Early Activation of Quorum Sensing

More than 30 years ago microbiologists found a way to induce luminescent bacteria to emit light prematurely. By adding cell-free culture fluid from a bright bioluminescent culture at high cell density to a nonluminous low-cell-density culture, they were able to eliminate the characteristic lag in bioluminescence (5, 12, 16). Nealson et al. proved that the “light switch” that controls the bioluminescence genes in the marine bacterium Vibrio (formerly Photobacterium) fischeri was cell density dependent, and they hypothesized that this light switch was controlled by a bacterially produced autoinducer signal (16).

Bioluminescence could occur only after the autoinducer accumulated in cultures to a threshold level that was attained at high cell density (>10⁷ CFU/ml) (16). Eberhard et al. went on to elucidate the structure of the autoinducer molecule that was responsible for this effect: the first known acyl-homoserine lactone, 3-oxo-hexanoyl-homoserine lactone (3-oxo-C6-HSL) (6). 3-oxo-C6-HSL was found to freely diffuse from bacterial cells into the surrounding medium and vice versa (11). This phenomenon of cell-density-dependent autoinduction of specific bacterial genes is now referred to as quorum sensing and involves two conserved regulatory gene products: (i) a LuxI-type acyl-HSL synthase and (ii) a LuxR-type transcriptional activator whose activity requires a particular acyl-HSL made by the cognate LuxI enzyme (9; for recent reviews, see references 8 and 28). In addition to the LuxR/LuxI control, a second quorum-sensing system regulates the luminescence (lux) genes in V. fischeri. This second system consists of an acyl-HSL synthase, AinS, which directs the synthesis of octanoyl-HSL (C8-HSL) (13). V. fischeri ainS mutants exhibited early luminescence, whereas the addition of C8-HSL delayed luminescence in cultures of wild-type cells (13). The above examples illustrate how the timing of a quorum-sensing-controlled (QSC) process can be advanced merely by early exposure of the cells to a critical concentration of acyl-HSL. QSC phenomena include antibiotic production, virulence gene expression, and other processes in many diverse bacteria. The focus of this commentary will be to highlight a number of recently identified gene products that modulate the timing of QSC gene expression in Pseudomonas aeruginosa. One such regulator, encoded by mvaT, is described in a paper in this issue (4). In most cases these gene products serve to prevent the early activation of quorum sensing.

Two intertwined quorum-sensing systems have been shown to be involved in virulence, biofilm development, and many other processes in P. aeruginosa. The first system (Las) was discovered by Iglewski and colleagues and consists of a lasI-encoded acyl-HSL synthase and the lasR-encoded transcriptional activator (10, 19). The second system (Rhl) was found by a number of investigators and consists of an rhlI-encoded acyl-HSL synthase and an rhlR-encoded transcriptional activator (2, 14, 17, 18). The respective quorum-sensing systems each produce and respond to a specific acyl-HSL: LasI directs the synthesis of 3-oxo-dodecanoyl-HSL (3-oxo-C12-HSL) (20), and RhlI directs the synthesis of butyryl-HSL (C4-HSL) (33). Further details of the complex regulation of the two quorum-sensing systems and how they are thought to control expression of the several genes in P. aeruginosa has been reviewed elsewhere (22). Recently Whiteley et al. using a P. aeruginosa lasI rhlI double mutant identified nearly 40 qsc (quorum sensing controlled) genes that showed a fivefold or greater response to exogenously added acyl-HSL signals (29). The qsc genes were classified based on the temporal pattern of their responses in cells grown in the presence of the Las signal, 3-oxo-C12-HSL, and/or the Rhl signal, C4-HSL (29). A number of “early” qsc genes were found that responded immediately to exogenously added signals (29), suggesting that these genes behave like the lux genes of V. fischeri (described above) and the carbapenem biosynthesis genes of Erwinia (32). Another group of qsc genes (called “late-response”) were able to respond to the signals only during stationary growth phase (29). Whiteley et al. hypothesized that these genes might be under the control of some unknown control mechanism(s) preventing their “early” expression. Since these seminal observations were made a number of other proteins have been found that support this hypothesis, including the stationary-phase sigma factor RpoS (30), a third LuxR homolog (QscR) (3), a secondary metabolite regulator, RsmA (24), and the stringent response protein RelA (27), all of which are involved in modulating expression of qsc genes. P. aeruginosa rpoS, qscR, and rsmA deletion mutants each abolish the lag in expression of the cyanide biogenesis genes (hcn) and pyocyanin genes (phz) (3, 24, 30). All of these studies have begun to dissect the mechanisms responsible for the observed phenotypes. In the case of RpoS, it was found to negatively regulate the expression of the C4-HSL synthase gene, rhlI (30). QscR was found to negatively regulate expression of both rhlI and of the 3-oxo-C12-HSL synthase gene, lasI (3). P. aeruginosa qscR mutants showed early activation of a number of qsc genes and premature synthesis of both signals C4-HSL and 3-oxo-C12-HSL (3). Overexpression of the rsmA gene product resulted in decreased production of QSC virulence factors and acyl-HSLs, whereas deletion of rsmA led to early activation of lasI and thus early synthesis of 3-oxo-C12-HSL (24).

RsmA, RpoS, and QscR all negatively regulate the Rhl or Las quorum-sensing systems, thus preventing early activation of these systems. The RelA protein synthesizes the nucleotide guanosine 3′,5′ bisdiphosphate (ppGpp). Under amino acid starvation conditions, the ppGpp synthetase activity of RelA is induced. van Delden et al. found that overexpression of RelA led to early induction of a number of QSC processes, including elastase expression and acyl-HSL production (27). Further work will be needed to determine the precise role of RelA in quorum sensing.

In this issue, Diggle et al. (4) report results of a screen of P. aeruginosa transposon mutants for genes that alter the expression of a qsc reporter gene, lecA, which encodes the PA-IL lectin. They report that two genes, clpA and mvaT (previously not shown to be involved in quorum sensing), when mutated, result in a twofold increase in lecA expression. One of their most interesting findings was that addition of exogenous acyl HSLs to the P. aeruginosa mvaT mutant results in early activation (or “advancement”) of lecA expression. The authors found that C4-HSL and 3-oxo-C12-HSL concentrations (determined at a single time point in stationary phase) were increased approximately 1.4-fold in the mvaT mutant compared to the wild type, whereas in the clpA mutant the C4-HSL and 3-oxo-C12-HSL concentrations were both increased, 4-fold and 1.3-fold, respectively. Further time course studies with the P. aeruginosa mvaT mutant following rhlI and lasI expression as well as acyl-HSL concentration measurements will be necessary to determine the precise role of mvaT in quorum sensing. MvaT is a homolog of a subunit of a Pseudomonas mevalonii heteromeric transcription factor (25), whereas ClpA is a chaperone that functions as an ATP-dependent “unfoldase” which assists the intracellular protease ClpP in protein turnover in the cell (31). More work will be needed to determine how ClpA affects lecA expression. Are the effects of MvaT and ClpA exerted directly on the lecA promoter region, or does MvaT exert indirect control on lecA via other factors?

mvaT is one of many transcriptional control proteins in P. aeruginosa, encoded by vfr, gacA, lasR, relA, rsaL, rpoS, and rsmA, that modulate the expression of the Las and Rhl quorum-sensing systems (for a recent model, see reference 15); there is also an additional signal molecule, 2-heptyl-3-hydroxy-4-quinolone (termed Pseudomonas quinolone signal or PQS) (23). As this list of known P. aeruginosa regulators increases, the need for further systematic and global approaches to examining the relative roles of these and unknown factors becomes increasingly apparent. High-throughput efforts will be needed to examine each mutant cultivated in identical conditions. Multigene knockout mutants also need to be generated, and the mutants will need to be tested in relevant animal models to determine the effects of the various mutations on virulence.

Diggle et al.'s mutagenesis was performed on a strain of P. aeruginosa carrying a chromosomally encoded lecA-lux reporter. Whiteley et al. showed early- and late-response qsc genes with a wide range of expression intensity (29). Thus, it seems possible that future screens using specific late-response qsc genes may lead to the identification of additional novel regulators of quorum sensing. One interesting screen would be to use DNA microarrays to study the expression profile of P. aeruginosa in the presence and in the absence of signals and using knockout mutants of the various regulatory genes such as rpoS, qscR, gacA, rsmA, mvaT, and clpA. One important consideration in designing these experiments will be to examine the transcriptional profile at multiple time points throughout the entire growth curve from low cell density, through mid-log phase, to high cell density and beyond into stationary phase. A comparison of the relative impacts of each mutation on the cell's transcriptional profile may uncover the most important points in the QSC pathways. In addition to studying the temporal pattern of global changes in transcription in these mutants, the temporal pattern of acyl-HSL expression by these mutants could be readily examined by de novo radiolabeling of the signals and subsequent high-performance liquid chromatography analysis (26). This method was described using P. aeruginosa but has recently been used to determine temporal patterns of multiple acyl-HSLs synthesized by Rhizobium leguminosarum (1). Interestingly Blosser-Middleton and Gray showed that addition of exogenous 3-OH-C14:1-HSL to cultures resulted in early and increased synthesis of 3 acyl-HSLs. These global and temporal methodologies offer more unbiased approaches to understanding quorum sensing than focusing only on a single or small set of qsc target genes and acyl-HSLs.

Even though the above methods are helpful, a careful functional analysis of each new gene or effect is still required. In the Diggle et al. study (4) one of the P. aeruginosa genes that was found encodes an efflux pump in the acriflavin resistance protein family (www.pseudomonas.com) and when knocked out causes a 90% decrease in lecA expression. Whiteley et al., using a P. aeruginosa lasI rhlI double mutant to screen for qsc genes, had previously identified this gene, which they termed qsc133 (ORF PA4207). The qsc133 gene behaved like a late-response gene whose expression was induced ninefold in response to a combination of both C4-HSL and 3-oxo-C12-HSL but not at all to either signal alone (29). This result suggests that qsc133 expression is likely to be controlled by additional unknown regulatory systems. One hypothesis is that the PQS could be involved. Quinolones contain two fused six-membered rings, and interestingly, acriflavins such as acridine orange contain three fused rings. It is possible that the putative qsc133 pump is involved in export of PQS out of the cell. Why would an efflux pump have an effect on a qsc gene like lecA? P. aeruginosa was shown previously to use the MexAB-OprM multidrug pump to actively export 3-oxo-C12-HSL (7, 21). Other pumps were not explored in those studies. Perhaps one of the acyl-HSL signals or even the PQS or some unidentified signal is involved in lecA expression and requires a functional qsc133 efflux system.

Based on the original findings advancing the timing of bioluminescence gene expression by exogenous addition of acyl-HSL, which shows a dramatic and immediate stimulation of light production in wild-type V. fischeri, and based on the recent findings for P. aeruginosa by Chugani et al. which show that the 3-O-C12HSL synthase gene, lasI, can be immediately induced upon exogenous addition of a signal (3), it is apparent that there still remain additional layers of control that prevent lecA and other late-response genes from being expressed early. What other prevention factors are stopping P. aeruginosa from expressing these late-response genes? The exciting answers could hold keys to unlocking other secrets about this pathogen and the many other bacteria with homologous quorum-sensing systems.

ADDENDUM IN PROOF

Early activation of quorum sensing has been demonstrated with other types of quorum-sensing systems besides the LuxR/LuxI-type system. In Vibrio harveyi, Lilley and Bassler have shown that LuxO represses early activation of the bioluminescence operon (B. N. Lilley and B. L. Bassler, Mol. Microbiol. 36:940-954, 2000). Recently, in Vibrio cholerae, LuxO was found to regulate virulence gene expression at low cell density and luxO mutants were less virulent in vivo in a mouse model (J. Zhu, M. B. Miller, R. E. Vance, M. Dziejman, B. L. Bassler, and J. J. Mekalanos, Proc. Natl. Acad. Sci. USA 99:3129-3134, 2002). For further review of this and other non-LuxR/LuxI quorum-sensing systems including those found in gram-positive bacteria, see the paper by Miller and Bassler (M. B. Miller and B. L. Bassler, Annu. Rev. Microbiol. 55:165-199, 2001).