QscR, a modulator of quorum-sensing signal synthesis and virulence in Pseudomonas aeruginosa

Abstract

The opportunistic pathogenic bacterium Pseudomonas aeruginosa uses quorum-sensing signaling systems as global regulators of virulence genes. There are two quorum-sensing signal receptor and signal generator pairs, LasR–LasI and RhlR–RhlI. The recently completed P. aeruginosa genome-sequencing project revealed a gene coding for a homolog of the signal receptors, LasR and RhlR. Here we describe a role for this gene, which we call qscR. The qscR gene product governs the timing of quorum-sensing-controlled gene expression and it dampens virulence in an insect model. We present evidence that suggests the primary role of QscR is repression of lasI. A qscR mutant produces the LasI-generated signal prematurely, and this results in premature transcription of a number of quorum-sensing-regulated genes. When fed to Drosophila melanogaster, the qscR mutant kills the animals more rapidly than the parental P. aeruginosa. The repression of lasI by QscR could serve to ensure that quorum-sensing-controlled genes are not activated in environments where they are not useful.

Pseudomonas aeruginosa is a versatile bacterium that can be found in many different environments. For example, it can be found in lakes, soils, and on plants (1). It is also an emerging opportunistic pathogen of humans (2, 3). The genome sequence of this bacterium was completed recently. The sequence analysis revealed a great diversity of genes, including a high number of putative transcription regulators. Nearly 10% of the 5,570 predicted P. aeruginosa genes appeared to be transcription factors (4).

A large number of genes (perhaps as many as 200–300), including virulence factor genes and genes involved in biofilm development, are activated by two homologous acyl-homoserine lactone (AHSL) quorum-sensing systems. These two systems are the LasR–LasI and RhlR–RhlI systems. LasR is a transcriptional activator that responds to the product of the LasI protein, N-3-(oxododecanoyl)homoserine lactone (3OC₁₂-HSL). At sufficient environmental concentrations of this AHSL signal, a number of genes are activated, including rhlR, which codes for the N-butyrylhomoserine lactone (C₄-HSL) receptor, and rhlI, which codes for the C₄-HSL signal generator. RhlR and C₄-HSL activate many other genes. As might be expected in a bacterium with so many lifestyles, and so many regulatory genes, the elements of the two quorum-sensing systems are controlled by other factors (5–8).

The analysis of the Pseudomonas genome revealed a gene coding for a homolog of LasR and RhlR but no additional genes coding for LasI and RhlI homologs. The predicted ORF for the LasR homolog, PA1898, begins at bsae pair 2,069,490 and terminates at base pair 2,070,203 (see http://www.pseudomonas.com). The gene is linked to a group of genes with similarity to the phz operon, which is required for production of the phenazine pigment, pyocyanin, itself a virulence factor (8, 9). However, this operon cannot substitute for the phz operon in the synthesis of pyocyanin (8). Here we investigate the role of the LasR, RhlR homolog in P. aeruginosa quorum sensing and virulence. Our evidence indicates that the homolog is a negative regulator of quorum-sensing-controlled genes and that it likely exerts its affect by repressing lasI. Thus we have termed the gene qscR for quorum-sensing-control repressor.

Materials and Methods

Bacterial Strains, Plasmids, and Culture Conditions.

The strains and plasmids we used are shown in Table 1. Escherichia coli and P. aeruginosa were routinely grown at 37°C in Luria–Bertani broth or Luria–Bertani agar. For virulence experiments, bacteria were grown in Brain Heart Infusion broth (Difco) for 8 h. Pyocyanin assays were performed with P. aeruginosa grown in pyocyanin production medium (PPM) (10). Media were supplemented with antibiotics for selection as follows: for E. coli, ampicillin (Ap) at 100 μg/ml, gentamicin (Gm) at 15 μg/ml, chloramphenicol (Cm) at 30 μg/ml, and spectinomycin (Sp) at 100 μg/ml; for P. aeruginosa, carbenicillin (Cb) at 300 μg/ml, Gm at 100 μg/ml, nalidixic acid at 20 μg/ml, and streptomycin (Sm) at 500 μg/ml. For promoter activity assays, pyocyanin production measurements, and AHSL production studies, the OD₆₀₀ at the start of an experiment was 0.005–0.05. Starter cultures were in the mid-log phase. Analyses were performed on samples taken at the indicated points during culture growth. For experiments with tac promoter-controlled genes, we included isopropyl-β-d-thiogalactopyranoside (1 mM) in the culture medium.

Table 1.

Bacterial strains and plasmids used

Strain or plasmid	Genotype	Source/ref.
P. aeruginosa
PAO1	Wild-type prototroph	30
PAOR3	qscR mutant of PAO1, GmR	This study
E. coli
DH5α	F− φ80dlacZΔM15Δ(lacZYA-argF)U169endA1 recA1 hsdR17 deoR gyrA96 thi-1 relA1 supE44	31
S17-1	thi pro hsdR recA RP4-2 (Tet∷Mu) (Km∷Tn7)	32
CM830	E. coli K-12 derivative
Plasmids
pGmΩ1	Broad-host-range vector containing the aacC1 gene, GmR	33
pCR2.1	TA cloning vector, ApR	Invitrogen
pEX1.8	Broad-host-range expression vector, ApR	34
pQF50	Broad-host-range lacZ transcriptional fusion vector, ApR	35
pMP7	E. coli cloning vector, (ColE1) ori, ApR	36
pKL13	pMP7 containing 3.1 kb qscR fragment, ApR	This study
pKL14	pKL13 with 1.6 kb aacC1 interrupting qscR ApRGmR	This study
pKV69	Vector carrying mobRP4 CmRTcR	K. Visick, Loyola Univ., Chicago
pMW16	pKL14 containing 1.6 kb mobRP4 from pKV69, ApRGmRCmR	This study
pKL8	pCR2.1 containing 714 bp qscR fragment, ApR	This study
pKL9	pEX1.8 containing ptac-qscR, ApR	This study
pMW303	Broad-host-range phzABC-lacZ reporter, ApR	15
pMW301	Broad-host-range hcnAB-lacZ reporter, ApR	15
pSC11	Broad-host-range lasI-lacZ reporter, ApR	This study
pSC10	Broad-host-range lasR-lacZ reporter, ApR	This study
pMW305	Broad-host-range rhlI-lacZ reporter, GmR, SpR, SmR	15
pMW304	Broad-host-range rhlR-lacZ reporter, GmR, SpR, SmR	15
pMW105B	Broad-host-range qsc105-lacZ reporter, ApR	This study

Construction and Complementation of a P. aeruginosa qscR Mutant.

A qscR mutant was constructed as follows. A 3.1-kb P. aeruginosa PAO1 chromosomal DNA fragment containing qscR (from base pair 2,067,814 to base pair 2,070,999 on the P. aeruginosa chromosome) was amplified by Expand long-template PCR (Boehringer Mannheim), and cloned into BamHI-digested pMP7. The resulting plasmid was pKL13. A 1.6-kb SmaI fragment of pGmΩ1 that contained the aacC1 gene (encoding Gm acetyltransferase) was ligated with EcoRV-digested pKL13 to create pKL14. This construct has the aacC1 gene inserted 19 bp downstream of the translation start site of qscR, and aacC1 is flanked by 1.7 kb of upstream and 1.5 kb of downstream P. aeruginosa DNA. A 1.6-kb KpnI fragment carrying a mobilizing (mob) cassette from pKV69 was cloned into KpnI-digested pKL14 to create pMW16, a suicide plasmid with a ColE1 ori. pMW16 was mobilized from E. coli S17–1 into P. aeruginosa PAO1 by conjugation. Selection for Gm^R colonies followed by screening for Cb^S yielded a qscR insertion mutant, P. aeruginosa PAOR3. We confirmed that P. aeruginosa PAOR3 contained an insertion of the Gm cassette in qscR at the expected location by Southern blotting (Genius, Boehringer Mannheim) with aacCI and qscR gene probes.

The qscR mutation in strain PAOR3 was complemented by introduction of pKL9, which contains qscR under the control of the tac promoter. To construct pKL9, a 714-bp DNA fragment carrying qscR was amplified from pKL13 with Expand long-template PCR and T/A cloned using the Original TA cloning kit (Invitrogen) to form pKL8, which was digested with MfeI and HindIII, and cloned into EcoRI–HindIII-digested pEX1.8 to generate pKL9.

Construction of Reporter Gene Fusions.

To create the lasR–lacZ reporter plasmid, pSC10, a 623-bp PCR fragment with engineered SalI and BamHI sites flanking bp −410 to +213 relative to the lasR translational start was ligated to SalI–BamHI-digested pQF50. To construct the lasI–lacZ reporter plasmid, pSC11, a 505-bp PCR fragment with engineered SalI and BamHI sites flanking base pairs −282 to +223 relative to the lasI translational start site was cloned into SalI–BamHI-digested pQF50. A 289-bp PCR-generated fragment flanking base pairs −273 to +16 relative to the translational start site of the ORF PA2587 (http://www.pseudomonas.com) was polished with T4 polymerase and cloned into SmaI-digested pQF50 to construct the qsc105-lacZ reporter plasmid, pMW105B. All constructs were verified by DNA sequencing.

AHSL, Pyocyanin, and β-Galactosidase Measurements.

Concentrations of 3OC₁₂-HSL and C₄-HSL were measured with bioassays as described (11, 12). Synthetic 3OC₁₂-HSL and C₄-HSL (Quorum Sciences, Coralville, IA) were used to generate standard curves. Pyocyanin was extracted from culture fluid with chloroform and then extracted from the chloroform with 0.2 M hydrochloric acid in water. The absorbance at 520 nm was a measure of the amount of extracted pyocyanin as described elsewhere (10). β-Galactosidase activity was measured as described (8). Results are given in units of β-galactosidase, activity per OD₆₀₀.

Fruit Fly Virulence Studies.

Approximately 8 × 10⁹ viable bacterial cells were pelleted from cultures by centrifugation. These cells were suspended in 170 μl of 5% sucrose solution (13) and added to a filter disk (2.3-cm diameter Whatman paper) that completely covered the agar surface (5 ml of 2.4% agar with 5% sucrose) at the bottom of a standard glass fly culture vial. Adult male 3- to 5-day-old fruit flies (Drosophila melanogaster Canton S) that had been starved for food and water for 5 h were added to each vial (9–12 flies per vial). The vials were capped with cotton, inverted over a tray with water for humidity, and incubated at 25°C. To determine the number of viable bacterial cells associated with individual flies, single flies were ground with a Teflon pestle in an Eppendorf tube with 100 μl of 10 mM MgSO₄, and serial dilutions of the homogenate were spread on Luria–Bertani agar.

Results

The Relationship of QscR to AHSL Signal Receptors.

The qscR ORF consists of 714 nucleotides, and it codes for a polypeptide of 27,236 Da. A sequence comparison shows that QscR is most similar to Ralstonia solanacearum SolR (33% identity, 49% similarity), P. aeruginosa RhlR (32% identity, 53% similarity), P. aeruginosa LasR (29% identity, 46% similarity), and Vibrio fischeri LuxR (27% identity, 45% similarity). An alignment of these polypeptides is shown in Fig. 1. QscR has the seven residues conserved among all of the members of this family of transcription factors (14). The qscR gene is linked to a cluster of genes that is nearly identical to the phz operon (Fig. 1). The DNA immediately upstream of this gene cluster bears no relation to the phz promoter region. Furthermore, P. aeruginosa strain PAO1 with a mutation in the phz operon does not produce pyocyanin (10). Thus, the genes linked to qscR do not direct the synthesis of pyocyanin in P. aeruginosa strain PAO1.

Figure 1.

Relationship of the qscR gene and gene product to other genes and their products. (Upper) Alignment of the QscR sequence with SolR, RhlR, LasR, and LuxR. Conserved amino acids are shaded in black. Gray shading indicates that 100% of the residues are similar at that position. The alignment was constructed by using the clustal w multiple alignment program and the degree of residue shading was determined by using Boxshade at 1.0 setting (37). The sequences used in the alignment are R. solanacearum SolR (GenBank accession no. AF021840), P. aeruginosa RhlR (L08962), P. aeruginosa LasR (M59425), and V. fischeri LuxR (M96844). (Lower) A map of qscR and the genes flanking qscR in the P. aeruginosa genome. In the P. aeruginosa genome annotation, genes designated A2–G2 are called phzA2-G2 based on similarity to genes in the phz operon (http://www.pseudomonas.com). The genes in the phzA2-G2 operon are each over 98% identical at the DNA sequence level to the bona fide phzA-G genes. Other phenazine-producing species of Pseudomonas have operons with related genes. The qscR gene is called phzR in the P. aeruginosa genome annotation.

A qscR Mutant Overexpresses Quorum-Sensing-Controlled Virulence Factors.

To study the role of qscR in P. aeruginosa, we constructed a null mutant, PAOR3. Cultures of this qscR mutant grew at the same rate as the parent strain PAO1 (data not shown); however, they were noticeably different from the parent in that they were blue. This is a characteristic of strains that overproduce the phenazine pigment pyocyanin (15). Direct measurements confirmed that the qscR mutant produced more pyocyanin than the parent. It also produced pyocyanin at a lower culture density than the wild type. Complementation of the qscR mutation in PAOR3 restored the parental pyocyanin levels and timing of pyocyanin synthesis (Fig. 2A).

Figure 2.

The influence of qscR on quorum-controlled factors in P. aeruginosa. (A) Pyocyanin levels in P. aeruginosa PAO1 (□), PAOR3 (▴), and PAOR3 with the qscR plasmid pKL9 (○). β-Galactosidase in P. aeruginosa PAO1 (□), and PAOR3 (▴) containing the phzABC-lacZ plasmid, pMW303 (B), the hcnAB-lacZ plasmid, pMW301 (C), and the qsc105-lacZ plasmid, pMW105B (D).

The influence of QscR on pyocyanin production could be at the level of phz operon transcription or at a posttranscriptional level. It could be direct or it could be a general repression of quorum-sensing-controlled genes. To address these questions, we tested the influence of QscR on the expression of β-galactosidase in strains with the lacZ gene fused to promoters of quorum-controlled genes (including a phzABC–lacZ fusion). Besides the phz operon, we examined transcription of the hcn operon. This operon appears to be required for production of hydrogen cyanide, which has been established as a virulence factor for P. aeruginosa (16), and like the phz operon, we believe it is controlled by C₄-HSL and RhlR, directly (8). We also examined a lacZ fusion to an ORF controlled by 3OC₁₂-HSL and LasR, qsc105 (8). The phzC–lacZ fusion was expressed earlier in the qscR mutant than the wild-type (Fig. 2B). This indicates that the early induction of pyocyanin synthesis in the qscR mutant can be attributed to derepression of phz operon transcription. Early expression of the phz operon could also account for increased levels of pyocyanin in cultures of the qscR mutant. The patterns of β-galactosidase expression in the parent and the qscR mutant carrying an hcnAB-lacZ fusion were similar to the patterns with the phzC-lacZ fusion (Fig. 2C). Thus, we conclude that both the hcn and phz promoters are affected similarly by QscR. Our analysis of expression of the LasR-3OC₁₂-HSL-controlled qsc105–lacZ fusion (Fig. 2D) showed that a mutation in qscR resulted in premature transcription from the qsc105 promoter. The expression of qsc105–lacZ was even earlier than expression from the phz or hcn promoters (Fig. 2).

Regulation of Quorum-Sensing Signal Generator and Signal Receptor Gene Expression by QscR.

One hypothesis for the finding that transcription from each of the quorum-sensing-controlled promoters we examined was influenced by QscR is that QscR functions to regulate components of the LasR–LasI or RhlR–RhlI quorum-sensing signal generator and signal receptor pairs. To test this hypothesis, we examined expression of lacZ fusions to the genes coding for each of these four polypeptides in the P. aeruginosa qscR mutant and the parental strain. Expression of the lasR–lacZ fusion in the parent and mutant was indistinguishable (Fig. 3A). A rhlR-lacZ fusion showed slightly elevated expression in early logarithmic phase (Fig. 3B). Expression of both the lasI–lacZ fusion and the rhlI–lacZ fusion occurred early in the qscR mutant as compared with the parent (Fig. 3 C and D).

Figure 3.

Expression of the lasR–lacZ (pSC10) (A), rhlR–lacZ (pMW304) (B), lasI–lacZ (pSC11) (C), and the rhlI–lacZ (pMW305) (D) transcriptional fusions in P. aeruginosa PAO1 (□), PAO1 with exogenously added 2 μM 3OC₁₂-HSL (○), and PAOR3 (▴).

The results described above lead to the hypothesis that a mutation in qscR leads to early production of the acyl-HSL signals, which attain a critical concentration for activation of quorum-sensing-controlled genes at a lower culture density in the mutant compared with the parent. In one test of this hypothesis, we added 3OC₁₂-HSL to the parent with the lasI–lacZ reporter. Consistent with the hypothesis, the exogenous addition of the signal compound resulted in a premature induction of β-galactosidase that was similar to the induction in the qscR mutant (Fig. 3C). In another test of the hypothesis we measured 3OC₁₂-HSL and C₄-HSL in cultures of the mutant and parent (Fig. 4). Both of the AHSLs were synthesized earlier in the qscR mutant than the parent. The LasI-generated signal, 3OC₁₂-HSL was synthesized earlier than the RhlI-generated signal, C₄-HSL (Fig. 4).

Figure 4.

The concentrations of 3OC₁₂-HSL (A) and C₄-HSL (B) in cultures of P. aeruginosa PAO1 (□) and PAOR3 (▴).

Virulence of a qscR Mutant.

Unlike its homologs, LasR and RhlR, which are required for virulence gene activation (17–20), QscR serves to repress the synthesis of at least two virulence factors, pyocyanin and hydrogen cyanide. Thus, we would predict that QscR is not required for virulence. To test this, we used an insect model for P. aeruginosa pathogenesis. We tested the role of QscR in virulence by feeding the mutant and parent to the fruit fly D. melanogaster; a particularly convenient model host among the various insects in which P. aeruginosa causes disease (21–23). We fed fruit flies with P. aeruginosa cells suspended in a sucrose solution (see Materials and Methods) and monitored the flies for 2 weeks. E. coli was nearly completely harmless and the parent, P. aeruginosa PAO1 was relatively benign, slowly killing a minority of the flies (Fig. 5). The qscR mutant, however, was considerably more virulent than the parent, killing 100% of the flies (114 in total) by 12.5 days (Fig. 5). At least in this model, QscR functions as a governor that reduces the virulence of P. aeruginosa.

Figure 5.

Virulence of the P. aeruginosa qscR mutant in fruit flies. Death of flies over time when fed P. aeruginosa PAO1 (□), PAOR3 (▴), or E. coli CM830 (○). The values are the averages from 10 replicate experiments, each with 9–12 flies, and the SEM is shown for each value plotted. The parent and mutant bacteria survived equally well in the dry culture vials during the course of the experiments (data not shown). The number of bacteria recovered from flies just after death was 12 (± an SEM of 2) × 10⁶ PAO1, and 8 (±3) × 10⁶ PAOR3.

Discussion

Quorum sensing controls virulence and biofilm formation in P. aeruginosa. There are two well-studied quorum-sensing signal generator-receptor pairs in this bacterium, LasI–LasR, and RhlI–RhlR (for reviews, see refs. 24–26). As a result of the P. aeruginosa genome-sequencing project, a gene coding for a homolog of LasR and RhlR was discovered (http://www.pseudomonas.com). There is no apparent cognate signal generator gene for this putative signal receptor gene. We show that in a wild-type strain of P. aeruginosa, the LasR–RhlR homolog represses transcription of three quorum-sensing-controlled genes, two of which are primarily activated by RhlR and one activated by LasR. This repression is in the logarithmic phase of growth (Fig. 2). The LasR-controlled genes are expressed earlier than the RhlR-controlled genes in a mutant in which the LasR–RhlR homolog has been inactivated (Fig. 2). Furthermore, we show that lasI and rhlI (both quorum-sensing-controlled themselves) are both prematurely transcribed in a mutant with an inactivated gene for this newly described transcription factor (Fig. 3). The mutant also produces the AHSL signals prematurely with the LasI-generated 3OC₁₂-HSL produced earlier than the RhlI-produced C₄-HSL (Fig. 4). This leads to a model whereby this newly described transcription factor represses all quorum-sensing-controlled genes by repressing transcription of lasI, and thus retarding production of 3OC₁₂-HSL in the early logarithmic phase of growth. Because 3OC₁₂-HSL and LasR are activators of rhlR and rhlI, activation of lasI commences a cascade of activation of all quorum-sensing-controlled genes. We have called this transcription factor QscR, a quorum-sensing-control repressor.

We do not yet know whether QscR responds to a signal itself, and we do not know how it affects repression of lasI transcription. QscR could bind to the promoter region of lasI itself or it could form inactive heteromultimers with LasR. It could require binding of a signal, perhaps even 3OC₁₂-HSL for its release from the lasI promoter or from LasR, or 3OC₁₂-HSL binding to LasR could overcome the activity of QscR. There are examples of proteins in other bacteria that serve to govern the activity of transcription factors related to LasR. Perhaps the closest analogy comes from the plant pathogen Agrobacterium tumefaciens, which has a gene called traS. The traS gene codes for a truncated polypeptide that resembles the quorum-sensing activator TraR, and TraS blocks the function of TraR, presumably by forming inactive heterodimers with TraR (27).

It may be of relevance that AHSLs with long acyl tails do not diffuse through the cell membrane as readily as those with shorter acyl tails (28). In fact efflux pumps appear to be involved in export of 3OC₁₂-HSL (28, 29). It is conceivable that the diffusion limitation could create a short circuit in the extracellular signaling process, and that transcription factors like QscR may circumvent this problem.

Our analysis of virulence (Fig. 5) shows that a mutation in QscR gives a hypervirulence phenotype. We hypothesize that there are disadvantages of this phenotype for P. aeruginosa. It is conceivable that in the initial stages of some infections, the display of quorum-sensing-controlled virulence factors allows the host the opportunity to mount a response. We also suspect that a qscR mutation could render P. aeruginosa less fit for survival in environments like soils or lakes where quorum-sensing-controlled virulence factors might be of no selective advantage.

Abbreviations

AHSL

acyl-homoserine lactone

3OC₁₂-HSL

N-(3-oxododecanoyl)homoserine lactone

C₄-HSL

N-butyrylhomoserine lactone

Ap

ampicillin

Gm

gentamicin

Cm

chloramphenicol

Sp

spectinomycin

Cb

carbenicillin

Sm

streptomycin