Introduction

Members of the genus Cronobacter spp. are considered opportunistic pathogens associated with rare but severe neonatal systemic infections predominantly in pre-term and/or low birth weight infants and thus have attracted the attention of public health authorities and researchers in the past1,2,3.

Epidemiological investigation of outbreaks of Cronobacter spp. infections in hospitals indicated powdered infant formula as a source of contamination, when these organisms were isolated from both reconstituted milk as well as from milk feeding equipment and utensils. The latter may be enhanced by the organism’s ability to adhere and form biofilms on many surfaces, including silicone, latex, polycarbonate (used in the feeding bottle manufacture) and stainless steel4,5.

Bacterial processes involved in biofilm formation and virulence, are often controled by quorum sensing (QS), a mechanism based on the production, release and detection of signaling molecules of low molar mass. Extracellular concentrations of signal molecules are sensed by the bacteria and, upon reaching a population density-dependent threshold, they are detected by the cells, which in turn induce target gene expression in a coordinated fashion6,7,8,9.

To date many structurally unrelated signal molecules have been identified, including N-acyl-homoserine lactones (AHLs) in Gram-negative bacteria, oligopeptides in many Gram-positive bacteria and autoinducer-2 (AI-2), which is thought to serve as a signal for interspecies communication7,8,9.

Another group of signal molecules are the cis-2-unsaturated fatty acids, often referred to as DSF (diffusible signal factor) family signals10. The first fatty acid signal, cis-11-methyl-2-dodecenoic acid, was identified in the culture supernatant of the phytopathogen Xanthomonas campestris pv. campestris (Xcc)11. Subsequently, fatty acid-based QS-systems were also identified and in members of the genera Xylella and Stenotrophomonas12,13 where they were shown to control the production of virulence factors11. More recent work showed that Burkholderia cenocepacia produces the signal molecule cis-2-dodecenoic acid, which was named BDSF (Burkholderia diffusible signal factor)14. BDSF is synthesized by the enoyl-CoA hydratase RpfFBc15 and is sensed by the receptor protein RpfR, which contains PAS-GGDEF-EAL domains16. Binding of BDSF to the PAS domain stimulates the c-di-GMP phosphodiesterase activity of RpfR, which in turn lowers the intracellular c-di GMP level. This signal transduction relay is very different from the one originally described for X. campestris, in which the DSF receptor RpfC is a hybrid sensor kinase that phosphorylates its cognate response regulator RpfG. This regulator contains in addition to a REC domain a HD-GYP domain, which is responsible for the c-di-GMP phosphodiesterase activity of the protein17.

Interestingly, homologs of RpfFBc and RpfR are present not only in many Burkholderia species but also in strains belonging to the genera Achromobacter, Yersinia, Serratia, Enterobacter and Cronobacter16, suggesting that RpfF/R type signaling systems may by far more widespread than anticipated. In this study we analysed the RpfF/R system of the clinical strain Cronobacter turicensis LMG 23827T and show that it is involved in the regulation of biofilm formation, macrocolony morphology, proteolytic activity and virulence.

Results and Discussion

RpfF directs the synthesis of a DSF family signal molecule and negatively regulates intracellular c-di-GMP levels in C. turicensis

Previous work identified homologs of both RpfR and RpfF from B. cenocepacia in C. turicensis LMG 23827T16. To investigate the role of this putative QS system in this organism we constructed defined mutants as well as genetically complemented derivatives thereof. We tested the strains for the production of DSF family signal molecules by the aid of the Burkholderia-based biosensor H111–rpfFBc pAN-L15 both in cross-streaking and liquid culture experiments. Under the conditions tested the wild type strain did not induce the biosensor. However, the complemented rpfF mutant, in which the wild type allele is expressed from a plasmid, clearly induced the biosensor (Fig. 1), suggesting that RpfF directs the biosynthesis of a cis-2 fatty acid signal molecule. We hypothesize that under standard laboratory conditions the amount of signal released by the wild type strain is below the detection limit of our bioassay but that the complemented strain, in which rpfF is expressed from a plasmid, produces sufficiently high amounts to induce the biosensor.

Figure 1
figure 1

Overexpression of rpfF activates the biosensor B. cenocepacia H111 –rpfFBc/pAN-L15, which is capable of detecting various DSF family signals.

(a) The C. turicensis LMG 23827T wild type (wt), the mutants (ΔrpfF, ΔrpfR), the complemented mutants (ΔrpfF + rpfF, ΔrpfR + rpfR) and the mutants carrying the empty vector (ΔrpfF/pCCR9, ΔrpfR/pCCR9) (vertical) were tested in cross-streak experiments against the biosensor (horizontal). The biosensor was clearly induced by the complemented rpfF mutant (ΔrpfF + rpfF). (b) The strains were also tested for the production of DSF family molecules in liquid assays. As with the cross streaking, induction of the biosensor was only observed with the ΔrpfF + rpfF strain. Error bars indicate SEM, n = 4; *P < 0.05 (ANOVA, oneway).

Full size image

RpfR family proteins contain a GGDEF as well as an EAL domain which are associated with the synthesis and degradation of c-di-GMP respectively18. The C. turicensis LMG 23827T RpfR homolog CBA31265 exhibits an identical domain structure. In order to evaluate the role of RpfR in this strain the intracellular c-di-GMP levels were determined in the wild type, the rpfR and rpfF mutants and in the complemented strains ΔrpfR + rpfR and ΔrpfF + rpfF. The intracellular c-di-GMP level of the rpfR and rpfF mutants was found to be 3.9-fold or 3.4-fold increased relative to the wild type. Genetic complementation of the mutant reduced the c-di-GMP level to the level of the wild type. These results suggest that both RpfR and RpfF have a negative effect on the intracellular c-di-GMP level (Fig. 2). This is in agreement with the finding that RpfR in B. cenocepacia exhibits a net phosphodiesterase activity16.

Figure 2
figure 2

RpfR and RpfF affect the intracellular c-di-GMP level.

The intracellular c-di-GMP level was significantly increased in the rpfR and rpfF mutants relative to the wild type. Detection was performed by by LC-MS/MS. Error bars indicate SEM, n = 2; *P < 0.05 (ANOVA, oneway).

Full size image

RpfF/R plays a role in quorum sensing regulated phenotypes in C. turicensis LMG 23827T

We next investigated whether the RpfF/R system is involved in the regulation of typically QS-associated phenotypes. In contrast to the wild type we observed a rough colony morphology of the rpfR and to a lesser degree with the rpfF mutant on Congo red agar plates (Fig. 3). The strong pinkish colour of the rpfR mutant relative to the wild type may suggest an increased production of cellulose and/or curli19. In order to support this hypothesis we performed expression studies targeting the gene coding for the catalytic subunit of the cellulose synthase bscA as well as the major curli subunit csgA. Expression of both genes was considerably increased in the mutants compared to the wild type. However, complementation only partially restored their expression (Fig. 4).

Figure 3
figure 3

The RpfF/R QS-system controls colony morphology and protease production.

Deletion of rpfR induced a rough, wrinkly colony morphology and increased EPS production on Congo red agar plates (CRA, upper panel). Both the ΔrpfR and the ΔrpfF mutant showed reduced protease production on skim-milk plates compared to the wild type (lower panel).

Full size image
Figure 4
figure 4

RT-qPCR analysis of csgA, bcsA and flhE gene expression in C. turicensis LMG 23827T (wt), the mutants (ΔrpfF, ΔrpfR), the complemented mutants (ΔrpfF + rpfF, ΔrpfR + rpfR) and the mutants carrying the empty vector (ΔrpfF/pCCR9, ΔrpfR/pCCR9).

The respective mRNA levels were normalized to the 16S rRNA reference gene. Error bars indicate SEM, n = 3.

Full size image

In addition, we observed that both mutants showed reduced proteolytic activity (Fig. 3). Swarming motility on NYG +0.4% agar was not significantly affected by inactivation of the RpfF/R system. However, the complemented mutants exhibited increased swarming motility (Fig. 5). These results are supported by the results of the RT qPCR experiments targeting the flagellar regulon-associated gene flhE, which was unaltered in the rpfF/R mutants but significantly higher in the complemented mutants (Fig. 4). This finding may be explained by a dose effect due to the additional copies of this gene in the complemented mutants.

Figure 5
figure 5

Overexpression of the RpfF/R QS-system increased swarming motility.

Strains were spot inoculated on 0.4% NYG agar and plates were photographed after 24 h incubation.

Full size image

Both mutants formed significantly more biofilm under static conditions in microtiter plates (Fig. 6a) than the parental strain. Complementation of the mutants partially restored the wild type phenotype. In the study by Hartmann et al. (2010)20 genes involved in biofilm formation in the closely related species Cronobacter sakazakii were identified using a transposon mutagenesis approach. BscA and flhE were – amongst others – two of the genes that were found to contribute to biofilm formation. Our expression analysis performed in this study suggests that bcsA but not flhE is regulated by the rpfF/R system (Fig. 4).

Figure 6
figure 6

The RpfF/R system regulates biofilm formation under static conditions.

(a) Deletion of either rpfR or rpfF increased biofilm formation in microtiter plates. (b) Partial restoration was obtained by genetic complementation or by supplementing the medium with BDSF (0.001 μM–20 μM) or DSF (20 μM). Error bars indicate SEM, n > 2. (c) Both RpfR and RpfF are involved in pellicle formation tested in NYG broth at room temperature for 48 h.

Full size image

Importantly, the strains ΔrpfR + rpfR and ΔrpfF/pCCR9 showed growth defects and did not reach the same OD as the other strains, which may explain the poor complementation of the ΔrpfR mutant. The growth curves of wild type strains, complemented mutants and mutants carrying the pCCR9 vector are depicted in Supplementary Figure S1. Partial restoration was also observed when the rpfF mutant was supplemented with at least 1 μM BDSF or 20 μM DSF (Fig. 6b). We also tested the various strains for pellicle formation, i.e. biofilm formation at the liquid-air interface. Both mutants showed increased pellicle formation and complementation restored the wild type behavior (Fig. 6c). In a study by Lehner et al. (2005)21 it has been reported that cellulose is one of the major components present in pellicles formed in Cronobacter spp. strains. The increased expression levels of bcsA in the mutants as observed in our study suggest a negative influence of the RpfF/R regulon in C. turicensis biofilm formation. This is in contrast to the homologous system of B. cenocepacia16 but similar to the genetically different DSF-dependent RpfCG systems of Stenotrophomonas maltophilia E77 or X. campestris pv. campestris22,23.

Zebrafish infection studies

We tested the ΔrpfF and the ΔrpfR mutants for pathogenicity in a zebrafish infection model. The dsRed-labeled wild type strain C. turicensis LMG23827T (wt::dsRed) served as control. The mortality rate of the zebrafish larvae at 48 hpi increased to approximately 90% for injection with wt::dsRed whereas the mortality rate decreased to 50% when the larvae were infected with the mutant ΔrpfF (Fig. 7a). Furthermore, the bacterial load was significantly lower with the rpfF mutant when compared with the wild type control (Fig. 7b) indicating a role of the RpfF/R system in the expression of virulence factors required for pathogenicity in the zebrafish model. Injection experiments using the complemented mutant strain (ΔrpfF + rpfF) resulted in higher mortality rate and higher bacterial load, whereas control experiments using the mutant strain transformed with the vector alone (ΔrpfF/pCCR9) yielded mortality rates and bacterial loads comparable to the ones observed in the ΔrpfF mutant experiments (data not shown).

Figure 7
figure 7

(a) Survival rates of zebrafish larvae injected with 50 CFU C. turicensis LMG 23827T/pRZT3::dsRed (wt::dsRed), the mutants ΔrpfF and ΔrpfR as well as the the complemented mutants ΔrpfF + rpfF and ΔrpfR + rpfR. A set of DPBS injected as well uninjected embryos served as controls. Error bars indicate SEM, n = 3. (b) Mean growth curve of C. turicensis LMG 23827T/pRZT3::dsRED (wt::dsRed), the mutants ΔrpfF and ΔrpfR as well as the complemented mutants (ΔrpfF + rpfF, ΔrpfR + rpfR) inside infected zebrafish larvae with a starting inoculum of approx. 50 CFU. A set of DPBS injected as well uninjected embryos served as controls. Error bars indicate SEM, n = 3

Full size image

Intriguingly, the mortality rate of larvae injected with the ΔrpfR strain was virtually indistinguishable from the wild type.

Here we have shown that C. turicensis posesses a RpfF/R family QS system which relies on a cis-2-unsaturated fatty acid signal molecule. RpfF/R-type QS systems are particularly widespread among members of the genus Burkholderia10. We analysed for the first time a RpfF/R family QS system in a bacterium not belonging to the genus Burkholderia and demonstrated that despite the phylogenetic distance (β versus γ subdivision of proteobacteria) both the molecular mechanism as well as the regulated phenotypes are very similar. Like in B. cenocepacia, the RpfF/R system was found to affect swarming motility, biofilm formation and virulence in C. turicensis. Furthermore, in both organisms the QS system modulates the intracellular secondary messenger c-di-GMP and this in turn appears to regulate the observed QS-dependent phenotypic traits. The finding that the rpfF but not the rpfR mutant reduced the virulence of C. turicensis suggests that an alternative signal receptor may be present in this strain. This is not unprecedented, as in B. cenocepacia an alternative BDSF receptor, BCAM0227, has been identified that is used by some strains as a parallel signaling system to control a subset of functions24. However, a bioinformatic analysis neither identified a homolog of BCAM0227 nor of rpfC, the DSF receptor of Xcc25.

In conclusion, our data provide evidence that RpfF/R-type QS systems are not restricted to Burkholderia sp. but may be widespread among Gram-negative bacteria, in which they influence surface colonization and virulence through modulation of the intracellular c-di-GMP levels. It will be of interest to investigate if homologous systems in other bacteria will control the same phenotypes.

Material and Methods

Bacterial strains and culture conditions

C. turicensis LMG 23827T 26, a clinical isolate responsible for two fatal sepsis cases in neonates in Zurich in 2006 was used in the study. Strains C. turicensis LMG 23827T_NalR as well as C. turicensis LMG 23827T/pRZT3::dsRed were described previously27,28.

For selection purposes, during zebrafish embryo infection experiments, C. turicensis LMG 23827T_ΔrpfF/pCCR9 as well as C. turicensis LMG 23827T_ΔrpfR/pCCR9 were constructed by transformation of the strains with the vector using standard methods.

Strains were grown in Luria–Bertani (LB) broth over night at 37 °C with gentle shaking. Where appropriate, culture medium or agar was supplemented with nalidixic acid at 256 mg L−1 (C. turicensis LMG 23827T_NalR), chloramphenicol at 30 mg L−1 (strains harbouring pDS132) or both (transconjugant strains) or tetracyclin at 50 mg L−1 (strains harbouring pCCR9 or pRZT3::dsRed).

For microinjection experiments, the bacteria were harvested by centrifugation at 5000 × g for 10 min and washed once in 10 ml of Dubelcco’s phosphate buffered saline (DPBS, Life Technologies, Switzerland.) After a second centrifugation step, the cells were resuspended in DPBS and appropriate dilutions were prepared in DPBS.

DNA extraction and manipulations

Chromosomal DNA was isolated using the DNeasy Blood and Tissue kit, plasmids were extracted with the QIAprep Spin Miniprep or Plasmid Midi kits following the manufacturer’s instructions. For purification purposes (PCR, restriction digest, agarose gel purification) the Qiagen MinElute PCR Cleanup kit or MinElute Gel Purification kit was employed. Enzymes and respective buffers were obtained from Roche Molecula Diagnostics (Rotkreuz, Switzerland) and used according to the manufacturer’s instructions.

Construction of C. turicensis LMG 23827T in frame deletion mutants

Bacterial strains, plasmids and primers used for the construction of mutants are listed in Supplemental Table S1. Deletion mutants of C. turicensis LMG 23827T rpfF (CTU_23310) and rpfR (CTU_23300) genes were constructed following the protocol described by Philippe et al. (2004)29. Details are provided in the supplementary material.

Phenotypic assays

Colony morphology: Overnight cultures grown in LB were adjusted to an OD600 = 1.0 in AB minimal medium30. 5 μl of this cell suspension was spottet on CRA plates (2 g Casamino acids, 0.3 g yeast extract, 80 μl of 1 M MgSO4, 4 g agar, dH2O ad 200 ml, supplemented with 1.6 ml congo red (0.5% in 50% EtOH), 0.65 ml coomassie blue (0.3% in 50% EtOH). The plates were incubated at room temperature for six days before colonies were photographed.

Protease production: Cells of an overnight culture were resuspended in LB and 5 μl cell suspension was spottet on skim milk plates (1% LB agar, 2% w/v skim milk powder). Plates were incubated at 37 °C for two nights and then kept at room temperature.

Swarming motility: Analysis was performed as previously described by Deng et al. (2012)16, except motility was monitored on NYG plates containing 0.5% peptone, 0.3% yeast extract, 2% glycerol and 0.4% agar.

Biofilm formation: Overnight cultures were washed and diluted to an OD600 = 0.01 in AB minimal medium supplemented with 0.4% glucose and 0.5% casamino acids30. 100 μl samples were added to 96 well plates incubated statically for 18 h at 30 °C. Growth was measured at 550 nm in a plate reader (Synergy HT; Bio-Tek, Germany). Surface attached cells were stained by the addition of 100 μl of 1% crystal violet for 30 min at room temperature. The plate was washed thoroughly with tap water and air-dried. To solubilize the stain, 120 μl DMSO was added to each well, incubated for 20 min at room temperature and OD at 570 nm was measured. Data are based on at least 2 independent experiments with 7 technical replicates each.

Bioassays for the production of cis-2 fatty acids by using the biosensor B. cenocepacia H111 –rpfFBc/pAN-L15. This sensor is sensitive to nM levels of synthetic BDSF and is suitable to detect a wide range of cis-2 fatty acid molecules (Suppiger et al. submitted). In cross-streaking experiments both the test- and the sensor strain were streaked on LB agar plates close to each other to form a T. The plates were incubated overnight at 37 °C. Following the addition of 10 μl decanal to the lid of the plate the bioluminescence of the sensor strain was visualized using the NightOWL LB 983 (Berthold Technologies, Zug, Switzerland). In liquid bioassays the biosensor was grown in LB broth containing kanamycin 100 μg ml−1 to an OD600 of 2.0. Overnight cultures of the strains to be tested were centrifuged at 6000 rpm, 5 min and the supernatant (SN) was centrifuged again. 100 μl of this cell-free SN was mixed with 100 μl sensor and incubated for 20 hours at 30 °C. Relative luminescence units (RLU) were obtained by adding 1–2 μl Decanal (Sigma Aldrich, Buchs, Switzerland) to each well and detection was performed using a plate reader (Synergy HT; Bio-Tek, Germany).

Intracellular cyclic-di-GMP level

Bacterial overnight cultures were subcultured in LB medium and 5 ml were harvested at an OD600 = 2.0 by centrifugation at 5000 rpm, 4 °C. Nucleotide extraction was performed as described by Spangler et al. (2010)31 with slight modifications: cXMP was omitted and the solvent was evaporated in Speedvac at 60 °C. Quantification was performed by LC-MS/MS32.

Expression analysis of selected genes by RT-qPCR

The expression levels of the 16S rRNA, csgA, bcsA, and flhE genes in Cronobacter turicensis LMG23827T wild type and its respective rpfR and rpfF mutants that were grown in AB medium supplementd with 0.4% glucose and 0.5% casamino acids at 30 °C to early stationary phase were determined using reverse transcription quantitative-PCR (RT-qPCR). 1.5 ml of the above bacterial suspension was re-suspended in 0.5 ml of the lysis buffer of the RNeasyPlus Mini Kit (Qiagen, Hilden, Germany). The samples were transferred on to the lysing bead matrix in MagNA lyser tubes and mechanically disrupted (1 min at 6500 rpm) using the MagNA Lyser Instrument (Roche Molecular Diagnostics, Rotkreuz, Switzerland). RNA was isolated from the bacterial lysates following the RNeasyPlusMini Kit protocol (Qiagen). Genomic DNA was removed by using a genomic DNA binding column and carrying out an on column DNAse I digestion. RNA was eluted in 50 μl of RNAse-free water and subsequently quantified and quality controlled using the Nanodrop and BioAnalyzer instruments, respectively. 100 ng of RNA were reverse transcribed to cDNA using the Quantitect Reverse Transcription Kit (Qiagen). Residual DNA contamination was ruled out in each RNA sample by including a control in which the RT enzyme was omitted. Quantitative PCR was performed on 2.5 ng cDNA using the SYBR green I kit (Roche Molecular Diagnostics) and primers that are listed in Supplemental Table S2 in the LC480 (Roche Molecular Diagnostics) instrument. Following RT PCR conditions were applied for all four genes: 5 min 95 °C followed by 40 cycles of 95 °C for 10 seconds, 50 °C for 20 seconds, 72 °C for 20 seconds, 78 °C for 1 second. Quantification was performed using the Light Cycler 480 Relative Quantification Software (Roche Molecular Diagnostics). The csgA, bcsA, flhE mRNA levels were normalized using 16S rRNA as reference gene27.

Zebrafish infection studies

Zebrafish (Danio rerio) strains used in this study were albino lines. Husbandry, breeding and microinjection of approx. 50 CFU of bacteria into the yolk sac of 2 dpf embryos was performed following the procedure described in the study by Fehr et al. (2015)28.

A set of uninjected embryos, incubated in E3 maintenance medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4) was included in order to determine the quality of the embryos; embryos injected with DPBS served as controls. Injected embryos were transferred into 24-well plates (1 embryo per well) in 1 ml E3 medium per well, incubated at 28 °C and observed for signs of disease and survival under a Leica M165 C stereomicroscope twice a day. In order to follow the course of infection embryos or larvae were collected at several time points, namely at 0, 24, 48 and 72 h post infection (hpi) and individually treated for bacterial enumeration.

Research was conducted with approval (NO 216/2012) from the Veterinary Office, Public Health Department, Canton of Zurich (Switzerland). The applied methods were carried out following the approved guidelines.

Bacterial enumeration by plate counting

The larvae were transferred to 1.5 ml centrifuge tubes and disintegrated by repeated pipetting and vortexing for 3 min in 1 mL of DPBS supplemented with 1% Triton X- 100 (Sigma-Aldrich, Buchs, Switzerland). Subsequently, serial dilutions of this mixture were plated onto LB plates supplemented with tetracycline 50 mg L−1 (strains harboring pCCR9 or pRZT3::dsRed). The plates were incubated up to 48 h at 37 °C.

Survival assay

Embryos were microinjected as mentioned above and maintained individually in 24-well plates in E3 medium at 28 °C. The number of dead larvae was determined at different time points visually based on the absence of a heartbeat.

Statistical analysis

Statistics and graphs were performed using GraphPad Prism 6 (GraphPad Software, San Diego, USA). Experiments were executed at least three times, unless stated otherwise. The CFU counts of individual larva at different time points and under different conditions were verified for significant variances by one-way ANOVA with Bonferroni’s post-test.

Additional Information

How to cite this article: Suppiger, A. et al. The DSF type quorum sensing signalling system RpfF/R regulates diverse phenotypes in the opportunistic pathogen Cronobacter. Sci. Rep. 6, 18753; doi: 10.1038/srep18753 (2016).