LuxR-family 'solos': bachelor sensors/regulators of signalling molecules.

N-Acylhomoserine lactone (AHL) quorum-sensing (QS) signalling is the best-understood chemical language in proteobacteria. In the last 15 years a large amount of research in several bacterial species has revealed in detail the genetic, molecular and biochemical mechanisms underlying AHL signalling. These studies have revealed the role played by protein pairs of the AHL synthase belonging to the LuxI family and cognate LuxR-family AHL sensor-regulator. Proteobacteria however commonly possess a QS LuxR-family protein for which there is no obvious cognate LuxI synthase; these proteins are found in bacteria which possess a complete AHL QS system(s) as well as in bacteria that do not. Scientists are beginning to address the roles played by these proteins and it is emerging that they could allow bacteria to respond to endogenous and exogenous signals produced by their neighbours. AHL QS research thus far has mainly focused on a cell-density response involving laboratory monoculture studies. Recent findings on the role played by the unpaired LuxR-family proteins highlight the need to address bacterial behaviour and response to signals in mixed communities. Here we review recent progress with respect to these LuxR proteins, which we propose to call LuxR 'solos' since they act on their own without the need for a cognate signal generator. Initial investigations have revealed that LuxR solos have diverse roles in bacterial interspecies and interkingdom communication.


Introduction
Bacteria in the wild mainly live in close association with many different bacteria and eukaryotic hosts, meaning that they constantly need to monitor and communicate with their neighbours. It is now established that bacteria produce signalling molecules which are believed to allow them to monitor their population density in a process called quorum sensing (QS) (Fuqua et al., 1994). The predominant QS signalling molecules produced and detected by proteobacteria are N-acylhomoserine lactones (AHLs). QS research has thus far largely concentrated on monocultures and has shown that AHLs are constantly produced and accumulate as the bacterial population increases. When the bacterial population reaches a certain size (a quorum), the bacteria respond to AHLs and synchronize community behaviour through regulation of their gene expression. These investigations have linked AHL QS with phenotypes that benefit the AHL-producing community, including production of virulence factors, motility, nodulation, plasmid transfer, antibiotic production, bioluminescence and biofilm formation (Bassler, 2002;Waters & Bassler, 2005;Whitehead et al., 2001).
A typical AHL QS system is most commonly mediated by two proteins, belonging to the LuxI and LuxR protein families (Fig. 1a). LuxI-type proteins synthesize AHLs which interact directly with the cognate LuxR-type protein, and this complex usually allows them to bind specific promoter sequences called lux-boxes affecting expression of QS target genes (Fuqua & Greenberg, 2002). AHLs can also be synthesized by the AinS/LuxM (belonging to Vibrio fischeri and Vibrio harveyi respectively) family, and these do not display any similarity to the LuxI protein family (Bassler et al., 1993;Gilson et al., 1995). The AHLs they produce differ in length of the acyl-chain moiety and substitution at position C-3, which can either be part of a methylene group or carry an oxo-or hydroxyl group. QS LuxR proteins bind and respond optimally to the AHL produced by the cognate LuxI-family protein, guaranteeing a good degree of selectivity. In most cases luxI/luxR pairs are actually genetically linked; however there are examples where the luxI/luxR functional pairs are distantly located in the bacterial chromosome or plasmid. LuxR proteins which belong to a LuxI/LuxR functional pair can however sometimes bind structurally closely related AHLs, albeit with less efficiency.
AHL QS research now needs to more insistently address the role in bacterial interspecies communication and understand whether AHLs take part in interkingdom signalling. A recent observation related to these roles is that numerous sequenced proteobacterial genomes have QS-related LuxR AHL sensors/regulators which lack a cognate LuxI AHL synthase (Case et al., 2008). These unpaired QS LuxR-family proteins have been called orphans (Fuqua, 2006) and possess the typical modular structure having an AHL-binding domain at their Nterminus and a helix-turn-helix DNA-binding domain at their C-terminus. Some, however, differ in their length and others lack conserved amino acids within the AHL-binding domain (Fuqua & Greenberg, 2002;Fuqua, 2006). In this mini-review we assess current knowledge and discuss the possible roles of these proteins, which we propose to designate as LuxR-family 'solos'. We believe 'solo' is a more appropriate term than 'orphan', since in bacterial genetics 'orphan' is often used to describe a putative gene of unknown function which does not have any homologues. The name 'solo' is therefore more suitable to describe this growing novel family of LuxR proteins (Table 1).

Working models for LuxR solos
Genetically, thus far AHL QS has been found exclusively in proteobacteria, and analysing 265 proteobacterial genomes revealed that 68 of them possess at least one LuxI and one LuxR homologue (Case et al., 2008). Interestingly, however, as many as 45 of these 68 genomes possess more LuxR than LuxI homologues, indicating that they probably possess LuxR solos (as depicted in Fig. 1a). A further 45 genomes do not possess an AHL synthase gene, but they contain LuxR solos. Nevertheless, it cannot be excluded that some bacteria which do not possess luxI homologues are able produce AHLs. LuxR solos are therefore widespread in bacteria which possess a complete AHL QS system(s) as well as in bacteria that apparently do not synthesize AHLs. LuxR solos present in AHL-producing bacteria could be sensing the endogenously produced AHL(s) and thus extending the AHL QS regulon to other gene targets. Alternatively, they could have a more relaxed or different specificity towards AHLs or possibly other ligands produced by different bacterial neighbours. The presence of LuxR solos in non-AHL-producing bacteria, on the other hand, could allow them to respond to AHLs produced by other bacteria, possibly allowing them to team up or maybe switch to a competitive behaviour towards their neighbours. It cannot be excluded that some LuxR solos which lack conserved amino acid residues in the AHL domain are functioning in a ligand-independent way. Then again, these imperfect LuxR solos could be binding to signals which are not AHLs but are produced by nearby bacteria or eukaryotic cells. If they respond to compounds produced by eukaryotes, they are not necessarily involved in QS but rather in informing bacteria of their whereabouts. Recent investigations have revealed that the processes described by these different models are actually taking place (summarized in Figs 1 and 2), highlighting the diverse roles in bacterial interspecies and interkingdom communication played by these proteins.

LuxR solos in AHL-producing bacteria
ExpR of Sinorhizobium meliloti, BisR of Rhizobium leguminosarum bv. viciae and QscR of Pseudomonas aeruginosa are LuxR solos in AHL-producing bacteria which have been well characterized. All of these bacterial species actually possess multiple complete AHL QS systems and produce several types of AHL molecules. The functional features of these three LuxR solos are integrated with the resident AHL QS regulatory networks (as depicted in Fig. 1a). A better understanding of the role of these LuxR solos could provide insights into the rationale for production of several AHLs by these bacteria and other closely related species that occupy a similar ecological niche.

ExpR of Sinorhizobium meliloti
ExpR was initially identified as being essential for the synthesis of symbiotically active galactoglucan (EPSII) in S.   LuxR-family solos in proteobacteria meliloti Rm1021. The wild-type strain produces negligible EPSII due to an inactive copy of expR with an insertion element (ISRm2011-1); precise spontaneous excision of this insertion element results in functional expR, inducing production of EPSII (Pellock et al., 2002). The SinI/R system of S. meliloti Rm1021 produces several long-chain AHLs ranging from C 12 -to C 18 -HSL and the presence of another system (designated Mel) has been hypothesized (i.e. this system has not been identified) which could produce short-chain AHLs . ExpR is not associated with a LuxI homologue; hence it is a solo and regulates expression of its targets in response to SinI-AHLs McIntosh et al., 2008).
Comparison of S. meliloti strains differing in the functional status of expR, sinI or sinR in various combinations showed that the regulation of the majority of target genes by the SinI/R system requires the presence of an active ExpR (Hoang et al., 2004(Hoang et al., , 2008. Significantly, ExpR regulates a separate set of genes, including those involved in nitrogen fixation and metal transport; expression of several of these genes is independent of the SinI-AHLs (Hoang et al., 2004).
ExpR is different from other typical QS LuxR regulators in being soluble and stable in the absence of AHLs (LuxRfamily proteins when overexpressed are usually highly insoluble in the absence of the AHL ligand) as well as not binding to a typical lux-box; in fact it has been reported to bind to a 24 bp sequence located upstream of the putative lux-box of the sinI promoter (Bartels et al., 2007;McIntosh et al., 2008). ExpR also specifically binds to the promoter regions of galactoglucan biosynthesis genes wgaA and wgeA (previously known as exp genes) in a 3-oxo-C14-HSLdependent manner, while binding to exoI and exsH (genes involved in succinoglycan biosynthesis) promoters is independent of the presence of AHL (McIntosh et al., 2008). Also, succinoglycan production is induced in the presence of SinI-AHLs and repressed without these AHLs in an expR + strain. It is however produced independently of AHLs in the absence of a functional ExpR (Glenn et al., 2007). ExpR is very versatile and either activates or represses target genes depending on the presence or absence of AHL molecules. Interestingly S. meliloti Rm1021 possesses four additional putative LuxR solos; recently one of these, designated NesR, has been shown to be involved in nutritional and stress responses as well as in nodulation (Patankar & Gonzalez, 2009). NesR is not integrated with the SinI/R system and its homology to the xanthomonad LuxR solos OryR and XccR suggests that the ligand for NesR might be an uknown plant compound (see below). Understanding ExpR-mediated regulation and the identification of signal molecules sensed by the other LuxR solos present in S. meliloti would help in deciphering this rather complex QS signalling network.

BisR of Rhizobium leguminosarum bv. viciae
BisR (bacteriostasis induction sensor) is a LuxR solo of R. leguminosarum bv. viciae strain A34 (a symbiont of pea and vetch) which contains the symbiotic plasmid pRL1J1. Strain A34 has four complete QS systems and produces a number of different AHLs .
The bisR gene is located on symbiotic plasmid pRL1J1 and is responsible for controlling growth inhibition in the presence of 3OH-C14 : 1-HSL (initially called small bacteriocin; Wilkinson et al., 2002) and transfer frequency of pRL1JI (Wilkinson et al., 2002). BisR represses the expression of cinI, which is responsible for the synthesis of 3OH-C14 : 1-HSL as well as inducing the expression of triR (encoding the LuxR regulator TriR, now designated TraR) in response to 3OH-C14 : 1-HSL (Danino et al., 2003;Wilkinson et al., 2002). Danino et al. (2003) revealed an elegant system in which BisR represses the expression of cinI and thereby maintains minimal amounts of 3OH-C14 : 1-HSL. Upon sensing of a critical concentration of 3OH-C14 : 1-HSL produced by an optimum number of recipient cells, the plasmid transfer functions are induced by BisR. In contrast to its role in plasmid transfer, not much is known about the mechanism of BisR-mediated growth inhibition of R. leguminosarum bv. viciae in the presence of 3OH-C14 : 1-HSL. Future work will probably focus on identification of additional target genes of BisR and its interaction with the resident complete AHL QS systems.

QscR of Pseudomonas aeruginosa
QscR is a well-studied LuxR solo of P. aeruginosa, which has two complete AHL QS systems, the LasI/R and RhlI/R systems. LasI directs the synthesis of 3-oxo-C12-HSL whereas RhlI synthesizes C4-HSL. The two systems are intimately connected, the LasI/R system regulating rhlI/R forming a cascade, and together are involved in the regulation of many virulence factors (Smith & Iglewski, 2003).
The QscR solo responds to endogenously LasI-produced 3oxo-C12-HSL and one of its roles is to time and prevent premature expression of endogenous AHL signals and virulence factors (Chugani et al., 2001;Lequette et al., 2006). In response to endogenous 3-oxo-C12-HSL, QscR also regulates an independent set of target genes distinct from those regulated by the LasI/R and RhlI/R systems . This role of QscR extends the AHL QS regulon to other targets, probably allowing a more timely and efficient response to cell density by P. aeruginosa. However, QscR can also respond well to 3oxo-C10-HSL, demonstrating a more relaxed specificity for AHL binding than LasR and RhlR . P. aeruginosa produces 3-oxo-C10-HSL via LasI; however, 3oxo-C12-HSL is the dominant signal produced. The more relaxed specificity towards AHLs is an indication that QscR could also respond to exogenous AHLs produced by neighbouring bacteria.

Other LuxR solos in AHL-producing bacteria
Other LuxR solos such as AviR and AvhR of Agrobacterium vitis have been reported (Hao et al., 2005;Zheng et al., 2003). A. vitis strain F2/5 most likely possesses two complete QS systems: the AvsI/R system responsible for the production of long-chain AHLs and another system that directs the production of short-chain AHLs (Hao & Burr, 2006). Both AviR and AvhR solos are required for induction of necrosis in grapes and hypersensitive response on non-host tobacco plants. Mutants in aviR produced lower levels of many AHLs while mutations in avhR resulted in higher intensity of the signals, suggesting contrasting effects of these two regulators on genes responsible for AHL production (Hao et al., 2005;Zheng et al., 2003). In AvhR, three of the six conserved amino acid residues of the AHL-binding domain have substitutions; these include the important residues Trp57, Asp70 and Trp85 designated with reference to TraR, where they are involved in AHL binding (Hao et al., 2005). Therefore the possibility that AvhR binds to a non-AHL molecule or functions without a signalling molecule cannot be ruled out.
CarR of Serratia marcescens is an example of a LuxR solo that regulates target gene expression in a ligand-independent manner (Cox et al., 1998). S. marcescens CarR functionally complements an Erwinia carotovora subsp. carotovora carI carR double mutant without the need for 3oxo-C6-HSL, whereas its homologue CarR of E. carotovora regulates production of carbapenem antibiotic in a 3-oxo-C6-HSL-dependent manner. S. marcescens is an AHL producer; the LuxI/LuxR pair designated SmaI/R constitutes the QS system, which produces and responds to C4-HSL and C6-HSL, regulating several cellular functions. Current understanding of these proteins indicates that SmaR acts as a repressor of carR and carbapenem biosynthetic genes at low cell densities; repression is possibly released by binding of S. marcescens AHLs to SmaR (Slater et al., 2003).
Finally, many members of the genera Pseudomonas, Burkholderia, Rhizobium and Agrobacterium which have complete AHL QS systems possess LuxR solos (Case et al., 2008). The function of most of these is currently unknown; however, among those LuxR solos that have been studied up to now, many are interconnected with the resident AHL QS systems. No clear indications have yet emerged as to whether they are involved in interspecies signalling; interestingly, several members of these genera share similar habitats.

LuxR solos in non-AHL-producing bacteria
Since many non-AHL-producing proteobacteria (almost 20 %) possess LuxR solos, this could imply that in communities with AHL producers and non-producers some of the members can detect and respond to molecular signals produced by others. Surprisingly, to our knowledge only two LuxR solos in non-AHL-producing bacteria have thus far been investigated in some detail. One, called SdiA, is present in members of the genera Salmonella, Escherichia and Klebsiella and is able to bind and detect AHLs produced by others (as depicted in Fig. 1b). The other is from plant-pathogenic Xanthomonas spp. and is also very well conserved in other plant-associated bacteria. In two Xanthomonas species it is involved in virulence upon binding an unknown plant-produced low-molecular-mass compound which is not an AHL. This extends the role of LuxR solos to interkingdom signalling by detecting non-AHL compounds produced by eukaryotes (as depicted in Fig. 1b).
The function of SdiA is best understood in Salmonella enterica serovar Typhimurium, where it has been determined that it regulates two Salmonella-specific loci in response to AHLs (Ahmer, 2004;Michael et al., 2001;Smith & Ahmer, 2003). They are the rck operon (resistance to complement killing), which is located on the S. Typhimurium virulence plasmid, and the srgE (sdiA regulated gene) locus located in the chromosome (Ahmer, 2004;Smith & Ahmer, 2003). The rck operon consists of six genes (pefI, srgD, srgA, srgB, rck and sgrC), most of which are believed to be involved in adhesion and resistance to complement killing. The chromosomal srgE locus on the other hand consists of a single gene which encodes a predicted protein of unknown function containing a coiledcoil domain. Salmonella sdiA mutants are not attenuated in virulence in several animal models and thus SdiA is believed to regulate only accessory virulence factors (Ahmer, 2004).
Scientists now need to determine whether SdiA is sensing AHLs in vivo since the intestinal environment contains a large number of different microbial species. Along these lines, Smith et al. (2008) recently observed that SdiAmediated AHL detection did not occur in the intestinal tract of several animals (e.g. mouse, chicken, pig, rabbit and cow); however, SdiA of Salmonella did detect AHLs during transit through the intestinal tract of turtles which harboured AHLproducing Aeromonas hydrophila.
Unlike the SdiA of Salmonella, no experiments have been performed on the sdiA gene of Klebsiella spp., whilst the data currently available for Escherichia coli are rather difficult to interpret. The sdiA gene in E. coli has been isolated on three occasions for conferring different phenotypes when overexpressed in a plasmid (Ahmer, 2004). It increases transcription of the cell division operon ftsQAZ and increases resistance to certain antibiotics and quinolones, probably through activation of efflux pumps (Ahmer, 2004;Rahmati et al., 2002;Sitnikov et al., 1996). E. coli SdiA has been shown to be able to activate Salmonella and E. coli gene promoters in an AHLdependent manner (Ahmer, 2004;Sitnikov et al., 1996).
Functionally, however, no E. coli gene has thus far been shown to be regulated by the single chromosomal copy of sdiA. Overexpression of SdiA in an enterohaemorrhagic E. coli strain also resulted in repression of certain virulence factors (Kanamaru et al., 2000). Apparently this repression is dependent on an uncharacterized E. coli-produced factor present in spent culture supernatants. This opens the door to the prospect that SdiA also binds to an as yet unidentified factor produced by E. coli.
SdiA was the first LuxR solo in non-AHL-producing bacteria shown to be able to bind AHLs. However, open questions remain, including whether SdiA also binds/ responds to an E. coli-produced factor and what are its functional roles. These questions might be answered by studying the interaction of E. coli and Salmonella with other bacterial species, since waste products, metabolites or molecular signals produced by neighbours can influence the functioning of SdiA. This approach makes sense since in nature both Salmonella spp. and E. coli live in close association with many other bacterial species. Finally it must be noted that recently, indole has been identified as a signalling compound produced by E. coli (reviewed by Walters & Sperandio, 2006). Indole signalling has been shown to be important for regulating biofilm formation and research is also now focusing on determining whether indole and SdiA interact and if they are in some way connected (Lee et al., 2007). SdiA could possibly be involved with the response to at least two different kinds of signalling molecules -AHLs (exogenous) and indole (endogenous) -and modulate cellular functions appropriately. Further studies are needed to confirm these possible roles of SdiA in intra-and interspecies signalling in E. coli.
The ability of the SdiA solo to bind and respond to AHLs interferes with the practical use of E. coli as an AHL detector/biosensor strain. E. coli has been engineered by several research groups via the introduction of plasmids harbouring heterologous luxR genes, target gene promoters and reporter systems in order to detect and respond to the presence of exogenously provided AHLs (Steindler & Venturi, 2007). It has been observed that SdiA interferes with these experiments and hence the best results are obtained when an E. coli sdiA knockout mutant is used for these experiments (Lindsay & Ahmer, 2005).

The LuxR solo of xanthomonads and other plantassociated bacteria
The plant-pathogenic Xanthomonas oryzae pv. oryzae (Xoo) and Xanthomonas campestris pv. campestris (Xcc) possess a LuxR solo, designated OryR and XccR, which is required for full virulence to rice and cabbage, respectively. Both OryR and XccR do not bind or respond to AHLs and in their AHL-binding domain they lack two of the several conserved residues involved in AHL binding (Ferluga et al., 2007;Zhang et al., 2007). XccR regulates in planta the neighbouring proline iminopeptidase (pip) virulence gene.
It associates with a plant factor and functions as a transcriptional activator by binding to a lux-box present in the promoter of the pip gene (Zhang et al., 2007). Similarly, OryR binds a compound present in the rice plant. This was concluded following the observation that OryR was not solubilized by many of the structurally different AHLs but OryR solubilization was achieved in the presence of rice extract (Ferluga et al., 2007). The concentration of the molecule in rice which interacts with OryR increases when rice is infected with Xoo, meaning that rice most probably increases its production upon pathogen attack (Ferluga & Venturi, 2009). Just like XccR, OryR also positively regulates expression of the adjacent pip gene upon exposure to the rice signal molecule. It is probable that OryR and XccR solos participate in interkingdom signalling, whereby they facilitate successful infection by detecting and binding to a host signal molecule(s).
Orthologues of OryR and XccR are present in the genomes of several other plant-associated bacteria which are both beneficial and pathogenic, for example Pseudomonas fluorescens and Pseudomonas syringae (Zhang et al., 2007).
Since some strains of these species do produce AHLs, it cannot be excluded that these solos can also respond to AHLs. It is postulated however that probably a subclass of LuxR solos are employed in interkingdom signalling between bacteria and plant; this is not so surprising, as bacteria have co-evolved with plants for many years. Plants have been reported to produce compounds that are able to act as agonists or antagonists to bacterial AHL QS systems and these have been called AHL mimics (Bauer & Mathesius, 2004). AHL mimics from several plants, including rice, are able to stimulate gene expression via LuxR-family AHL sensors/regulators. To date the structures of these plant compounds are unknown and it cannot be excluded that similar molecules are involved in interkingdom signalling with the XccR/OryR subgroup of LuxR solos.
Other LuxR solos in non-AHL-producing proteobacteria In Brucella melitensis, two LuxR solos called VjbR and BlxR have been shown to be involved in regulating virulence loci as well as cross-regulating their gene expression (Delrue et al., 2005;Rambow-Larsen et al., 2008;Uzureau et al., 2007). Surprisingly, B. melitensis has been shown to produce C12-HSL even though it does not have a luxI homologue in its genome (Taminiau et al., 2002). C12-HSL is believed to interact with VjbR and control virulence-associated loci (Delrue et al., 2005). As mentioned previously, it is therefore possible that some bacteria which do not possess luxI homologues are able produce AHLs.
Predicted LuxR solos in non-AHL-producing bacteria are present in several other bacterial genera, including Pseudomonas and Shigella (Case et al., 2008). It is probable that scientists are currently studying these regulators and we will therefore have insights into their mode of action in the near future. Most often LuxR solos within a genus are closely related, indicating a probable conserved function.

Concluding remarks
Bacteria most probably are constantly vigilant of the living organisms present in the surroundings as this will affect decisions on their programme of gene expression. Consequently, an important question that we need to address in the future is how often bacteria are involved in interspecies and interkingdom signalling. Initial studies with LuxR solos indicate that some of these may be important players in this process (Table 1 summarizes the LuxR solos studied to date).
The few LuxR solos studied so far highlight that they could be highly diverse in the type of ligands they respond to, as well as their mechanism of regulation of target genes (Fig.  2). Generally, this probably depends on their occurrence in an AHL-producing or non-AHL-producing bacterium and the immediate surrounding environment. It is clear that LuxR solos extend bacterial participation to processes other than just cell-density-dependent signalling. As QS research now enters the next phase of studying mixed bacterial communities, it is important to recognize the functional resourcefulness of LuxR solos. Any future intervention strategies based on disruption/manipulation of bacterial communication networks will require a better understanding of these proteins.