Vfm a new quorum sensing system controls the virulence of Dickeya dadantii

bs_bs_banner

Environmental Microbiology (2013) 15(3), 865–880

doi:10.1111/1462-2920.12049

Vfm a new quorum sensing system controls the virulence of Dickeya dadantii William Nasser,1 Corinne Dorel,1 Julien Wawrzyniak,1 Frédérique Van Gijsegem,2 Marie-Christine Groleau,3 Eric Déziel3 and Sylvie Reverchon1* 1 UMR5240 CNRS/INSA/UCB, Université de Lyon, F-69003; INSA-Lyon, Villeurbanne, F-69621, France. 2 UMR217 INRA/AgroParisTech/UPMC P6, Paris, France. 3 INRS-Institut Armand-Frappier, Laval, Québec, H7V 1B7, Canada. Summary Dickeya dadantii is a plant pathogen that secretes cell wall-degrading enzymes (CWDE) that are responsible for soft-rot symptoms. Virulence genes are expressed in a concerted manner and culminate when bacterial multiplication slows. We identify a 25 kb vfm cluster required for D. dadantii CWDE production and pathogenesis. The vfm cluster encodes proteins displaying similarities both with enzymes involved in amino acid activation and with enzymes involved in fatty acid biosynthesis. These similarities suggest that the vfm genes direct the production of a metabolite. Cell-free supernatant from the D. dadantii wild-type strain restores CWDE production in vfm mutants. Collectively, our results indicate that vfm genes direct the synthesis of an extracellular signal and constitute a new quorum sensing system. Perception of the signal is achieved by the two-component system VfmH– VfmI, which activates the expression of the vfmE gene encoding an AraC regulator. VfmE then activates both the transcription of the CWDE genes and the expression of the vfm operons. The vfm gene cluster does not seem to be widespread among bacterial species but is conserved in other Dickeya species and could have been laterally transferred to Rahnella. This work highlights that entirely new families of bacterial languages remain to be discovered.

Received 3 October, 2012; revised 30 October, 2012; accepted 10 November, 2012. *For correspondence. E-mail [email protected]; Tel. (+33) 472 43 26 95; Fax (+33) 472 43 15 84.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd

Introduction Dickeya dadantii (formerly Erwinia chrysanthemi) is a softrot pathogen that causes severe diseases in a wide range of fruit and vegetable crops. It is listed in the top 10 of the most important bacterial plant pathogens (Mansfield et al., 2012). The plant infection by Dickeya proceeds in two distinct steps. Dickeya first resides and grows in the vegetal intercellular spaces causing a latent infection lacking symptoms (Murdoch et al., 1999; Fagard et al., 2007). Then, if the bacteria encounter favourable conditions such as mild temperatures (30°C), high humidity levels and poor oxygen availability, a shift to disease expression occurs (Perombelon, 2002). A successful infection relies on the coordinated expression of numerous virulence factor-encoding genes that occurs when a threshold bacterial population density is reached. The virulence of Dickeya is mainly correlated with their ability to synthesize and secrete plant cell wall-degrading enzymes (CWDE), including pectinases (Robert-Baudouy et al., 2000), proteases (Letoffe et al., 1990) and a cellulase (Py et al., 1991). Many other factors reviewed in Sepulchre and colleagues (2007) are required for the full expression of the disease (lipopolysaccharide, exopolysaccharide, the Hrp type III secretion system, iron uptake, etc.) but are not directly responsible for the development of soft-rot symptoms. CWDE are massively produced only during an attack of the host plants, whereas the latent infection is characterized by a moderate expression of the virulence genes (Lebeau et al., 2008; Reverchon et al., 2010). CWDE are therefore excellent biomarkers of the Dickeya pathogenic process. This property has been widely used to identify the regulatory components controlling bacterial virulence through screening of mutant libraries with simple plate assays for the production of pectinases, proteases and cellulases. Ten major regulators were identified (KdgR, PecS, PecT, CRP, H-NS, FIS, Fur, GacA, SlyA and MfbR) (Reverchon et al., 1991; 1994; 1997; 2010; Castillo et al., 1998; Franza et al., 2002; Nasser and Reverchon, 2002; Lautier et al., 2007; Lebeau et al., 2008; Haque et al., 2009; Ouafa et al., 2012) and were reviewed in Charkowski and colleagues (2012). Most of these regulators control not only CWDE production but also other virulence factors such as exopolysaccharide production, motility and synthesis of compounds involved in

866 W. Nasser et al. resistance to oxidative burst (Castillo et al., 1998; Condemine et al., 1999; Reverchon et al., 2002; Lautier and Nasser, 2007; Hommais et al., 2008). The coordination of the expression of the pectinase genes has been mathematically modelled (Sepulchre et al., 2007; Kepseu et al., 2010b). Regulation of the pectinases genes at the promoter level appears to mainly act as a security device to prevent the premature expression of these genes. In addition, modelling studies suggest the existence of a quorum sensing control of pectinase production (Sepulchre et al., 2007; Kepseu et al., 2010a). However, although specific interactions have been observed in vitro between the AHL-related quorum sensing regulator ExpR and pel gene promoters (Nasser et al., 1998), there is no specific phenotype associated to the expI or expR mutations. In addition to ExpI-ExpR, the LuxS-based quorum sensing system of D. dadantii does not play a pivotal role in the cell density-dependent control of virulence gene expression in vitro or in planta (Reverchon et al., 1998; Mhedbi-Hajri et al., 2011). The nature of the Dickeya quorum sensing system thus remains to be discovered. To identify such a system, we looked for mutants displaying defects in the synthesis of all of the CWDE that could be complemented by the cell-free culture supernatant of D. dadantii. Two mutants satisfying these criteria were obtained. The corresponding mutated genes belong to a 25 kb cluster that we named vfm for Virulence Factor Modulating cluster. Here, we report the characterization of the vfm locus that directs the synthesis of an extracellular signal that controls Dickeya virulence. We elucidate how the signal is sensed and transduced into CWDE expression at a molecular level. Results Synthesis of CWDE is dependent on the production of an extracellular signal To determine whether the synthesis of CDWE (pectinases, cellulase and proteases) is dependent on an extracellular signal in D. dadantii, we screened for mutants with decreased production of all of these enzymes that could be complemented by cell-free culture supernatant. Out of approximately 5000 insertion mutants generated using transposon Tn5-B21, strains A3470 and A3473 that displayed reduced synthesis of proteases, cellulase and pectinases based on in situ plate assays (Fig. S1A) and whose protease– phenotype can be restored by addition of a cell-free supernatant of the D. dadantii prtE protease-deficient strain A3997 were isolated (Fig. S1B). Protease activity was selected as representative of CWDE production because it is the strongest affected phenotype in the in situ plate assays. The sites of the Tn5-B21 insertions were determined

by sequencing the flanking chromosomal DNA. The two insertions were 25 kb apart from each other and were located between the expR gene, which encodes a quorum sensing regulator, and the indigoidine biosynthesis genes (Fig. 1A). The insertion from strain A3470 is located at the end of a gene (vfmD, ID16075) encoding a putative membrane anchored esterase (Fig. 1A, Table S1). The insertion from strain A3473 is located between two divergently transcribed genes: expR (ID19414) and a gene (vfmK, formerly virA ID19411) encoding either a putative decarboxylase that acts on ornithine, diaminopimelate, arginine or an alanine racemase (Fig. 1A, Table S1). New insertions were constructed in different open reading frames from this region (Fig. 1A) and the resulting mutants were analysed for CWDE production after recombination of these insertions in the D. dadantii chromosome (Fig. 1B). Most of these mutants displayed decreased CWDE production except the insertion in ID16066 (Fig. 1B). Therefore, we designate this cluster vfm for Virulence Factor Modulating cluster. The decreased CWDE production in vfm mutants is correlated to a decreased expression of CWDE genes as revealed by real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis (Fig. S2). This indicates that the vfm cluster controls the transcription of the CWDE genes. Transcriptional organization of the vfm locus and occurrence of this locus in other bacteria RT-PCR analysis revealed the existence of three polycistronic messages vfmAZBCD, vfmFGHIJ, vfmKLMNOPQRSTUVW and two monocistronic mRNAs vfmE, ID16066 in the vfm locus (Fig. S3). Interestingly, the vfm gene cluster does not seem to be widespread among bacterial species. It is conserved in other Dickeya species (Dickeya zeae: Ech586; Dickeya paradisiaca: Ech703; Dickeya chrysanthemi: Ech1591). However, although the expI-expR-vfm-indC synteny is conserved in D. dadantii, D. zeae and D. chrysanthemi, it is not conserved in the D. paradisiaca: Ech703 strain, which lacks expI and indC genes. Surprisingly, the vfm cluster is absent from the related pectinolytic enterobacteria of the Pectobacterium genus. The recent sequencing of plasmid pRAHAQ01 from the plant-associated enterobacterium Rahnella sp. Y9602 revealed the presence of a cluster homologous to vfm. This plasmid contains many genes related to Dickeya including genes involved in pectin catabolism. A previous analysis of the diversity of small cryptic plasmids in the genus Rahnella revealed that some of the cryptic plasmids contained regions homologous to the chromosomes of Erwinia tasmaniensis and Photorhabdus luminescens. These

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya

867

Fig. 1. A. Genetic organization of the vfm region. Three polycistronic mRNAs (vfmAZBCD, vfmFGHIJ and vfmKLMNOPQRSTUVW) and two monocistronic mRNAs (vfmE and ID16066) were detected by RT-PCR in the vfm cluster (Fig. S3). The two insertions from mutants A3473 and A3470 are 25 kb apart. The new insertion mutations constructed in different ORFs of the vfm region are indicated by a flag (for uidA-KmR cassette insertion) or a flower (for nptI-sacB-sacR KmR cassette insertion) or a circle (for CmR cassette insertion). For uidA-KmR cassette insertions, the transcriptional direction of the uidA reporter gene is shown by the orientation of the flags. The uidA gene is expressed when the flag is black. B. The pectinase (Pel), cellulase (Cel) and protease (Prt) phenotypes of the resulting mutants were analysed using agar plate assays. R: reduced production, NA: no affected production.

observations highlight the importance of plasmids for lateral gene transfer to distinct genera (Rozhon et al., 2010) and lead us to suggest that the vfm cluster was recently transferred from Dickeya to Rahnella.

type of the other vfm mutants in cross feeding tests. No cross feeding could be obtained with any cell-free supernatant in any background indicating that inactivation of any vfm transcriptional unit abolishes signal production (Table 1).

The four vfm transcriptional units are required for extracellular signal production

The vfm mutants are affected in their virulence abilities

The complete vfm cluster was cloned in vivo on plasmid pSR2895 (R′ plasmid pULB110 carrying the expI::CmRvfm-indC::KmR region). This plasmid complemented all of the vfm mutants. Moreover, cell-free culture supernatant from the Escherichia coli strain harbouring plasmid pSR2895 restored protease production of the majority of vfm mutants except vfmE and vfmH (Fig. 2). In contrast, cell-free culture supernatant from the E. coli strain harbouring the empty plasmid pULB110 did not restore protease production (Fig. 2). These results indicate that the vfm cluster directs the synthesis of an extracellular signal that controls D. dadantii CWDE production. Cell-free supernatants from four vfm mutants (vfmA, vfmE, vfmH and vfmK) representative of the four transcriptional units of the vfm cluster (Fig. S3) were tested for their ability to complement protease- pheno-

The virulence of the vfmA, vfmE, vfmH and vfmK mutants, which are representative of the vfmAZBCD, vfmE, vfmFGHIJ and vfmKLMNOPQRSTUVW transcriptional units, was tested on Saintpaulia ionantha (African violet), the plant from which strain 3937 has been isolated. In our experimental conditions, approximately one half of the plants inoculated with the wild-type strain showed maceration symptoms 24 h after inoculation, and most of the plants harboured systemic symptoms 6 days after infection (Fig. 3). Compared with the wild-type strain, the virulence of all of the vfm mutants was strongly affected because only a few of the inoculated plants exhibited symptoms (Fig. 3). Thus, the vfm cluster is required for the development of soft-rot disease on plants, indicating a key role of this locus in D. dadantii virulence.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

868 W. Nasser et al. Fig. 2. Synthesis of CWDE is dependent on the production of an extracellular signal by the vfm locus. Protease activity was detected in various vfm mutants in the presence of cell-free culture supernatants. A. A milk plate supplemented with E. coli HB101/pULB110 culture supernatant was used as a negative control. B. A milk plate supplemented with culture supernatant from E. coli HB101/pSR2895 carrying the vfm cluster was used to visualize complementation. Culture supernatants were incorporated into the agar medium at a final concentration of 10% v/v.

Expression of the vfm cluster is induced by an extracellular signal

VfmH is a transcriptional regulator of vfmAZBCD operon and vfmE gene

The expression of the vfmA, vfmE, vfmH and vfmK genes was further analysed by qRT-PCR in the wild-type background and in the vfmA, vfmE, vfmH and vfmK mutants. Inactivation of any of these vfm genes results in the decreased expression of the other vfm genes relative to the wild-type strain (Fig. 4A). Thus, expression of the vfm genes is interdependent. The expression of vfmH is the least affected by the other vfm mutations (Fig. 4A). The expression of the vfm genes was also analysed in the presence of cell-free culture supernatant from E. coli strains containing either the pSR2895 plasmid harbouring the vfm region or the empty pULB110 plasmid. Variable behaviours were observed for the different vfm genes in response to exogenous addition of the extracellular signal. In the vfmA background, the expression of vfmE and vfmK was induced by the extracellular signal, whereas there was no significant change in vfmH expression (Fig. 4B and C). Similarly, in the vfmK background, the expression of vfmA and vfmE was induced by the extracellular signal, whereas vfmH expression was not induced (Fig. 4B and C). In the vfmE and vfmH backgrounds, none of the vfm genes was induced by cell-free supernatant from HB101/pSR2895-Vfm+ (Fig. 4B and C). Thus, the proteins from the vfmFGHIJ and vfmE transcriptional units seem to play a pivotal role in signal perception. Moreover, as vfmE expression is induced by the extracellular signal and as this induction is lost in the vfmH mutant, we postulated that the vfmFGHIJ operon is on the top of the hierarchy in signal perception.

The vfm cluster contains two putative transcriptional regulators: (i) the vfmE gene, which encodes a protein of the AraC-type family, and (ii) vfmH, which encodes a protein of the NtrC s54-dependent response regulator family (Table S1). Furthermore, VfmH is associated with the VfmI sensor histidine kinase, suggesting that VfmH–VfmI constitutes a two-component system (Table S1). Overproduction and purification of both VfmH and VfmE regulators have been attempted. A stable VfmH-His6 protein was

Table 1. Production of Vfm signal by different mutants. Vfm signal Mutation

Production

Response

A3997 prtE vfmA vfmE vfmH vfmK

+ -

+ + +

Production of the signal was determined by assaying cell-free culture supernatant from the indicated mutant for its ability to complement protease- phenotype of the other vfm mutants as described in Fig. S1B. A plus indicates that signal activity in cell-free culture supernatant from the indicated mutant was similar to that from the signalproducing A3997 strain. A minus indicates that there was no detectable signal activity. The ability of the indicated mutant to respond to the signal present in the cell-free culture supernatant from strain A3997 is also summarized. A plus indicates that the protease- phenotype of the indicated mutant can be restored by the cell-free culture supernatant from strain A3997. A minus indicates that the protease- phenotype could not be restored.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya

869

Fig. 3. Virulence of the vfm mutants on Saintpaulia ionantha. Pathogenicity tests on 1-month-old potted S. ionantha cv. Blue Rhapsody cuttings were performed as described in Lebeau and colleagues (2008). Eighteen plants were inoculated for each bacterial strain. The experiment was performed in triplicate; one representative experiment is shown. The occurrence of symptoms was scored daily for 6 days. Symptoms are classified in four stages: stage 0, no symptoms; stage 1, rotting confined to the infiltrated zone; stage 2, maceration of the leaf limb; and stage 3, maceration of the whole leaf including the petiole. The significance of differences in symptom severity between wild-type and mutant strains was evaluated (P < 0.05) with a Fisher-exact test. No symptoms were observed for the negative control plants treated with sterile buffer.

obtained, but no stable VfmE derivatives could be purified because of the intrinsic tendency of VfmE to aggregate. In vitro gel retardation assays were performed in the presence of the phosphorylated form of VfmH-His6 (VfmHHis6~P) and DNA fragments containing the regulatory regions of the vfmA, vfmE, vfmF, vfmK, pelE and celZ genes (Fig. 5). The pelE and celZ genes encode the pectinase PelE and the cellulase CelZ respectively. VfmHHis6~P bound to the vfmA and vfmE DNA fragments, but no interaction could be detected between this protein and

the vfmF, vfmK, pelE and celZ genes. Competition assays performed with a specific (unlabelled vfmE or vfmA promoter regions) or a non-specific competitor, demonstrate the specificity of VfmH binding to the vfmE and vfmA regulatory regions (Fig. 5). Thus, VfmH regulates the expression of the vfmAZBCD operon and of the vfmE gene by specifically binding to their promoter regions. The absence of an interaction between VfmH and pelE or celZ suggested that VfmE could be responsible for the direct activation of CWDE genes.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

870 W. Nasser et al.

Fig. 4. Expression of the vfm genes. The impact of the vfmA, vfmE, vfmH and vfmK mutations on vfmA, vfmE, vfmH and vfmK transcript levels was monitored by qRT-PCR. A. The expression levels of the vfmA, vfmE, vfmH and vfmK genes in the vfmA, vfmE, vfmH and vfmK mutants were measured as transcript levels relative to wild-type strain (WT) at early stationary growth phase from cultures in M63 minimal medium supplemented with sucrose. The values represent the average gene expression levels ⫾ SD from three independent experiments. The results obtained with the vfm mutants are different from the wild-type results, with P-values < 0.05 in a one sample t-test. B. The expression of the vfmA, vfmE, vfmH and vfmK genes was measured in the different vfm mutants in the presence of cell-free culture supernatant from the E. coli strains HB101/pULB110 or HB101/pSR2895 carrying the vfm region. Bacteria were treated with 10% (v/v) cell-free supernatant, and the samples were removed 15 min post treatment for gene expression analysis. The effect of the HB101/pSR2895 supernatant on the expression of vfmA, vfmE and vfmH was normalized to the amount of the transcript obtained for the same gene in the vfmK mutant in the presence of the HB101/pULB110 supernatant (normalized to 1). The effect of the HB101/pSR2895 supernatant on the vfmK expression was normalized to the amount of vfmK transcript obtained in the vfmA mutant in the presence of the HB101/pULB110 supernatant (normalized to 1). Each value represents the mean of three independent experiments. The bars indicate the standard deviation. The results obtained with the HB101/pSR2895 supernatant are different from those with the HB101/pULB110 supernatant (P-values < 0.05 in a sample t-test) except for vfmH expression. C. For each mutant, the fold changes (FC) are expressed as the ratio of the specific gene expression level in the presence of HB101/pSR2895 supernatant relative to that obtained in the presence of HB101/pULB110 supernatant.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya

871

Fig. 5. Binding of VfmH at the vfmA, vfmE, vfmF, vfmK, pelE and celZ promoters. 30 fmol of labelled DNA probes corresponding to promoter regions of vfmA, vfmE, vfmF, vfmK, pelE and celZ genes were incubated with increasing concentrations of VfmH-His6~P, as indicated on the top. The position of free DNA (F) and of the main VfmH-His6/DNA complexes (C) are indicated. Specificity of VfmH binding to vfmE and vfmA regulatory regions was evaluated by competition assays performed with 50-fold molar excesses of a specific competitor (unlabelled vfmE or vfmA promoter regions respectively) or with 50-fold molar excesses of a non-specific competitor. The unlabelled mfhR non-specific fragment derived from the coding region of a transcriptional regulator not controlled by VfmH was used as unspecific competitor.

VfmE is a transcriptional regulator of CWDE genes and all vfm operons Repeated attempts to detect an in vitro interaction between VfmE and the regulatory regions of the vfmA, vfmE, vfmF, vfmK, pelE and celZ genes most likely failed because of VfmE instability. To overcome this difficulty, VfmE binding to gene promoters was tested in vivo using chromatin immunoprecipitation (ChIP). A fusion of the VSVG epitope tag to the N-terminus of VfmE was used for these experiments. Dickeya dadantii A5461 strain, carrying chromosomic vfmE VSVG-tagged allele was constructed by marker exchange-eviction procedure (Ried and Collmer, 1987). This strain A5461 displayed a wildtype phenotype indicating that the construction was functional and that the VSVG-tag did not perturb the DNAbinding activity of VfmE. A vfmA::uidA-Kan mutation which abolishes the Vfm signal production was then also introduced in the VSVG-tagged vfmE strain in order to analyse the DNA-binding activity of VfmE in the absence of Vfm signal. Specific enrichment of the vfmA, vfmE, vfmF, vfmK, pelE and celZ regulatory regions were observed with immunoprecipitated DNA using Anti-VSVG antibodies in the A5461 VSVG-tagged vfmE strain. The amplified

products were confirmed as promoters of the interested genes by DNA sequencing (data not shown). These results demonstrated the specific binding of VfmE to the vfmA, vfmE, vfmF, vfmK, pelE and celZ operators (Fig. 6). In the vfmA background, binding of VfmE to its target operators was strongly reduced (Fig. 6). However, Western blot experiments, performed with the same culture extracts and the anti-VSVG antibodies, revealed that in the vfmA mutant, the VfmE cellular content was strongly reduced (data not shown) in accordance with the interdependent expression of the vfm genes. Therefore, in the absence of the Vfm extracellular signal, VfmE cellular content was low and did not allow activation of CWDE genes. In the wild-type background, when the extracellular signal accumulated, VfmE was produced and promoted activation of both CWDE genes and all vfm operons leading to rapid amplification of the Vfm signal and high production of virulence factors. Production of the Vfm signalling molecule is growth phase dependent The kinetics of the Vfm signal production were analysed by measuring the expression of the signal-sensitive

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

872 W. Nasser et al. Fig. 6. ChIP analysis of VfmE binding on the vfmA, vfmE, vfmF, vfmK, pelE and celZ promoters. Independent VSVG-VfmE ChIP experiments were performed with the wild-type strain and its VSVG-vfmE derivative, as well as with the vfmA mutant and its vfmA, VSVG-vfmE derivative. qPCR was then used to quantify the fold enrichment of regions immediately upstream of vfmA, vfmE, vfmF, vfmK, pelE and celZ genes. ChIP fold enrichment scores are calculated relative to the non-target rpoA internal amplicon and the signals obtained in the corresponding non-tagged VfmE+ strains (VSVG-VfmE/VfmE). These scores were averaged across three biological replicates. Bars indicate the standard deviation.

vfmE::uidA reporter fusion from strain A5243. This fusion was retained because it was the most sensitive to the signal. In addition, the vfmE mutation also abolished signal production because of the lack of activation of the vfmAZBCD and vfmKLMNOPQRSTUVW operons. Therefore, this strain was suitable to monitor signal activity in culture spent supernatants and to appreciate the impact of exogenous signal addition. In the presence of cell-free supernatants from the vfmA mutant strain, no induction of the vfmE::uidA fusion was observed confirming a role of VfmA in signal production (Fig. S4A). In contrast, increased expression of the vfmE::uidA fusion was observed in the presence of wild-type strain supernatants; the Vfm signal production increases from the middle until the end of the logarithmic phase and then declines (Fig. S4A). A dose-dependent response of the vfmE::uidA fusion was observed towards the wild-type supernatant containing the Vfm signal (Fig. S4B). Therefore, the extracellular accumulation of a self-induced Vfm signal during growth and the dose-dependent response of specific gene expression are strong arguments in favour of a quorum sensing system that coordinates cellular behaviour at the population level. The signal appeared to be highly polar because it is not extractible with organic solvents such as ethyl acetate, dichloromethane, chloroform or ethyl ether in either acidic or alkaline conditions. We also noticed that the signal molecule is heat resistant because it retains its activity after heating at 100°C for 5 min. Functional characteristics of vfm gene products Putative functions were assigned to vfm gene products based on the results of BLASTP searches. Most of the predicted amino acid sequences show moderate but significant similarity (< 45% identity) to functionally characterized or putative gene products; most of these products were from diverse bacteria phylogenetically distant from the genus Dickeya (Table S1). Besides the VfmE and VfmH transcriptional regulators, the vfm cluster encodes a wide range of proteins potentially involved in the biosyn-

thesis and exportation of a signalling molecule composed of both amino acid and fatty acid moieties. The VfmO and VfmP proteins encode amino acid activating enzymes and display similarities to the adenylation (A) domains of the non-ribosomal peptide gramicidin synthetase (NRPS). The NRPSpredictor2 tool that analyses A domain selectivity (Rottig et al., 2011) suggests that VfmO and VfmP would activate alanine. VfmK most likely catalyses L-alanine to D-alanine racemization. VfmN could function as an amino-acyl carrier protein because it contains a potential phosphopantetheine attachment site, and VfmJ is a 4-phosphopantetheinyl transferase. Thus, VfmO, VfmP, VfmN and VfmJ could act similarly to a NRPS loading module, selectively activating an amino acid for subsequent coupling. VfmQ could be involved in C–C bond formation on the amino acid tethered to the VfmN peptidyl carrier protein because of its similarity with the CmaC enzyme, which is involved in coronamic acid biosynthesis in Pseudomonas (Kelly et al., 2007). The second building block of the Vfm signal, fatty acids, are synthesized by the VfmA, VfmT and VfmW proteins, which display similarity to 3-oxoacyl-ACP synthases. These fatty acids are then most likely dehydrogenated by the acyl-CoA dehydrogenases VfmR and VfmU. Condensation between fatty acids and modified amino acids could be catalysed by the VfmM protein, which is an AMP-dependent ligase of the coumarate-CoA ligase subfamily of adenylating enzymes. The vfm locus contains two putative transport systems (VfmF–VfmG and VfmC) that could be involved in signal exportation. Given the putative nature of the signal, we favour the ABC-type transport system VfmF–VfmG, which displays similarities with the Lol system involved in the export of lipoproteins to the outer membrane of Gramnegative bacteria. Discussion Plant-pathogenic bacteria represent a threat impacting economy due to their ability to cause crop losses. Further-

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya more, an increased attention is given to their potential to be employed as biological weapons. Understanding of the mechanisms used by plant-pathogenic bacteria to infect their host is therefore of utmost importance. The infection is a sequential process involving coordinated changes in genetic activity facilitating on the one hand the survival of bacteria under hostile conditions encountered in the host and on the other hand, enabling the timely production of virulence factors. Every pathogen is unique and uses distinctive combinations of specific mechanisms to exploit its hosts. However a common theme in most pathogenic bacteria is the cell–cell communication process called quorum sensing to regulate pathogenicity. Quorum sensing systems represent attractive targets to combat infection diseases and it is of extreme significance to study these processes in a wide range of bacteria including plant pathogenic bacteria. It has been suggested that an as yet unidentified quorum sensing system might exist to allow D. dadantii to modulate the expression of virulence factors in a cell density-dependent process (Sepulchre et al., 2007; Kepseu et al., 2010a). The central aim of this study was to characterize this quorum sensing system. Starting from a screening for D. dadantii mutants that have decreased production of CWDE, we identified a virulence-factormodulating gene cluster that has the characteristics of a quorum sensing system. The vfm locus directs the production of an extracellular signal that regulates the expression of virulence genes and the expression of the vfm cluster itself. The production of a signalling molecule by the vfm locus is supported by two sets of data. First, a vfm mutation impacts the expression of a whole subset of virulence genes in a pleiotropic manner. Moreover, cellfree culture supernatants from strains containing a functional vfm cluster can restore the reduced expression of the target genes in the vfm mutants. The vfm-induced signal accumulates in culture medium as cells grow to high density and then declines. This transient production of the Vfm signal and its ability to induce its own synthesis are both traits associated with quorum sensing signalling molecules. The functional characteristics of the proteins encoded by the vfm locus suggest that the Vfm signal results from the condensation of at least two types of building blocks consisting of modified amino acids and fatty acids. The vfmE::uidA fusion was used to monitor the purification of the signal molecule from culture supernatant, but our attempts to characterize the signal by liquid chromatography/mass spectrometry (LC/MS) were unsuccessful probably in regard to its low concentration in the active fraction or because the molecule did not ionize. Our primary goal was to elucidate the mechanism by which proteins encoded by the vfm cluster mediate the response to the extracellular signal. As expected, the two mutants that are not complemented by the cell-free

873

culture supernatants are the vfmH and vfmE regulatory mutants indicating that both VfmH and VfmE regulators are required for the induction of target genes by the extracellular signal. However, because the expression of vfmE is itself induced by the Vfm signal and because this induction is lost in the vfmH mutant, we concluded that VfmH mediates the response to the Vfm extracellular signal. This result led us to postulate a hierarchical organization in which the two-component system VfmH–VfmI senses the signal and controls vfmE expression. This conclusion is supported by in vitro experiments showing that the response regulator VfmH specifically binds to the vfmE and vfmA promoters and activates the transcription of the regulatory gene vfmE and the expression of the vfmAZBCD operon involved in signal biosynthesis. VfmH does not interact with the regulatory region of its own operon (vfmFGHIJ), which explains the absence of modulation of vfmH expression by supernatants containing the signal. Furthermore, VfmH does not interact with the regulatory regions of CWDE genes, whereas the VfmE regulator specifically binds to the promoters of CWDE genes. These observations confirm the regulation cascade in which VfmH responds to the signal and activates the synthesis of the regulator VfmE, which then activates the expression of virulence genes (Fig. 7). ChIP experiments revealed that VfmE also interacts with the regulatory regions of all vfm operons. On the basis of the decreased expression of the vfmAZBCD, vfmFGHIJ and vfmKLMNOPQRSTUVW operons in the vfmE mutant, we conclude that VfmE also activates these operons. This control leads to a positive autoactivation feedback loop responsible for the rapid accumulation of the extracellular signal. What remains to be determined is how the vfm regulatory pathway works together with the other systems to control virulence factor production in D. dadantii. We previously showed that the global regulator of virulence, PecS, represses expression of the vfm cluster (Hommais et al., 2008). Thus, overproduction of the Vfm signal in pecS mutant may explain the advanced induction of virulence factors observed in this mutant and the early development of disease symptoms (Mhedbi-Hajri et al., 2011). This assertion is supported by the stronger complementation of vfm mutations by the cell-free supernatant of the pecS mutant, an indication of a higher concentration of the Vfm signal compared with wild-type strain. In contrast to what is observed for the pecS strain, the vfm mutants are strongly disturbed in virulence. Some genes in the vfm cluster (vfmH and vfmQ) are also regulated by the HrpY or HrpL regulators of the secretion type III system (Yap et al., 2008; Yang et al., 2010), suggesting additional cross-talk between the regulatory networks controlling CWDE production and hrp expression. Furthermore, a recent transcriptomic analysis reveals that vfm genes are induced by antimicrobial peptides produced by plant in response to

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

874 W. Nasser et al.

Fig. 7. Model for the vfm function. The vfmAZBCD and vfmKLMNOPQRSTUVW operons are involved in the biosynthesis of the Vfm signal which is then exported to the external medium via the ABC-transporter system VfmF–VfmG. This signal accumulates in culture medium as cells grow to high density. When a threshold concentration of the signal is reached in the extracellular medium, it initiates a phosphorelay between VfmI and VfmH. The phosphorylated VfmH regulator then binds to vfmA and vfmE regulatory regions and activates the RNA polymerase. The VfmE regulator is then produced and activates in turn the genes encoding CWDE. VfmE also activates the vfmAZBCD, vfmFGHIJ and vfmKLMNOPQRSTUVW operons. This thereby increases the cellular levels of mRNA transcripts encoding Vfm signal synthesis functions and related enzymes production, leading to rapid amplification of the Vfm signal.

bacterial infection (Rio-Alvarez et al., 2012). This induction of vfm genes will allow the subsequent expression of virulence genes. Taken together, these data demonstrate that D. dadantii can intercept compounds released during host stress and integrate them into quorum sensing circuitry leading to enhanced virulence. A similar situation exists in the opportunistic human pathogen Pseudomonas aeruginosa, which activates its genetic virulence programme in response to dynorphin, a host stress produced opioid peptide, through its quorum sensing circuitry (Zaborina et al., 2007). Future research will concern the structural characterization of the signalling molecule synthesized by the vfm gene products. Based on the functional characteristics of the VfmO and VfmP proteins, which are predicted to activate alanine, it seems reasonable to speculate that the Vfm signal contains an alanine residue. Therefore, a

genomisotopic approach using isotopically labelled alanine (Gross et al., 2007) might be considered for the characterization of the Vfm signal. The Vfm signal characterization will help us to design structural antagonists and develop anti-quorum sensing strategies against Dickeya species. Recently, inhibitors of the LuxO quorum sensing response regulator of Vibrio cholera have been characterized (Ng et al., 2012). NtrC-type response regulators including LuxO and VfmH bind to s54-dependent promoters and hydrolyse ATP to activate transcription. The LuxO inhibitors are 6-thio-5-azauracil derivatives that bind to the central ATPase domain (AAA+ type) of the pre-formed LuxO–ATP complex to prevent ATP hydrolysis (Ng et al., 2012). It will be fascinating to investigate whether the thio-azauracil derivatives could also inhibit ATPase activity of VfmH. Antagonizing quorum sensing in Dickeya represents a promising approach to combat

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya Dickeya infection diseases and is of upmost importance because of the emergence in Europe of a new Dickeya pathogen, Dickeya solani, which is highly destructive and represents a severe threat to potato production (Toth et al., 2011). In conclusion, we have identified a new cluster of genes, termed vfm genes, which are involved in D. dadantii virulence. Our data indicate that these genes constitute a new quorum sensing system specific to Dickeya species. Thus, we identified the elusive missing link between quorum sensing and virulence in Dickeya.

Experimental procedures Bacterial strains, plasmids, phage, culture conditions and DNA manipulations Bacterial strains, plasmids, phage are described in Table S2. Dickeya dadantii and E. coli cells were grown at 30°C and 37°C, respectively, in Luria–Bertani (LB) medium or M63 minimal medium (Miller, 1972) supplemented with 0.2% carbon source, except for 0.4% polygalacturonate and, when required, with amino acids (40 mg ml-1) and antibiotics at the following concentrations: ampicillin (Amp), 100 mg ml-1; kanamycin (Km) and chloramphenicol (Cm), 50 mg ml-1; tetracycline (Tc) 20 mg ml-1; Streptomycin (Sm), 20 mg ml-1. Medium were solidified with 1.5% (w/v) Difco agar. Liquid cultures were grown in a shaking incubator (220 r.p.m.). DNA manipulations were performed using standard methods (Sambrook and Russell, 2001).

Agar plate detection tests for pectinases, cellulase, proteases and enzyme assays Detection of protease activity was performed on medium containing Skim Milk (12.5 g l-1) and detection of cellulase activity was performed using the Congo red procedure (Teather and Wood, 1982). Detection of pectinase activity was performed on medium containing PGA using the copper acetate procedure, as previously described by Reverchon and colleagues (1994). Assays of b-glucuronidase were performed on toluenized cell extracts. b-Glucuronidase activity was measured by following the degradation of p-nitrophenylb-D-glucuronide into p-nitrophenol that absorbs at 405 nm (Bardonnet and Blanco, 1992). Specific activity is expressed as nmol of products liberated per min per mg of bacterial dry weight. For growth in synthetic medium, bacterial concentration was estimated by measuring turbidity at 600 nm given that an OD600 of 1 correspond to 109 bacteria ml-1 and to 0.47 mg of bacterial dry weight ml-1.

Screening for D. dadantii mutants with reduced production of pectinases, cellulase and proteases Wild-type D. dadantii strain 3937 was mutated with transposon Tn5-B21, which carries a promoterless lacZ gene and a tetracycline resistance gene (Simon et al., 1989). This transposon is present on the mobilizable E. coli vector

875

pSUP102-Gm that does not replicate in D. dadantii. Equal volumes (0.2 ml) of a late logarithmic phase culture of E. coli S17-1 carrying pTn5-B21 and a late logarithmic phase culture of D. dadantii 3937 were mixed together onto a M63 agar plate without carbon source. After 5 h at 30°C, the cells were suspended in 500 ml of M63 medium. This suspension was diluted 100-fold and 100 ml aliquots of the dilution were spread onto M63 agar plates containing sucrose as carbon source and tetracycline, (E. coli S17-1 was counterselected by auxotrophy). After 48 h at 30°C, colonies were checked for pectinase, cellulase and protease activities using agar plate detection tests. Mutants with decreased production of all of these extracellular enzymes were collected, and tested for protease restoration by extracellular complementation using cell-free culture supernatant prepared from early stationaryphase cultures of the D. dadantii protease-deficient strain A3997. Supernatants from the signal-producing strain A3997 were incorporated into the agar medium at the final concentration of 10% v/v. Two mutants A3470 and A3473 that could be complemented by cell-free culture supernatant were retained for further analysis. To confirm that the observed phenotype resulted from transposon insertion, Tn5-B21 was transduced with phage FEC2 as previously described (Resibois et al., 1984).

Construction of vfm mutants Mutations in vfm genes were obtained by introducing uidA-Km cassettes, nptI-sacB-sacR (KmR) cartridge or Cm resistance cartridge into various restriction sites (Fig. 1A, Table S2). The uidA-Km cassettes (Bardonnet and Blanco, 1992) include a promoterless uidA gene that conserves its Shine Dalgarno sequence. Insertion of this cassette in a gene in the correct orientation generates a transcriptional fusion. In addition, insertion of this cassette can cause polar effects on downstream genes in an operon through transcription termination. The different insertions were introduced into the D. dadantii chromosome by marker exchange recombination between the chromosomal allele and the plasmid-borne mutated allele. The recombinants were selected after successive cultures in low phosphate medium in the presence of the suitable antibiotic, conditions in which pBR322 derivatives are very unstable (Roeder and Collmer, 1985). Correct recombination was confirmed by PCR. Phenotype of the resulting mutants was then analysed for pectinase (Pel), cellulase (Cel) and protease (Prt) activities using agar plate assays (Fig. 1B).

In vivo cloning of R⬘ plasmids harbouring the complete vfm region Plasmid pULB110, which is a RP4 derivative containing a Mu3A insertion was used to generate R-prime derivatives containing bacterial DNA (Van Gijsegem and Toussaint, 1982). Mating between recipient (HB101 SmR) and donor strain (D. dadantii expI::CmR indC::KmR/pULB110) was performed by spreading 0.2 ml of overnight cultures of the strains on M63 minimal medium plates and incubating for 5 h at 30°C. Bacteria were resuspended in 1 ml of M63 minimal medium and spread on the appropriate selective media (Sm,

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

876 W. Nasser et al. Cm, Km) to isolate pSR2895 (R′ plasmid carrying the expI::CmR- vfm-indC::KmR region).

Preparation of cell-free culture supernatants Cell-free culture supernatants were usually prepared from early stationary-phase cultures of various donor strains grown in M63 supplemented with 0.2% sucrose for D. dadantii strains (wild-type, prtE, vfmA, vfmE, vfmH and vfmK) and in M63 supplemented with 0.2% glucose as well as with leucine and proline for E. coli strains (HB101/pULB110 and HB101/ pSR2895 carrying the vfm region). Spent culture supernatants were sterilized using Millipore filters (0.22 mm pore size), and sterility was confirmed by testing aliquots for growth. For the kinetics of signal production, samples were collected throughout the growth and treated as above. Supernatants were conserved at -20°C.

RNA extraction, cDNA synthesis and quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis Dickeya dadantii cultures (10 ml) were taken from late exponential (10 h, OD600 = 1.8) growth in M63 sucrose minimal medium at 30°C. RNA was isolated by the frozen-phenol method (Maes and Messens, 1992) followed by Turbo DNase digestion (Ambion). RNA was quantified using a Nanodrop ND1000 (NanoDrop Technologies) and the quality evaluated using agarose gel electrophoresis. The cDNA was synthesized, using random hexamers and Fermentas reverse transcriptase, and qPCR was performed using the LightCyclerR faststart DNA masterplus SYBR Green I kit from Roche (Roche Applied Science). Minus RT controls were performed as a control for DNA contamination. Oligonucleotides used as primers for quantitative PCR are listed in Table S3. The statistical program used to analyse the data was the Relative Expression Software Tool (REST) (Pfaffl et al., 2002). The genes lpxC and hemF were selected as reference genes for real-time RT-PCR to provide an accurate normalization (Hommais et al., 2011). We confirmed that similar lpxC and hemF expressions were observed in the D. dadantii parental strain and in its various vfm derivatives in the different growth conditions used in this work.

unique restriction sites, so that the resulting fragment contained a SphI site at the ATG initiation codon and n BamHI site before the stop codon of vfmH. The resulting 1.4 kb SphI–BamHI restriction fragment PCR product was cloned into the pQE70 vector (Qiagen) to generate pSR3378 (Table S2). In the resulting plasmid pSR3378, the vfmH gene was placed under the control of the T5 promoter and fused to a His6tag on its C-terminus. Overproduction of the VfmH-His6 was carried out in E. coli M15/pREP4. Purification of the recombinant VfmH protein was achieved from cells grown at 30°C in LB medium containing ampicillin and kanamycin to maintain both pSR3378 and pREP4. When the optical density at 600 nm reached 0.3, cells were transferred at 20°C for 1 h before IPTG was added to a final concentration of 500 mM to induce VfmH synthesis. Then the cells were grown for an additional 2 h at 30°C. Cells were collected by centrifugation at 5000 g for 10 min and resuspended in an appropriate volume of lysis buffer (Qiagen). The bacteria were disrupted at 7000 kPa in a French press (Aminco) and crude extracts were subsequently centrifuged at 20 000 g, for 15 min, to remove the subcellular fractions. The supernatants obtained were used for purification. Protein purification was performed under native conditions at 4°C, according to the QIA expressionist handbook. Fractions containing the VfmH protein with more than 95% purity, as measured by SDSPAGE, were pooled and dialysed twice for 2 h against 2 l of desalting buffer (20 mM Tris-HCl pH 7.9, 1 mM EDTA, 1 mM DTT, 10% glycerol), then overnight against storage buffer (20 mM Tris-HCl pH 7.5, 500 mM NaCl, 0.1 mM EDTA, 0.2 mM DTT, 50% glycerol). The final preparation was stored at -20°C. Similarly, the coding region of vfmE was amplified by PCR using primers vfmE-30f and vfmE-30r (Table S3), containing unique restriction sites, so that the resulting fragment contained a BamHI site before the ATG initiation codon and a HindIII site after the stop codon of vfmE. The resulting 0.6 kb BamHI–HindIII restriction fragment PCR product was cloned into the pQE30 vector (Qiagen) to generate pSR3416 (Table S2). In the resulting plasmid pSR3416, the vfmE gene was placed under the control of the T5 promoter and fused to a His6tag on its N-terminus. Overproduction of the His6-VfmE was carried out in E. coli M15/pREP4 as previously described for VfmH excepted that the lysis buffer contained 500 mM NaCl.

In vitro DNA/protein interaction

Virulence assays Pathogenicity tests on 1-month-old potted S. ionantha cv. Blue Rhapsody cuttings were performed using 5 ¥ 106 bacteria as inoculum as reported by Lebeau and colleagues (2008). Plants were incubated in tropical conditions (day/ night temperature of 28°C/26°C; 16 h light; relative humidity of ⫾ 100%). Progression of the symptoms was scored daily for 6 days. Eighteen plants were tested for each bacterial strain. The assay was carried out in triplicate. Negative controls were performed using sterile buffer.

VfmH-His and VfmE-His purification The coding region of vfmH was amplified by PCR using primers vfmH-70f and vfmH-70r (Table S3), containing

The vfmA, vfmE, vfmF, vfmK regulatory regions were specifically amplified using the couple primers RRvfmAf -RRvfmAr, RRvfmEf-RRvfmEr, RRvfmFf-RRvfmFr, RRvfmKf-RRvfmKr respectively (Table S3). The resulting PCR fragments were cloned into the pGEM-T plasmid, using the TA cloning kit from Promega, to generate plasmids pSR3291, pSR3292, pSR3293, pSR3294 harbouring the vfmA, vfmE, vfmF, vfmK regulatory regions respectively (Table S2). These regulatory regions as well as those of pelE and celZ genes were recovered from plasmids pSR3291, pSR3292, pSR3293, pSR3294, pSR1175 and pWN2965, respectively, by a XmaI– EcoRI digestion for vfmA, vfmE, vfmF, vfmK, a SalI–SacII digestion for celZ, and an EcoRI–HindIII digestion for pelE. The DNA fragments obtained were further end-labelled by filling up the XmaI, SalI and HindIII extremities in the

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya 32

-1

presence of (a- P)dCTP (3000 Ci mmol , GE HealthCare) and the Klenow fragment of DNA polymerase. The labelled DNA fragments were purified, after electrophoresis, on agarose gel using the Qiagen gel extraction kit. Binding reactions were carried out in 20 ml of 10 mM Tris-HCl (pH 7.4), 50 mM KCl, 5 mM MgCl2, 1 mM dithiothreitol (DTT), 1 mM EDTA, 5% glycerol, 6 mg of bovine serum albumin, 1 mg of dI-dC DNA and 20 mM acetyl phosphate as a phosphodonor molecule for VfmH. 30 fmol of radiolabelled DNA probe were incubated with various amounts of purified VfmH-His6 for 30 min at 30°C prior to loading onto a 4% non-denaturing polyacrylamide gel (ratio of acrylamide to bisacrylamide, 80:1). Electrophoresis was carried out in 44 mM Tris-borate (pH 8)-0.5 mM EDTA.

Construction of strains for ChIP experiments The VSVG epitope, an 11-amino-acid sequence (YTDIEMNRLGK) from residues 501–511 of the Vesicular Stomatis Virus Glycoprotein, a viral transmembrane protein was fused to the N-terminus of VfmE. The VSVG-tagged vfmE allele was synthesized by the GENECUST company and cloned in pBluescript resulting in plasmid pSR3548 (Table S2). Dickeya dadantii A5461 strain, carrying chromosomic VSVGtagged vfmE allele was constructed by marker exchangeeviction procedure. Briefly, the D. dadantii A5460 strain, sucrose-sensitive and vfmE-deficient because it carries the nptI-sacB-sacR (KmR) cartridge into the chromosomic vfmE gene, was transformed with plasmid pSR3548 bearing the VSVG-tagged vfmE allele. Then, the tagged allele was exchanged for the chromosomal allele by selecting for sucrose tolerance and sensitivity to kanamycin. A correct recombination of the VSGV-tagged vfmE allele into the chromosome was checked by PCR using oVSVG and vfmE-30r primers (Table S3). This construction is functional because it generated a wild-type phenotype. A vfmA::uidA-Kan mutation was also introduced in the VSVG-tagged vfmE strain in order to analyse the DNA-binding activity of VfmE in the absence of Vfm signal synthesis. Production of the VSVG-VfmE protein was confirmed by Western blot analysis with an anti-VSVG antibody (Bethyl Laboratories).

Chromatin immunoprecipitation of VSVG epitope-tagged VfmE protein VSVG-VfmE bound DNA was isolated from D. dadantii using a protocol modified from Castang and colleagues (2008) (Castang et al., 2008). The wild-type strain and its VSVGvfmE derivative as well as the vfmA mutant and its vfmA,VSVG-vfmE derivative were grown until the late exponential phase (OD600 = 1.5) in M63 minimal medium supplemented with sucrose 0.2%. Proteins bound to DNA were cross-linked by addition of formaldehyde (at a final concentration of 1%, v/v) and chromatin isolated (Castang et al., 2008). VSVG epitope-tagged VfmE protein bound to chromatin was isolated using a VSVG immunoprecipitation Kit (Bethyl Laboratories) according to manufacturer’s instructions. Chromatin bound proteins were de-cross-linked to isolate only the DNA fragments for further analysis. Following DNA purification using a PCR purification Kit (Qiagen), qPCR were performed in the presence of 1 ml of immunoprecipitated

877

DNA samples to determine the relative amounts of certain DNA regions. Oligonucleotides flanking the promoter regions of vfmA, vfmE, vfmF, vfmK, pelE and celZ and, as a negative control, primers specific to the rpoA coding region were used in a standard qPCR reaction of 25 cycles. These oligonucleotides were described tin Table S3. Experiments were performed in triplicate.

Acknowledgements This work was supported by grants from the Centre National de la Recherche Scientifique (CNRS), and from French ‘ANR blanc Régupath 2007 Program, N°ANR-07-BLAN-0212’. The funders were not involved in the study design, data collection and analysis, decision to publish or preparation of the manuscript.

References Bardonnet, N., and Blanco, C. (1992) ′uidA-antibioticresistance cassettes for insertion mutagenesis, gene fusions and genetic constructions. FEMS Microbiol Lett 93: 243–248. Castang, S., McManus, H.R., Turner, K.H., and Dove, S.L. (2008) H-NS family members function coordinately in an opportunistic pathogen. Proc Natl Acad Sci USA 105: 18947–18952. Castillo, A., Nasser, W., Condemine, G., and Reverchon, S. (1998) The PecT repressor interacts with regulatory regions of pectate lyase genes in Erwinia chrysanthemi. Biochim Biophys Acta 1442: 148–160. Charkowski, A., Blanco, C., Condemine, G., Expert, D., Franza, T., Hayes, C., et al. (2012) The role of secretion systems and small molecules in soft-rot Enterobacteriaceae pathogenicity. Annu Rev Phytopathol 50: 425–450. Condemine, G., Castillo, A., Passeri, F., and Enard, C. (1999) The PecT repressor coregulates synthesis of exopolysaccharides and virulence factors in Erwinia chrysanthemi. Mol Plant Microbe Interact 12: 45–52. Fagard, M., Dellagi, A., Roux, C., Perino, C., Rigault, M., Boucher, V., et al. (2007) Arabidopsis thaliana expresses multiple lines of defense to counterattack Erwinia chrysanthemi. Mol Plant Microbe Interact 20: 794–805. Franza, T., Michaud-Soret, I., Piquerel, P., and Expert, D. (2002) Coupling of iron assimilation and pectinolysis in Erwinia chrysanthemi 3937. Mol Plant Microbe Interact 15: 1181–1191. Gross, H., Stockwell, V.O., Henkels, M.D., NowakThompson, B., Loper, J.E., and Gerwick, W.H. (2007) The genomisotopic approach: a systematic method to isolate products of orphan biosynthetic gene clusters. Chem Biol 14: 53–63. Haque, M.M., Kabir, M.S., Aini, L.Q., Hirata, H., and Tsuyumu, S. (2009) SlyA, a MarR family transcriptional regulator, is essential for virulence in Dickeya dadantii 3937. J Bacteriol 191: 5409–5418. Hommais, F., Oger-Desfeux, C., Van Gijsegem, F., Castang, S., Ligori, S., Expert, D., et al. (2008) PecS is a global regulator of the symptomatic phase in the phytopathogenic bacterium Erwinia chrysanthemi 3937. J Bacteriol 190: 7508–7522.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

878 W. Nasser et al. Hommais, F., Zghidi-Abouzid, O., Oger-Desfeux, C., PineauChapelle, E., Van Gijsegem, F., Nasser, W., and Reverchon, S. (2011) lpxC and yafS are the most suitable internal controls to normalize real time RT-qPCR expression in the phytopathogenic bacteria Dickeya dadantii. PLoS ONE 6: e20269. Kelly, W.L., Boyne, M.T., 2nd, Yeh, E., Vosburg, D.A., Galonic, D.P., Kelleher, N.L., and Walsh, C.T. (2007) Characterization of the aminocarboxycyclopropane-forming enzyme CmaC. Biochemistry 46: 359–368. Kepseu, W.D., Woafo, P., and Sepulchre, J.A. (2010a) Dynamics of the transition to pathogenicity in Erwinia chrysanthemi. J Biol Syst 18: 173–203. Kepseu, W.D., Sepulchre, J.A., Reverchon, S., and Nasser, W. (2010b) Toward a quantitative modeling of the synthesis of the pectate lyases, essential virulence factors in Dickeya dadantii. J Biol Chem 285: 28565–28576. Lautier, T., and Nasser, W. (2007) The DNA nucleoidassociated protein Fis co-ordinates the expression of the main virulence genes in the phytopathogenic bacterium Erwinia chrysanthemi. Mol Microbiol 66: 1474–1490. Lautier, T., Blot, N., Muskhelishvili, G., and Nasser, W. (2007) Integration of two essential virulence modulating signals at the Erwinia chrysanthemi pel gene promoters: a role for Fis in the growth-phase regulation. Mol Microbiol 66: 1491– 1505. Lebeau, A., Reverchon, S., Gaubert, S., Kraepiel, Y., SimondCote, E., Nasser, W., and Van Gijsegem, F. (2008) The GacA global regulator is required for the appropriate expression of Erwinia chrysanthemi 3937 pathogenicity genes during plant infection. Environ Microbiol 10: 545– 559. Letoffe, S., Delepelaire, P., and Wandersman, C. (1990) Protease secretion by Erwinia chrysanthemi: the specific secretion functions are analogous to those of Escherichia coli alpha-haemolysin. EMBO J 9: 1375–1382. Maes, M., and Messens, E. (1992) Phenol as grinding material in RNA preparations. Nucleic Acids Res 20: 4374. Mansfield, J., Genin, S., Magori, S., Citovky, V., Sriariyanum, M., Ronald, P., et al. (2012) Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol 13: 614–629. Mhedbi-Hajri, N., Malfatti, P., Pedron, J., Gaubert, S., Reverchon, S., and Van Gijsegem, F. (2011) PecS is an important player in the regulatory network governing the coordinated expression of virulence genes during the interaction between Dickeya dadantii 3937 and plants. Environ Microbiol 13: 2901–2914. Miller, J.H. (1972) Experiments in Molecular Genetics. New York: Cold Spring Harbor Laboratory Press. Murdoch, L., Corbel, J.C., Reis, D., Bertheau, Y., and Vian, B. (1999) Differential cell wall degradation by Erwinia chrysanthemi in petiole of Saintpaulia ionantha. Protoplasma 210: 54–74. Nasser, W., and Reverchon, S. (2002) H-NS-dependent activation of pectate lyases synthesis in the phytopathogenic bacterium Erwinia chrysanthemi is mediated by the PecT repressor. Mol Microbiol 43: 733–748. Nasser, W., Bouillant, M.L., Salmond, G., and Reverchon, S. (1998) Characterization of the Erwinia chrysanthemi expI-

expR locus directing the synthesis of two N-acylhomoserine lactone signal molecules. Mol Microbiol 29: 1391–1405. Ng, W.L., Perez, L., Cong, J., Semmelhack, M.F., and Bassler, B.L. (2012) Broad spectrum pro-quorum-sensing molecules as inhibitors of virulence in vibrios. PLoS Pathog 8: e1002767. Ouafa, Z.-A., Reverchon, S., Lautier, T., Muskhelishvili, G., and Nasser, W. (2012) The nucleoid-associated proteins H-NS and FIS modulate the DNA supercoiling response of the pel genes, the major virulence factors in the plant pathogen bacterium Dickeya dadantii. Nucleic Acids Res 40: 4306–4319. Perombelon, M.C.M. (2002) Potato diseases caused by soft rot erwinias: an overview of pathogenesis. Plant Pathol 51: 1–12. Pfaffl, M.W., Horgan, G.W., and Dempfle, L. (2002) Relative expression software tool (REST) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucleic Acids Res 30: e36. Py, B., Bortoli-German, I., Haiech, J., Chippaux, M., and Barras, F. (1991) Cellulase EGZ of Erwinia chrysanthemi: structural organization and importance of His98 and Glu133 residues for catalysis. Protein Eng 4: 325–333. Resibois, A., Colet, M., Faelen, M., Schoonejans, E., and Toussaint, A. (1984) Phi-Ec2, a new generalized transducing phage of Erwinia chrysanthemi. Virology 137: 102– 112. Reverchon, S., Nasser, W., and Robert-Baudouy, J. (1991) Characterization of kdgR, a gene of Erwinia chrysanthemi that regulates pectin degradation. Mol Microbiol 5: 2203– 2216. Reverchon, S., Nasser, W., and Robert-Baudouy, J. (1994) pecS: a locus controlling pectinase, cellulase and blue pigment production in Erwinia chrysanthemi. Mol Microbiol 11: 1127–1139. Reverchon, S., Expert, D., Robert-Baudouy, J., and Nasser, W. (1997) The cyclic AMP receptor protein is the main activator of pectinolysis genes in Erwinia chrysanthemi. J Bacteriol 179: 3500–3508. Reverchon, S., Bouillant, M.L., Salmond, G., and Nasser, W. (1998) Integration of the quorum-sensing system in the regulatory networks controlling virulence factor synthesis in Erwinia chrysanthemi. Mol Microbiol 29: 1407–1418. Reverchon, S., Rouanet, C., Expert, D., and Nasser, W. (2002) Characterization of indigoidine biosynthetic genes in Erwinia chrysanthemi and role of this blue pigment in pathogenicity. J Bacteriol 184: 654–665. Reverchon, S., Van Gijsegem, F., Effantin, G., ZghidiAbouzid, O., and Nasser, W. (2010) Systematic targeted mutagenesis of the MarR/SlyA family members of Dickeya dadantii 3937 reveals a role for MfbR in the modulation of virulence gene expression in response to acidic pH. Mol Microbiol 78: 1018–1037. Ried, J.L., and Collmer, A. (1987) An nptI-sacB-sacR cartridge for constructing directed, unmarked mutations in gram-negative bacteria by marker exchange-eviction mutagenesis. Gene 57: 239–246. Rio-Alvarez, I., Rodriguez-Herva, J.J., Cuartas-Lanza, R., Toth, I., Pritchard, L., Rodriguez-Palenzuela, P., and Lopez-Solanilla, E. (2012) Genome-wide analysis of the

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

New quorum sensing signal in Dickeya response of Dickeya dadantii 3937 to plant antimicrobial peptides. Mol Plant Microbe Interact 25: 523–533. Robert-Baudouy, J., Nasser, W., Condemine, G., Reverchon, S., Shevchik, V.E., and Hugouvieux-Cotte-Pattat, N. (2000) Pectic Enzymes of Erwinia chrysanthemi, Regulation and Role in Pathogenesis. St. Paul, MN, USA: The American Phytopathological Society. Roeder, D.L., and Collmer, A. (1985) Marker-exchange mutagenesis of a pectate lyase isozyme gene in Erwinia chrysanthemi. J Bacteriol 164: 51–56. Rottig, M., Medema, M.H., Blin, K., Weber, T., Rausch, C., and Kohlbacher, O. (2011) NRPSpredictor2 – a web server for predicting NRPS adenylation domain specificity. Nucleic Acids Res 39: W362–W367. Rozhon, W., Petutschnig, E., Khan, M., Summers, D.K., and Poppenberger, B. (2010) Frequency and diversity of small cryptic plasmids in the genus Rahnella. BMC Microbiol 10: 56. Sambrook, J., and Russell, D.W. (2001) Molecular Cloning: A Laboratory Manual. New York, USA: Cold Spring Harbor Laboratory Press. Sepulchre, J.A., Reverchon, S., and Nasser, W. (2007) Modeling the onset of virulence in a pectinolytic bacterium. J Theor Biol 244: 239–257. Simon, R., Quandt, J., and Klipp, W. (1989) New derivatives of transposon Tn5 suitable for mobilization of replicons, generation of operon fusions and induction of genes in gram-negative bacteria. Gene 80: 161–169. Teather, R.M., and Wood, P.J. (1982) Use of Congo redpolysaccharide interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl Environ Microbiol 43: 777–780. Toth, I., van der Wolf, J.M., Saddler, G., Lojkowska, E., Hélias, V., Pirhonen, M., et al. (2011) Dickeya species: an emerging problem for potato production in Europe. Plant Pathol 60: 385–399. Van Gijsegem, F., and Toussaint, A. (1982) Chromosome transfer and R-prime formation by an RP4:mini-Mu derivative in Escherichia coli, Salmonella typhimurium, Klebsiella pneumoniae, and Proteus mirabilis. Plasmid 7: 30–44. Yang, S., Peng, Q., Zhang, Q., Zou, L., Li, Y., Robert, C., et al. (2010) Genome-wide identification of HrpL-regulated genes in the necrotrophic phytopathogen Dickeya dadantii 3937. PLoS ONE 5: e13472. Yap, M.N., Yang, C.H., and Charkowski, A.O. (2008) The Response regulator HrpY of Dickeya dadantii 3937 regulates virulence genes not linked to the hrp cluster. Mol Plant Microbe Interact 21: 304–314. Zaborina, O., Lepine, F., Xiao, G., Valuckaite, V., Chen, Y., Li, T., et al. (2007) Dynorphin activates quorum sensing quinolone signaling in Pseudomonas aeruginosa. PLoS Pathog 3: e35.

Supporting information Additional Supporting Information may be found in the online version of this article: Table S1. Characteristics of the predicted gene products of the vfm cluster. Table S2. Bacterial strains, plasmids and phage.

879

Table S3. Oligonucleotides. Fig. S1. Identification of a new locus modulating extracellular enzymes production in D. dadantii. A mutagenesis of the D. dadantii wild-type 3937 strain was performed by using transposon Tn5-B21. Screening for extracellular enzymes production was made on agar plates allowing detection of pectinases, cellulases and proteases respectively. A. Comparison of enzymes production by the wild-type strain 3937 and the two mutants A3470 and A3473 is shown for each type of enzyme. B. Restoration of the protease- phenotype of mutants A3470 and A3473 by complementation using cell-free supernatant of the D. dadantii signal producing strain A3997 which is protease deficient due to a prtE mutation. Cell-free supernatant was incorporated into the agar medium at a final concentration of 10% v/v. Fig. S2. The vfm genes are required for high transcription of CWDE genes. Impact of the four vfm mutations (vfmA, vfmE, vfmH and vfmK) representative of the four transcriptional units of the vfm cluster on pelE, pelB, celZ and prtC transcript levels was monitored by qRT-PCR. The pelE and pelB genes encode pectate lyases (a family of pectinase); celZ gene encodes a cellulase and prtC gene encodes a protease. Expression of the pelE, pelB, celZ and prtC genes in the vfmA, vfmE, vfmH and vfmK mutants was measured as transcript levels relative to wild-type strain (WT) at early stationary growth phase from cultures in M63 minimal medium supplemented with sucrose and PGA. Gene expression was normalized using lpxC and hemF mRNAs as controls. Values represent average gene expression ⫾ SD from three independent experiments. Results obtained with the vfmA, vfmE, vfmH and vfmK mutants are different from the wild-type results, with P-values < 0.05 in a one sample t-test. Fold changes (FC) are expressed as the ratio of the specific gene expression in the studied mutant, compared with that in the wild-type strain. Negative FC values represent genes downregulated in the studied vfm mutant compared with the wildtype strain. Fig. S3. Transcriptional organization of the vfm locus. Transcription units were analysed by RT-PCR using the Access RT-PCR System kit from Promega. For each pair of primers, three reactions were performed. The first reaction (1) corresponds to the RT-PCR assay. The second reaction (2) corresponds to a negative control without reverse transcriptase. The third reaction (3) corresponds to a positive control using genomic DNA instead of cDNA. For the vfmKLMNOPQRSTUVW transcriptional unit, the amplification of two overlapping fragments A and B confirmed the presence of a long polycistronic RNA. No RT-PCR amplification product (fragment C) could be detected using the vfmW-ID16066 pair of primers, whereas a RT-PCR amplification product (fragment D) was observed using the pair of primers spanning ID16066. These results indicate that ID16066 is a monocistronic transcriptional unit. The lane M corresponds to DNA ladder. Fig. S4. Kinetic of Vfm signal production in D. dadantii 3937 wild-type strain and Vfm signal dose-dependent response of vfmE gene expression. A. Accumulation of the Vfm signal in culture supernatant from D. dadantii. Kinetic of the Vfm signal production was analysed by measuring the expression of the strongly signal-

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880

880 W. Nasser et al. sensitive vfmE::uidA reporter fusion from strain A5243. Cellfree supernatants from the D. dadantii wild-type strain and vfmA mutant grown in M63 minimal medium supplemented with sucrose 0.2% were collected throughout the growth curve. Bacterial growth was estimated by OD600 nm. The D. dadantii vfmE::uidA strain A5243 was grown overnight in M63 sucrose 0.2% medium in the presence of 10% (v/v) cell-free culture supernatants from each time point of the wild-type strain and vfmA mutant growth curve. Following A5243 strain growth, b-glucuronidase assays were

performed in triplicate. Values represent average of the vfmE gene expression ⫾ SD. B. Expression of vfmE gene is dependent of the Vfm signal concentration. Dickeya dadantii A5243 vfmE::uidA mutant was grown overnight in M63 sucrose 0.2% medium in the presence of various amounts of cell-free supernatant prepared from early stationary-phase culture of the 3937 wildtype strain. Following growth, b-glucuronidase assays were performed in triplicate. Values represent average of the vfmE gene expression ⫾ SD.

© 2012 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 15, 865–880