A Unique Regulator Contributes to Quorum Sensing and Virulence in Burkholderia cenocepacia

Burkholderia cenocepacia causes chronic and life-threatening respiratory infections in immunocompromized people. The B. cenocepacia N-acyl-homoserine lactone (AHL)-dependent quorum sensing system relies on the production of AHLs by the synthases CepI and CciI while CepR, CciR and CepR2 control expression of many genes important for pathogenesis. Downstream from, and co-transcribed with cepI, lies BCAM1871 encoding a hypothetical protein that was uncharacterized prior to this study. Orthologs of B. cenocepacia BCAM1871 are uniquely found in Burkholderia spp and are conserved in their genomic locations in pathogenic Burkholderia. We observed significant effects on AHL activity upon mutation or overexpression of BCAM1871, although these effects were more subtle than those observed for CepI indicating BCAM1871 acts as an enhancer of AHL activity. Transcription of cepI, cepR and cciIR was significantly reduced in the BCAM1871 mutant. Swimming and swarming motilities as well as transcription of fliC, encoding flagellin, were significantly reduced in the BCAM1871 mutant. Protease activity and transcription of zmpA and zmpB, encoding extracellular zinc metalloproteases, were undetectable in the BCAM1871 mutant indicating a more significant effect of mutating BCAM1871 than cepI. Exogenous addition of OHL restored cepI, cepR and fliC transcription but had no effect on motility, protease activity or zmpA or zmpB transcription suggesting AHL-independent effects. The BCAM1871 mutant exhibited significantly reduced virulence in rat chronic respiratory and nematode infection models. Gene expression and phenotypic assays as well as vertebrate and invertebrate infection models showed that BCAM1871 significantly contributes to pathogenesis in B. cenocepacia.


Introduction
Diverse pathogenic bacteria often utilize cell-cell communication (quorum sensing, QS) systems to communicate with neighboring cells. These QS systems enable organisms to produce and perceive chemical signals resulting in altered gene expression and phenotypes contributing to pathogenesis. In Gram-negative bacteria, the most widely-studied QS system involves production of N-acyl-homoserine lactones (AHLs) by synthases of the LuxI protein family (for review see [1,2] . During bacterial growth these AHLs are released from bacterial cells into the surrounding medium. As cell density increases and the AHL concentrations become relatively high, sufficient numbers of these molecules are found inside bacterial cells where they reversibly bind cognate transcriptional regulators that are members of the LuxR protein family. AHL binding induces a conformational change activating the transcriptional regulator. Transcriptional activation or repression can be exerted by LuxR proteins in the AHL-bound or unbound state. Additionally, there are LuxR homologs that are not genetically linked to an AHL synthase gene and these are called orphan or solo LuxR homologs [3]. The Burkholderia cepacia complex (Bcc) is a group of 17 closelyrelated bacteria that are commonly found in soil, water and agricultural environments and have importance in medicine, agriculture and biotechnology [4,5]. Bcc cause nosocomial infections as well as severe illnesses in neonates, patients with cystic fibrosis (CF) or cancer, transplant recipients and other immunocompromized patients [6]. B. cenocepacia is transmissible, intrinsically resistant to antibiotics and contributes to lung and bloodstream infections resulting in significantly decreased survival rates in CF and chronic granulomatous disease patients [7,8]. Some CF patients succumb to ''cepacia syndrome'', a rapid deterioration in lung function associated with necrotizing pneumonia, bacteremia and sepsis [8,9]. Bloodstream dissemination of B. cenocepacia can also occur in individuals receiving hemodialysis [10].
The B. cenocepacia AHL-dependent QS system is composed of two AHL synthases and three regulatory proteins enabling finetuned regulation of gene expression [11,12,13]. The CepIR system is present in all species of the Bcc [11,14,15,16] while the CciIR system is only present in B. cenocepacia containing the cenocepacia island (cci) found in highly transmissible strains [15]. All B. cenocepacia strains contain the orphan LuxR homolog CepR2 [13]. CepI is primarily responsible for the synthesis of N-octanoyl-Lhomoserine lactone (OHL) and minor amounts of N-hexanoyl-Lhomoserine lactone (HHL) [11]. CciI primarily synthesizes HHL with lesser amounts of OHL produced [12]. The functional significance of the B. cenocepacia AHL-mediated QS system is underscored by the fact that AHL activity is maintained in the majority of sputum isolates recovered from chronically-infected CF patients [17].
There is a complex regulatory interrelationship between the B. cenocepacia QS systems. Transcription of cepI and the cciIR operon depends on CepR, but CciR negatively regulates the expression of cepI and the cciIR operon thus allowing regulatory feedback [12,16]. Furthermore CepR, CciR and CepR2 are negative autoregulators [12,13,16]. Additionally, CepR activity can be inhibited by excess amounts of HHL which is the primary product of cciI [18]. For the most part, CepR is a positive regulator while CciR is a negative regulator of gene expression, with the vast majority of commonly-regulated genes being reciprocally regulated [19]. CepR2 positively or negatively influences gene expression and can influence gene expression in the absence of AHLs [13].
The B. cenocepacia global regulators CepIR and CciIR control protease, lipase and chitinase activities, swimming and swarming motilities, siderophore production, biofilm formation and maturation, as well as expression of genes encoding components of efflux pumps, flp type pili, lectins and types II, III and VI secretion systems [11,12,19,20,21,22,23,24,25]. The B. cenocepacia CepIR system contributes to virulence in murine, Caenorhabditis elegans, Galleria mellonella, Danio rerio and alfalfa infection models [23,26,27,28]. The CciIR system influenced virulence in the rat chronic respiratory infection model but did not affect virulence in C. elegans or G. mellonella [15,27]. It has been shown that B. cenocepacia QS influences bacterial dissemination from initial infection sites in mouse and D. rerio infection models although the specific mechanism(s) have not been revealed [23,28].
Despite having a detailed understanding of the core regulatory features that govern B. cenocepacia QS, to date only four additional regulators have been identified that induce or repress QS [29,30]. The B. cenocepacia genome is of considerable size (.8 Mbp) and contains many uncharacterized genes. We previously showed CepR positively regulated expression of the BCAM1871 ORF located immediately downstream from cepI [19]. We hypothesized the conserved hypothetical protein encoded by BCAM1871was involved in B. cenocepacia virulence based on its genomic location, protein conservation and regulation by QS. In this study, we used gene expression and phenotypic assays as well as vertebrate and invertebrate infection models to assess the contribution of BCAM1871 to pathogenesis in B. cenocepacia.

BCAM1871 is co-transcribed with cepI and directly regulated by CepR
In B. cenocepacia the cepI and cepR ORFs are divergently transcribed and separated by BCAM1869 (Fig. 1). Fifty-two bp downstream from cepI lies the BCAM1871 ORF that we previously showed was positively regulated by CepR using microarrays [19]. To further study the regulation of BCAM1871 we compared expression of BCAM1871 in the cepI mutant (K56-dI2) which contains a 290 bp deletion in cepI using qRT-PCR [12]. Expression of BCAM1871 was substantially reduced (31.8610.4-fold) in the cepI mutant compared to that of wild type. When we exogenously added 300 rM OHL to cultures of the cepI mutant BCAM1871 expression was almost fully restored to wild type levels in that is was reduced only 4.762.0-fold. We determined that cepI and BCAM1871 were co-transcribed using RT-PCR with a primer pair designed to span the intergenic region ( Fig. 1). BCAM1871 expression appeared to be driven from the cepI promoter as no other promoter was identified upstream of the BCAM1871 ORF using promoter prediction software (www. softberry.com or http://www.fruitfly.org, data not shown). CepR directly binds the cepI promoter to positively regulate its transcription [18,30]. Our data suggested CepIR-dependent regulation of BCAM1871occurred through CepR binding to the cepI promoter driving expression of the cepI-BCAM1871 operon.
Orthologs of the ORF encoding BCAM1871 are conserved in their genomic locations, in that each lies downstream from an AHL synthase, in at least six species of the Bcc (B. ambifaria, B. dolosa, B. lata, B. multivorans, B. ubonensis and B. vietnamiensis) as well as Burkholderia pseudomallei and Burkholderia mallei [31]. In an analysis of 1657 genomes, orthologs of BCAM1871 were found in 41 genomes, all of which belonged to Burkholderia spp. indicating this protein is unique to this genus [31]. BCAM1871 encodes a conserved hypothetical protein that has limited homology (E value of 4.55 204 ) to hydroxymethylglutaryl-CoA reductase [32]. BCAM1871 is predicted to encode a protein of 254 aa that has a cytoplasmic localization [31].

BCAM1871 enhances AHL activity
We hypothesized BCAM1871 may affect AHL activity based on its genomic location, protein conservation and regulation by QS. A BCAM1871 mutant was constructed by removing an internal 365 bp portion of the ORF. Mutation of BCAM1871 did not affect growth of the strain at 29uC or 37uC compared to wild type in LB (data not shown). Using a real-time co-culture AHL activity assay, the BCAM1871 mutant had significantly reduced AHL activity throughout growth compared to wild type ( Fig. 2A). AHL activity was not restored to wild type levels when BCAM1871 was expressed on a lower copy plasmid (pM1871). This was likely due to insufficient expression of BCAM1871 which is expressed at relatively high levels in stationary phase in wild type (data not shown). Consistent with a relatively high expression level, AHL activity was restored to wild type levels by expression of BCAM1871 in trans on a high copy plasmid (p28T-M1871) ( Fig. 2A). Mutation of BCAM1871 did not completely abolish AHL production. In fact, AHL activity of the BCAM1871 mutant was significantly higher than that observed for the cepI mutant ( Fig. 2A).
Considering BCAM1871 mutation had a more subtle effect on AHL activity than a cepI mutation, we decided to examine potential effects of overexpressing BCAM1871 in wild type cells. Figure 1. Genetic organization of the BCAM1871 locus. BCAM1871 is located downstream from cepI with 52 bp separating cepI and BCAM1871 which were determined to be co-transcribed (solid line) using RT-PCR. Size (bp) of each ORF is indicated. The cepI transcription start site is located 28 bp upstream of the ATG start codon [18]. One promoter (arrow) was identified for the cepI-BCAM1871 operon, and this was located upstream of cepI and included in the cepI promoter::lux fusion (pCP300). Nomenclature used to describe K56-dI2 and K56-2DBCAM1871 mutants. doi:10.1371/journal.pone.0037611.g001 Overexpression of BCAM1871 significantly enhanced AHL activity above wild type levels in early stages of growth, albeit to a lower level than observed for overexpression of CepI (Fig. 2B). These data suggested the effect of BCAM1871 on AHL activity was likely related to enhancing AHL production rather than being directly involved in AHL synthesis. Supporting the hypothesis that BCAM1871 indirectly affected AHL production, and with the predicted protein function indicating no relationship to AHL synthases, expression of BCAM1871 in trans in the heterologous host E. coli resulted in no detectable AHL activity. In contrast, high levels of AHL activity were observed when cepI was expressed in E. coli (data not shown). To further explore the AHL activity relationship between BCAM1871 and cepI, we performed crosscomplementation experiments. AHL activity was increased in the BCAM1871 mutant expressing cepI in trans to levels higher than that those observed in wild type (Fig. 2C). In contrast, AHL activity was not altered in the cepI mutant expressing BCAM1871 in trans (Fig. 2C). We also sought to determine if a secreted factor(s) from wild type could fully restore AHL activity to the BCAM1871 mutant. AHL activity in a 1:1 mixed culture of wild type and the BCAM1871 mutant was significantly increased compared to that in the BCAM1871 mutant but still significantly reduced compared to wild type (data not shown). These data suggested lower AHL activity in the BCAM1871 mutant was not attributable to absence of a secreted factor in this strain. Together, these data indicate BCAM1871 acts as an enhancer of AHL activity rather than being essential or sufficient for AHL activity in B. cenocepacia.
We considered the possibility that BCAM1871 could exert a regulatory influence in another bacterial species with an AHLdependent QS system. There was no difference in AHL activity or expression of a lasI::lux fusion when B. cenocepacia BCAM1871 was expressed in trans in Pseudomonas aeruginosa PAO1 compared to vector control (data not shown). These data suggest the effect of BCAM1871 may not be observed in other organisms but may be specific to Burkholderia spp or B. cenocepacia.

BCAM1871 promotes transcription of cepIR and cciIR
Since BCAM1871 appears to enhance AHL activity, the effects of BCAM1871 mutation on expression of promoter::lux fusions were determined to measure transcription of QS genes. Expression of cepI, cepR and cciIR was significantly reduced in the BCAM1871 mutant compared to wild type (Fig. 3). Addition of 300 rM exogenous OHL to cultures immediately prior to maximal cepR promoter activity in wild type (8 h) resulted in significantly increased cepI and cepR expression in the BCAM1871 mutant (Fig. 3AB). In contrast, expression of cciIR was not restored to wild type levels at any point during growth of the BCAM1871 mutant in cultures supplemented with exogenous OHL (Fig. 3C). Addition of 30 rM exogenous OHL had a significant effect on cepR but not cepI or cciIR transcription (Fig. 3). Furthermore, we found the timing of OHL addition was critical since supplementation with Figure 2. Influence of BCAM1871 on AHL activity. AHL activity was monitored using A. tumefaciens A136 (pCF218) (pMV26) in a real-time liquid co-culture assay. (A) Significantly decreased in K56-2DM1871 (pUCP28T) compared to that in K56-2 (pUCP28T) from 16-30 h. Significantly increased in K56-2DM1871 (p28T-M1871) compared to that in K56-2DM1871 (pUCP28T) from 18-30 h. Significantly decreased in K56-dI2 (pUCP28T) compared to that in K56-2DM1871 (pUCP28T) from 16-30 h. (B) Significantly increased in K56-2 (p28T-M1871) compared to that in K56-2 (pUCP28T) from 2-10 h and 18-30 h. exogenous OHL at the start of the culture (0 h) had no effect on cepI, cepR or cciIR transcription (data not shown).

BCAM1871 promoted motility independently of AHL activity or CepIR function
A number of phenotypic traits are controlled by QS in B. cenocepacia including bacterial motility and production of extracellular enzymes. We previously showed motility is reciprocally regulated by CepIR and CciIR in B. cenocepacia [19]. Different modes of motility, such as swimming and swarming, can be differentiated using L-agar plates supplemented with 0.25% and 0.5% agar, respectively. In each mode of motility, the BCAM1871 mutant showed significantly reduced translocation compared to that of wild type (Fig. 4AB). Furthermore, the BCAM1871 mutant was unresponsive to exogenous OHL in terms of motility. A significant reduction in swimming and swarming motility was observed for the cepI mutant compared to wild type (Fig. 4AB), adding to and confirming previous data that showed reduced swarming motility [12] for an independently constructed cepI insertion mutant (K56-I2) [11]. Addition of exogenous OHL to the medium significantly stimulated motility in wild type and restored motility to wild type levels in the cepI mutant, in contrast to the unresponsive BCAM1871 mutant (Fig. 4AB). We also noted differences between the cepI and BCAM1871 mutants in that swarming motility, but not swimming motility, was significantly decreased in the BCAM1871 compared to the cepI mutant in absence of OHL (Fig. 4AB).
Reduced swimming and swarming motilities of the BCAM1871 mutant were significantly increased by expressing BCAM1871 in trans, confirming the involvement of BCAM1871 in regulating this motility phenotype (Fig. 4CD). In a series of cross-complementation experiments, we found expressing BCAM1871 in trans in the cepI mutant significantly increased motility to levels similar to that seen for expression of cepI in trans in the cepI mutant (Fig. 4CD). We also employed the alternative strategy of introducing a cepI expression construct into the BCAM1871 mutant. Expression of cepI in trans did not alter the motility defect of the BCAM1871 mutant corroborating data above that showed the BCAM1871 mutant was unresponsive to exogenous OHL in terms of motility (Fig. 4CD). Furthermore, expression of cepR in trans in the BCAM1871 mutant had no effect on motility in medium with or without exogenous OHL ( Fig. 4CD and data not shown).
Swimming and swarming motility is influenced by flagellar function. Expression of fliC, encoding flagellin, was significantly reduced (up to 15-fold) in the BCAM1871 mutant compared to wild type (Fig. 4E). Addition of 30 or 300 rM exogenous OHL significantly increased fliC expression in the BCAM1871 mutant to wild type levels or greater (Fig. 4E) despite having no effect on the reduced motility phenotype described above (Fig. 4CD). Together these data show that the BCAM1871 mutant was significantly affected for motility and that exogenous OHL or CepIR function was insufficient to alter the reduced motility phenotype of the BCAM1871 mutant.
BCAM1871 promoted protease activity independently of AHL activity or CepIR function B. cenocepacia possesses a T2SS that is responsible for the secretion of several extracellular enzymes such as zinc metalloproteases ZmpA and ZmpB that influence virulence in a rat chronic respiratory infection model [22,33]. Expression of zmpA and zmpB is carefully regulated by CepIR and CciIR leading us to hypothesize that AHL activity differences in the BCAM1871 mutant may also affect protease activity. We were unable to detect  protease activity surrounding cultures of the BCAM1871 mutant on skim milk agar up to 40 h incubation and this phenotype was not altered by exogenous OHL (Fig. 5A). Additionally, we observed the BCAM1871 mutant had a more severe protease activity defect than previously observed for the cepI mutant, from which some protease activity was detected (Fig. 5A), albeit at significantly reduced levels compared to wild type as previously determined [11]. Supplementation of the medium with OHL restored protease activity to wild type levels in the cepI mutant as expected (Fig. 5A).
Expression of BCAM1871 in trans significantly enhanced protease activity in the BCAM1871 mutant but did not affect protease activity in the cepI mutant (Fig. 5B). Confirming these data, expression of BCAM1871 in trans did not affect protease activity in an independently constructed cepI insertion mutant (K56-I2) (data not shown). Significantly increased protease activity was detected by expressing cepI in trans in the cepI mutant. However, expressing cepI in trans in the BCAM1871 mutant had no effect on its protease negative phenotype (Fig. 5B) confirming the results above obtained with addition of exogenous OHL to the medium (Fig. 5A). Expression of cepR in trans in the absence or presence of OHL also did not affect the protease negative phenotype of the BCAM1871 mutant ( Fig. 5B and data not  shown). Together, these data suggest BCAM1871 and CepIR independently promote protease activity.
Consistent with the protease negative phenotype of the BCAM1871 mutant, expression of zmpA and zmpB was reduced to background levels (i.e. similar to the promoterless lux fusion plasmid pMS402) in the BCAM1871 mutant compared to that of wild type (Fig. 5CD). Consistent with the phenotypic data, expression of zmpA and zmpB was unaffected by addition of exogenous OHL to the medium (Fig. 5CD). Since addition of exogenous OHL to the medium, or expression of cepI or cepR in trans in the BCAM1871 mutant, did not alter either motility or protease activity, it suggests that unidentified factor(s) controlled by BCAM1871, other than AHL activity or CepIR function, are required for expression of these phenotypes in B. cenocepacia.

BCAM1871 influences expression of additional virulencerelated genes
Our laboratory is also interested in other B. cenocepacia genes that contribute to virulence. The B. cenocepacia LysR-type transcriptional regulator, ShvR, influences colony morphology, biofilm formation, cell surface hydrophobicity, QS, type II secretion, antifungal activity, as well as plant and animal virulence [30,34,35]. Some of these characteristics are controlled through direct regulation by ShvR of the adjacent afcA operon that contains twenty-two genes, some of which influence colony morphology, biofilm formation and virulence [35]. Expression of shvR and the afcA operon was significantly decreased in the BCAM1871 mutant compared to wild type (Fig. 6), in a manner similar to that observed for indirect CepR regulation of shvR and afcA [30].

BCAM1871 contributes to persistence and inflammation in a rat chronic respiratory infection model
Considering the effect of mutating BCAM1871 on the above virulence-related phenotypes and gene expression we chose to assess the contribution of BCAM1871 to virulence in a rat chronic respiratory infection model. Rats were infected with approximately 10 8 CFU of the wild type and BCAM1871 mutant. The number of bacteria recovered from lungs seven days post infection was significantly reduced in animals infected with the BCAM1871 mutant compared to wild type indicating the BCAM1871 mutant had reduced ability to persist in the lungs (Fig. 7A). Lung sections were obtained and stained with hematoxylin and eosin to quantify pathological changes associated with infection. Consistent with reduced bacterial burden, significantly reduced inflammation was observed in rats infected with the BCAM1871 mutant compared to wild type (Fig. 7B).

The BCAM1871 mutant displayed reduced expression of a gene encoding a nematocidal protein and caused reduced nematode death
AidA is a relatively uncharacterized protein in B. cenocepacia that significantly contributes to virulence of the nematode C. elegans [36]. Nematode survival was significantly higher following feeding infection with the BCAM1871 or cepI mutant compared to feeding infection with wild type (Fig. 8A). Expression of BCAM1871 in trans significantly decreased nematode survival to a level similar to that observed for wild type (Fig. 8A). Along with cepI, CepRdependent expression of aidA has been consistently identified in numerous studies [18,19,37,38]. We found aidA transcription was significantly decreased in the BCAM1871 mutant compared to wild type and its expression was significantly influenced by 300 rM exogenous OHL in the medium (Fig. 8B).

Discussion
In this study, we have characterized multiple deleterious effects of mutating the ORF encoding BCAM1871 on pathogenesisrelated traits in B. cenocepacia. Our data suggests BCAM1871 is an integral component of the B. cenocepacia QS network and its inactivation disrupts optimal timing and magnitude of cell-density dependent gene expression and virulence factor production, resulting in significantly reduced pathogenesis in rat and nematode infection models (Fig. 9).
The predicted protein encoded by BCAM1871 appears to be unique and conserved in its genetic context with cepIR homologs in Burkholderia spp suggesting it has been conserved evolutionarily and retains an important function. In the predicted protein sequence for BCAM1871, there exists a partial domain with similarity to hydroxymethylglutaryl-CoA reductases (HMGR) (e value 4.6 24 ) despite absence of an N-terminal membrane region. HMGR function as rate-limiting enzymes in cholesterol biosynthesis in eukaryotes (class I) and isoprenoid biosynthesis in prokaryotes (class II) [39]. B. cenocepacia contains a putative class II HMGR encoded by mvaA (BCAM0531) that has little or no similarity at the amino acid level to BCAM1871 (data not shown). Our attempts to inhibit activity of these HMGR using mevastatin [39] had no effect on AHL or protease activity of B. cenocepacia (data not shown).
Apart from the synthases and transcriptional regulators that form the core of the B. cenocepacia AHL-mediated QS system, only four regulators were previously demonstrated to positively or negatively influence B. cenocepacia QS. Three regulators were identified that promoted AHL activity and biofilm formation in B. cenocepacia H111 [29]. These mutants had defects in biofilm formation that were not restored by exogenous OHL. Furthermore, the mutants exhibited defects in biofilm formation and AHL activity that could not be rescued by expressing cepR in trans [29]. Based on similarities in results obtained for these regulators and BCAM1871, it appears each of them exerts activity upstream of CepR in the B. cenocepacia QS hierarchy, although we additionally demonstrated cepR is affected at the transcriptional level by BCAM1871 mutation. We also determined mutation of BCAM1871 had no effect on biofilm formation (data not shown) indicating additional differences upon mutation of the three previously described regulators [29] or BCAM1871. Our results showed BCAM1871 is positively autoregulated by enhancing AHL activity and cepIR expression which in turn influences CepRdependent regulation of the cepI-BCAM1871 operon. In addition to these positive regulators, we showed mutation of BCAM1871 reduced expression of the LysR-type transcriptional regulator ShvR. ShvR is a negative regulator of cepIR and cciIR transcription  that affects AHL activity and secretion of virulence factors [30]. Our laboratory previously showed ShvR was important for biofilm formation, colony morphology and virulence in plant and animal infection models [34]. The fact that ShvR and CepR independently control biofilm formation [30], and that mutation of BCAM1871 had no effect on biofilm formation, despite reduced shvR and cepR expression, suggests additional factors regulating biofilm formation in B. cenocepacia have yet to be uncovered.
We observed a more severe defect in the BCAM1871mutant compared to the cepI mutant for swarming motility and protease activity despite the fact that BCAM1871 mutant retains more AHL activity than the cepI mutant. Previous studies attributed defects in motility, protease activity or zmpA/zmpB gene expression to mutation of cepR or cepI which could be complemented by providing cepR or cepI in trans, or by supplying exogenous OHL [11,12,21,23,24]. The current study refines our understanding to show that cepR or cepI mutation would in fact have adversely affected expression of the cepI-BCAM1871 operon. The transcriptional coupling of cepI and BCAM1871 means that BCAM1871 expression is reduced, but not abolished, in a cepI mutant and can be rescued to near wild type levels by expressing cepI in trans or by providing exogenous OHL. We also observed that significantly reduced transcription of BCAM1871 in a cepIcciI mutant was increased by providing exogenous OHL, but not HHL (data not shown). In contrast, reduced motility or protease activity in the BCAM1871 mutant could not be altered by providing cepI or cepR in trans, or by supplying exogenous OHL, but only by expressing BCAM1871 in trans.
We cannot fully explain why exogenous OHL stimulated fliC transcription but neither swimming nor swarming motility. This result could be due to deregulation of one or more of the approximately thirty flagellar-associated motility genes or the production of a biosurfactant that has not yet been identified in B. cenocepacia. Exogenously-supplied biosurfactants restored swarming motility to a B. cenocepacia H111 cepR mutant [21]. The positive effect of OHL-mediated QS on indirect regulation of flagellarassociated motility has been studied in more detail in the rice plant pathogen Burkholderia glumae [40]. In fact, lack of flagella formation in a B. glumae AHL synthase (tofI) mutant can be overcome by incubation at 28uC indicating other factors besides QS are involved [40]. We previously found reciprocal regulation of swimming motility by CepR and CciR was apparent at 22uC and 28uC [19]. It was interesting to note that expressing BCAM1871 in trans promoted motility but did not affect protease activity in cepI mutant. These different outcomes on motility and protease activity in the cepI mutant occurred despite the fact that expressing BCAM1871 in trans does not affect AHL activity of the cepI mutant, indicating BCAM1871 can affect motility independently of altering AHL activity in this strain. We also demonstrated addition of exogenous OHL to cultures of the wild type promoted motility but not protease activity. This suggests protease activity is controlled by factors in addition to QS or the effects of QS can be saturated or subject to negative feedback by CciIR as previously described [19,22]. Together, these data suggest there are additional factors directly or indirectly affected by BCAM1871 mutation that regulate these genes and the phenotypes they affect.
Upon exogenous OHL supplementation, different responses were observed in the BCAM1871 mutant with respect to transcription of affected genes. For the most part, exogenous supplementation with 300 rM, but not 30 rM, OHL was sufficient to restore expression of those genes that were responsive to exogenous OHL in the BCAM1871 mutant. It was previously shown that cepR expression was increased in a cepR mutant using a promoter::lacZ fusion indicating cepR was negatively autoregulated [11]. In our study, cepR transcription could be increased by exogenous OHL supplementation immediately prior to maximal cepR transcription using a cepR::lux fusion in wild type (data not shown) or the BCAM1871 mutant. It appears that optimal timing of OHL supplementation is critical since these effects were not observed for wild type or the BCAM1871 mutant when exogenous OHL was added at the start of the culture (data not shown). Consistent with increased cepR transcription upon exogenous OHL supplementation in the BCAM1871 mutant, we also observed increased expression of two genes (cepI and aidA) that are directly regulated at the promoter level by CepR [18]. Although it has not been demonstrated that CepR-dependent regulation occurs at the promoter level for fliC, we observed fliC transcription was restored by exogenous OHL in the BCAM1871 mutant. In contrast, exogenous OHL had no effect on transcription of zmpA, zmpB and cciIR. It is noteworthy that zmpA is not directly regulated at the promoter level by CepR [41], and zmpB and cciIR do not contain a cep box motif in their promoters [20] suggesting they are also not directly regulated by CepR. We did not investigate potential effects on posttranscriptional control of cepR, stability of CepR or its ability to bind and retain OHL since mutation of BCAM1871 had a significant effect on cepR transcription.
The BCAM1871 mutant had a similar effect on nematode survival and aidA gene expression as previously demonstrated for a cepI mutant [26,36]. In the rat chronic respiratory infection model, reduced persistence was observed for the BCAM1871 mutant at seven days postinfection which was not identified for cepI or cepR mutants at ten days postinfection [23]. Despite decreased lung persistence we detected no defect in intracellular survival of the BCAM1871 mutant in RAW264.7 murine macrophages or in resistance to buffered solutions at low pH (6.4 or 4.5) (data not shown), i.e. pH values that have been quantified inside macrophages [42].
Recent studies identified a small regulatory protein CinS that is cotranscribed with the AHL synthase CinI in R. leguminosarum [43]. In contrast to the effect of BCAM1871 on cepI transcription, mutation of cinS did not affect cinI transcription or synthesis of the CinI gene product, but resulted in reduced expression of a second AHL synthase, RaiI, and the RaiI gene product [43]. A swarming defect in R. leguminosarum that was attributed to cinI mutation [44] was in fact due to polar effects of cinI mutation on cinS expression [43]. These swarming defects in R. leguminosarum cinI or B. cenocepacia cepI mutants can each be restored by expressing the downstream gene (cinS or BCAM1871, respectively) in trans. This indicates the primary defect in these strains results from lower expression of the downstream gene, through effects of the cinI-cinS or cepI-BCAM1871 operons. Further work determined CinS interacted with an AHL-independent transcriptional repressor PraR to relieve the effects of this repressor on the raiIR and rhiIR QS systems in R. leguminosarum [45]. The predicted protein product of BCAM1871 (254 residues) does not show homology to CinS (67 residues), and neither CinS nor RpaR homologs are found in B. cenocepacia (data not shown).
Further experiments are required to characterize the mechanism of action of the regulator encoded by BCAM1871 which could include DNA-, RNA-or protein-binding capabilities. Such regulatory mechanisms have previously been identified for other QS systems that lead to modulation of QS activity thereby enabling fine-tuned control of gene expression (for review see, [46][47][48][49]). Mutation of BCAM1871 resulted in significant effects on gene expression, phenotypes and virulence. While BCAM1871 is an enhancer of AHL synthesis and it transcription is CepIRregulated, some phenotypes were affected independently of AHL activity or CepIR function suggesting BCAM1871 operates by a novel mechanism that has not yet been described elsewhere and potentially involves a sophisticated coordination of AHL-dependent and AHL-independent activities.

Strains, plasmids and growth conditions
The bacterial strains and plasmids used in this study are listed in Table 1. Cultures were routinely grown at 37uC, in Miller's Luria broth (LB) (Invitrogen, Burlington, ON) with shaking or on 1.5% Lennox LB agar plates. For RT-PCR or qRT-PCR experiments, cultures were grown to stationary phase (16 h) in 10 ml LB in 125 ml erlenmeyer flasks. For promoter::lux assays, strains were grown in LB or trypticase soy broth (TSB) (Difco, Franklin Lakes, NJ) in black, clear-bottom, 96-well plates (Corning, Inc., Corning, NY) and assays performed as previously described [19,20]. Cultures entered stationary phase at 18 h (growing at 37uC) or 30 h (growing at 29uC) [30]. For selected qRT-PCR or promoter::lux experiments, OHL (dissolved in 20% acetonitrile), was added in concentrations ranging from 30-300 rM and results compared to addition of solvent control. AHL activity was monitored using Agrobacterium tumefaciens A136 (pCF218) (pMV26) in a real-time liquid co-culture assay with B. cenocepacia or P. aeruginosa grown in 96-well plates in TSB at 29uC [50]. Swimming and swarming motility assays were performed using L-agar plates supplemented with 0.25% and 0.5% agar, respectively [51]. Overnight cultures were normalized to an OD 600 of 0.4 prior to inoculation and growth was assessed at 37uC. Protease activity was determined using D-BHI with 1.5% skim milk agar plates [52] using strains grown in 0.25% trypticase soy broth (Difco) with 5% Bacto-Peptone (Difco) (PTSB). Cell pellets were washed twice with PTSB and normalized to an OD of 0.3 prior to inoculation [30]. When appropriate, 2500 rM exogenous OHL was spread on the surface of the protease agar and briefly air-dried prior to inoculation. When appropriate, the following concentrations of antibiotics were used: 100 mg/ml of trimethoprim (Tp) or 200 mg/ ml tetracycline (Tc) for B. cenocepacia; 200 mg/ml Tc for P. aeruginosa; 100 mg/ml of Tp and 50 mg/ml of kanamycin (Km) for E. coli; and 25 mg/ml of Km and 4.5 mg/ml of Tc for A. tumefaciens. Antibiotics were purchased from Sigma-Aldrich Canada Ltd. (Oakville, ON).

DNA manipulations
DNA manipulations were performed using standard techniques as described [53] and genomic DNA was isolated as described [54]. Oligonucleotide primers ( Table 2) were designed with Primer3 [55] and were synthesized by the University of Calgary Core DNA and Protein Services (Calgary, Alberta, Canada). Plasmids were introduced into B. cenocepacia by electroporation [56].

Construction of the K56-2DM1871 mutant
The K56-2DM1871 mutant was constructed following a described method [57] using primers listed in Table 2. PCR confirmed deletion of a 365 bp fragment from the BCAM1871 ORF in this mutant.

RNA manipulations and Quantitative RT-PCR (qRT-PCR)
RNA samples were prepared using a RiboPure bacterial RNA isolation kit (Ambion, Streetsville, Ontario, Canada) from B. cenocepacia strains grown to stationary phase in LB [19]. DNase treatment was performed, and samples were confirmed to be free of DNA by PCR using Taq polymerase (Invitrogen) prior to cDNA synthesis using an iScript Select cDNA synthesis kit (Bio-Rad, Mississauga, ON). Quantification and melting curve analyses were performed with iQ SYBR green Supermix on an iCycler (Bio-Rad) according to manufacturer's instructions. qRT-PCRs were performed in triplicate and normalized to the control gene, sigE (BCAM0918) as described previously [13]. Data were calculated as previously described [58] and data shown below are from three independent experiments.

Animal studies
Animal infections were performed using the rat agar bead respiratory infection model [59]. Adult male Sprague-Dawley rats (150-180 g; Charles River, QC, Canada) were inoculated transtracheally with approximately 10 8 CFU of the appropriate strain. At seven days postinfection, infected lungs were aseptically removed, and lungs from four to five animals per group were used for quantitative bacteriology or histopathology analysis as previously described [51]. Infiltration of hematoxylin-and-eosin-stained lung sections with inflammatory cells and exudates was quantitated using Image Pro Plus (Media Cybernetics, Bethesda, MD). Animal experiments were conducted according to the guidelines of the Canadian Council of Animal Care for the care and use of experimental animals under protocol M08089 approved by the University of Calgary Animal Care Committee.

Nematode studies
Nematode infections were performed using C. elegans strain Bristol N2 as previously described [26]. Briefly, nematodes were maintained on nematode growth (NG) agar [60] with Escherichia coli OP50 as a food source at 25uC in a humidified atmosphere [61]. Killing assays were performed by culturing the appropriate E. coli or B. cenocepacia strains overnight in LB and inoculating NG agar with approximately 100 ml of cultures adjusted to 1610 4 CFU/ml on 2.0-cm diameter plates. Plates were incubated overnight at 37uC and allowed to cool to before transferring at least 25 L4 larvae or adult worms per plate and incubating plates at 25uC. Using a stereomicroscope (Fischer) at 506 magnification, nematodes were enumerated as alive or dead by gentle prodding with a platinum wire and assessing responsiveness.

Statistical analyses
All statistical analyses were performed using GraphPad Prism software (GraphPad Software, San Diego, California). For AHL activity assays, motility assays, protease activity assays or promoter::lux fusions, all data were analyzed by two-way analysis of variance (ANOVA) and all values are the means 6 SD of results from triplicate cultures and are representative of results from at least two individual trials. For motility and protease activity assays, values were recorded at several timepoints and values at one representative timepoint are shown. For rat infections, statistical analysis was performed using an unpaired t test with Welch's correction. For nematode infections, statistical analysis was performed using Kaplan-Meier survival kinetics and the log-rank (Mantel-Cox) test.