Advertisement
Research Article

The PlcR Virulence Regulon of Bacillus cereus

  • Michel Gohar mail,

    michel.gohar@ jouy.inra.fr

    Affiliations: INRA, Génétique microbienne et Environnement, Guyancourt, France, INRA, Microbiologie et Génétique Moléculaire, Thiverval-Grignon, France

    X
  • Karoline Faegri,

    Affiliation: Department of Pharmaceutical Biosciences, University of Oslo, Oslo, Norway

    X
  • Stéphane Perchat,

    Affiliation: INRA, Génétique microbienne et Environnement, Guyancourt, France

    X
  • Solveig Ravnum,

    Affiliation: Department of Pharmaceutical Biosciences, University of Oslo, Oslo, Norway

    X
  • Ole Andreas Økstad,

    Affiliation: Department of Pharmaceutical Biosciences, University of Oslo, Oslo, Norway

    X
  • Myriam Gominet,

    Affiliation: Institut Pasteur, Paris, France

    X
  • Anne-Brit Kolstø,

    Affiliation: Department of Pharmaceutical Biosciences, University of Oslo, Oslo, Norway

    X
  • Didier Lereclus

    Affiliation: INRA, Génétique microbienne et Environnement, Guyancourt, France

    X
  • Published: July 30, 2008
  • DOI: 10.1371/journal.pone.0002793

Abstract

PlcR is a Bacillus cereus transcriptional regulator, which activates gene expression by binding to a nucleotidic sequence called the ‘PlcR box’. To build a list of all genes included in the PlcR regulon, a consensus sequence was identified by directed mutagenesis. The reference strain ATCC14579 sequenced genome was searched for occurrences of this consensus sequence to produce a virtual regulon. PlcR control of these genes was confirmed by comparing gene expression in the reference strain and its isogenic Δ-plcR strain using DNA microarrays, lacZ fusions and proteomics methods. The resulting list included 45 genes controlled by 28 PlcR boxes. Forty of the PlcR controlled proteins were exported, of which 22 were secreted in the extracellular medium and 18 were bound or attached to cell wall structures (membrane or peptidoglycan layer). The functions of these proteins were related to food supply (phospholipases, proteases, toxins), cell protection (bacteriocins, toxins, transporters, cell wall biogenesis) and environment-sensing (two-component sensors, chemotaxis proteins, GGDEF family regulators). Four genes coded for cytoplasmic regulators. The PlcR regulon appears to integrate a large range of environmental signals, including food deprivation and self cell-density, and regulate the transcription of genes designed to overcome obstacles that hinder B. cereus growth within the host: food supply, host barriers, host immune defenses, and competition with other bacterial species. PlcR appears to be a key component in the efficient adaptation of B. cereus to its host environment.

Introduction

In pathogenic bacteria, the production of virulence factors is often coordinately regulated in response to changes in the bacterial cell environment, with various types of regulatory processes being employed. In Gram-positive bacteria, these processes may involve two-component systems [1], [2], alternative sigma factors [3] or stand-alone transcription regulators [4]. In some cases, the three regulatory mechanisms act together, each controlling a part in the production of virulence factors. This situation is found for instance in the nosocomial infection agent Staphylococcus aureus, in which more than 40 cell-surface or secreted proteins involved in bacterial virulence are controlled by a complex network involving the transcriptional regulator SarA, the two component regulator Agr and the general stress response regulator SigB [5]. In some species, a master regulator controls most of the virulence factors, which are therefore members of the same regulon. Virulence regulons may include a large number of genes: for example, the PrfA regulon of the food-borne pathogen Listeria monocytogenes includes 73 genes located on the chromosome [6]. Functional analysis of genes included in virulence regulons and a precise understanding of their regulation provide means to determine how environmental signals are integrated by virulence regulators and which strategies are used by bacterial cells to survive and develop within their host environment.

In Bacillus cereus, the transcriptional regulator PlcR (Phospholipase C Regulator) controls most known virulence factors [7]. B. cereus is a sporulating low-GC Gram-positive bacterium widely distributed in the environment and genetically close to two other pathogens: the human pathogen B. anthracis, which is the cause of anthrax and was implicated in the killing of five people in the US in the fall of 2001, and the insect pathogen B. thuringiensis. B. cereus is a food-poisoning pathogen frequently diagnosed as the causative agent of gastroenteritis [8] but it may also cause more severe diseases such as endophthalmitis [9] or meningitis [10]. PlcR controls the expression of several enterotoxins, haemolysins, phospholipases and proteases [7], [11]. PlcR has been shown to bind to DNA on a specific sequence called the ‘PlcR box’, located upstream from controlled genes, and at various distances ahead of the −35 box of the sigma A promoter [11], [12]. The transcription of plcR starts shortly before the onset of the stationary phase t0 and reaches a plateau two hours later (t2) [13]. plcR transcription is autoinduced [13], and is repressed by the sporulation factor Spo0A [14]. PlcR needs PapR to be active: this peptide is expressed as a propeptide under the control of PlcR, is exported out of the cell, is processed to form the active peptide either during export or in the extracellular medium, and is captured back by the cell through the oligopeptide permease system OppABCDF [12], [15], [16]. Thus, the three partners PlcR, OppABCDF and PapR function as a quorum-sensing system. Therefore, PlcR integrates at least two classes of signals: cell growth state through Spo0A and self cell density through PapR [12], [14].

Although several B. cereus genes have been demonstrated to be controlled by PlcR, no detailed study of the whole PlcR regulon has been undertaken until now. Moreover, several B. cereus group genomes have now been sequenced, presenting the possibility of building a virtual PlcR regulon by searching for matches with the PlcR box consensus sequence. Using this method, a virtual regulon was in fact proposed after the sequences of B. cereus ATCC14579 and other strains were published [17][19]. However the presence of a PlcR box is not sufficient to classify a gene as PlcR-regulated: experimental evidence is also required. In addition, PlcR may recognize sequences diverging from the previously defined consensus sequence, as has been reported in a study of metalloprotease gene inhA2 regulation [20]. In order to define all factors involved in the coordinated PlcR-based virulence response in B. cereus, we have undertaken an extensive study to map the complete PlcR regulon in the ATCC 14579 reference strain, utilizing mutagenesis experiments and in silico predictions in combination with proteomics and transcriptomics analyses.

Results

Directed mutagenesis defines a new PlcR box

A consensus sequence was previously determined by alignment of the promoter regions of 13 PlcR-controlled genes [11]. However, the PlcR box located upstream from inhA2, a gene known to be under PlcR control, diverged from this original consensus sequence by one base. Therefore, we investigated which substitution could be introduced in the consensus sequence while still maintaining significant PlcR-dependent activity. Thus, we mutated half of the palindromic nucleotide sequence of the PlcR box located upstream from plcA, a PlcR-regulated gene coding for a phosphatidyl-inositol specific phospholipase C (PI-PLC) and used to report PlcR activity [12]. The recognition of the mutated PlcR boxes by PlcR was investigated in the ATCC14579 strain by a transcriptional fusion between the modified promoter region of plcA and lacZ, carried on the pHT304-18Z plasmid [21]. The results, displayed in Table 1, showed that the first and the last nucleotide of the 16 bp consensus sequence can be replaced by any base without a greater than five-fold drop in expression of plcA-lacZ. However, deletion of A or T or their replacement by C or G in position 7–8 in the middle of the sequence, or a replacement of C by G in position 5 of the sequence, leads to a dramatic loss of activity. Similarly A2, T3 or G4 could not be replaced by another base. Therefore, the PlcR target sequence identified in the mutagenesis experiment was ‘ATGhAwwwwTdCAT’, were h, w and d stand for, respectively: C, A or T; A or T; G, A or T. In addition to the previous consensus sequence ‘TATGnAnnnnTnCATA’, this sequence was retained for the subsequent in silico analysis step.

thumbnail

Table 1. Effect of base mutation on the plcA PlcR box activity.

doi:10.1371/journal.pone.0002793.t001

Searching for PlcR boxes in the sequenced B. cereus genome

We searched for the two PlcR target sequences TATGnAnnnnTnCATA and ATGhAwwwwTdCAT in the B. cereus ATCC14579 genome sequence. We identified a total of 69 boxes located at least 35 bp, but not more than 700 bp, upstream from a putative coding sequence. These boxes may control as many as 138 genes, as the same box could act on several genes putatively organized into an operon and/or that were divergently transcribed (see supplementary material, Table S1). Included in this list are genes that have not been annotated in the published genome sequence of the ATCC14579 strain (Ivanova et al., 2003; http://www.ncbi.nlm.nih.gov/genomes/lpro​ks.cgiview1), but for which expression was confirmed by proteomic or genetic means (Bc0361a, this study; BC3763, [7]; Bc2463a, Bc3185a, Bc5101a, [22]).

Microarray analysis

We used microarrays to determine the ratio of expression between the wild type ATCC14579 strain and the isogenic ΔplcR strain for the 138 genes identified in the in silico procedure. For this determination, we chose two time points in the growth curve: the onset of the stationary phase (t0), because PlcR expression increases sharply at this point, and two hours later (t2), after PlcR expression reaches a plateau. Most of the genes, which on the basis of genome annotation were expected to be transcribed as part of an operon structure, displayed similar expression ratio values. Only genes with a relative expression ratio greater than 2.5 at t0 or at t2, and a significance value (p) smaller than 0.2, were considered for subsequent analysis. Consequently, 75 genes (Figure 1, blue box) were considered as not controlled by PlcR under standard culture conditions and were discarded. No microarray data were available for 25 genes. For most of the 38 remaining genes, transcription was enhanced by PlcR both at t0 and at t2 (Figure 1). Noticeably, the clO hemolysin, the cytK cytotoxin and the enterotoxins hblC, hbLD, hblA and nheA, nheB, nheC, were the genes most strongly induced by PlcR with relative expression ratios of 10 to 50 in the ΔplcR mutant versus wildtype cells. The expression of genes coding for other secreted proteins, including proteases (sfp, nprB, nprC, mpbE, colA, and colC) and phospholipases (plcA, plcB and smase) was also induced by PlcR, with a ratio of induction ranging from 3 to 30 (see supplementary Table S1). The expression of a high number of genes coding for cell-surface proteins appeared also to be induced by PlcR, although at a lower level than for secreted proteins. By contrast, only a few genes coding for cytosolic proteins, including four transcriptional regulators, had their expression significantly induced by PlcR. Data relating to plcR and papR were discarded, because insertion of the KmR cassette into the plcR gene introduced a promoter upstream from the regions recognized by the microarray 70-mer oligos. These genes have however previously been shown to be PlcR-regulated [11], [13]. Finally, 6 genes appeared to be repressed by PlcR at a ratio of 2 to 6 in this microarray analysis, either at t0 (BC4986) or at t2 (BC0069, BC1736, BC3520, BC4982, BC4983).

thumbnail

Figure 1. ΔplcR-wt expression ratios as determined by microarray experiments.

Ratios of expression between the wildtype strain and the delta plcR strain as determined by microarrays. The log2 of these ratios were plotted at t2 vs t0. Each red circle represents the values obtained for one gene. Inside the blue square at the center of the figure are the genes for which the expression ratios were equal or less than 2. Genes for which the transcription was induced by PlcR both at t0 and at t2 are in the yellow square ‘a’, whereas genes induced only at t2 or only at t0 are in the yellow squares ‘b’ or ‘d’, respectively. Genes repressed by PlcR both at t0 and at t2 are in the yellow square ‘c’.

doi:10.1371/journal.pone.0002793.g001

Transcriptional fusions

Results from the microarray analysis were then crossed with data from previous proteomic or genetic analysis [7], [11][13], [22][24]. All genes encoding secreted proteins and identified by DNA microarray analysis were confirmed as belonging to the PlcR regulon, including BC2463a, BC3185a and BC5101a for which microarray results could not be produced, except for colC, for which no data from previous reports was available. Some cell surface and cytosolic proteins (inhA2, prp2 and plcR) were also confirmed as part of the PlcR regulon. However, for 39 genes, some of which were in operon, no previous results were available regarding their control by PlcR. Therefore, we constructed 30 transcriptional fusions between the genes promoter region including the PlcR box and a lacZ gene, to determine if they were truly controlled by PlcR. Two genes that were not previously predicted to be in the ATCC14579 strain genome sequence, but which were located downstream from PlcR boxes, were included in the analysis: cwh (BC3763) and a small open reading frame located downstream from BC0361, which we named BC0361a. The ratios of β-galactosidase (encoded by lacZ) activity between the wild type strain and the ΔplcR strain was plotted at t0 vs t2 (Figure 2; the kinetics of expression obtained between t−1 and t4 are shown in supplementary, Figure S1). Among the 30 promoter regions assayed by lacZ fusions, 14 were controlled by PlcR whereas the remaining 16 were not. Microarray results were missing or gave low ratio values for these 16 PlcR-independent genes.

thumbnail

Figure 2. ΔplcR-wt expression ratios as determined by lacZ fusions.

Ratios of expression between the wildtype strain and the delta plcR strain as determined by lacZ fusions. The log2 of these ratios were plotted at t2 vs t0. Each red circle represents the values obtained for one gene. Inside the blue square at the center of the figure are the genes for which the expression ratios were equal or less than 2. The transcription of all the other genes was induced by PlcR both at t0 and at t2 (yellow square ‘a’).

doi:10.1371/journal.pone.0002793.g002

Proteomic study

We have observed in a previous two-dimensional protein gel electrophoresis analysis of the B. cereus ATCC14579 secretome at t2 that most of the extracellular proteins disappeared upon inactivation of plcR [7]. However, in the same time a large number of spots appeared on the gel obtained from the mutant strain. This suggests that PlcR is possibly a repressor for some extracellular proteins. We identified the protein content of 103 of these spots by peptide mass fingerprints and by N-terminal sequencing (see supplementary material, Figure S2 and Table S2), and only one of the proteins was encoded by a gene preceded by a PlcR box. This protein, the fructose bisphosphate aldolase FbaA, was shown by lacZ fusion not to be controlled by PlcR (Figure S1). Therefore no repressor role for PlcR acting on genes coding for secreted proteins was identified. The appearance of cytosolic proteins in the culture supernatant if plcR is inactivated was due to a greater cell lysis in the mutant strain than in wildtype: this lysis, determined by measuring isocitrate dehydrogenase activity in the bacterial cells and culture supernatant, was 1% in the wildtype strain and 15% in the mutant strain at t2.

Final list of the PlcR-controlled genes

We built a list of PlcR-controlled genes using the data generated. We added three genes coding for antibacterial peptides to this list, which were previously shown to be controlled by PlcR: sppc1, sppc2 and sppc3 [22]. The final list included 45 genes controlled by 28 PlcR boxes, as the same PlcR box may control several genes (Table 2 and Figure S4). Genes coding for secreted proteins made up 49% of the regulon, whereas genes coding for proteins associated to the membrane or to the peptidoglycan (cell wall proteins) represented 40%. The 22 secreted proteins were toxins, phospholipases, proteases, peptides with antibacterial activity and included one cell-cell communication peptide (PapR); however, the 18 cell wall proteins were annotated as being involved in cell immunity, drug efflux transport, cell wall biogenesis, and environment-sensing (in connection with regulation systems). Environmental sensors included two chemotaxis proteins, McpA and TlpA, the two-component system sensor YufM and one GGDEF family protein. The GGDEF family protein displays three conserved domains: a dinucleotide cyclase and a phosphodiesterase domain, involved in regulating the intracellular level of cyclic dinucleotide diguanylate, a second messenger, in response to ligands detected by the PAS domain [25]. One protein, InhA2, was possibly involved in cell immunity. InhA2 is a member of the Immune Inhibitor A metalloprotease family, previously shown to specifically degrade antibacterial peptides [26] and involved in bacterial virulence [20]. The cytosolic proteins controlled by PlcR were PlcR itself, a TetR family regulator, a two-component response regulator, a protein of unknown function and a protein homologous to the RimL ribosomal alanine acetyl transferase. Therefore, all cytoplasmic PlcR-controlled proteins of known function are likely to be regulators. The TetR family regulators are transcriptional repressors involved in the biosynthesis of antibiotic efflux pumps and the response to osmotic stress [27]. The two-component response regulator YufM is in an operon with its sensor component YufL and a chemotaxis transducer protein McpA, all under PlcR transcriptional control. RimL belongs to the GNAT superfamily of acetyltransferases [28] and acetylates a ribosomal protein interacting with elongation factors EF-Tu and EF-G [29], [30].

thumbnail

Table 2. List of the PlcR-controlled genes in the ATCC14579 strain.

doi:10.1371/journal.pone.0002793.t002

Analysis of the nucleotidic sequences of the active PlcR boxes

The 28 active PlcR boxes that we determined were scattered all along the chromosome (Figure S3, supplementary material). Consequently, no pathogenicity island could be found in the B. cereus chromosome. Alignment of the active PlcR boxes led us to a new consensus sequence, shown as a logo in Figure 3. To investigate whether nucleotide sequences surrounding active PlcR boxes could exhibit additional properties required for the box to be recognized by PlcR, a comparison of sequences upstream or downstream from the active and inactive PlcR boxes was performed. We found that in the vicinity of the active PlcR boxes, the AT-content was much higher than in the vicinity of inactive boxes (Figure 4). Downstream from all the active PlcR boxes, we identified a putative −10 σA binding sequence (Figure S5), suggesting that the PlcR-regulated genes may be transcribed by a σA-associated RNA polymerase.

thumbnail

Figure 3. PlcR consensus sequence.

The height of the letter representing a base is proportional to its frequency at each position in the alignment. For each position, the most frequent base is drawn in blue, followed by green and pink for less frequent bases.

doi:10.1371/journal.pone.0002793.g003
thumbnail

Figure 4. Percentage of A+T in the vicinity of PlcR boxes.

Active boxes are plotted in blue whereas inactive boxes are plotted in red. The dashed line represents the average A+T percentage for the ATCC14579 chromosome. The difference between active and inactive boxes for the A+T percentage is highly significant (Qui-square test, p<0.001).

doi:10.1371/journal.pone.0002793.g004

Discussion

Building a list of PlcR-controlled genes

In 1999, Agaisse and colleagues used a genetic screen to identify PlcR-regulated genes. They reported 13 genes encoding exported proteins, mostly toxins and degradative enzymes [11]. As a consequence, PlcR appeared to be a pleiotropic virulence regulator controlling extracellular factors. That study also led to the definition of a PlcR target sequence, to which the active complex PlcR/PapR binds [12]. Later, the sequencing of B. cereus genomes in combination with proteomic and genetic studies revealed that PlcR may control a much higher number of genes, all of which were not, at least not directly, involved in virulence [7], [17], [24], [31]. Indeed, the role of PlcR in virulence has been extensively documented [31][34]. Various studies have also suggested that PlcR could acts on other functions, including sporulation [35] and biofilm formation [36]. We therefore systematically investigated the PlcR regulon to understand better the role of PlcR during bacterial infection, and have provided the first comprehensive, genome-wide characterization of the complete PlcR virulence regulon based on functional experiments. A virtual PlcR regulon was constructed in silico using the PlcR DNA target sequence defined through mutagenesis experiments, and was investigated by transcriptional studies using DNA microarrays and lacZ fusions. The resulting data were cross-analyzed with data from proteomic studies, to build a list of 45 genes positively controlled by PlcR under standard culture conditions. The genes were scattered along the chromosome, and did not form a pathogenicity island. Aligning the sequences of the PlcR boxes located upstream from these genes led to the identification of the PlcR consensus binding sequence wTATGnAwwwwTnCATAw.

Inactive PlcR boxes

A high number of PlcR boxes turned out to be inactive under our culture conditions. Alignment of the sequences located upstream and downstream from the PlcR boxes revealed that, for active boxes, these sequences are significantly more AT-rich. Therefore, in addition to the consensus sequence, the genetic environment of the PlcR box could be critical for the binding of PlcR to its box, and/or could be important for the transcription activity of the promoter. Also, we found cases in which a PlcR box is placed between two divergently transcribed genes, and PlcR controls the transcription of only one of these genes (for example, Bc0555/Bc0556). Thus, the binding of PlcR to its box is required but not sufficient to activate the transcription of genes located downstream. A putative σA -10 region consensus sequence was found downstream from all the active PlcR boxes. In various conditions, including the host, the transcription of genes not controlled by PlcR despite the binding of the regulator to the PlcR box may require alternative sigma factors, such as σB, σH, or ECF factors. In B. anthracis [37], L. monocytogenes [38], [39] and S. aureus [5], σB has been shown to be involved in virulence under some culture conditions. Similarly, σH is required for toxin gene expression in B. anthracis [40]. These two sigma factors are likely to be expressed in early stationary phase under standard culture conditions [37], [40]. Finally, it seems unlikely that we may have missed any PlcR-regulated genes in this study due to no expression in LB medium because no additional genes were identified in a screen for genes specifically induced during growth in vivo [31].

Role of the PlcR regulon

Ninety percent of the genes included in the final list of PlcR-controlled genes encode proteins either secreted or located at the cell wall, i.e. at the interface between the bacterial cell and its environment – including the eukaryotic host. Proteases and phospholipases, in addition to enterotoxins and hemolysins, have been found located at this interface. These enzymes are likely to be involved in host tissue degradation. Phosphatidylcholine-specific phospholipase C (PC-PLC) and sphingomyelinase were previously shown to induce hemolysis [41], [42] and the InhA2 metalloprotease is involved in protecting the bacterial cell from host immune defenses [26]. Proteins involved in peptidoglycan synthesis and modification (four genes) are also likely to be involved in bacterial cell protection by strengthening the cell wall, as suggested by the significantly greater tendency of cell lysis observed for the plcR mutant strain than the wildtype strain (this study). Furthermore, three secreted antibacterial peptides and four drug efflux transporters shown to be controlled by PlcR for the first time here may protect the cell from competition with other bacterial species and their bacteriocins.

Thus, these functions may work together to provide nutrition and bacterial cell protection in a hostile host environment (Figure 5). The bacterium may feed on host tissues by producing toxins, phospholipases and proteases. Proteins, peptides and amino acids have been suggested as the preferred nutrient sources for B. cereus [17], possibly linked to the growth of the bacterium as a human and animal pathogen. Meanwhile, other functions of the regulon may inhibit the growth of other bacterial species in the same niche, inactivate host antibacterial peptides, and increase cell wall resistance to lysis.

thumbnail

Figure 5. Overview of the PlcR regulon organization.

PlcR positively controls (dark line) the transcription of a vast array of genes coding for proteins located in the cell wall or in the extracellular space. Cell wall proteins are designed in green. Secreted proteins are exported through the SecA machinery designed in pink. Environmental signals are sensed by cell-wall sensors and act via (dashed line) regulators on undetermined genes or proteins. PlcR requires PapR to be active (dashed line). Signals integrated by PlcR and PlcR-controlled regulators are designed in red.

doi:10.1371/journal.pone.0002793.g005

Sensing the host environment

Sensing the surrounding environment is necessary for a bacterium to react appropriately to changes. Bacterial pathogens often use two-component systems to sense their host environment, and promote or repress the transcription of genes in response to changes in this environment [1], [2]. Interestingly, as shown here for the first time, four sensors are included in the B. cereus PlcR virulence regulon, only one of which (YufL) is part of a two-component system. The other sensors are chemotaxis proteins (McpA and TlpA) or a GGDEF-family regulator producing a second messenger. This variety in sensor types is likely to reflect a variety in the types of signals providing input to the cell. Furthermore, genes controlled by PlcR-dependent transcriptional regulators could add to the list of genes controlled by PlcR, and extend the regulon size. However, these regulators could also recruit genes already controlled by PlcR. If so, PlcR-controlled sensors and their regulators could modulate the transcription of subgroups of PlcR-controlled genes. It was recently suggested that regulators other than PlcR could act on the expression of PlcR-controlled genes [43], [44]. Accordingly, L. monocytogenes internalins, which constitute a subgroup of the PrfA virulence regulon, are simultaneously controlled by σB, itself acting on gene transcription in response to stress signals [39]. These observations have led us to propose the following hypothesis for the integration of PlcR/PapR-related environmental signals in B. cereus: PlcR triggers the transcription of its regulon, including the sensor proteins, in response to food deprivation sensed via transition state regulators and in response to its own cell density sensed via PapR. In turn, the PlcR-controlled sensors repress (or promote) transcription of subgroups of genes in response to host signals. The PlcR/PapR quorum-sensing system thus provides an efficient way to integrate several environmental signals and produce a gene expression profile continuously adapted to a changing host environment, such as that experienced by the bacterium during infection.

Materials and Methods

Strains and culture conditions

The B. cereus strains used in this study were the type strain, ATCC14579, and the isogenic ΔplcR strain [32], obtained by insertion of a KmR cassette in plcR. The two strains were grown in Luria Bertani broth (LB) at 30°C. Cultures were harvested at the onset of the stationary phase (t0) or two hours later (t2). The onset of the stationary phase (t0) was defined as the breakpoint in the vegetative phase slope.

Directed mutagenesis of the PlcR box

Point mutations were introduced into the PlcR box of the plcA promoter region by PCR amplification with primer Bc-plc matching the 5′ end of the plcA gene, and primers pRX1 to pRX18 carrying a modified PlcR box (supplementary Table S3). Each PCR product (a 390-bp DNA fragment) was digested with XbaI and HindIII enzymes and cloned between the XbaI and HindIII sites in pHT304-18Z [21]. The nucleotide sequence of each DNA fragment was determined and analyzed by Genome express (France) by using oligonucleotides UP and OVG flanking the DNA fragments cloned into pHT304-18Z. Plasmids carrying the plcA'-lacZ transcriptional fusions were introduced into B. cereus ATCC14579 by electroporation, and ß-galactosidase activity produced by the recombinant clones was measured two hours after entry in the stationary phase in LB medium.

Microarray analysis

Harvested bacterial cells were incubated in an equal volume of ice-cold methanol for 5 minutes before centrifugation at 4°C and 4000 rpm. RNA isolation was performed with the RNeasy Midi Kit (Qiagen, Germany) together with the RNase-Free DNase Set (Qiagen, Germany). For microarray slide preparation, 70-mer oligos from the whole genomic B. cereus ATCC14579 ORFs (released at NCBI in 2003) were designed and synthesized by Qiagen-Operon (Germany). The oligos were printed in 50% DMSO on UltraGAPS™ gamma amino silane-coated slides from Corning (USA), at the Norwegian Radiumhospital (DNR). The microarray slides were prehybridized before use for 30–60 minutes in a 5×SSC/0.1 % SDS/0.1 % BSA solution at 42°C, according to the UltraGAPS™ Coated Slides instruction manual from Corning. The slides were then washed three times in MQ H2O, once in isopropanol and finally spun dry.

cDNA synthesis, labeling and purification was carried out with the FairPlay™ microarray labeling kit (Stratagene, CA, USA), using 500 ng random hexamers (Applied Biosystems, CA, USA) on 20 μg of RNA, and with amino-allyl coupling of Cy3 and Cy5 dyes from Amersham Biosciences (GE Healthcare Bio-Sciences AB, Sweden). After purification, the samples were concentrated with a Microcon column (Millipore, MA, USA) and hybridization solution was added to a final concentration of 30 % formamide, 5×SSC, 0.1 % SDS and 0.1 mg/mL sperm DNA, based on the UltraGAPS™ Coated Slides instruction manual from Corning (USA). Labeled DNA were denatured at 95°C for 2 minutes, and incubated at 42°C before hybridization. The samples were hybridized in a hybridization chamber (Monterey Industries, CA, USA), humidified with 5×SSC for 16 hours in a 42°C water bath. After hybridization, the slides were washed at 42°C in 0.5×SSC/0.01 % SDS and in 0.06×SSC, and finally at room temperature in isopropanol before they were spun dry.

The slides were scanned with an Axon 4000B scanner. Gridding, spot annotation and calculation was carried out using GenePix Pro 6.0 software. The R platform [45] and LIMMA [46], [47] were used for filtering, normalization and further analysis (for details, see supplementary material). P-values were computed using false discovery rate correction of 0.05.

Transcriptional fusions

Transcriptional fusions were constructed in the pHT304-18Z plasmid, between the XbaI and PstI or HindIII and BamHI cloning sites of the plasmid [48]. Primers used for PCR-amplification of the promoter regions cloned are listed in supplementary Table S4. The resulting plasmids were then transferred into B. cereus strains ATCC14579 or ATCC14579 ΔplcR by electroporation. For β-galactosidase activity measurement, bacterial cells were lysed using the FastPrep 120 system (Savant), and β-galactosidase-specific activities were measured as described previously [49]. The specific activities are expressed in units of β-galactosidase milligram−1 of protein (Miller units). Two to four assays were performed for each transcriptional fusion.

Two-dimensional electrophoresis

Protein extracts were prepared from the culture supernatants and subjected to two-dimensional electrophoresis as described earlier [7]. The spots were immediately excised and stored at −70°C until use. Proteins were identified by peptide mass fingerprint and by N-terminal sequencing. Peptide mass fingerprints were generated after trypsin-digestion and MALDI-TOF analysis (Biobac, INRA, Jouy-en-Josas, France), and proteins were identified using ProteinProspector or Mascot programs. N-terminal sequencing was performed by Prof. K. Sletten at the Biotechnology Center, University of Oslo, Norway.

Sequence analysis

PlcR boxes were searched in the sequenced genome of the ATCC14579 strain using the ‘find sequence’ tool of Vector NTI (Invitrogen). The same method was used to find σA −10 boxes in the promoter regions of PlcR-controlled genes. The consensus sequence for PlcR binding (‘PlcR box’) was drawn as a logo where, at each nucleotide position, the letter height is proportional to the frequency of the base and to the weight of the position in the sequence [50]. The content in A+T bases in sequences upstream and downstream from the inactive and active PlcR boxes were compared using a chi-square test.

Supporting Information

Table S1.

Microarray results for genes with a PlcR box in their promoter region

doi:10.1371/journal.pone.0002793.s001

(0.26 MB PDF)

Table S2.

Proteins identified in the culture supernatant of the delta-PlcR ATCC14579 strain harvested at t2

doi:10.1371/journal.pone.0002793.s002

(0.19 MB PDF)

Table S3.

Primers used for the directed mutagenesis of the PlcR box

doi:10.1371/journal.pone.0002793.s003

(0.04 MB PDF)

Table S4.

Primers for transcriptional fusions

doi:10.1371/journal.pone.0002793.s004

(0.05 MB PDF)

Figure S1.

Results from lacZ fusions

doi:10.1371/journal.pone.0002793.s005

(0.13 MB PDF)

Figure S2.

Two-dimensional gel electrophoresis of the Δ-plcR ATCC14579 supernatant

doi:10.1371/journal.pone.0002793.s006

(0.97 MB PDF)

Figure S3.

Location of PlcR boxes on the ATCC14579 chromosome

doi:10.1371/journal.pone.0002793.s007

(0.02 MB PDF)

Figure S4.

Genetic environment of the 45 PlcR-regulated genes

doi:10.1371/journal.pone.0002793.s008

(0.26 MB PDF)

Figure S5.

Putative −10 σA boxes located downstream of PlcR boxes for PlcR-controlled genes

doi:10.1371/journal.pone.0002793.s009

(0.60 MB PDF)

Acknowledgments

We wish to thank Professor Knut Sletten (Biotechnology Centre, University of Oslo, Norway) for N-terminal sequencing, Dr. Nicolas J. Tourasse (Department of Pharmaceutical Biosciences, University of Oslo, Norway) for microarray bioinformatics support, and Jean-Claude Huet and Céline Henry (BioBac, INRA, France) for mass spectrometry.

Author Contributions

Conceived and designed the experiments: MG ABK DL. Performed the experiments: MG KF SP MG. Analyzed the data: MG KF SP SR OA MG ABK DL. Contributed reagents/materials/analysis tools: KF. Wrote the paper: MG.

References

  1. 1. Standish AJ, Stroeher UH, Paton JC (2007) Pneumococcal Two-Component Signal Transduction System RR/HK06 Regulates CbpA and PspA by Two Distinct Mechanisms. J Bacteriol 189: 5591–5600.
  2. 2. Lamy MC, Zouine M, Fert J, Vergassola M, Couve E, et al. (2004) CovS/CovR of group B streptococcus: a two-component global regulatory system involved in virulence. Mol Microbiol 54: 1250–1268.
  3. 3. Matamouros S, England P, Dupuy B (2007) Clostridium difficile toxin expression is inhibited by the novel regulator TcdC. Mol Microbiol 64: 1274–1288.
  4. 4. Hamon M, Bierne H, Cossart P (2006) Listeria monocytogenes: a multifaceted model. Nat Rev Microbiol 4: 423–434.
  5. 5. Novick RP (2003) Autoinduction and signal transduction in the regulation of staphylococcal virulence. Mol Microbiol 48: 1429–1449.
  6. 6. Milohanic E, Glaser P, Coppee JY, Frangeul L, Vega Y, et al. (2003) Transcriptome analysis of Listeria monocytogenes identifies three groups of genes differently regulated by PrfA. Mol Microbiol 47: 1613–1625.
  7. 7. Gohar M, Økstad OA, Gilois N, Sanchis V, Kolstø AB, et al. (2002) Two-dimensional electrophoresis analysis of the extracellular proteome of Bacillus cereus reveals the importance of the PlcR regulon. Proteomics 2: 784–791.
  8. 8. Kotiranta A, Lounatmaa K, Haapasalo M (2000) Epidemiology and pathogenesis of Bacillus cereus infections. Microbes Infect 2: 189–198.
  9. 9. Callegan MC, Gilmore MS, Gregory M, Ramadan RT, Wiskur BJ, et al. (2007) Bacterial endophthalmitis: therapeutic challenges and host-pathogen interactions. Prog Retin Eye Res 26: 189–203.
  10. 10. Gaur AH, Patrick CC, McCullers JA, Flynn PM, Pearson TA, et al. (2001) Bacillus cereus bacteremia and meningitis in immunocompromised children. Clin Infect Dis 32: 1456–1462.
  11. 11. Agaisse H, Gominet M, Økstad OA, Kolstø AB, Lereclus D (1999) PlcR is a pleiotropic regulator of extracellular virulence factor gene expression in Bacillus thuringiensis. Mol Microbiol 32: 1043–1053.
  12. 12. Slamti L, Lereclus D (2002) A cell-cell signaling peptide activates the PlcR virulence regulon in bacteria of the Bacillus cereus group. EMBO J 21: 4550–4559.
  13. 13. Lereclus D, Agaisse H, Gominet M, Salamitou S, Sanchis V (1996) Identification of a Bacillus thuringiensis gene that positively regulates transcription of the phosphatidylinositol-specific phospholipase C gene at the onset of the stationary phase. J Bacteriol 178: 2749–2756.
  14. 14. Lereclus D, Agaisse H, Grandvalet C, Salamitou S, Gominet M (2000) Regulation of toxin and virulence gene transcription in Bacillus thuringiensis. Int J Med Microbiol 290: 295–299.
  15. 15. Gominet M, Slamti L, Gilois N, Rose M, Lereclus D (2001) Oligopeptide permease is required for expression of the Bacillus thuringiensis plcR regulon and for virulence. Mol Microbiol 40: 963–975.
  16. 16. Declerck N, Bouillaut L, Chaix D, Rugani N, Slamti L, et al. (2007) Structure of PlcR: Insights into virulence regulation and evolution of quorum sensing in Gram-positive bacteria. Proc Natl Acad Sci U S A 104: 18490–18495.
  17. 17. Ivanova N, Sorokin A, Anderson I, Galleron N, Candelon B, et al. (2003) Genome sequence of Bacillus cereus and comparative analysis with Bacillus anthracis. Nature 423: 87–91.
  18. 18. Read TD, Peterson SN, Tourasse N, Baillie LW, Paulsen IT, et al. (2003) The genome sequence of Bacillus anthracis Ames and comparison to closely related bacteria. Nature 423: 81–86.
  19. 19. Rasko DA, Ravel J, Økstad OA, Helgason E, Cer RZ, et al. (2004) The genome sequence of Bacillus cereus ATCC 10987 reveals metabolic adaptations and a large plasmid related to Bacillus anthracis pXO1. Nucleic Acids Res 32: 977–988.
  20. 20. Fedhila S, Nel P, Lereclus D (2002) The InhA2 metalloprotease of Bacillus thuringiensis strain 407 is required for pathogenicity in insects infected via the oral route. J Bacteriol 184: 3296–3304.
  21. 21. Agaisse H, Lereclus D (1994) Structural and functional analysis of the promoter region involved in full expression of the cryIIIA toxin gene of Bacillus thuringiensis. Mol Microbiol 13: 97–107.
  22. 22. Brillard J, Lereclus D (2007) Characterization of a small PlcR-regulated gene co-expressed with cereolysin O. BMC Microbiol 7: 52.
  23. 23. Fedhila S, Gohar M, Slamti L, Nel P, Lereclus D (2003) The Bacillus thuringiensis PlcR-regulated gene inhA2 is necessary, but not sufficient, for virulence. J Bacteriol 185: 2820–2825.
  24. 24. Hajaij-Ellouze M, Fedhila S, Lereclus D, Nielsen-Leroux C (2006) The enhancin-like metalloprotease from the Bacillus cereus group is regulated by the pleiotropic transcriptional activator PlcR but is not essential for larvicidal activity. FEMS Microbiol Lett 260: 9–16.
  25. 25. Romling U, Amikam D (2006) Cyclic di-GMP as a second messenger. Curr Opin Microbiol 9: 218–228.
  26. 26. Dalhammar G, Steiner H (1984) Characterization of inhibitor A, a protease from Bacillus thuringiensis which degrades attacins and cecropins, two classes of antibacterial proteins in insects. Eur J Biochem 139: 247–252.
  27. 27. Ramos JL, Martinez-Bueno M, Molina-Henares AJ, Teran W, Watanabe K, et al. (2005) The TetR family of transcriptional repressors. Microbiol Mol Biol Rev 69: 326–356.
  28. 28. Vetting MW, de Carvalho LP, Yu M, Hegde SS, Magnet S, et al. (2005) Structure and functions of the GNAT superfamily of acetyltransferases. Arch Biochem Biophys 433: 212–226.
  29. 29. Vetting MW, de Carvalho LP, Roderick SL, Blanchard JS (2005) A novel dimeric structure of the RimL Nalpha-acetyltransferase from Salmonella typhimurium. J Biol Chem 280: 22108–22114.
  30. 30. Miao L, Fang H, Li Y, Chen H (2007) Studies of the in vitro Nalpha-acetyltransferase activities of E. coli RimL protein. Biochem Biophys Res Commun 357: 641–647.
  31. 31. Fedhila S, Daou N, Lereclus D, Nielsen-Leroux C (2006) Identification of Bacillus cereus internalin and other candidate virulence genes specifically induced during oral infection in insects. Mol Microbiol 62: 339–355.
  32. 32. Salamitou S, Ramisse F, Brehelin M, Bourguet D, Gilois N, et al. (2000) The plcR regulon is involved in the opportunistic properties of Bacillus thuringiensis and Bacillus cereus in mice and insects. Microbiol 146: 2825–2832.
  33. 33. Callegan MC, Kane ST, Cochran DC, Gilmore MS, Gominet M, et al. (2003) Relationship of plcR-regulated factors to Bacillus endophthalmitis virulence. Infect Immun 71: 3116–3124.
  34. 34. Ramarao N, Lereclus D (2006) Adhesion and cytotoxicity of Bacillus cereus and Bacillus thuringiensis to epithelial cells are FlhA and PlcR dependent, respectively. Microbes Infect 8: 1483–1491.
  35. 35. Mignot T, Denis B, Couture-Tosi E, Kolstø AB, Mock M, et al. (2001) Distribution of S-layers on the surface of Bacillus cereus strains: phylogenetic origin and ecological pressure. Environ Microbiol 3: 493–501.
  36. 36. Hsueh YH, Somers EB, Lereclus D, Wong AC (2006) Biofilm Formation by Bacillus cereus Is Influenced by PlcR, a Pleiotropic Regulator. Appl Environ Microbiol 72: 5089–5092.
  37. 37. Fouet A, Namy O, Lambert G (2000) Characterization of the operon encoding the alternative sigma(B) factor from Bacillus anthracis and its role in virulence. J Bacteriol 182: 5036–5045.
  38. 38. Nadon CA, Bowen BM, Wiedmann M, Boor KJ (2002) Sigma B Contributes to PrfA-Mediated Virulence in Listeria monocytogenes. Infect Immun 70: 3948–3952.
  39. 39. McGann P, Wiedmann M, Boor KJ (2007) The alternative sigma factor sigma B and the virulence gene regulator PrfA both regulate transcription of Listeria monocytogenes internalins. Appl Environ Microbiol 73: 2919–2930.
  40. 40. Hadjifrangiskou M, Chen Y, Koehler TM (2007) The alternative sigma factor sigmaH is required for toxin gene expression by Bacillus anthracis. J Bacteriol 189: 1874–1883.
  41. 41. Beecher DJ, Wong AC (2000) Cooperative, synergistic and antagonistic haemolytic interactions between haemolysin BL, phosphatidylcholine phospholipase C and sphingomyelinase from Bacillus cereus. Microbiol 146: 3033–3039.
  42. 42. Pomerantsev AP, Kalnin KV, Osorio M, Leppla SH (2003) Phosphatidylcholine-specific phospholipase C and sphingomyelinase activities in bacteria of the Bacillus cereus group. Infect Immun 71: 6591–6606.
  43. 43. Duport C, Zigha A, Rosenfeld E, Schmitt P (2006) Control of enterotoxin gene expression in Bacillus cereus F4430/73 involves the redox-sensitive ResDE signal transduction system. J Bacteriol 188: 6640–6651.
  44. 44. Gilois N, Ramarao N, Bouillaut L, Perchat S, Aymerich S, et al. (2007) Growth-related variations in the Bacillus cereus secretome. Proteomics 7: 1719–1728.
  45. 45. R Development Core Team (2005) R: A Language and Environment for Statistical Computing. http://www.r-project.org.
  46. 46. Smyth GK, Speed T (2003) Normalization of cDNA microarray data. Methods 31: 265–273.
  47. 47. Smyth GK (2004) Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3: Article3 Epub.
  48. 48. Agaisse H, Lereclus D (1994) Expression in Bacillus subtilis of the Bacillus thuringiensis cryIIIA toxin gene is not dependent on a sporulation-specific sigma factor and is increased in a spo0A mutant. J Bacteriol 176: 4734–4741.
  49. 49. Bouillaut L, Ramarao N, Buisson C, Gilois N, Gohar M, et al. (2005) FlhA Influences Bacillus thuringiensis PlcR-Regulated Gene Transcription, Protein Production, and Virulence. Appl Environ Microbiol 71: 8903–8910.
  50. 50. Schneider TD, Stephens RM (1990) Sequence logos: a new way to display consensus sequences. Nucleic Acids Res 18: 6097–6100.