INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The expression of many virulence factors produced by the gram-negative opportunistic pathogen Pseudomonas aeruginosa is regulated by a cell density-dependent mechanism known as quorum sensing (9, 10, 38). Quorum sensing utilizes a transcriptional activator protein (R protein) which acts in concert with a small, signaling molecule, known as the autoinducer (AI), to stimulate the expression of target genes. At low cell density, AI is produced at a basal level. As the cell population increases, so does the concentration of AI, providing a chemical signal with which to monitor cell density. Once a threshold level of AI is reached, it binds to and activates the R protein, resulting in expression of quorum-sensing-controlled genes. A number of AI molecules from various organisms have been identified and characterized (see reference 8 for a review). Most of these molecules are N-acylated homoserine lactones and differ only in the length and substitution of their acyl side chain. The phenomenon of quorum sensing appears to be nearly ubiquitous among gram-negative bacteria (8), and a regulatory system employing a -butyrolactone signaling molecule has been identified in gram-positive Streptomyces spp. (14).

P. aeruginosa contains two known quorum-sensing systems, each with its own R protein-AI pair. The better understood of these is the las system, which consists of the transcriptional activator protein, LasR, and an AI called PAI-1. The synthesis of PAI-1 (N-3-oxo-dodecanoyl homoserine lactone) is directed by the autoinducer synthase LasI (23, 25). LasR and PAI-1 work in concert to increase the expression of a number of virulence genes, including those for several proteases (lasA, lasB, and aprA) and exotoxin A (toxA) (9, 10, 23, 33, 38). LasR-PAI-1 is also required for expression of the AI synthase gene, lasI, creating a positive feedback loop (35). Studies done with Escherichia coli demonstrated that 10-fold-less PAI-1 was required to activate lasI, as compared to lasB (35). Therefore, it appears that a hierarchy of gene regulation exists, with particular genes activated or turned on earlier than others. In this hierarchy, low levels of PAI-1 apparently bind to LasR, which then induces expression of lasI and increases PAI-1 production. After sufficient PAI-1 has accumulated, secondary activation of virulence genes such as lasB occurs. Maintaining tight control of lasI expression is therefore a key feature of the P. aeruginosa quorum-sensing process.

The second P. aeruginosa quorum-sensing system consists of the regulatory protein RhlR and PAI-2 (N-butyryl homoserine lactone) and has been shown to regulate the biosynthesis of rhamnolipid, RpoS, and, to some extent, the lasB-encoded elastase (4, 16, 21, 22, 26, 27, 39). Interestingly, LasR and PAI-1 have been found to regulate the expression of RhlR, making LasR and PAI-1 the dominant regulators in P. aeruginosa (16, 28).

It is apparent that quorum sensing plays a key role in the regulation of virulence gene expression in P. aeruginosa, and recently quorum sensing has been shown to be involved in the differentiation of P. aeruginosa biofilms (6). Despite the fact that this cellular signaling mechanism plays a critical role in P. aeruginosa virulence, as well as facilitating its survival in hostile environments, very little is known about how the regulators of these systems are themselves regulated. All experiments reported to date have focused on positive gene regulation, and specific repressor molecules have not been reported for the P. aeruginosa quorum-sensing systems. Here we report the identification of a novel gene, rsaL, which encodes a negative regulatory protein. We show that RsaL completely represses lasI transcription, thereby blocking activation of the P. aeruginosa quorum-sensing cascade.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. Bacterial strains are listed in Table 1. Plasmids are shown in both Table 1 and Fig. 1B. Oligonucleotide primers were synthesized by the University of Rochester Medical Center Nucleic Acid Core Facility. In all cases in which PCR was used to generate plasmid constructs, DNA sequencing was performed to verify the sequences of the PCR products. Transformants were selected on agar medium containing the appropriate antibiotic(s).

Nucleic acid manipulation. Purification, cloning, electrophoresis, and other manipulations of nucleic acid fragments and constructs were performed using standard techniques (12, 34).

PCR. PCR was performed under standard conditions as presented by Gibco Life Technologies (Gaithersburg, Md.) on data sheets supplied with their Taq polymerase. MgCl₂ was used at a concentration of 3 mM. In a 100-µl reaction mixture, 100 pmol of each primer was used. DNA fragments were isolated from agarose gels by using the GeneClean system (Bio101, La Jolla, Calif.).

-Galactosidase assays. E. coli strains were grown at 37°C in supplemented A medium (28) with ampicillin (100 µg/ml). In addition, synthetic PAI-1 (final concentration, 100 nM) was added to the flasks before addition of culture medium. P. aeruginosa strains were grown at 37°C in peptone Trypticase soy broth (PTSB) medium supplemented with carbenicillin (200 µg/ml), and where stated, synthetic PAI-1 (generated as previously described [24]) was added to the flasks to a final concentration of 1 µM before addition of culture medium. For both E. coli and P. aeruginosa, overnight cultures were diluted 1 to 100 and allowed to grow until they had reached an optical density of 0.5 to 0.8. Optical density was measured at 540 nm for P. aeruginosa and 600 nm for E. coli. -Galactosidase activity was assayed in triplicate as described by Miller (19).

Protein expression. To purify RsaL protein, rsaL was ligated into a commercially available expression vector that generates a protein bearing six tandem histidine residues at its amino terminus. rsaL was ligated into pTrcHisA, -B, and -C (Promega Corp., Madison, Wis.) to ensure that all three translational reading frames were examined. DNA sequencing verified that rsaL is in the correct frame for translation in pTrcHisB/rsaL. Overnight cultures of E. coli JM109 harboring these plasmids were subcultured 1 to 20 in Luria-Bertani medium supplemented with ampicillin (100 µg/ml) and allowed to grow for 30 min before addition of 2 mM IPTG (5-bromo-4-indol-3-chloro-isopropyl -D-galactopyranoside). Following 4 h of growth, 1-ml aliquots of the cultures were sedimented (12,000 × g) and the cell pellets were lysed in buffer B (8 M urea, 0.1 M Na-phosphate, 0.01 M Tris-HCl [pH 8.0]). A 50-µl aliquot of Ni-nitrilotriacetic acid (NTA) resin (Qiagen, Valencia, Calif.) was added to the cell lysates, which were incubated for 30 min at room temperature with frequent inversion. The resin was washed three times with buffer C (8 M urea, 0.1 M Na-phosphate, 0.01 M Tris-HCL [pH 6.3]) and then resuspended in 20 ml of buffer C containing 100 mM EDTA to elute any proteins which had bound to the resin.

Elastolytic activity determination. Cultures were streaked onto elastin agar containing carbenicillin (200 µg/ml) and incubated at 30°C for 48 h as previously described (32). The zones of clearing (elastin solubilization) were measured, and the average of four independent test streaks was reported for each bacterial strain.

Extraction of PAI-1 from P. aeruginosa culture supernatants. Cultures were grown to late log phase (optical density at 540 nm of 0.8) in PTSB medium. Extraction of PAI-1 was done as described previously (25). As a control, synthetic PAI-1 was added to PTSB medium at a final concentration of 1 µM. This sample was treated identically to the supernatants. A 1-ml aliquot of an E. coli MG4(pKDT17) (23, 25) overnight culture diluted 1:100 in modified A medium containing 100 µg of ampicillin per ml was added to each tube, which contained either experimental extract or control PAI-1. The strains were grown for 5 h at 32°C and assayed for -galactosidase. Comparison of -galactosidase values obtained against those of a standard curve plotted with the synthetic PAI-1 allowed estimation of the PAI-1 content of each sample.

Western blot analysis of LasR. Cell pellets from cultures to be tested were obtained by centrifugation (6,000 × g, 5 min) of 1 to 2 ml of the culture. Each pellet was solubilized in lysis buffer (6 M urea, 1% sodium dodecyl sulfate [SDS], 10% glycerol, and 1% -mercaptoethanol in 50 mM Tris-HCl [pH 7.5]) by incubation for 30 min or until samples were clear. The protein content of each sample was determined using the commercially available bicinchoninic acid method (Pierce, Rockford, Ill.). Equal amounts of protein (usually 25 µg) were prepared in a total volume of 25 µl. The samples were separated on an SDS-10 to 12% polyacrylamide gel and transferred to a nitrocellulose membrane according to standard techniques (34).

Computer analysis of nucleic acids and proteins. Computer analysis of both nucleic acids and proteins was carried out using the Genetics Computer Group (GCG) Wisconsin Package software version 8 (Genetics Computer Group, Madison, Wis.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

A novel gene is contained in the intergenic region between lasR and lasI. In P. aeruginosa, lasR and lasI are separated by a 367-bp intergenic region and are transcribed in the same direction from independent promoters (Fig. 1A and C) (9, 23, 35). Preliminary studies indicated that regulatory elements controlling lasI expression exist within the lasR-lasI intergenic region, outside the lasI promoter. While studying this region, we discovered that a novel gene was located on the strand opposite to lasR. Two plasmids were constructed which contained a 658-bp EcoRI fragment (Fig. 1A) encompassing the lasRI intergenic region fused in both the forward (pPCS2001) and reverse (pPCS2002) orientations relative to lacZ. Both the plasI-lacZ fusion (pPCS2001) and the reverse lacZ fusion (pPCS2002) were expressed in the wild-type P. aeruginosa strain PAO1 (data not shown), suggesting the presence of a novel gene, which we have termed rsaL.

View larger version (27K): [in this window] [in a new window]   FIG. 1.   Plasmid inserts and DNA sequence of the rsaL gene. (A) The region including and surrounding the lasR, rsaL, and lasI genes is depicted. Restriction sites include EcoRV (RV), EcoRI (E; E* indicates that this restriction site was located in the multiple cloning site of vector pLP170 [29] and not in P. aeruginosa DNA), SfiI (S), and HphI (H). The 658-bp EcoRI DNA insert found on pPCS2001 and pPCS2002 is indicated. The 649-bp EcoRV fragment that was deleted and replaced with a tetracycline resistance (Tetr) cassette during the generation of strain PAO-R1 (lasR) (9) is also shown. The 307-bp SfiI-HphI fragment encodes a functional RsaL protein; thus, the Tetr cassette insertion in strain PAO-R1 does not interrupt the rsaL gene. (B) DNA inserts of the plasmids used in this study are indicated. a, plasmids containing both lasR and rsaL in their native conformation on a single DNA fragment; b, plasmids that contain the lasR and rsaL genes on separate DNA fragments. (C) lasR-lasI intergenic region. The deduced amino acid sequence of RsaL is indicated in bold. The rsaL translational start codon (ATG) is underlined, and potential lux-box-like elements are enclosed by boxes.

rsaL encodes a unique protein. To determine whether the rsaL open reading frame (ORF) encodes a protein, an in-frame translational fusion to lacZ (pSWRL2) was constructed and introduced into P. aeruginosa PAO1. Significant -galactosidase expression was observed in strain PAO1 containing the translational fusion (9,246 ± 322 Miller units). As a control, an out-of-frame rsaL-lacZ translational fusion (pSWRL) was also examined. The strain carrying the out-of-frame fusion expressed only slightly more -galactosidase (41 ± 3 Miller units) than strain PAO1 carrying the vector control (3 ± 1 Miller units). These findings suggested that a protein is produced from the rsaL transcript.

View larger version (44K): [in this window] [in a new window]   FIG. 2.   Expression of RsaL as a fusion protein. Lysates from E. coli JM109 carrying either the control vector (pTrcHisB) (lane 1) or the vector with an in-frame fusion of rsaL (pTrcHisB/rsaL) (lane 2) were separated by SDS-PAGE. Ni-NTA resin was used to purify the fusion protein from E. coli JM109(pTrcHisB/rsaL) (lane 3) and from E. coli JM109 carrying rsaL cloned onto pTrcHisA (lane 4) to serve as an out-of-frame control. Molecular mass standards are shown in daltons. The arrow indicates the protein observed in lane 3, which is believed to be RsaL.

Expression of rsaL requires LasR and not RhlR for expression. To assess which, if any, of the quorum-sensing systems regulate rsaL expression, cultures of P. aeruginosa wild-type strain PAO1 or strains that contain either a LasR (PAO-R1 [9]), RhlR (PDO111 [4]), or LasI (PAO-JP1 [26]) null phenotype and that carry an rsaL-lacZ translational fusion (pSWRL2) were assayed for expression of the fusion. Those strains that were LasR or LasI null mutants demonstrated a lack of rsaL expression (Fig. 3). Conversely, an RhlR null mutant exhibited wild-type levels of rsaL expression. Addition of exogenous PAI-1 to a LasI null mutant carrying the rsaL-lacZ fusion resulted in the restoration of rsaL expression to approximately wild-type levels. Taken together, the data indicate that rsaL expression requires both LasR and PAI-1. Furthermore, RhlR does not significantly affect rsaL expression.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Effect of the P. aeruginosa quorum-sensing regulators on expression of rsaL. Strains carrying a plasmid-borne rsaL-lacZ translational fusion (pSWRL2) or the control plasmid (pSW205) were monitored for expression of -galactosidase. Expression of the fusion was analyzed in the wild-type strain PAO1, the LasR null mutant PAO-R1, the RhlR null mutant PDO111, and the LasI null mutant PAO-JP1. PAO-JP1 grown in the presence of 1 µM exogenous PAI-1 was also examined.

Effects associated with rsaL overexpression. To assess the function of rsaL, multiple plasmids containing lasR alone or in combination with rsaL were examined in P. aeruginosa PAO-R1. Expression of the lasR and rsaL genes on these constructs is driven by either their native promoters or the lac promoter. In addition, the plasmids carry a lasB-lacZ translational fusion that functions as a reporter system for LasR-PAI-1-dependent activation. Strains PAO-R1(pLasR) and PAO-R1(pKDT17), which contain lasR under control of the wild-type promoter and the lac promoter, respectively, but no rsaL, showed high levels of lasB expression (Fig. 4). The increased lasB expression in pLasR as compared to that in pKDT17 is most likely because lasR expression is positively regulated in P. aeruginosa, albeit modestly, by LasR-PAI-1 (28). Plasmid pPCS15 is a derivative of pKDT17 that contains both lasR_plac and rsaL under control of its own promoter. In strain PAO-R1(pPCS15 lasR_plac rsaL_pwt lasB-lacZ), lasB expression was markedly lower than in strain PAO-R1(pKDT17 lasR_plac lasB-lacZ), suggesting that the presence of rsaL decreased lasR expression or activity (Fig. 4). The decrease in lasB expression is even more pronounced when lasR is expressed from its native promoter in the presence of rsaL, as seen in strains PAO-R1(pTS4001.7 lasR_pwt rsaL_pwt lasB-lacZ) and PAO-R1(pPCS16 lasR_pwt rsaL_plac lasB-lacZ), particularly when rsaL is constitutively expressed from the lac promoter [strain PAO-R1(pPCS16 lasR_pwt rsaL_plac lasB-lacZ)] (Fig. 4). These findings indicated that in P. aeruginosa, rsaL caused a strong repressive effect on lasB expression.

View larger version (27K): [in this window] [in a new window]   FIG. 4.   Overexpression of rsaL inhibits expression of lasB in P. aeruginosa. -Galactosidase activity from a plasmid-borne lasB-lacZ translational fusion was monitored in the presence of either lasR alone or both lasR and rsaL in P. aeruginosa PAO-R1. Expression of lasR and rsaL was directed by either their native promoters (wt) or the lac promoter (plac) as indicated under each plasmid. Plasmids in which the RsaL start methionine is removed (pPCS16/NS) or in which a stop codon is prematurely introduced (pPCS16/ES), but which otherwise are identical to pPCS16, were also examined for -galactosidase production. To monitor the level of rsaL expression in plasmids pPCS15, pTS4001.7, and pPCS16, the rsaL gene located on these constructs was replaced by an rsaL-lacZ translational fusion. -Galactosidase activity from the rsaL-lacZ fusions in P. aeruginosa PAO-R1 is indicated under each plasmid. n.d., not determined.

RsaL directly affects lasI expression. The finding that lasB, which is dependent on LasR for expression, is downregulated by RsaL led to the question of how RsaL is able to mediate this regulation. At least two possibilities exist. First, RsaL may bind to sequence upstream of, or within, lasB, thereby directly blocking transcription or translation. Second, RsaL may affect a common mediator required for lasB activation, the two most obvious candidates being LasR and PAI-1. To address each of these possibilities, we made use of E. coli single-copy  lysogens carrying either a lasI-lacZ (E. coli MG4 I₁4) or lasB-lacZ (E. coli MG4 B₂) transcriptional fusion on the prophage (35). Plasmids containing either lasR under its own promoter (pEXR), lasR under its own promoter and rsaL under the tac promoter (pEXRR), rsaL under the tac promoter (pEXRL), or a vector control (pEX1.8) were mobilized into the two E. coli  lysogens. Plasmid pEXRR/NS, which contains rsaL with its start codon removed and a functional copy of lasR, was also examined. In all cases, the E. coli cultures were grown in the presence of 100 nM exogenous PAI-1, which was required for LasR activation. In the E. coli MG4 B₂ cells expressing both rsaL and lasR (pEXRR lasR_pwt rsaL_ptac), -galactosidase activity was similar to that in cells expressing only lasR (pEXR lasR_pwt), indicating that rsaL does not directly affect lasB expression (Fig. 5). However, rsaL did have a significant effect on lasI expression. In the E. coli MG4 I₁4 cells containing both rsaL and lasR (pEXRR lasR_pwt rsaL_ptac), very low -galactosidase levels were observed compared to those in cells containing lasR alone (pEXR lasR_pwt) (Fig. 5). These findings indicate that RsaL is able to act directly on lasI to downregulate its expression. In the lasI  lysogen containing pEXRR/NS (lasR_pwt rsaL_ptac [no start codon]), lasI expression was significantly increased compared to that in the same lysogen containing pEXRR (lasR_pwt rsaL_ptac). However, the repressive activity associated with RsaL does not appear to be totally abolished by removal of the RsaL start methionine. Inspection of the nucleotide sequence of the pEXRR/NS (lasR_pwt rsaL_plac [no start codon]) construct revealed the presence of an in-frame GTG codon 19 amino acids upstream of the native RsaL start methionine. Data from previous studies have shown that a GTG codon can be used as a translational start codon approximately 8% of the time (11). Therefore, the partial repression seen from pEXRR/NS (lasR_pwt rsaL_ptac [no start codon]) may be due to the synthesis of a modified RsaL protein using this GTG start codon. To verify that the 81-amino-acid ORF believed to encode RsaL is sufficient to produce a functional protein, plasmid pEXRR2 was created. Plasmid pEXRR2 is identical to pEXRR (lasR_pwt rsaL_ptac) except that it contains the rsaL gene on a 307-bp DNA fragment (Fig. 1A) instead of an 800-bp fragment (pEXRR). Similar to pEXRR (lasR_pwt rsaL_ptac), pEXRR2 inhibited lasI expression in the lasI  lysogen strain (data not shown), confirming that this smaller fragment encodes a functional RsaL protein.

View larger version (21K): [in this window] [in a new window]   FIG. 5.   Effect of RsaL on lasI expression. The expression of lasB-lacZ (A) and lasI-lacZ (B) fusions carried as prophages in E. coli MG4 was examined in the presence of lasR under control of its own promoter (pEXR), rsaL overexpressed from the tac promoter (pEXRL), or lasR expressed from its own promoter and rsaL overexpressed from the tac promoter (pEXRR). A construct identical to pEXRR but which has the start codon of RsaL removed through mutagenesis (pEXRR/NS) was also examined. Plasmid pEX1.8 was used as a control vector in these experiments. In all cases, 100 nM PAI-1 and 1 mM IPTG were added to the cultures at the initiation of growth.

View larger version (43K): [in this window] [in a new window]   FIG. 6.   RsaL's repressive effect is overcome by addition of exogenous PAI-1. -Galactosidase activity from a plasmid-borne lasB-lacZ translational fusion was monitored in the presence of either lasR alone or both lasR and rsaL in P. aeruginosa PAO-R1 after growth in either the presence or absence of 1 µM exogenous PAI-1. Expression of lasR and rsaL was directed by either their native promoters (wt) or the lac promoter (plac) as indicated under each plasmid.

RsaL does not inhibit LasR expression. To determine whether RsaL has any effect on LasR expression, Western blot analysis of total protein obtained from cells collected in the early stationary phase of growth was performed. Representative results of three separate blots are shown in Fig. 7. In each assay, strains PAO-R1(pKDT17 lasR_plac lasB-lacZ), PAO-R1(pPCS15 lasR_plac rsaL_pwt lasB-lacZ), PAO-R1(pTS4001.7 lasR_pwt rsaL_pwt lasB-lacZ), and PAO-R1(pPCS16 lasR_pwt rsaL_plac lasB-lacZ) demonstrated similar signals for LasR antigen, indicating that the presence of RsaL does not inhibit LasR production. Strain PAO-R1(pLasR lasR_pwt lasB-lacZ) exhibited slightly elevated levels of LasR, as noted by the broader, more intense signal, and PAO-R1(pTS400 lasB-lacZ), which does not produce LasR, did not exhibit any signal in the assays.

View larger version (11K): [in this window] [in a new window]   FIG. 7.   Western blot analysis of LasR expression when rsaL is overexpressed in P. aeruginosa PAO-R1. Anti-LasR antibody (1:5,000) was used to visualize LasR in total cell lysates of strain PAO-R1 in the presence of lasR alone or both lasR and rsaL. Expression of lasR and rsaL was directed by either their native promoters (wt) or the lac promoter (plac) as indicated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

While details regarding the mechanism of quorum sensing in P. aeruginosa are continuously coming to light, very little is known about how the regulators themselves are regulated. Studies have revealed a variety of factors that affect the positive regulation of autoinduction (1, 30); however, no negative regulators of the P. aeruginosa quorum-sensing systems have been identified. This study describes a newly identified P. aeruginosa protein, RsaL, which negatively affects the las quorum-sensing system. Expression of rsaL resulted in decreased lasB expression, and mutations abolishing RsaL synthesis abrogated this effect, thereby confirming that RsaL was responsible for the repression.

The observation that rsaL decreased lasB expression led to the formulation of two hypotheses for its target of repression. First, RsaL could interact with a specific operator sequence upstream of the lasB gene. Evidence against this hypothesis was obtained from our data demonstrating that in the presence of LasR and PAI-1, lasB gene expression in an E. coli lysogen was not decreased when rsaL was expressed. Data from the lasI lysogen experiments supported the second hypothesis, in which RsaL directly affects one or more of the regulatory factors of the las quorum-sensing regulon. In E. coli, in the presence of LasR and added PAI-1, RsaL inhibited expression of lasI (Fig. 5). While lasI expression requires both LasR and PAI-1 (35), our Western blot analysis (Fig. 7) indicated that RsaL did not decrease LasR expression. Together these data indicated that RsaL specifically inhibits the expression of the lasI gene, preventing the production of PAI-1. In P. aeruginosa, in the absence of PAI-1, LasR would remain inactive and lasB would not be transcribed.

There have been only four reports to date of proposed negative regulators in quorum sensing. In Vibrio harveyi, LuxO was shown to repress autoinducible luminescence at the receptor portion of a two-component regulatory system (2). In Pantoea stewartii, the transcriptional activator EsaR acts as a repressor of quorum sensing at low cell density until sufficient AI becomes available for derepression to occur (3). TraM of Agrobacterium tumefaciens is proposed to repress autoinduction through protein-protein interactions with the transcriptional activator TraR (7, 15), and finally, in Erwinia carotovora, rsmA encodes a 6.8-kDa protein which has been shown to decrease synthesis of many extracellular enzymes (5, 20). The mode of action of the Erwinia protein, RsmA, is believed to be through repression of AI synthesis, a function similar to that predicted for RsaL. Cui and coworkers (5) demonstrated that multicopy rsmA suppressed the level of HSL [N-(3-oxohexanoyl)-L-homoserine lactone] in Erwinia culture supernatants and decreased the levels of the hslI transcript, which is required for HSL production. RsmA is a homolog of the E. coli CsrA protein, a negative regulator of carbon storage (31), and the predicted protein products of the two genes are 95% identical (5). CsrA is proposed to bind mRNA of its target gene, glg, to accelerate decay of the transcript (17, 18). The finding that the Erwinia rsmA gene was able to suppress glycogen synthesis in E. coli, together with the fact that, like CsrA, RsmA contains a putative RNA-binding domain, has led to the suggestion that RsmA may regulate gene expression by affecting mRNA stability (5). Analysis of the RsaL sequence revealed no similarities with either CsrA or RsmA. Furthermore, no RNA-binding motif equivalent to that described for RsmA or CsrA was identified in RsaL.

Whether RsaL binds directly to an operator element upstream of lasI and disrupts its expression or whether it affects the stability of the lasI mRNA has not been determined. However our  lysogen experiments suggest that it is through the former mechanism. In the E. coli lysogen, the lasI transcriptional fusion contains an RNase III cleavage site between the lasI and lacZ genes; consequently, the mRNA is cleaved into two separate messages. Destabilization of the lasI message should therefore have no effect on stability of the lacZ mRNA. Thus, the absence of -galactosidase activity in the lasI  lysogen containing RsaL suggests that RsaL affects lasI transcription and not lasI message stability. An interaction between RsaL and an element upstream of lasI would be a novel finding in quorum-sensing systems. It is interesting to note the proximity of the rsaL promoter/operator region to that of lasI. The small region defining these two promoters suggests that overlapping operators, or perhaps even the same operator, may be used for expression of both the rsaL and lasI transcripts by LasR and PAI-1. Intriguingly, two putative lux boxes are located between the rsaL and lasI genes (Fig. 1C). One of the lux boxes is approximately centered in the rsaL-lasI intergenic region, and the second encompasses the lasI start of translation. Whether one or both of these regions is used for regulating the expression of rsaL and/or lasI has yet to be established. Recently, Fuqua and coworkers described the arrangement of the traI and traC genes of A. tumefaciens whereby the divergently transcribed genes share an operator targeted by TraR, a LasR homologue, and AAI, the Agrobacterium autoinducer (7). Previous studies have proposed that the LuxR homologues may function as activators by stabilizing and positioning the RNA polymerase complex at target gene promoters (36, 37). The commonly shared operator between two independent and divergent promoters implies that if LuxR-like proteins interact directly with RNA polymerase, they may dimerize and promote transcription in opposite directions.

In P. aeruginosa, LasR and PAI-1 globally regulate many products associated with virulence, as well as the second P. aeruginosa quorum-sensing system. We theorize that at low cell density, RsaL inhibits transcription of lasI by binding to the lasI operator region, thereby blocking activation by LasR-PAI-1. As the cell density increases, so does the basal level of PAI-1, which enables sufficient LasR-PAI-1 formation to outcompete RsaL for binding to the lasI operator. Thus, it appears that during the early stages of growth, RsaL blocks the quorum-sensing cascade by inhibiting the transcription of lasI. We speculate that in an RsaL mutant, lasI would be turned on much earlier in the growth cycle, resulting in premature activation of the las quorum-sensing-controlled genes. Studies are currently in progress to generate an RsaL null mutant and assess the phenotypic effects associated with this mutation. During the infection process, the repression of virulence factor expression may be critical for minimizing both the immunogenicity and host damage associated with these factors, allowing the organism to achieve a high population density prior to dissemination. Given the essential role of many of the quorum-sensing-controlled genes in P. aeruginosa virulence, RsaL appears as a pivotal regulator of P. aeruginosa pathogenicity. Further studies in progress should clarify the role of this unique regulator in P. aeruginosa virulence.