Bacillus cereus causes two types of gastrointestinal diseases: emesis and diarrhea. The emetic type of the disease is attributed to the heat-stable depsipeptide cereulide and symptoms resemble Staphylococcus aureus intoxication, but there is no rapid method available to detect B. cereus strains causing this type of disease. In this study, a polymerase chain reaction (PCR) fragment of unknown function was identified, which was shown to be specific for emetic toxin producing strains of B. cereus. The sequence of this amplicon was determined and a PCR assay was developed on this basis. One hundred B. cereus isolates obtained from different food poisoning outbreaks and diverse food sources from various geographical locations and 29 strains from other species belonging to the B. cereus group were tested by this assay. In addition, 49 non-B. cereus group strains, with special emphasis on food pathogens, were used to show that the assay is specific for emetic toxin producing B. cereus strains. The presented PCR assay is the first molecular tool for the rapid detection of emetic toxin producing B. cereus strains.
Bacillus cereus is a causative agent of gastrointestinal and non-gastrointestinal diseases. It can cause two types of food poisoning syndromes: emesis and diarrhea. Besides its food poisoning potential, B. cereus has been shown to be responsible for wound and eye infections, as well as systemic infections . Recently, it has been reported that systemic complications of B. cereus infections in premature neonates might be at least partly related to enterotoxins . However, in general the role of the diverse toxins and virulence factors of B. cereus in systemic infections is poorly studied. The development of molecular tools will be necessary to allow a rapid characterization of virulence mechanisms of clinical B. cereus isolates.
Diarrheal poisoning is caused by heat-labile enterotoxins produced during vegetative growth of B. cereus in the small intestine whereas the emetic type of food poisoning is caused by the small, heat- and acid-stable cyclic dodecadepsipeptide cereulide [3,4]. While enterotoxins are comparatively well characterized at the molecular and the expression level (for review see ), far less is known about the emesis causing toxin. The chemical structure and characteristics of cereulide have been studied in some detail but the molecular basis for its synthesis remains unknown. Cereulide causes cellular damaging effects in animal models [4,6], is toxic to mitochondria by acting as a potassium ionophore  and it was involved in fulminant liver failure in a human case . Recently, it has been reported that cereulide inhibits human natural killer cells and might therefore have an immunomodulating effect .
In general, the incidence of B. cereus food poisoning is underestimated since B. cereus is not a reportable disease and reporting procedures vary between countries. There is a tendency for many more B. cereus food poisoning cases to be reported in northern countries. In Norway B. cereus was the most common microbe isolated from food-borne illnesses in 1990  and it was responsible for 14% of the outbreaks in Finland in which the causative agent was identified . B. cereus is a major problem in convenience food and mass catering. Due to heat and acid resistance of its spores it is not eliminated by pasteurization or sanitation procedures. Investigation of food-borne outbreaks in the German Federal Armed Forces showed that B. cereus was by far the most frequently isolated pathogen in the retained food samples. It was responsible for 42% of the outbreaks reported between 1985 and 2000 . While in Norway, Finland and Hungary the diarrheal type of food poisoning was predominant, the emetic type was prevalent in the UK, Japan and the USA [13,14]. The true incidence of B. cereus food poisoning may also be underestimated due to misdiagnosis of the illness, which is symptomatically similar to some other types of food poisoning. While the emetic type of food poisoning caused by B. cereus resembles staphyloenterotoxicosis, the diarrheal syndrome of B. cereus resembles the symptoms of a Clostridium perfringens infection. In general, both types of food-borne illness caused by B. cereus are relatively mild and usually do last not more than 24 h. Nevertheless, more severe forms have occasionally been reported, including one death after the ingestion of food contaminated with high amounts of emetic toxin and three deaths caused by a necrotic enterotoxin [15,8].
Since B. cereus is a ubiquitous spore former that cannot be totally avoided, it is necessary to develop rapid methods to discriminate hazardous strains from non-toxic strains. The utility of polymerase chain reaction (PCR) based methods is evident by the 1999 guidelines issued by NCCLS , encouraging the use of molecular methods in clinical laboratories performing bacterial identification assays. Such an assay would also be advantageous for quality control in the food industry and could improve food safety substantially. While for enterotoxic B. cereus strains molecular diagnostic PCR assays have been described (e.g., [17–19]) and commercial immunological assays are available, for emetic strains such tools are still missing. The presented PCR system may fill that gap by providing a molecular assay to rapidly detect emetic toxin producing B. cereus strains.
2Materials and methods
Reference strains for the development of the PCR assay were the emetic reference strain F4810/72 , three isolates from recent emetic outbreaks in Germany and three emetic isolates from foods (MHI 280, MHI 297, MHI 1305, and MHI 87, MHI 135, MHI 294 ), as well as the non-emetic culture collection strains, B. cereus ATCC 14579T and ATCC 27877, and two non-emetic isolates from foods (MHI 13, MHI 124 ). For evaluation of the PCR assay B. cereus strains were obtained from the project culture collection of an ongoing EU project (QLK1-CT-2001-00854). The strain set covered isolates from diverse geographical origins; it included 60 clinical isolates and isolates from food remnants connected to food-borne outbreaks, as well as 30 isolates from diverse food stuffs and the environment (for details on strains see ). The origins of emetic strains included in this study are shown in Table 1. Twenty-nine strains from members of the B. cereus group closely related to B. cereus, type strains of other Bacillus species, as well as 41 isolates from non-Bacillus species (see Table 2) were used to assess the specificity of the PCR assay.
Table 1. Origin of emetic B. cereus used in this study
aPublic Health Laboratory Service, London, UK.
bM. Salkinoja-Salonen, University of Helsinki, Department of Applied Chemistry and Microbiology, Helsinki, Finland.
cE. Märtlbauer, Ludwig-Maximilians-Universität, Lehrstuhl für Hygiene und Technologie der Milch, Munich, Germany.
dNagoya City Public Health Research Institute, Nagoya, Japan.
eF. van Leusden, Rijksinstituut voor Volksgezondheid en Milieu, Bilthoven, The Netherlands.
fA. Christiansson, Svensk Mjolk, Swedish Dairy Association, Lund, Sweden.
gDeutsche Sammlung für Mikroorganismen und Zellkulturen, Braunschweig, Germany.
Number of emetic B. cereus
Strains connected to emetic type of food poisoning
Chicken and rice
Table 2. Bacterial species used to test the specificity of PCR assay
aIncluding 63 clinical isolates and isolates from food remnants connected to food poisoning, 35 isolates from food and environment and the culture collection strains ATCC 145979, ATCC 27877.
Bacterial species (number of species)
Number of strains tested
Bacillus cereus group (6)
Other Bacillus sp. (4)
Other non-Bacillus species (9)
Escherichia coli (incl. serovar O157)
2.2Isolation of DNA
DNA from Gram-positive bacteria was prepared using the AquaPure Genomic DNA Isolation kit (Bio-Rad, Germany). DNA from Gram-negative bacteria was prepared by suspending cells from one colony in sterile water. The suspension was heated at 95°C for 3 min and then placed on ice. After centrifugation the supernatant was used as template for PCR or stored at −20°C. For Southern analysis, total chromosomal DNA was isolated by phenol–chloroform extraction. In brief, after lysis (0.5% sodium dodecyl sulfate (w/v), 0.1 mg ml−1 proteinase K, 37°C, 3 h), cell wall debris, denatured proteins and polysaccharides were complexed to hexadecyltrimethylammonium bromide and removed by phenol–chloroform extraction. DNA was precipitated with 2-propanol, washed with ethanol and dissolved in water.
Degenerate oligonucleotide primers targeting highly conserved motifs of known non-ribosomal peptide synthetases (NRPS [23,24]) were used to amplify and identify putative NRPS gene fragments in B. cereus. Based on the sequence information derived from an amplified DNA fragment, specific primers and probes for detection of emetic B. cereus strains were designed. The primers EM1F: 5′-GACAAGAGAAATTTCTACGAGCAAGTACAAT-3′ and EM1R: 5′-GCAGCCTTCCAATTACTCCTTCTGCCACAGT-3′ amplify a fragment of 635 bp from emetic B. cereus genomic DNA. The primer HpaIF (5′-GCCAGAAGATGCAATGATTCCAGTATG-3′) that is located inside the 635-bp amplicon was used in combination with EM1R to generate a probe for Southern analysis.
2.4Cloning of PCR amplicons and DNA sequencing
Amplification products from PCR using the emetic B. cereus reference strain F4810/72 were subcloned in Topo TA vector (Invitrogen, Germany) and sequenced. The resulting sequences were searched against the sequenced genomes of B. cereus (strains ATCC 14579T and ATCC 10987), B. anthracis (strains A2012, Ames, Kruger B, and Western NA) and S. aureus (Mu50, MW2, N315), and against NCBI's non-redundant protein database using BLAST . PCR amplicons obtained with the specific primers (EM1F+EM1R) from 10 emetic B. cereus strains were purified using the QIAquick PCR purification kit (Qiagen) and were directly sequenced using a DNA dyedeoxy terminator cycle sequencing kit (Sequiserve, Germany).
Chromosomal DNA was digested with HpaI. The fragments were separated on a 1% agarose gel and blotted onto nitrocellulose. Southern analysis  was performed using a 260-bp digoxigenin labelled probe (Roche, Germany) directed against the gene fragment of unknown function. This probe was obtained from the emetic reference strain F4810/72 by PCR using the oligonucleotide primers HpaIF and EM1R described above. In order to ensure that the bacterial cell lysis protocol was efficient for all species tested hybridizations with a 16S rDNA universal DNA probe were performed. For this purpose, a 241-bp digoxigenin labelled probe (Roche, Germany) directed against a conserved region of the 16S rDNA gene  was used.
2.6Emetic B. cereus specific PCR
PCR mixture (50 μl) contained 0.8 mM dNTP mix, 1.5 mM MgCl2, 50 pM of each primer, 0.5 U ThermoStart Taq DNA polymerase (ABgene, Epsom, UK), 5 μl 10×polymerase buffer and 1 μl template DNA. The PCR protocol started with a denaturation step for 15 min at 95°C, followed by 30 cycles of 30 s 95°C, 30 s 60°C, 60 s at 72°C each and ended with a final elongation step at 72°C for 5 min. All strains used for the evaluation of the assay were checked with the primer set EM1F+EM1R. Selected strains were amplified in parallel with 16S primers as a positive control. For amplification of 16S rDNA the general primers derived from E. coli (8–26/56: 5′-AGAGTTTGATCCTGGCTCA-3′; 1511–1493: 5′-CGGCTACCTTGTTACGAC-3′) were used .
2.7Microbial DNA mixtures
The specificity of the designed primers was determined using mixtures of chromosomal DNA isolated from different bacterial food-borne pathogens. One DNA mixture contained all members of the B. cereus group (B. cereus ATCC 14579T, B. anthracis Sterne CIP7702, B. thuringiensis WS2614, B. weihenstephanensis WSBC10204T, B. mycoides WSBC10276, B. pseudomycoides WS3118 T), a second mixture contained different types of S. aureus (WS2604 (SEA); WS2606 (SEB); WS2608 (SEC); WS2610 (SED); WS2612 (SEE)) and a third mixture contained B. licheniformis WSBC23001; C. perfringens ATCC 3628; S. aureus WS2604; S. equorum WS2733; L. monocytogenes WSLC10209; E. coli WS2577; Salmonella enterica serovar Dublin WS2692; Y. enterocolitica WS2589. Each DNA mixture was prepared with and without DNA from the emetic reference strain F4810/72. Equal amounts of DNA from the different bacteria were provided as template DNA for the PCR reaction. To further assess the influence of background flora on the PCR assay 200 ng DNA (corresponding to approx. 108 CFU) from the non-emetic B. cereus type strain ATCC 14579T was mixed with 10-fold dilutions of DNA from B. cereus F4810/72. PCR was carried out as described above.
3.1Identification of an emetic B. cereus specific genomic DNA fragment
In a PCR based screening assay, using degenerate primers targeting conserved sequence motifs of NRPS genes, two distinct bands of 700 bp and 800 bp were amplified from emetic B. cereus strains. Only one band (700 bp) was obtained for non-emetic B. cereus strains (Fig. 1). The 800-bp fragment that was amplified from the cereulide producing B. cereus strain F4810/72 was cloned and sequenced. A database search using BLAST revealed weak homologies of this genomic DNA fragment to AMP binding genes and acetyl-CoA synthase, but no significant homologies to non-ribosomal peptide synthetases. The sequence of this genomic DNA fragment was used to design oligonucleotide primers for the amplification of the identified DNA fragment from emetic toxin producing B. cereus strains. Sequencing of PCR amplicons from 10 emetic strains revealed identical sequences for all strains (data not shown), although these strains were epidemiologically unrelated and were obtained from diverse habitats.
Subsequently, this sequence information was used to generate a 260-bp probe for Southern analysis and to develop a PCR based assay. Selected emetic and non-emetic strains were tested for the presence or absence of the identified DNA fragment by Southern hybridization. The digoxigenin labelled DNA probe hybridized only to DNA from emetic B. cereus isolates, whereas no hybridization signals were observed from non-emetic B. cereus strains or DNAs from other bacterial species (Fig. 2).
3.2Evaluation of the PCR assay
The specificity of the assay was assessed using a panel of 100 B. cereus isolates from clinical cases, foods and environments. Cross-reactivity of the oligonucleotide primers designed was tested for closely related members of the B. cereus group as well as for other Bacilli and known food pathogens. In total, 178 bacterial isolates from different bacterial species were analyzed (see Table 1). The assay turned out to be highly specific for emetic strains. Thirty out of thirty cereulide producing B. cereus isolates were detected by the assay while no non-cereulide producing isolates gave any signal in the PCR assay. To prove the absence of PCR inhibiting substances or any inadequacies of the PCR assay, selected isolates were amplified in parallel with universal primers targeting the 16S rDNA gene (Fig. 3). No cross-reaction was observed with emetic-like strains which share biochemical properties with emetic strains. Recently, it has been shown that these strains, which do not produce cereulide, cluster together with emetic strains by genetic as well as by phenotypic methods .
The sensitivity of the assay was determined using 10-fold serial dilutions from purified genomic DNA of the emetic reference strain F4810/72. The detection limit was found to be 2.5 pg template DNA for a 30-cycle PCR protocol, which corresponds to approximately 500 copies of the B. cereus genome (deduced from the genome size of B. cereus ATCC 14579T). In order to enhance the sensitivity of the assay the number of cycles was increased. For PCR assays with 40 cycles, the sensitivity was increased 100-fold (approx. five genome copies) while the time for the PCR reaction was increased by approximately 30 min (data not shown).
When the target nucleic acid consisted of chromosomal DNA from B. cereus group members or from mixtures of food-borne pathogens, the primers EM1F and EM1R amplified a fragment only in the presence of emetic B. cereus DNA (Fig. 4). Finally, PCRs using constant amounts of template DNA from a non-emetic strain and 10-fold serial dilutions of emetic B. cereus template DNA were performed. We observed a detection limit of 2 pg DNA for emetic B. cereus DNA in the presence of excessive amounts of DNA (200 ng) from a non-emetic B. cereus strain in a 40-cycle PCR.
Based on its structure and function, the emetic toxin cereulide may be synthesized enzymatically by a NRPS, similar to other natural peptides and depsipeptides . Therefore, a PCR based screening assay using degenerate primers was targeted to known NRPS sequences. In this approach we identified a genomic DNA fragment that was specifically present in emetic toxin producing B. cereus strains. Surprisingly, the sequence of this fragment showed no homologies to NRPS sequences from databases. The coding potential of this genomic DNA fragment remains unknown since a homology search using BLAST  revealed no clear homologies to other sequences in databases. Nevertheless, Southern blot hybridization, analysis of sequenced B. cereus and B. anthracis genomes and PCR tests of 178 bacterial strains showed that the genomic DNA fragment is specific for emetic toxin producing B. cereus isolates (Fig. 2). Its sequence was therefore used to develop a PCR based molecular assay for the detection of emetic toxin producing B. cereus isolates.
This simple and rapid PCR assay represents an attractive alternative to used methods for the detection of emetic toxin producing B. cereus. Currently, there are three possibilities to detect the emetic toxin: via cytotoxicity, HPLC-MS and using a sperm based bioassay [20,30,31]. These assays are rather difficult to perform on a routine basis and require one day to one week with precultivation and laborious sample preparations. With the exception of HPLC-MS, these assays do not specifically detect cereulide; e.g., the bioassay is also sensitive to other mitochondrial toxins like gramicidin .
Using the PCR test, emetic B. cereus was specifically detected in high background microbial flora. The observed tolerance of the primers is important for direct detection of B. cereus in foods. B. cereus food poisoning is often the consequence of improper food storage, therefore a high microbial background bacterial flora is expected to occur in food samples. The potential of the developed PCR system for a direct detection of B. cereus in foods is currently under investigation and the general applicability of the primer system described is being tested in a routine lab.
The gastrointestinal diseases caused by B. cereus are symptomatically similar to other types of food poisonings. Emesis caused by B. cereus strains cannot be differentiated symptomatically from intoxications with enterotoxigenic S. aureus. Only in 2% (1 out of 50) of food poisoning cases in which S. aureus had been suspected to be the causative agent was S. aureus identified. In the majority of the cases the causative agent remained unknown (Dietrich Mäde, Halle, Germany, personal communication). However, B. cereus has been reported to be the main cause of food-borne disease in the mass catering of the German Federal Armed Forces from 1985 to 2000. It was responsible for 30% of the cases and 42% of the outbreaks . While for the detection of S. aureus suitable molecular assays using various genes as targets have been described (e.g., [32,33]), the presented PCR assay is the first molecular tool for identification of emetic B. cereus.
In conclusion, we have developed a highly sensitive and test strain specific molecular assay that allows a rapid detection of emetic toxin producing B. cereus strains. In addition, the PCR based detection system presented will contribute to the development of multiplex PCR systems to detect the full set of toxin genes involved in gastrointestinal diseases caused by B. cereus. Such molecular assays might also be useful to clarify the role of enterotoxins and the emetic toxin producing B. cereus strains in non-gastrointestinal infections.
This work was supported by the European Commission (QLK1-CT-2001-00854). The technical assistance of Verena Pfund and help from Natasa Vukov in the PCR based screening assay is greatly appreciated. We are grateful to Erwin Märtlbauer (Institute of Hygiene and Technology of Food of Animal Origin, Ludwig-Maximilians-Universität, Munich, Germany) who provided us with cultures from recent food poisoning cases in Germany. In addition, we thank Gilles Vernaud (Institut de Génétique et Microbiologie, Paris, France) for the gift of B. anthracis DNA and Dietrich Mäde (Landesanstalt für Verbraucherschutz, Halle, Germany) for making available unpublished observations.