverse group of genes, ranging from extracellular proteases, periplasmic proteins like DsbA and intracellular enzymes (Firoved ; Firoved & Deretic, 2003; Malhotra ). Mucoid P. aeruginosa isolates from CF patients frequently have a defective mucA allele, a gene downstream of algT/U (Martin ). The mucA gene product is an anti-sigma factor that negatively regulates the activity of AlgT/U (Hughes & Mathee, 1998).

P. aeruginosa is intrinsically resistant to most β-lactam antibiotics. One of the factors contributing to the resistance is the existence of enzymes that can deactivate β-lactams, known as β-lactamases (Kong ; Rolinson, 1998). Two inducible chromosome-encoded β-lactamases, AmpC and PoxB (Oxa-50), have been identified in P. aeruginosa (Girlich ; Kong ; Lodge ). Expression of the ampC and poxB genes is tightly controlled by AmpR, a global LysR-like transcriptional regulator (Kong ). In addition, inactivation of ampR in the prototypic non-mucoid PAO1 (henceforth referred to as Alg− PAO1) resulted in high constitutive production of β-lactamases and pyocyanin, increased LasA staphylolytic protease activity and decreased LasB elastase activity (Kong ).

The production of virulence factors in P. aeruginosa is under the control of quorum-sensing (QS) systems mediated by diffusible chemical signalling molecules such as acylhomoserine lactones and quinolones. P. aeruginosa has three QS systems – las, rhl and Pseudomonas quinolone system that controls many virulence mechanisms (Ng & Bassler, 2009). Transcriptome studies have led to the identification of a large number of virulence factors that are under QS regulation in P. aeruginosa, these include proteases and toxins (Hentzer ; Schuster & Greenberg, 2006; Wagner ).

It has long been established that the production of proteases is inversely correlated with alginate production (Mathee ; Mohr ; Ohman & Chakrabarty, 1982). Previous comparison of Alg− PAO1 and its isogenic ampR mutant strain, Alg− PAOampR, showed differential regulation of virulence factors, including the las QS system (Kong ). In the present study, we sought to understand the regulatory network between alginate production, protease activity, β-lactam resistance and QS in P. aeruginosa. We hypothesized that AmpR may be differentially regulated in alginate-producing strains with consequent effects on the protease activities. To address this, ampR was inactivated in an alginate constitutive producer, Alg+ PDO300, generating an Alg+ PDOampR mutant strain. This mutant produced exceedingly high levels of β-lactamase, extracellular proteases and pyocyanin suggesting that AmpR either directly or indirectly suppresses the expression of many other virulence factors.

METHODS

Bacterial strains, plasmids and media.

Table 1 shows the bacterial strains, plasmids and primers used in this study. The bacterial strains of Escherichia coli and P. aeruginosa were routinely cultured in Luria–Bertani medium. Pseudomonas isolation agar (Difco) was used in triparental mating experiments for the selection of P. aeruginosa. Antibiotics, when used, were at the following concentrations unless indicated otherwise: ampicillin at 50 μg ml−1, tetracycline at 20 μg ml−1, gentamicin at 30 μg ml−1 for E. coli; and carbenicillin at 300 μg ml−1, gentamicin at 300 μg ml−1, tetracycline at 60 μg ml−1 for P. aeruginosa. For induction, 500 μg benzyl-penicillin ml−1 was used.

Table 1.

Bacterial strains, plasmids and primers used in this study

Strain/plasmid	Genotype	Reference
E. coli
DH5α	F−φ80dlacZΔM15 Δ(lacZYA–argF) U169 deoR recA1 endA1 hsdR17(rk−, mk+) phoA supE44 λ- thi-1 gyrA96 relA1	New England Biolabs
TOP10F′	F′[lacIq, Tn10(TetR)] mcrA Δ(mrr-hsdRMS-mcrBC)φ80dlacZΔM15 ΔlacX74 deoR recA1 araD139 D(ara-leu)7697 galU galK rpsL (StrR) endA1 nupG	Invitrogen
P. aeruginosa
PAO1	Prototype; Alg−	Holloway & Morgan (1986)
PDO300	PAOmucA22; Alg+	Mathee et al. (1999)
PKM805	PAOmucA22algT24-1; Alg−	PDOalgT; PDO300 derivative; Ramos et al. (2003)
PKM300	PAOampR : : aacCI; Alg−	PAOampR; Kong et al. (2005b)
PKM307	PAOmucA22 ampR : : aacCI; Alg+	PDOampR; this study
PKM308	PAOmucA22 attB : : PampC–lacZ; TcR, Alg+	PDO300 derivative; this study
PKM309	PAOmucA22 attB : : PampR–lacZ; TcR, Alg+	PDO300 derivative; this study
PKM310	PAOmucA22 ampR : : aacCI attB : : PampC-lacZ; GmR, TcR, Alg+	PDOampR derivative; this study
PKM311	PAOmucA22 ampR : : aacCI attB : : PampR–lacZ; GmR, TcR, Alg+	PDOampR derivative; this study
Plasmids
Mini-CTX-lacZ	TcR; integration-proficient vector for single-copy chromosomal lacZ fusion	Becher & Schweizer (2000)
pEX100T	ApR; sacB oriT	Schweizer & Hoang (1995)
pGEMEX-1	ApR; ColE1ori lacZα	Promega
pKMG37	ApR; pQF50 containing PalgT-lacZ transcriptional fusion	Mathee et al. (1997)
pLP170	ApR; lacZ transcriptional fusion vector that contains an RNase III splice sequence positioned between the MCS and lacZ	Preston et al. (1997)
pLPLA	ApR; pLP170 containing PlasA-lacZ transcriptional fusion	Preston et al. (1997)
pLPLB	ApR; pLP170 containing PlasB-lacZ transcriptional fusion	Preston et al. (1997)
pLPR1	ApR; pLP170 containing PrhlI-lacZ transcriptional fusion	Van Delden & Iglewski (1998)
pME6030	TcR; oriVpVS1oriVp15AoriT	Heeb et al. (2000)
pPCS223	ApR; pLP170 containing PlasI-lacZ transcriptional fusion	Van Delden & Iglewski (1998)
pPCS1001	ApR; pLP170 containing PlasR-lacZ transcriptional fusion	Pesci et al. (1997)
pPCS1002	ApR; pLP170 containing PrhlR-lacZ transcriptional fusion	Pesci et al. (1997)
pQF50	ApR; broad-host-range vector with promoterless lacZ	Farinha & Kropinski (1990)
pRK2013	KmR; ColE1ori-Tra (RK2)+	Figurski & Helinski (1979)
pSJ01	ApR; pGEMEX-1 with a 1220 bp EcoRI–BamHI flanked fragment containing ampR	Kong et al. (2005b)
pSJ06	ApR; pME6030 with a 1220 bp EcoRI–BamHI flanked fragment containing ampR (referred to as pAmpR)	Kong et al. (2005b)
pSJ07	ApR; pEX100T derivative with ampR : : aacCI	Kong et al. (2005b)
pSJ09	ApR, GmR; pGEMEX-1 with a 330 bp EcoRI–BamHI flanked fragment containing ampC-ampR intergenic region	Kong et al. (2005b)
pSJ10	TcR; CTX-lacZ fused with ampC promoter, PampC	Kong et al. (2005b)
pSJ11	TcR; CTX-lacZ fused with ampR promoter, PampR	Kong et al. (2005b)
pUCGm	ApR, GmR; pUC19 derivative containing gentamicin cassette	Schweizer (1993)
Primers
SBJ01ampRFor*	5′-GGAATTCTGGCGAACAGCAGTGTGGAAGCGG-3′
SBJ02ampRRev*	5′-CGGGATCCATTCCAATCACAACCCCAACGCC-3′
SBJ03ampCRFor*	5′-GGAATTCTGAGGCCGCGCGGCAGACGCTTGAACA-3′
SBJ04ampCRRev*	5′-CGGGATCCCATGAGGATTGGCGTCCTTTG-3′
DBS_QRTAmpRF	5′-CATTGGCCTTCATCACCGGTTGTA-3′
DBS_QRTAmpRR	5′-GGTTTCTCATGCAGCCCACGACAA-3′

*The italicized portion of the sequence indicates a restriction site in a PCR product prepared with the primer.

DNA manipulations.

All molecular techniques were performed according to standard protocols (Ausubel ).

Insertional inactivation of the ampR gene.

Inactivation of ampR in Alg+ PDO300 (PAOmucA22) was performed as previously reported using the same constructs (Kong ). The ampR : : aacCI fragment subcloned into pEX100T (Schweizer & Hoang, 1995) was introduced by conjugation into an alginate-overproducing P. aeruginosa, Alg+ PDO300 (Mathee ), with a helper strain harbouring pRK2013 (Figurski & Helinski, 1979). The merodiploids resulting from homologous recombination were selected with Pseudomonas isolation agar containing gentamicin. The colonies were then screened for gentamicin resistance and carbenicillin sensitivity by replica plating. The insertion was confirmed by PCR and restriction analysis of the PCR product. The Alg+ PDO300 isogenic strain with defective ampR (PAOmucA22ampR) is named Alg+ PDOampR (Table 1). Complementation studies were performed using plasmid pSJ06 that contains a PCR-amplified ampR on a low-copy-number, highly stable shuttle vector pME6030 to minimize the effects of gene dosage (Kong ). This plasmid is referred to as pAmpR.

Construction of promoter-lacZ fusions.

A 330 bp ampC–ampR intergenic region with the putative promoters was subcloned into the promoterless lacZ in the mini-CTX-lacZ reporter plasmid (Becher & Schweizer, 2000), creating pSJ10 (P) and pSJ11 (P) (Table 1) (Kong ). The resulting clones were mobilized into Alg+ PDO300 and Alg+ PDOampR (Table 1).

Quantitative real-time PCR (qPCR).

RNA extraction was performed with an RNeasy mini kit (Qiagen) following the manufacturer's protocols after treatment of cells with subMIC levels (200 μg ml−1) of penicillin G at OD600 0.6 for 1 h. The samples were stabilized with 5 % phenol/95 % ethanol mixture (pH 4.7) immediately after harvesting and during cell lysis (Brencic ). After determining RNA quantity spectrophotometrically (Beckman DU640; Beckman Coulter) and quality by denaturing agarose gel electrophoresis (Northern Max Gly; Ambion), cDNA was synthesized by annealing NS5 random primers to total purified RNA. Subsequent extension was carried out using SuperScript III reverse transcriptase (Invitrogen) (Brencic ). The cDNA was quantified and 10 ng cDNA was used per qPCR. We used the ABI 7500 cycler (Applied Biosystems) and Power SYBR Green PCR mastermix with ROX (Applied Biosystems) to test for expression of the ampR gene in these strains. The ATP-binding subunit clpX (PA1802) of the ATP-dependent protease was used as the internal control. Assays were performed in triplicate. Primer specificity was determined from dissociation profiles using melt curves. The cycling conditions for the qPCR were: 95 °C for 2 min (holding); 40 cycles of 95 °C for 15 s, 60 °C for 1 min (cycling); 95 °C for 15 s, 60 °C for 1 min (melt curve conditions).

Quantification of pyocyanin and LasA protease.

Extracellular pyocyanin was quantified as previously described (Kong ). LasA protease activity was measured by determining the ability of P. aeruginosa culture supernatants to lyse boiled Staphylococcus aureus, as described by Kessler .

β-Lactamase assay.

The assay of the P. aeruginosa chromosomal β-lactamase was performed as previously described using nitrocefin as the colorimetric substrate (Kong ).

β-Galactosidase assay.

Assays for β-galactosidase in P. aeruginosa were performed as previously described (Mathee ) and adapted into a high-throughput 96-well array (Griffith & Wolf, 2002).

P. aeruginosa–Caenorhabditis elegans paralysis assays.

The P. aeruginosa–C. elegans standard paralysis assay was modified from that of Gallagher & Manoil (2001). Bacterial cultures were grown overnight. A 1 : 1000 dilution was plated onto brain heart infusion agar plates. These plates were incubated for 18–24 h for the formation of bacterial lawns. Meanwhile, a synchronized culture of L4 stage larvae hermaphrodite Bristol N2 C. elegans was washed off an E. coli OP50-seeded nematode growth medium plate (1.7 % agar, 0.35 % peptone, 0.34 % K2HPO3, 0.3 % NaCl, 0.012 % MgSO4, 0.011 % CaCl2, 0.0005 % cholesterol). The worms were centrifuged at 1300  for 2 min and washed twice with M9 medium to remove residual E. coli bacteria. A total of 30 to 50 worms was then added to the P. aeruginosa bacterial lawns. Both live and paralysed worms were scored at 1, 2 and 4 h by microscopic observation. The analysis was performed in triplicate.

Statistical analysis.

All data were analysed with one-way ANOVA using the statistical software package spss (SPSS).

RESULTS

P expression in ampR mutants

We have previously reported that in the non-mucoid strain, AmpR positively regulates ampC expression but negatively controls the expression of poxB (Kong , b). To test whether such opposing controls remain true in the Alg+ background, strains were constructed with a single copy of the ampC promoter fused to a promoterless reporter gene, lacZ (P-lacZ). This was integrated into the Alg+ PDO300 and the Alg+ PDOampR chromosomes via attB-attP site-specific recombination, thus allowing mimicking of the chromosomal regulation. In the absence of inducer, the P-lacZ activity remained at a basal level in Alg+ PDO300 and Alg+ PDOampR strains (Table 2). A significant ninefold induction of the ampC promoter was observed in Alg+ PDO300 upon challenge with β-lactams (Table 2). However, the inducibility of the P was lost in Alg+ PDOampR.

Table 2.

β-Galactosidase activities of ampC and ampR promoters

Strain	PampC-lacZ (Miller units)		P value*	PampR-lacZ (Miller units)		P value*
	Non-induced	Induced		Non-induced	Induced
Alg− PAO1†	124.1±11.6	1644.2 ± 33.7	<0.05	77.1±8.7	123±1.2	ns
Alg− PAOampR†	113.2±7.5	122.3±7.4	ns	96.3±15.2	106.0±16.0	ns
P value‡	ns	<0.05		ns	ns
Alg+ PDO300	104.2±4.5	957.5±161.4	<0.05	79.5±26.3	332.8±14.3	<0.05
Alg+ PDOampR	142.4±4.9	143.8±6.9	ns	155.8±0.8	153.0±2.1	ns
P value§	ns	<0.05		ns	<0.05

ns, Not significant (P values >0.05).

*ANOVA compares the activity values between the presence (+) and absence (−) of inducers.

†These data are presented in a previous paper (Kong ; they are included here for comparison.

‡ANOVA compares the activity values between the Alg− PAO1 and the mutant Alg− PAOampR.

§ANOVA compares the activity values between the Alg+ PDO300 and the mutant Alg+ PDOampR.

Based on the above analysis, we expected to observe a loss of β-lactamase activity concomitant with the loss of ampR. However, the Alg+ PDOampR expressed a statistically significant sixfold higher β-lactamase compared to the parent Alg+ PDO300 in the absence of antibiotics (Fig. 1). No further induction was demonstrated in the presence of the inducer. This phenotype varied from the parental strain Alg+ PDO300, which showed only a threefold inducible phenotype (Fig. 1). The inducible phenotype was restored in Alg+ PDOampR mutant by complementation with pAmpR. The high β-lactamase activity in an ampR mutant has been shown previously to be due to the uninhibited expression of an oxacillinase poxB gene, rather than the elevated expression of ampC gene (Kong ).

Fig. 1.

β-Lactamase expression in the Alg+ PDOampR mutant. Assays were performed using the parent strain Alg+ PDO300, the mutant Alg+ PDOampR and Alg+ PDOampR (pAmpR) in the absence (−) and presence (+) of an inducer. The plasmid pAmpR carries the wild-type ampR gene on a broad-host-range low-copy-number plasmid pME6030 (Heeb ). Fresh cultures of OD600 0.6–0.8 were induced with 100 μg benzylpenicillin ml−1 for 3 h before harvesting. Assays were performed on sonicated lysate using nitrocefin as a chromogenic substrate. Assays were performed in triplicate. One Miller unit of β-lactamase is defined as 1 nmol nitrocefin hydrolysed min−1 (μg protein)−1.

ampR transcription in alginate-overproducing strains

The LysR family of transcriptional regulators is known to repress their own transcription as in the case of Citrobacter freundii AmpR (Lindquist ). However, we have previously reported that P. aeruginosa AmpR does not autoregulate in the prototypic strain Alg− PAO1 (Kong ). To determine if there is a change in the AmpR autoregulation in Alg+ strains, a single-copy fusion of P-lacZ was introduced at the attP site in Alg+ PDO300 and Alg+ PDOampR. In the absence of inducers, the ampR transcription remained at low levels in both strains (Table 2). In the presence of inducers, a significant increase in P expression was seen in Alg+ PDO300 (Table 2). Comparing the genotypes of the isogenic Alg− PAO1 and Alg+ PDO300 strains, this significant increase was likely due to the uninhibited activity of the ECF sigma factor AlgT/U in the latter. This suggests that AlgT/U activates the ampR promoter in the presence of inducers. Due to loss of ampR, no significant induction of P was seen in Alg+ PDOampR. In order to test AlgT/U regulation of ampR, mRNA levels of ampR were determined by qPCR with the Alg− PAO1, Alg+ PDO300 and Alg− PDOalgT strains. The Alg+ PDO300 strain showed an increase in the ampR mRNA levels (relative quantity of 2.1±0.2 compared to 1.0±0 in Alg− PAO1) indicating positive regulation of ampR by AlgT/U. Mutation in algT/U in Alg− PDOalgT led to a decrease in this expression (relative quantity 1.4±0.1 compared to 2.1±0.2 in Alg+ PDO300) supporting our hypothesis of positive regulation of ampR by AlgT/U and concurs with the transcriptional fusion assays. These results suggest that both AlgT/U and AmpR are required for the induction of the ampR promoter in the presence of inducers.

ampR mutation affects algT/U transcription

The loss of inducibility of ampR transcription in the Alg+ PDOampR background provided us with the first clue of the existence of a co-regulatory network involving β-lactam resistance and alginate production. To determine if this relationship is bidirectional, a P-lacZ fusion construct was introduced into Alg− PAO1, Alg+ PDO300 and the corresponding ampR mutant strains. As expected, the expression of algT/U promoter is constitutive in Alg− PAO1 and increased in Alg+ PDO300 (Fig. 2). Insertional inactivation of ampR in Alg− PAO1 and Alg+ PDO300 resulted in an approximately twofold increase in P activity in the absence of inducer. The effect of ampR mutation in Alg− PAOampR is the same as the known AlgT/U repressor mucA mutation (Alg+ PDO300) with respect to P expression, indicating negative regulation of P by AmpR. The AmpR-regulation of algT/U promoter in these strains was not significantly affected by β-lactam antibiotic. The consistent increase in P in the absence of ampR suggests that AmpR is a negative modulator of the ECF sigma factor, AlgT/U.

Fig. 2.

The effects of the ampR mutation on algT/U transcription. The promoter fusion P-lacZ was introduced into Alg− PAO1, Alg− PAOampR, Alg+ PDO300 and Alg+ PDOampR. Induction was carried out using 500 μg benzylpenicillin ml−1 and the β-galactosidase activity was determined in Miller units after 30 min incubation. The basal level of expression was detected in the promoterless lacZ vector, pLP170.

AlgT/U-dependent regulation of pyocyanin

Our quantitative analysis showed that the Alg+ PDO300 produced threefold less pyocyanin than Alg− PAO1 in the absence of β-lactam antibiotics (Table 3). This finding confirms that the AlgT/U sigma factor suppresses the production of pyocyanin. The presence of inducer resulted in an increase in pyocyanin production, albeit at low levels in Alg+ PDO300. However, the Alg+ PDOampR mutant produced a significantly high basal level of pyocyanin, which was inducible in the presence of β-lactam antibiotics (Table 3). Expressing ampR in trans in Alg+ PDOampR on a low-copy-number plasmid restored the phenotype to the parental strain, Alg+ PDO300 (data not shown). On the basis of this data we further argue that AmpR acts as a negative regulator of pyocyanin production.

Table 3.

Pyocyanin, LasA and P activities

Inducer*	Pyocyanin [μg (μg total protein)−1]†		LasA [ΔOD600 h−1 (μg protein)−1]‡		PlasA-lacZ (Miller units)§
	−	+	−	+	−
Alg− PAO1||	0.285±0.219	2.293±0.216¶	0.310±0.065	0.317±0.059	790.0±128.9
Alg− PAOampR||	2.934±0.761#	3.317±0.638	1.109±0.099#	0.951±0.045#	1970.5±312.6#
Alg+ PDO300	0.102±0.017**	0.324±0.051	0.135±0.011	0.147±0.024	246.5±26.5
Alg+ PDOampR	0.538±0.026††	2.518±0.640††	0.268±0.017	0.143±0.025	536.0±19.0

*Induction was carried out using 500 μg benzylpenicillin ml−1 for P. aeruginosa.

†Pyocyanin concentrations were expressed as μg pyocyanin produced (μg total protein)−1.

‡LasA activities were determined as the reduction of OD600 over a period of 1 h (μg total protein)−1.

§β-Galactosidase assays were performed in a high-throughput 96-well array and the results expressed in Miller units.

||These data are presented in a previous paper (Kong ; they are included here for comparison.

¶P<0.05 between non-induced and induced in the same strain.

#P<0.05 between Alg− PAO1 and Alg− PAOampR under the same conditions.

**P<0.05 between Alg− PAO1 and Alg+ PDO300 under the same conditions.

††P<0.05 between Alg+ PDO300 and Alg+ PDOampR under the same conditions.

LasA protease activity and lasA promoter expression in Alg+ PDOampR

The inverse relationship seen between alginate production and proteases is presumed to be AlgT/U-dependent (Mathee ; Mohr ; Ohman & Chakrabarty, 1982). Thus, a significant increase in algT/U expression in Alg+ PDOampR (Fig. 2) should result in downregulation of LasA protease expression. As expected, in comparison to the wild-type Alg− PAO1, Alg+ PDO300 produced 2.3-fold less LasA protease. However, loss of ampR resulted in a marginal increase in the LasA protease activity (Table 3) in an inducer-independent manner. To further confirm the above hypothesis, a P transcriptional fusion plasmid was introduced into Alg− PAO1, Alg+ PDO300 and the ampR mutant strains (Table 3). In concordance to the LasA activity analysis, the P levels were low in all mucoid strains, suggesting that the transcription of these promoters was suppressed (Table 3). Furthermore, the P fusion expression was increased twofold to threefold in Alg− PAOampR and Alg+ PDOampR, as compared to their respective parental strains (Table 3). These results suggest that AmpR is a negative regulator of lasA expression.

QS gene expression in mucoid ampR mutants

In line with previous observations (Kong ), we postulated that the slight increase of lasA expression in Alg+ PDOampR could be due to upregulation of the las system. To address this, all the four QS promoter fusions, P, P, P and P were introduced into Alg− PAO1, Alg− PAOampR, Alg+ PDO300 and Alg+ PDOampR. As we postulated, the Alg+ PDO300 exhibited significantly lower QS gene expression as compared to Alg− PAO1. There was no difference in the P expression in Alg+ PDO300 and Alg+ PDOampR (Fig. 3). However, the P activity was significantly increased in Alg+ PDOampR. Similar to the Alg− PAOampR, the loss of ampR in Alg+ PDO300 resulted in minimal alteration of P and P expression. Thus, AmpR negatively regulates lasI expression in alginate-overproducing strains.

Fig. 3.

The effects of the ampR mutation on QS las and rhl gene transcription. The alteration in the transcription of the QS systems in Alg− PAO1 (hatched bars), Alg− PAOampR (grey bars), Alg+ PDO300 (white bars) and Alg+ PDOampR (black bars) was monitored using four transcriptional fusions, P, P, P and P. The promoterless lacZ vector has a low basal level of activity of <20 Miller units.

Role of AmpR in virulence

The nematode C. elegans has been used as a bacterial pathogenesis model for the determination of virulence in P. aeruginosa (Gallagher & Manoil, 2001; Sifri ; Tan ). This simple host–pathogen interaction model was used to ascertain the virulence of Alg− PAO1, Alg− PAOampR, Alg+ PDO300 and Alg+ PDOampR. As expected, there was no observable paralysis in the negative control (E. coli OP50) plates (Fig. 4) and during the first hour of incubation with all the four isogenic P. aeruginosa strains. Consistent with the molecular and biochemical data, Alg− PAOampR paralysed C. elegans at a significantly (P<0.05) faster rate than the wild-type Alg− PAO1 (Fig. 4). The lowest survival was seen at the second hour, 19 % with Alg− PAOampR, as compared to 34 % with Alg− PAO1 (Fig. 4). In addition, Alg+ PDOampR also showed a higher virulence than Alg+ PDO300 with 85 and 98 % survival at 4 h post-incubation, respectively (Fig. 4). The increase in virulence in Alg− PAOampR and Alg+ PDOampR could be restored using pAmpR (Fig. 4), suggesting that AmpR acts as a negative regulator of P. aeruginosa virulence.

Fig. 4.

Kinetics of the paralysis of C. elegans by P. aeruginosa Alg− PAO1 wild-type strain (—), Alg− PAOampR (– . . –), complemented Alg− PAOampR(pAmpR) (– . –), Alg+ PDO300 (-.-), Alg+ PDOampR (--) and complemented Alg+ PDOampR(pAmpR) (…). L4 stage larval hermaphrodite Bristol N2 C. elegans were placed on each brain heart infusion agar plate containing a bacterial lawn and scored for dead worms by microscopic examination. E. coli OP50 (— —) was used as a negative control. Values are the mean±sd of triplicate analyses. Results were statistically significant (P<0.05 for PAO1 vs PAOΔampR at 2 h, and PDO300 and PDOampR at 4 h).

DISCUSSION

AmpR is the master transcriptional regulator involved in β-lactam antibiotic resistance. We have demonstrated previously that in addition to regulating AmpC and PoxB β-lactamases, P. aeruginosa AmpR plays a role in controlling the expression of some virulence factors (Kong ). In this study, we have shown that there is a complex regulatory network between β-lactam resistance, alginate production, and QS and virulence gene expression, factors determining the establishment of both acute and chronic P. aeruginosa infection.

ampR autoregulation requires AlgT/U

Previous studies in Enterobacteriaceae spp. have demonstrated that the transcription of ampR is autoregulatory (Lindquist ); but, we reported otherwise for the non-mucoid strain of P. aeruginosa (Kong ). Data presented here show that the autoregulatory mechanism of ampR could be seen in the presence of inducers in an alginate overproducing strain, Alg+ PDO300, but was lost in the absence of ampR (Table 2). This suggests that the regulation of ampR transcription requires AlgT/U, and is AmpR-dependent in Alg+ PDO300. The requirement of these factors for autoregulation explains the inconsistency seen with the Enterobacteriaceae models: in an in vivo system using the heterologous host E. coli, the P-lacZ activity was repressed threefold in the presence of Citrobacter freundii AmpR (Lindquist ). However, this mode of regulation was lost in a minicell system with Enterobacter cloacae AmpR (Lindberg & Normark, 1987).

The intriguing autoregulatory mechanism seen in the alginate-overproducing PAO1 derivative may have important clinical implications: the data with Alg− PAO1 suggest that this early colonizer is able to induce the production of β-lactamases upon β-lactam chemotherapy. However, this non-mucoid strain is unable to autoregulate ampR, indicating that the production of β-lactamases is induced only upon contact with β-lactam antibiotics. This phenomenon may be disadvantageous during antibiotic selections. Persistence of the organism in the lungs of patients with CF will ultimately result in the selection of mucoid strains that hyperproduce alginate (Høiby, 1975). This phenotypic alteration is accompanied by resistance to antibiotics and immune clearance (Giwercman ). Data from Alg+ PDO300 suggest that the selected mucoid P. aeruginosa strains are primed to β-lactam resistance by the increased production of AmpR, and hence β-lactamases, upon contact with β-lactam antibiotics. This observation should be further verified using clinical strains with commonly used β-lactams.

AmpR is a negative regulator of algT/U

The simultaneous presence of β-lactam resistance and alginate-overproduction suggests a possible co-regulation of these phenomena. We have shown here that autoregulation of ampR is AlgT/U-dependent. Loss of the ampR gene in Alg− PAO1 resulted in a significant increase in the promoter activity of algT/U operon (Fig. 2). However, this did not phenotypically alter Alg− PAO1 to an Alg+ phenotype due to the post-transcriptional control of AlgT/U by the anti-sigma factor, MucA, expressed downstream of algT/U (Hughes & Mathee, 1998). In Alg+ PDO300, like in Alg− PAO1, there is an increase of algT/U expression upon loss of ampR. Data from these two strain backgrounds suggest that AmpR suppresses the expression of algT/U.

The possible mechanistic interaction between the alg and amp regulons has been reported in Azotobacter vinelandii, where a mutation in ampDE, encoding negative regulators of β-lactamases, resulted in elevated expression of alginate biosynthetic genes (Núñez ). In addition, microarray data also have demonstrated that alginate production is induced upon antibiotic challenge (Bagge ), and a later study identified AlgT, AlgW and Prc proteases as being involved in this process (Wood ). Our results are further supportive of their findings in which β-lactam resistance and alginate production of P. aeruginosa are co-regulated. This co-regulation is likely mediated by AmpR-AlgT/U interaction. Future studies will address this potential interaction.

AlgT/U and AmpR are regulators of virulence factors

Multiple QS-dependent phenotypes, including LasA and pyocyanin production, are differentially regulated in an ampR mutant, and are probably an indirect effect of AmpR on the QS system. We have previously shown that deletion of ampR gene increased the production of LasA protease in an Alg− strain, suggesting that lasA expression is suppressed by AmpR (Kong ). We postulated that a similar observation should be obtained in an Alg+ strain. As expected, the absence of ampR in the presence of functional AlgT/U elevated the promoter expression of lasA and the production of LasA protease (Table 3). This alteration in LasA synthesis suggests that both AlgT/U and AmpR negatively impact transcription of the lasA gene. Although the inverse correlation between alginate and protease production has been repeatedly reported, our results establish that this correlation is mediated through the downregulation of QS in Alg+ strains. Comparing two isogenic strains, Alg− PAO1 and Alg+ PDO300, the see-saw effect is brought upon by the ECF sigma factor, AlgT/U. Since sigma factor, an essential component of RNA polymerase, is unlikely to be involved in the repression of gene expression, AlgT/U-mediated downregulation of QS genes is probably indirect.

To determine whether the in vitro alterations in virulence factor expression could be translated into significant in vivo killing, the C. elegans–P. aeruginosa interaction model was employed. As predicted, loss of ampR strongly correlated with an increase in virulence with both Alg− PAOampR and Alg+ PDOampR showing higher rates of C. elegans paralysis as compared to their parent strains (Alg− PAO1 and Alg+ PDO300, respectively). The significantly higher amounts of pigmentation produced by ampR mutants compared to the isogenic wild-type strain explains the higher killing rate, which is in agreement with other studies (Tan ).

Concluding remarks

The data presented here reveal a complex co-regulatory network between β-lactam resistance, alginate production, QS and virulence gene expression. We have previously shown that AmpR regulates AmpC and PoxB β-lactamases and QS-dependent proteases (Kong , b). In this paper, that observation is further extended to include the alginate master regulator, AlgT/U. Importantly, we show that the positive autoregulation of ampR requires AlgT/U, whereas AmpR negatively regulates algT/U expression (Fig. 2) serving as a negative feedback loop to limit the AlgT/U expression. We propose that this intimate crosstalk between these two global regulators provides a potential molecular framework for the simultaneous occurrence of β-lactam resistance and alginate-overproducing strains in chronic CF lung infections. Further studies on clinical isolates are warranted to understand the complex regulatory network linking all these critical factors in establishing infections. Delineating the interplaying factors and regulatory network is of fundamental significance to understanding the pathogenesis of P. aeruginosa.