Induction of natural competence in Bacillus cereus ATCC14579.

Natural competence is the ability of certain microbes to take up exogenous DNA from the environment and integrate it in their genome. Competence development has been described for a variety of bacteria, but has so far not been shown to occur in Bacillus cereus. However, orthologues of most proteins involved in natural DNA uptake in Bacillus subtilis could be identified in B. cereus. Here, we report that B. cereus ATCC14579 can become naturally competent. When expressing the B. subtilis ComK protein using an IPTG-inducible system in B. cereus ATCC14579, cells grown in minimal medium displayed natural competence, as either genomic DNA or plasmid DNA was shown to be taken up by the cells and integrated into the genome or stably maintained respectively. This work proves that a sufficient structural system for DNA uptake exists in B. cereus. Bacillus cereus can be employed as a model system to investigate the mechanism of DNA uptake in related bacteria such as Bacillus anthracis and Bacillus thuringiensis. Moreover, natural competence provides an important tool for biotechnology, as it will allow more efficient transformation of B. cereus and related organisms, e.g. to knockout genes in a high-throughput way.

Introduction

Genetic competence is the natural ability to take up exogenous DNA from the environment and integrate it into the own genome (Dubnau, 1999), a phenomenon that is widespread among eubacteria (Lorenz and Wackernagel, 1994). One of the best‐studied naturally competent bacteria is Bacillus subtilis. To enable incorporation of DNA from the environment, B. subtilis cells synthesize a specific DNA‐binding and uptake system. Several operons have been identified that are required for functional DNA uptake including comC, comE, comF, comG and nucA. A comG operon‐coded type IV pilus helps the DNA to pass the cell wall and reach the cell membrane (Dubnau, 1997; Chen ), where it is bound by ComEA and taken up by the ComEC permease with the involvement of ComFA and NucA (Chen and Dubnau, 2004). DNA is then integrated via recombination by the action of a protein complex consisting at least of RecA, SsbB, DprA and YjbF (Kramer ).

Competence development is controlled by a complex signal transduction network in B. subtilis, and is expressed under strict conditions (Hamoen ). ComK plays an important role in the regulation of expression of late‐competence genes under these conditions. The induction of ComK is strictly controlled at the level of comK transcription as well as post‐translationally. The transcription of comK is repressed by at least three different transcription factors: AbrB, Rok and CodY, but is also positively regulated by the DegSU two‐component system and by ComK itself (Hamoen ). ComK protein is trapped by the adaptor protein MecA, and targeted to the protease complex ClpCP by which it is degraded (Turgay ). In response to high cell density, quorum‐sensing pathways induce the expression of the small anti‐adaptor protein ComS, which destabilizes the protein complex and liberates ComK (D'Souza ; Hamoen ; Solomon ; Prepiak and Dubnau 2007). ComK then binds to its own promoter region and derepresses transcription resulting in high level of ComK in the cells and development of competence state. However, the enhanced level of ComK and the concomitant development of competence is only achieved in part of the cells (Nester and Stocker 1963; Maamar and Dubnau, 2005; Smits ). ComK is directly or indirectly responsible for the activation of more than 100 genes (Berka ; Hamoen ; Ogura ). Activation of late‐competence genes is achieved by binding of ComK to specific sequences, the so‐called K‐boxes, that consist of two AT‐boxes with the consensus sequence AAAA‐N5‐TTTT and separated by a spacer of a discrete number of helical turns (Hamoen ; Susanna ).

Within the Bacillus family natural DNA uptake has been described for B. subtilis, Bacillus licheniformis and Bacillus amyloliquefaciens (Spizizen, 1958; Thorne and Stull 1966; Koumoutsi ), but the phenomenon has been never demonstrated to occur in the Bacillus cereus sensu lato group.

Bacillus cereus is a rod‐shaped, Gram‐positive spore‐forming bacterium and an important food‐born human pathogen. Bacillus cereus occurs ubiquitously in soil and in a diversity of foods as milk, rice, dairy products and vegetables. As a member of the B. cereus sensu lato group, B. cereus is closely related to Bacillus anthracisand Bacillus thuringiensis (Helgason ; Rasko ). The former one is the ethiological agent of anthrax, and the latter is widely used as an insecticidal biopesticide (Helgason ).

The appearance of whole genome sequences of the Bacillus group showed that the presence of competence genes is not restricted to B. subtilisand closely related species, but is apparent throughout the Bacillus genus. Homologues of most structural proteins required for transformation in B. subtilishave been found in B. cereus, with the exception of the ComGE, ComGF, ComGG and ComFB homologues, suggesting that this strain has a potential for developing natural competence (Rasko ). Interestingly, two homologues of ComK could be identified in B. cereus ATCC14579, suggesting that the regulation of competence gene expression might differ from that in B. subtilis.

It is known that in the B. cereus sensu lato group, pathogens have adapted to various environments, because of insertion of DNA through homologous recombination, a process called horizontal gene transfer (Read ). However, uptake of genomic DNA has never been shown to occur in the B. cereus group. Here, we prove that a minimal system for functional DNA uptake exists and can be induced in B. cereus, which enables us to elucidate how natural transformation proceeds in this organism.

Results

Overexpression of B. subtilis ComK in B. cereus

Competence development in B. subtilis depends on the expression of a set of late‐competence genes that are transcribed only in the presence of the competence transcription factor ComK. Previously, it has been shown that induction of comK combined with positive auto‐stimulation of native comK results in an increased percentage of competence cells in the population (Maamar and Dubnau 2005; Smits ). Interestingly B. cereus ATCC14579 contains two homologues of the B. subtilis comK gene, i.e. comK1 (BC1134) and comK2 (BC5250). The first putative ComK protein (ComK1) shows 62% homology to B. subtilisComK, while ComK2 has only 48% homology and is 32 amino acids shorter at the C‐terminal than the B. subtilis ComK protein.

Natural transformation is an efficient tool for genetic manipulation, which has been used successfully in B. subtilis for many years. Therefore, we set out to determine whether B. cereus ATCC14579 is naturally transformable. Driven by the fact that many late‐competence genes and recombination genes are present in the genome of this strain (Table S1), the first step to enhance expression of late‐competence genes in B. cereus was to increase ComK levels in the cells. The regulation and function of the two comK genes are unknown in this bacterium. Here, we employed the comK gene of B. subtilis and investigated whether functional DNA uptake exists in B. cereus. For this purpose the B. subtilis comK gene (hereafter referred to as comK) was cloned behind the Phyperspank promoter that can be induced by isoporyl‐β‐d‐thiogalactoside (IPTG) addition. The resulting plasmid pNWcomKBsu (or the empty plasmid pNW33N used as a control) was introduced by electroporation into B. cereus ATCC14579. To verify the overexpression of ComKBsu following IPTG addition, Western blot analysis was performed using ComKBsu‐specific antibodies (Fig. 1). The most favourable condition was selected, where ComK protein production was most effectively induced. Cells collected at two different time points, 2 or 3 h after induction, showed high expression of ComKBsu in the induced B. cereus strain containing pNWcomKBsu plasmid. Slightly higher expression was observed with the samples taken after 3 h of induction. In contrast, no ComK could be detected either in the non‐induced strains or when the induced strain contained the empty plasmid pNW33N. The experiments above demonstrate that a reasonably high yield of IPTG‐dependent overexpression of ComKBsu can be achieved in B. cereus.

Figure 1

Detection of B. subtilis ComK protein. Equal amounts of proteins were loaded in each lane. Strains carrying an empty plasmid pNW33N or pNWcomKBsu were grown in MM, supplemented with chloramphenicol (5 µg ml−1). Samples were taken 2 h (A) and 3 h (B) after IPTG induction (+) (final concentration 1 mM) or without IPTG induction (−). Cells were separated from the growth medium by centrifugation, lysed and analysed by Western blotting using ComKBsu‐specific antibodies (see Experimental procedures). The arrow indicates ComKBsu‐specific signal.

The B. subtilisComK protein can induce transcription from PcomGA of B. cereus

As comG has been shown to be a suitable competence‐specific reporter (Smits ) and transcription of this gene is under strict control of ComK in B. subtilis, we constructed a fusion between the promoter region of B. cereus comGA (PcomGA) and the green fluorescence protein (gfp) gene. The resulting plasmid (pILcomGA–gfp) containing an in‐frame fusion of gfp with the first six codons of B. cereus comG was used to determine the effect of ComKBsu overexpression under various growth conditions. As the fusion contains only a few N‐terminal amino acid residues of ComGA protein, it should not specifically localize in the cells, similar to previous experiments (Maamar and Dubnau, 2005; Smits ). This construct was introduced into B. cereus by electroporation. Also the pIL253 control plasmid was introduced into B. cereus. Both fluorescence microscopy and flow cytometry were used to visualize the expression of GFP from the PcomGA in individual cells.

Experiments were performed with strains harbouring either pNWcomKBsu or the control pNW33N plasmid, both carrying the pILcomGA–gfp vector to determine whether the expression system used could result in the high ComKBsu level that is probably required for activation of the B. cereus competence cascade. The strains were grown in presumed competence stimulating medium in the presence of glucose and induced at different time points and with different concentrations of IPTG (varying from 0.1 to 1.0 mM). Subsequently, cultures were sampled every hour after induction and analysed. Fluorescence microscopy showed the presence of GFP signal from PcomGA in IPTG induced B. cereus samples carrying the pNWcomKBsu vector (Fig. 2). Interestingly, only a few percent of cells in the culture showed the expression of GFP. In agreement with the microscopy experiments, flow cytometric analysis also proved that only part of the population expressed high levels of GFP (Fig. 3). Flow cytometry experiments showed that highest PcomGA–gfp expression occurred after 3 h of induction with a 1 mM IPTG concentration. We did not notice GFP expression in B. cereus carrying the control plasmid pNW33N neither under non‐inducing or inducing conditions [1.6 ± 0.5 and 1.6 ± 0.6 arbitrary units (AU) in non‐induced and induced cells respectively]. In the strain harbouring pNWcomKBsu, either induced or not induced, the GFP signal was enhanced (3.4 ± 0.6 and 12.3 ± 1.0 AU in non‐induced and induced cells respectively). However, in the induced strain, the GFP signal was stronger. The low level fluorescence, which occurs in the non‐induced B. cereus pNWcomKBsu strain, is most likely originating from the low basal expression of comK gene from the Phyperspank promoter. In the ComK‐expressing samples, two subpopulation of cells could be separated in flow experiments, one that possesses GFP activity (2.3 ± 0.3 AU) similar to the control strain (B. cereus ATCC14579 with pNW33N), while in the other part of the cells high GFP activity could be detected (117 ± 11 AU in 23 ± 2.5% of the culture).

Figure 2

Visualization of PComGABce–GFP protein production in B. cereusby fluorescence microscopy.
A and B. Bacillus cereuspNWcomKBsu carrying pILcomGA–gfp.
C and D. Bacillus cereuscarrying pILcomGA–gfp.
(A) and (C) are phase contrast pictures of cells from (B) and (D) respectively. Strains were grown in MM medium, and samples were taken 3.5 h after IPTG induction. Cells producing GFP from the comGA promoter were visualized on (B) and (D). Production of GFP, whose expression was driven by activity of comGA promoter, was visualized by fluorescence microscopy, as described in Experimental procedures.

Figure 3

Flow cytometric analysis of PcomGABce–gfp in liquid MM medium. Analyses were performed as described in Experimental procedures. The panels present the different time points in hours relative to the induction point at OD600 of 0.75 (1 h after IPTG induction is T1, etc.). Strains carrying PcomGABce–gfp fusion were grown in liquid MM, supplemented with chloramphenicol (2.5 µl ml−1) and erythromycin (2.5 µl ml−1). Bacillus cereus strains contain the pNW33N control plasmid or the ComKBsu‐overexpressing plasmid, pNWcomKBsu. The plus and minus symbols denote samples incubated in the presence or absence of IPTG respectively. The numbers of cells are indicated on the y‐axis, and their relative fluorescence levels are indicated on the x‐axis on a logarithmic scale. For detailed description of the visualization, see also the WinMDI manual (http://facs.scripps.edu/software.html).

Our results show that high levels of ComKBsu in the cells can activate transcription of the comGAgene either directly or indirectly and probably also increases expression of the other late‐competence genes.

ComKBsu binds to the promoter regions of potential competence related genes in B. cereus

Previously, it has been shown that in B. subtilis ComK can bind directly to a region upstream of late‐competence genes (Hamoen ). Expression activation at the comGA promoter is solely dependent on ComK (Susanna ). Here, electrophoretic mobility shift assays (EMSAs) were performed to determine whether purified ComKBsu protein can bind to the comGAand comE promoter regions of B. cereus(hereafter referred to as PcomGA and PcomEA). The results from the gel shift analyses showed that ComKBsu can bind in vitro to both promoter fragments, leading to reduced mobility of the probes (Fig. 4A). These data confirm that ComKBsu may activate transcription at the PcomGA promoter region by directly binding to an operator region. Reaction mixtures contained poly(dI‐dC) that is known to eliminate non‐specific DNA binding of ComKBsu (Hamoen ).

Figure 4

Bacillus subtilis ComK binds to both B. cereus late‐competence genes promoters PcomGA and PcomEA (A), and the promoters of comK1 and comK2 genes (B). Assessment of binding of ComKBsu was performed as described in Experimental procedures. DNA fragments from the indicated promoter regions were prepared by PCR and end‐labelled with 33P. In the panel, the lanes marked with an X correspond to probe alone. The remaining lanes correspond to probe samples incubated with ComKBsu, at concentration ranging from 0.2 µM to 1.6 µM. Each lane, from left to right, corresponds to a doubling in the concentration of ComKBsu.

Analysis of the PcomGA sequence showed the presence of a putative ComK binding site upstream of the putative –35 and –10 promoter elements of comGA (Fig. S1).

Subsequently, we performed EMSA analysis with the comK1 and comK2 promoter regions to determine whether ComKBsu could also affect the expression of these genes. In both cases, DNA binding was observed by ComKBsu (Fig. 4B), showing it might also directly affect comK1 or comK2 expression, thereby passing by their putative natural regulation.

Occurrence of the competence state in B. cereus ATCC14579

The previous experiments showed that high expression of ComKBsucan activate PcomGA–gfp expression in B. cereus ATCC14579. Having constructed an inducible ComKBsu system, which efficiently activated transcription level of comGA, we next investigated whether the presumed competence machinery of B. cereus could be fully functional. For this, we tested the presence of natural transformation using B. cereus pNWcomKBsuand, as a control, a B. cereus strain carrying the empty plasmid pNW33N. Both B. cereus strains were grown in B. subtilis‐like competence minimal medium (see Experimental procedures). When they reached an optical density at 600 nm (OD600) of 0.7, they were induced with 1 mM IPTG. Either genomic DNA of B. cereus FM1400 (van Schaik ) containing a sigBnull mutation carrying an erythromycin cassette on the chromosome or the replicative plasmid pNG8048E (Kuipers ) was used as donor DNA during the transformation experiments. In these experiments, genomic or plasmid DNA was added to the cells, which were incubated for 30 min at 30°C with shaking. Subsequently, TY medium was added and the incubation was continued for an additional 1 h (see Experimental procedures). The transformation was monitored by plating cells on TY plates containing 2.5 µg ml−1 erythromycin. Colony counts are presented in Table 1. Although the transformation efficiency was relatively low, only the B. cereus strain containing the inducible comK could act as a recipient of plasmid or genomic DNA, being stably maintained or integrated respectively. In all transformants the presence of erythromycin‐ and chloramphenicol‐resistant markers was confirmed by PCR (data not shown), proving the uptake of the foreign DNA by the cells containing the pNWcomKBSU. Deletion in the sigB gene was also proven in specific PCR reactions (data not shown). Strains transformed with the genomic DNA of FM1400 were checked for their ability to survive a 60 min exposure to 50°C after pre‐treatment at 42°C for 30 min. The donor sigB mutant strain and the transformed strains showed the similar 103 times reduction in viable counts upon these treatments. Plasmid pNG8048E could be isolated and retransformed into Lactococcus lactis proving that it is stably maintained. These results confirm that ComKBsu is able to induce natural competence in B. cereusand above all, that B. cereus contains the minimal structural machinery required to bind and take up DNA.

Table 1

Natural transformation experiments.

Strain	Induction with 1 mM IPTG	cfu
		pNG8048e			FM1400
Bacillus cereus pNWcomKBsu	+	3	1	3	9	1	3
Bacillus cereus pNWcomKBsu	−	0	0	0	0	0	0
Bacillus cereus pNW33N	+	0	0	0	0	0	0
Bacillus cereus pNW33N	−	0	0	0	0	0	0

Numbers present colonies obtained after transformation with 1.5 µg of genomic DNA or plasmid. cfu, colony‐forming units.

Discussion

Competence development has been described for a variety of species belonging to both Gram‐negative and Gram‐positive bacteria. Natural DNA uptake presents a valuable tool for genetic studies. This beneficial property made B. subtilis the model organism of the Bacillus family. Since the description of natural transformability of B. subtilis (Spizizen, 1958), the phenomenon has been described only in closely related species [B. licheniformis and B. amyloliquefaciens (Thorne and Stull, 1966; Koumoutsi )] within Bacillaceae. In contrast, competence development has not been described for B. cereus or any of its close relatives, such as B. anthracis. The growing number of available genome sequences indicates the presence of several genes capable of coding for such a DNA uptake apparatus in these species. The presence of these orthologous genes raises the question whether these species can become competent.

To address this question, we tested natural DNA uptake in B. cereus ATCC14579 harbouring an inducible comK gene, using either genomic or plasmid DNA containing an erythromycin‐resistant marker. We showed the occurrence of a low level of transformation (one to six colonies per microgram of DNA) under the given condition. Considering the low efficiency, it is still needed to find the best environmental conditions for obtaining enhanced rates of DNA uptake. The presence of this low transformation frequency can be explained in several ways. Although many homologues for competence‐related genes are found in B. cereus ATCC14579, some genes are lacking or could be non‐functional. Two striking examples are the absence of the comF gene in the comF operon and the missing homology between the ComGEFG proteins of B. cereus and B. subtilis, expressed from the comG locus (Table S1). However, the comF gene is also missing in Streptococcus pneumoniae, another example of a natural competent Gram‐positive bacterium (Berge ). Interestingly, there are three non‐homologous genes present in the comG gene cluster of B. cereus at the analogous position as comGE, comGF and comGG in B. subtilis that could be the functional complements of these B. subtilis genes. Another possible explanation for the relatively low transformation efficiency might be related to a decreased recombination efficiency or the result of a low level of induction of recombination genes.

In spite of the presumably homogeneous overexpression of ComKBsu, the transcription initiation from the PcomGA region occurs in a low, bistable manner (Fig. 2 and 3). Bistable expression of genes in bacteria depends on a positive or double negative regulatory cascade (Smits ; Dubnau and Losick, 2006). In B. subtilis only 10–20% of the cells of a given culture express the late‐competence genes (Nester ; Hadden and Nester, 1968). This bistable outcome requires positive autoregulation at the PcomK, and the basal level of ComKBsu determines the fraction of cells that develop competence (Maamar and Dubnau, 2005; Smits ). The fraction of B. subtilis cells that develop competence is increased if the level of ComKBsu is enhanced (e.g. by inducing comKBsu expression from a heterologous promoter or by inactivating MecA involved in proteolytic degradation of ComKBsu) (Smits ). A bistable outcome of PcomGA upon monomodal overexpression of ComKBsu suggests that one or more regulators of B. cereus are also involved in the expression of comGA. It is interesting to note that expression from PcomGA was not enhanced by overexpressed ComKBsu in rich medium (Fig. S2), suggesting that medium‐dependent factors are also affecting this heterogeneous regulation of competence genes. In B. cereus ATCC14579 two comK genes could be identified based on the homology to ComKBsu sequence. The roles of these proteins are unknown in B. cereus, although this is not the only example of doubled regulators in the B. cereus group. Bacillus cereus G9241 was shown to contain two copies of the atxA gene, the regulator shown to be involved in pathogenicity of B. cereus (Hoffmaster ).

Electrophoretic mobility shift assay experiments demonstrated binding of ComKBsu to the promoter regions of two late‐competence genes (comGA and comEA), but also to PcomK1 and PcomK2. The overexpressed ComKBsu might bind and modulate the expression of comK1 and comK2 genes in B. cereus, thereby promoting the occurrence of a bistable expression of competence genes. Thus the ComK1 and/or ComK2 proteins might play a role in this regulation. As ComKBsu binds to PcomGA and PcomEA regions, the effect of ComKBsu might be independent of the B. cereus comK genes. The investigation of the ComK1 and ComK2 functions in the transcriptional regulation of late‐competence genes is the subject of future studies in our laboratory. The intent of our current study was to show the presence of a functional DNA uptake machinery in B. cereus ATCC14579.

The induction of competence in B. cereus resulted in uptake and homologous recombination of genomic DNA. The place of recombination was verified both with molecular biology techniques (e.g. PCR) and by phenotypic characterization (e.g. the effect of sigB mutation on heat resistance). This indicates that the recombination apparatus was also functional under these conditions. Thus, our natural transformation method should help us to introduce mutations into the genome or plasmids of B. cereus ATCC14579 using linear PCR fragments containing homologous fragments and resistance genes, i.e. the method described as allelic replacement mutagenesis using long‐flanking‐homology PCR (Wach, 1996). This biotechnologically important and convenient method would facilitate a rapid mutagenesis of target genes and by‐pass the time‐consuming steps of current methods using selection of single and subsequent double‐recombination events and electroporation.

In addition to conjugation and phage transduction, the observed natural transformation can also provide an explanation for the naturally occurring genome plasticity observed in B. cereus and related species such as B. anthracis and B. thuringiensis. Recent studies indicate that these pathogens have adapted to various environments by foreign insertion of DNA through homologous recombination (Ivanova ; Read ; Rasko ).

In this report, we have shown that a functional DNA uptake apparatus exists in B. cereus ATCC14579, which will further enhance B. cereus research. Our work presents a useful tool for up to now non‐transformable Bacilli. Establishing a competence regime for a wider range of Bacillus species will greatly enhance the genetic accessibility, and thus research speed, for medically and industrially relevant Bacilli.

Experimental procedures

Bacterial strains and media

The strains and plasmids used in this study are listed in Table 2. Bacillus cereus strains were grown in TY (1% Bacto‐Tryptone, 0.5% Bacto‐yeast extract and 1% NaCl) or in MM [62 mM K2HPO4, 44 mM KH2PO4, 15 mM (NH4)2SO4, 5.6 mM sodium citrate, 0.8 mM MgSO4, 0.02% of casamino acids, 27.8 mM glucose and 0.1 mM l‐tryptophan (Leskela )] at 30°C, supplemented with appropriate antibiotics, erythromycin (5 µg ml−1) or chloramphenicol (3–5 µg ml−1). Escherichia coli strains were grown at 37°C in TY, supplemented with chloramphenicol (15 µg ml−1) or erythromycin (150 µg ml−1). Lactococcus lactisstrains were grown in twofold‐diluted M17 broth (Terzaghi and Sandine, 1975) supplemented with 0.5% glucose (GM17) at 30°C and erythromycin (5 µg ml−1).

Table 2

Bacterial strains and plasmids.

Strains/plasmids	Relevant characteristics	Reference
Strains
ATCC14579	Type strain	Bacillus Genetic Stock Center
FM1400	sigB::ery	van Schaik et al. (2004)
MG1363	Lac‐ Prt‐; plasmid‐free derivative of NCDO712	Gasson (1983)
MC1061	F‐ araD139Δ(ara‐leu) 7696 galE15 galK16Δ(lac)X74 rpsL (Strr) hsdR2 (rK_mK+) mcrA mcrB1	Wertman et al. (1986)
Plasmids
pSG1151	ampr, cat	Lewis and Marston (1999)
pIL253	ermAM, eryr	Simon and Chopin (1988)
pNW33N		Genetic Stock Center
pNWcomKBsu	pNW33N containing comKBsu	This study
pSC1151‐comGA	pSG1151 containing PcomGA fused to gfp	This study
pIlcomGA–gfp	pIL253 containing PcomGA–gfp	This study

Construction of pILcomGA–gfp vector

The comGA promoter region including the ribosome bind site was amplified by PCR using primers comGA‐EcoRI‐fw (primer sequences are available in Table S2) and comGA‐ApaI‐rw. The PCR product was digested with EcoRI and ApaI ligated into the corresponding sites of pSG1151 (Terzaghi and Sandine, 1975; Lewis and Marston, 1999), resulting pSG1151‐PcomGA. This plasmid was used as a template to amplify out PcomGA–gfpby PCR using primers oAM17 and oAM18. The resulting 1391 bp PCR fragment was digested with XbaI and Eco47III, and inserted to into XbaI–SmaI cleaved pIL253 (Simon and Chopin, 1988).

The resulting plasmid, pILcomGA–gfp, was introduced into L. lactis MG1363 by electroporation (Holo and Nes, 1995). Correct construct was confirmed by sequencing. Subsequently, plasmid pILcomGA–gfp was introduced by electroporation (Bone and Ellar, 1989) into B. cereus and B. cereus pNWcomKBsu.

Construction of pNWcomKBsu

After amplification of B. subtilis comK gene with pDRK‐F and pDRK‐R, the 746 bp PCR fragment was digested with HindIII and SphI and cloned into the corresponding sites of pDR111 (kind gift of David Rudner), yielding pDRKwt. The sequence of the insert was verified with digestion and sequencing. The 2506 bp BamHI and PstI fragment of pDRKwt was inserted to the adequate sites of pNW33N (GenBank Accession No. AY237122), resulting pNWcomKBsu. The verified construct along with pNW33N as control was electroporated into B. cereus ATCC14579.

Natural competence in B. cereus ATCC14579

Bacillus cereus ATCC14579 strains carrying either pNW33N or pNWcomKBsu were grown overnight in TY supplemented with chloramphenicol (5 µg ml−1). For IPTG‐mediated induction of ComKBsu expression a fresh overnight culture of strains was inoculated into fresh minimal medium, supplemented with chloramphenicol (3 µg ml−1) with an OD600 of 0.15. When cultures achieved an OD600 of 0.75, they were induced by 1 mM IPTG. Cultures were grown further for 3.5 h before 1.5 µg of genomic or plasmid DNA was added. Genomic DNA was extracted from B. cereus FM1400 with the Sigma‐Aldrich GenEluteTM Bacterial Genomic DNA Kit and contains no detectable spores. Plasmid DNA was isolated from L. lactis with the Roche High Pure Plasmid Isolation Kit. It contains mono‐, di‐ or multimer forms of plasmid DNA as it is active in transformation of B. subtilis natural competence cells that requires di‐ or multimeric forms of plasmid. Cells were incubated at 30°C with shaking for 30 min before TY medium was added, and incubation was continued for a subsequent hour. Cells were plated out onto TY agar plate (1.5% agar) supplemented with erythromycin (2.5 µg ml−1). The presence of chloramphenicol resistance gene in transformants was proved by PCR using CmForw and CmRev; the presence of erythromycin was shown in the FM4100‐transformed stain using Ery2‐F and Ery2‐R; in pNG4048E‐transformed strain using Ery‐F and Ery‐R; the deletion in sigB was confirmed by PCR using primers SigB‐down‐R and SigB‐up‐F.

Flow cytometric analyses

Cultures were grown similar to the conditions described above expect that erythromycin was added (5 µg ml−1 in overnight cultures and 2.5 µg ml−1 in minimal medium). After induction with 1 mM IPTG at OD600 of 0.75, samples were taken every hour.

Cells were diluted in minimal salts and analysed on Coulter Epics XL‐MCL flow cytometer (Beckman Coulter Mijdrecht, NL) operating an argon laser at 488 nm. GFP signals were collected through an FITC filter with the photomultiplier voltage set between 700 and 800 V. Date were obtained using EXPO32 software (Beckman Coulter) and further analysed using WinMDI 2.8 (http://facs.scripps.edu/software.html). Figures were prepared using WinMDI 2.8 and Corel Graphics Suite 12.

Fluorescence microscopy

The fluorescence of the GFP reporter protein was visualized with a Zeiss Axiophot microscope, using appropriate filters. Imaging of pILcomGA–gfp in individual cells using fluores cence microscopy was performed as described by Smits and colleagues (2005).

Western blotting and protein detection

SDS‐PAGE and Western blotting of cell lysates was performed as described (Bolhuis ). Chemiluminescent detection of bound ComK‐specific antibodies (Kong and Dubnau, 1994) was performed with horseradish peroxidase‐conjugated anti‐rabbit IgG and the ECL Western blotting analysis system (Amersham).

Eelectrophoretic mobility shift assay

Electrophoretic mobility shift assays were carried out essentially as described by Susanna and colleagues (2004). The promoter regions of B. cereus competence genes, comK1, comK2, comEA and comGA, were obtained by PCR, using primers comK1‐F and comK1‐R for PcomK1; primers comK2‐F and comK2‐R for PcomK2; primers comGA‐F and comGA‐R for PcomGA; primers comEA‐F and comEA‐R for PcomEA. As a template chromosomal DNA of B. cereus ATCC14579 was used. The resulting fragments were end‐labelled with T4 polynucleotide kinase and [γ‐33P]‐ATP. Purified B. subtilis ComK protein and probe were premixed on ice in binding buffer (80 mM Tris HCl pH 8, 4 mM EDTA, 20 mM MgCl2, 2 MM DTT, 400 mM KCl, 40% glycerol) containing 0.5 µg µl−1 poly(dI‐dC) as non‐specific DNA and 2 µg µl−1 BSA. Samples were incubated at 30°C, and were loaded on 6% polyacrylamide gel after 20 min of incubation. Gels were run in 1× TBE buffer (0,089 mM Tris, 0.089 mM Boric Acid, 0,022 mM EDTA) at 90 V for 60 min, dried in a vacuum dryer and autoradiographed using phosphoscreens and a Cyclone PhosphorImager (Packard Instruments, Meridien, CT).