Cis-2-dodecenoic acid quorum sensing system modulates N-acyl homoserine lactone production through RpfR and cyclic di-GMP turnover in Burkholderia cenocepacia

Background Burkholderia cenocepacia employs both N-Acyl homoserine lactone (AHL) and cis-2-dodecenoic acid (BDSF) quorum sensing (QS) systems in regulation of bacterial virulence. It was shown recently that disruption of BDSF synthase RpfFBc caused a reduction of AHL signal production in B. cenocepacia. However, how BDSF system influences AHL system is still not clear. Results We show here that BDSF system controls AHL system through a novel signaling mechanism. Null mutation of either the BDSF synthase, RpfFBc, or the BDSF receptor, RpfR, caused a substantial down-regulation of AHL signal production in B. cenocepacia strain H111. Genetic and biochemical analyses showed that BDSF system controls AHL signal production through the transcriptional regulation of the AHL synthase gene cepI by modulating the intracellular level of second messenger cyclic di-GMP (c-di-GMP). Furthermore, we show that BDSF and AHL systems have a cumulative role in the regulation of various biological functions, including swarming motility, biofilm formation and virulence factor production, and exogenous addition of either BDSF or AHL signal molecules could only partially rescue the changed phenotypes of the double deletion mutant defective in BDSF and AHL signal production. Conclusions These results, together with our previous findings, thus depict a molecular mechanism with which BDSF regulates AHL signal production and bacterial virulence through modulating the phosphodiesterase activity of its receptor RpfR to influence the intracellular level of c-di-GMP.


Background
Quorum sensing (QS) is widely employed by bacterial pathogens to coordinate bacterial group behavior and regulate biological functions such as biofilm formation, motility, virulence, plasmid transfer, and antibiotic production [1,2]. This regulation mechanism depends on the production and perception of diffusible signal molecules in a cell-density dependent manner [2][3][4]. At low cell density, bacterial cells produce a basal level of QS signals, which are diffused or transported into extracellular environments. When the cell density reaches a critical concentration, the accumulated signals initiate a set of biological activities in a coordinated fashion. Several types of QS systems have been identified including the most-characterized acylhomoserine lactone (AHL) dependent QS system and the relatively newly identified diffusible signal factor (DSF) dependent QS system [3,5]. The AHL-and DSF-QS systems are mainly conserved in different Gram-negative bacteria pathogens.
While most bacterial pathogens employ either AHLor DSF-dependent QS systems in regulation of virulence and biofilm formation [3,6], the members of the Burkholderia cepacia complex were found to produce both AHL-and DSF-type QS signals [7][8][9]. In B. cenocepacia, which is an opportunistic pathogen in cystic fibrosis or immunocompromised patients, the AHL-type QS system comprises the AHL synthase CepI, which was shown to catalyze the synthesis of N-octanoyl homoserine lactone (C8HSL, also known as OHL) as a major AHL signal [10,11], and the AHL receptor CepR. The receptor CepR forms a complex with AHL signals to activate or repress a set of target genes, and thus control a range of biological functions, including virulence, swarming motility and biofilm formation [8,9].
In addition to the AHL-dependent QS system, a DSFdependent system has recently been identified in B. cenocepacia [12][13][14][15]. The QS signal synthase, RpfF Bc , catalyzes the production of BDSF signal (cis-2-dodecenoic acid), which is an analogue of the QS signal DSF (cis-11methyl-2-dodecenoic acid), originally identified in the plant bacterial pathogen Xanthomonas campestris pv. campestris [16]. Our recent study showed that BDSF acts by interacting with its receptor RpfR, which is a modular protein with PAS-GGDEF-EAL domains [14]. Perception of BDSF by RpfR sharply enhances its c-di-GMP phosphodiesterase activity and consequently causes a reduction in the intracellular level of the second messenger cyclic di-GMP (c-di-GMP) in B. cenocepacia, which consequently affects a range of biological activities, including swarming motility, biofilm formation and virulence [14].
It has become clear that both AHL and BDSF systems control similar biological functions. Recently, it was reported that there is a direct relationship between the two QS systems as inactivation of BDSF synthase reduces the production of AHL signals in B. cenocepacia [17,18]. However, how BDSF system affects AHL system remains obscure. In this study, by generating and analyzing single-and double-deletion mutants defective in QS signal production, we showed that BDSF signaling system plays a dominant role in the regulation of AHL QS system and various biological activities in B. cenocepacia. In addition, we have investigated the molecular mechanisms with which BDSF signaling system influencing AHL signal production and unveiled the involvement of the second messenger c-di-GMP. Furthermore, we have determined the relationships of these two QS systems in the cell-cell communication signaling cascade and their impacts on bacterial physiology and virulence.

BDSF system positively regulates AHL signal production
To further confirm whether the AHL and BDSF systems are functionally related, we determined the AHL and BDSF signal production levels in corresponding mutants. Consistently, we found deletion of either the AHL synthase gene cepI or the AHL receptor gene cepR had no effect on BDSF production (data no shown). However, we found that disruption of the BDSF synthase gene rpfF Bc in B. cenocepacia H111 caused a significant reduction of the total AHL signal level with the aid of AHL reporter strain ( Figure 1A). BDSF production was restored by in trans expression of the wild type rpfF Bc ( Figure 1A), confirming the role of BDSF system in regulation of AHL biosynthesis. In contrast, in trans expression of rpfF Bc in the cepI deletion mutant displayed no effect, suggesting that BDSF probably functions through modulation of CepI expression level or enzyme activity. Furthermore, we used the TLC method to analyze the different AHL signals produced by these strains. Results showed that deletion of rpfF Bc affected the production of both HHL and OHL signals in B. cenocepacia H111 ( Figure 1B).

BDSF system positively controls cepI expression at transcriptional level
To further study the regulation mechanism of the BDSF system on AHL signal production, we constructed the cepI reporter system in B. cenocepacia H111 strains to test whether BDSF system controls cepI expression at transcriptional level. In agreement with the above results, deletion of rpfF Bc resulted in a reduced expression of cepI at various growth stages ( Figure 2A). Exogenous addition of BDSF rescued the cepI expression in ΔrpfF Bc close to the wild-type level ( Figure 2A). In agreement with the above results, western blotting analysis showed that null mutation of RpfF Bc substantially decreased the CepI protein level ( Figure 2B). These data established a positive regulatory role of the BDSF-dependent QS system in modulation of the cepI transcriptional expression.
BDSF system controls AHL signal production through its receptor RpfR Previous studies showed that two BDSF sensors, BCAM0227 and RpfR (BCAM0580), are involved in the BDSF-mediated QS. Among them, BCAM0227, which was originally characterized in B. cenocepacia strain J2315, controls only a subset of the BDSF-regulated phenotypes and target genes [19], whereas RpfR was shown to be a major receptor of BDSF as null mutation of RpfR results in similar mutant phenotypes as the BDSF-minus mutants [14]. These results suggest that two BDSF signaling pathways may be operating in B. cenocepacia, which motivated us to investigate which BDSF signaling pathway plays a role in regulation of the cepI expression. Significantly, deletion of the BDSF receptor gene rpfR caused a similar reduction in AHL signal production as the deletion mutant of rpfF Bc that encodes a BDSF synthase ( Figure 3A). Analysis of the cepI expression profile using its promoter fused with the lacZ reporter gene showed that RpfR controlled the cepI expression at the transcriptional level ( Figure 3B). Importantly, in contrast to the deletion mutant of rpfF Bc , which could be rescued by addition of BDSF (Figure 2A), addition of BDSF to the rpfR mutant had no effect on the cepI expression ( Figure 3B). The data are consistent with the idea that BDSF modulates AHL signal production through its cognate receptor RpfR. Agreeable with our recent finding that BCAM0227 has a negligible role in BDSF signaling [14], deletion of this gene did not reveal any effect on cepI expression in B. cenocepacia H111 (Additional file 1: Figure S1).

BDSF system controls AHL signal production and biological functions through regulation of intracellular c-di-GMP level
RpfR is a modular protein with PAS-GGDEF-EAL domains. Among these domains, PAS is the domain interacting with BDSF, and GGDEF and EAL domains are associated with c-di-GMP metabolism [14]. To investigate the regulation of RpfR in AHL signal production in detail, critical catalytic residues of either the GGDEF or the EAL domain were mutagenized. In trans expression of RpfR harboring a mutation in the GGDEF motif (changed to GGAAF) complemented the AHL signal production defects of the rpfR mutant (Additional file 2: Figure S2). In contrast, mutation of the EAL motif (changed to AAL) failed to complement the AHL signal production of the rpfR mutant (Additional file 2: Figure S2), To further confirm the change of intracellular c-di-GMP level could affect AHL signal production, we expressed in trans the wspR gene from Pseudomonas aeruginosa, which encodes a well-characterized c-di-GMP synthase [20], and the DNA sequences encoding the GGDEF domain of RpfR in B. cenocepacia wild-type strain H111. Bioassay results showed that increasing intracellular level of c-di-GMP by expressing either the B A Figure 1 Influence of the BDSF system on AHL signal production. (A) AHL signal production was quantified with the aid of AHL reporter strain CF11 to test the β-galactosidase activity. (B) TLC assay of AHL signal production. For convenient comparison, the AHL signal production of wild-type strain was defined as 100% and used to normalize the AHL signal production of other strains. The data presented are the means of three replicates and error bars represents the standard deviation.
c-di-GMP synthase WspR or the GGDEF domain of RpfR in B. cenocepacia wild-type strain H111 caused a reduction of AHL signal production by about 34% and 18%, respectively, compared with the wild type control containing empty vector only ( Figure 4). We then in trans expressed the rocR gene from P. aeruginosa encoding a known c-di-GMP phosphodiesterase [21], and the DNA fragment encoding the EAL domain of RpfR in the BDSF-minus mutant ΔrpfF Bc , separately. The results showed that decreasing the intracellular c-di-GMP level by expression of c-di-GMP degradation proteins RocR and the EAL of RpfR increased AHL signal production by about 29% and 46%, respectively, compared with the parental strain ΔrpfF Bc (Figure 4). We have shown previously that in trans expression of the c-di-GMP synthase GGDEF domain of RpfR diminished the swarming motility, biofilm formation, and protease activity of △rpfF Bc , whereas in tans expression of RocR, a c-di-GMP phosphodiesterase, significantly increased the motility, biofilm formation and protease production of ΔrpfF Bc [14]. Similarly, we found that in trans expression of the c-di-GMP synthase WspR diminished the swarming motility (Additional file 3: Figure S3A), biofilm formation (Additional file 3: Figure S3B), and protease activity (Additional file 3: Figure S3C) of ΔrpfF Bc to the level of double deletion mutant ΔrpfF Bc ΔcepI, whereas in tans expression of RocR, a c-di-GMP phosphodiesterase, significantly increased the motility, biofilm formation and protease production of ΔrpfF Bc (Additional file 3: Figure S3A-C). Taken together, these results demonstrated that BDSF system controls AHL signal production and influences the bacterial physiology via modulation of the intracellular c-di-GMP level in B. cenocepacia H111.
The cumulative effect BDSF and AHL systems on regulation of bacterial motility, biofilm formation and protease activity To understand how AHL and BDSF systems function in regulation of bacterial biological activities, we compared the phenotype changes of the wild type strain H111, single deletion mutants of rpfF Bc and cepI, and the double deletion mutant of rpfF Bc and cepI, in the presence and B A Figure 3 Effect of RpfR on AHL system. (A) AHL signal production was quantified with the aid of AHL reporter strain CF11 to test the β-galactosidase activity. (B) The β-galactosidase activity of a cepI-lacZ transcriptional fusion in H111 wild-type (■), ΔrpfR (▲) and ΔrpfR supplemented with BDSF signal (◆). For convenient comparison, the AHL signal production of wild-type strain was defined as 100% and used to normalize the AHL signal production of other strains. The data presented are the means of three replicates and error bars represents the standard deviation. absence of BDSF signal and OHL signal, respectively. As shown in Figure 5A-C, exogenous addition of 5 μM OHL or BDSF showed no evident effect on the phenotypes of wild type strain, suggesting that both signals were produced by H111 at "saturated" levels under the experimental conditions used in this study. As expected, addition of the same amount of OHL or BDSF to the corresponding AHL-minus and BDSF-minus mutants restored the mutants phenotypes including swarming motility ( Figure 5A), biofilm formation ( Figure 5B), and protease activity ( Figure 5C). It was noticed that exogenous addition of BDSF to the AHL-minus mutant ΔcepI failed to rescue the changed phenotypes ( Figure 5A-C). This could be explained that the mutant ΔcepI produced a similar "saturated" level of BDSF as the wild type, thus extra addition of BDSF had no effect in phenotype restoration. Interestingly, two different responses were noticed when OHL was added to the BDSF-minus mutant ΔrpfF Bc . While exogenous addition of the OHL signal could partially or even largely restore the biofilm formation and protease activity of this BDSF-minus mutant ( Figure 5B, 5C), exogenous addition of OHL had no effect on the swarming motility of ΔrpfF Bc ( Figure 5A). One plausible hypothesis is that regulation of bacterial motility requires only a low level of AHL signals and the BDSF-minus mutant could still produce sufficient amount of AHL signal molecules above the "threshold" level for full activation of the AHL-dependent motility, whereas in the cases of biofilm formation and protease activity deletion of rpfF Bc dropped the AHL level below the "threshold" concentration for full activation so that extra AHL addition could partially rescue the changed phenotypes. Consisting with the involvement of both BDSF and AHL systems in regulation of bacterial physiology, a cumulative effect on motility, biofilm formation and protease activity became evident when both rpfF Bc and cepI were knocked out ( Figure 5A-C). Significantly, only addition of both BDSF and OHL together could fully rescue the changed phenotypes of the double deletion mutant ΔrpfF Bc ΔcepI ( Figure 5A-C).  Effect of intracellular c-di-GMP level on AHL signal production. In trans expression of the c-di-GMP synthases, WspR from P. aeruginosa or the GGDEF domain of RpfR, in wild type H111 led to decreased AHL signal production; while overexpression of the c-di-GMP phosphodiesterases, RocR from P. aeruginosa or the EAL domain of RpfR resulted in increased AHL signal biosynthesis in BDSF-minus mutant ΔrpfF Bc . Quantification of AHL signal production was performed with the aid of AHL reporter strain CF11. For convenient comparison, the AHL signal production of wild-type strain was defined as 100% and used to normalize the AHL signal production of other strains. The data presented are the means of three replicates and error bars represents the standard deviation.

The impact of BDSF and AHL signaling systems on B. cenocepacia H111 pathogenicity
The impact of BDSF and AHL systems on B. cenocepacia virulence was evaluated by using C. elegans infection models. Agreeable with the previous reports [14,22], deletion of either rpfF Bc or cepI led to an reduction of virulence in both slow killing and fast killing assays of C. elegans ( Figure 6A, 6B). Remarkably, deletion of both rpfF Bc and cepI completely or almost completely abolished the bacterial virulence against C. elegans ( Figure 6A, 6B). one of the exceptions which contain both BDSF and AHL signaling mechanisms [7,12,13,15,19,23]. In this study, we have investigated the relationship of the two QS systems in signaling modulation of bacterial physiology and virulence. Although the recently published results believe that the BDSF and AHL systems control overlap set and specific genes [17,18], we found that the two QS systems exert cumulative effect on bacterial motility, biofilm formation and virulence factor production ( Figure 5A-C). In addition, we showed that BDSF regulates AHL signal production by influencing the c-di-GMP phosphodiesterase activity of its receptor RpfR. Given that both QS systems are widely conserved in the members of B. cepacia complex [7,10], it would be of great interest to investigate whether the similar crosstalking mechanisms of the AHL and BDSF systems are conserved in other members of the Burkholderia species. The intracellular signal c-di-GMP is a widely conserved second messenger, which is known to be involved in the regulation of a range of biological activities, including bacterial motility, biofilm formation and virulence factor production [10,24,25]. The research progress over the last few years shows that c-di-GMP commonly controls various biological functions through interacting with different receptor or effector proteins, such as PilZ, FleQ, VpsT, LapD, FimX, PelD, and Clp [26][27][28][29][30][31][32]. Interestingly, different from this paradigm, the findings from this study have unveiled a new mechanism with which c-di-GMP could influence bacterial physiology. We showed that null mutation of RpfR, which is an onecomponent BDSF sensor/response regulator containing a BDSF-binding domain and the GGDEF-EAL domains associated with c-di-GMP metabolism [14], resulted in a similar level of reduction in AHL signal production as the BDSF-minus mutant ΔrpfF Bc ( Figure 3A). Given that binding of BDSF by RpfR could substantially increases its activity in c-di-GMP degradation [14], it is rational that increasing c-di-GMP level would lead to downregulation of the AHL signal production and that decreasing c-di-GMP level would promote AHL signal production. Consisting with the above reasoning, our results showed that in trans expression of the c-di-GMP synthases, WspR from P. aeruginosa or the GGDEF domain of RpfR, in wild type H111 led to decreased AHL production (Figure 4), and that reducing c-di-GMP level in the BDSF-minus mutant ΔrpfF Bc by overexpressing either RocR from P. aeruginosa or the EAL domain of RpfR resulted in increased AHL signal biosynthesis ( Figure 4). These findings have elucidated a signaling pathway with which the BDSF-type QS system regulates the AHL-type QS system in B. cenocepacia and, additionally, have also further expanded our understanding of the c-di-GMP signaling mechanisms in modulation of bacterial physiology. However, how c-di-GMP controls AHL signal production remains to be further investigated.

Discussion
Identification of the second messenger c-di-GMP as a key element in the BDSF/c-di-GMP/AHL signaling pathway is also critical for explanation of the seeming puzzling relationship between BDSF and AHL systems in regulation of bacterial physiology and virulence and for elucidation of the QS regulatory mechanisms in B. cenocepacia H111. Our data showed that both BDSF and AHL systems control similar phenotypes including bacterial motility, biofilm formation and protease production with an obvious cumulative effect ( Figure 5). How these two QS systems interact in regulation and coordination of various biological functions? Do they act in cascade or independently? Our data support a partial "cascade" and a partial "independent" signaling mechanisms. Firstly, knocking out BDSF production affects AHL production but only partially reduced the total AHL level ( Figure 1). Secondly, null mutation of RpfR, which acts as a net c-di-GMP degradation enzyme upon interaction with BDSF [14], showed an almost identical effect on AHL signal production as the BDSF-minus mutant ( Figure 3). Thirdly, double deletion of the BDSF synthase gene rpfF Bc and the AHL synthase gene cepI showed a more severe impact on bacterial physiology and virulence than the corresponding single-deletion mutants (Figures 5 and 6). Finally, exogenous addition of either BDSF or AHL could only partially rescue the changed phenotypes of the double deletion mutant ΔrpfF Bc ΔcepI but a combination of BDSF and AHL could completely restore the changed phenotypes ( Figure 5). These findings, together with the previous knowledge of the second messenger c-di-GMP, suggest a working model of QS network in B. cenocepacia H111 in which BDSF and AHL elements are linked through the second messenger c-di-GMP (Figure 7). Considering that c-di-GMP is widely associated with the regulation of various biological functions, including motility, biofilm formation and virulence factor production [10,24,25], it is highly likely that BDSF system could influence the downstream gene expression through modulating the intracellular levels of both c-di-GMP and AHL signals. On the other hand, the AHL system could also act independently in regulation of downstream genes in the absence or presence of BDSF as the AHL signal production is only partially controlled by the BDSF system. In summary, the findings presented in this study have outlined a novel and flexible multicomponent QS network, which consists of BDSF and AHL QS systems and the second messenger c-di-GMP, in B. cenocepacia H111. This regulatory network has an interesting feature that both BDSF and AHL systems could act either together or independently in modulation of bacterial physiology and virulence, which may offer competitive advantages and flexibility in pathogen-host and microbemicrobe interactions.

Conclusions
The QS signal BDSF controls AHL signal production through regulation of the AHL synthase CepI expression at transcriptional level by modulating the intracellular level of the second messenger c-di-GMP through its novel receptor RpfR. The two QS systems have a cumulative role in regulation of various biological functions, including swarming motility, biofilm formation and virulence factor production. Exogenous addition of either BDSF or AHL signal molecules could only partially rescue the changed phenotypes of the double deletion mutant defective in BDSF and AHL signal production.
Slow killing assays were performed on NGM medium and fast killing assays on high-osmolarity PGS medium (peptone-glucose-sorbitol) [22]. BDSF and OHL signal molecules were added to the medium at a final concentration of 5 μM unless indicated otherwise.

Construction of in-frame deletion mutants and complementation strains
The cepI deletion mutant of B. cenocepacia strain H111 was used as the parental strain to generate the in-frame double deletion mutant of rpfF Bc and cepI, following the methods described previously [12]. For complementation analysis, the coding region of WspR was amplified by PCR using the primers listed in Additional file 4: Table S1, and cloned under the control of the S7 ribosomal protein promoter in plasmid vector pMSL7. The resultant construct was conjugated into the rpfF Bc deletion mutant B. cenocepacia H111 using tri-parental mating with pRK2013 as the mobilizing plasmid.

Construction of reporter strains and measurement of β-galactosidase activity
The promoter of cepI was amplified using the primer pairs listed in Additional file 4: Table S1 with HindIII and XhoI restriction sites attached. The resulting products were digested with HindIII and XhoI, and ligated at the same enzyme sites in the vector pME2-lacZ [35]. These constructs, verified by DNA sequencing, were introduced into B. cenocepacia H111 using tri-parental mating with pRK2013. Transconjugants were then selected on LB agar plates supplemented with ampicillin and tetracycline. Bacterial cells were grown at 37°C and harvested at different time points as indicated, and measurement of β-galactosidase activities was performed following the methods as described previously [36].

Biofilm formation, swarming motility and proteolytic activity assays
Biofilm formation in 96-well polypropylene microtiter dishes was assayed essentially as described previously [23]. Swarming motility was determined on semi-solid agar (0.5%). Bacteria were inoculated into the center of plates containing 0.8% tryptone, 0.5% glucose, and 0.5% agar. The plates were incubated at 37°C for 18 h before measurement of the colony diameters. Protease assay was performed following the previously described method [37]. Protease activity was obtained after normalization of absorbance against corresponding cell density.

Analysis of AHL signals
Bacterial cells were grown in NYG medium to a same cell density in the late growth phase. The supernatants were acidified to pH = 4.0 and extracted using ethyl  Figure 7 Schematic representation of the QS signalling networks in B. cenocepacia. RpfR Bc and CepI are involved in synthesis of BDSF and AHL signals, respectively. Perception of BDSF by RpfR substantially enhances its c-di-GMP phosphodiesterases activity and causes a reduction of the intracellular c-di-GMP level, and consequently affects the cepI transcriptional expression level and a range of biological functions, including swarming motility, biofilm formation and virulence through an unknown c-di-GMP effector X. The AHL-dependent QS system is also implicated in regulation of motility, biofilm formation, and virulence through its cognate receptor CepR. Solid arrows indicate the signalling regulation or signal transport.
acetate in a 1:1 ratio. Following evaporation of ethyl acetate the residues were dissolved in methanol. Quantification of AHL signals was performed using β-galactosidase assay with the aid of the AHL reporter strain CF11 as described previously [38]. Briefly, the reporter strain was grown in minimal medium at 28°C with shaking at 220 rpm overnight. The cultures were inoculated in the same medium supplemented with extracts containing AHL signals. Bacterial cells were harvested and βgalactosidase activities were assayed as described in previous section. For TLC analysis, 5 μl of the concentrated AHL extracts were spotted onto 10 × 20 cm RP-18254 s plate (MERCK) and separated with methanol-water (60:40, v/v). The plates were subsequently air dried and overlaid with 50 ml minimal medium containing 0.8% agarose, 50 μg ml-1 X-gal, and 1 ml stocked CF11 culture.