Psychrotolerant species from the Bacillus cereus group are not necessarily Bacillus weihenstephanensis


  • Lotte P Stenfors,

    1. Department of Pharmacology, Microbiology and Food Hygiene, The Norwegian School of Veterinary Science, P.O. Box 8146 Dep., N-0033 Oslo, Norway
    Search for more papers by this author
  • Per Einar Granum

    Corresponding author
    1. Department of Pharmacology, Microbiology and Food Hygiene, The Norwegian School of Veterinary Science, P.O. Box 8146 Dep., N-0033 Oslo, Norway
      *Corresponding author. Tel.: +47 (22) 964845; Fax: +47 (22) 964850, E-mail:
    Search for more papers by this author

*Corresponding author. Tel.: +47 (22) 964845; Fax: +47 (22) 964850, E-mail:


Twenty-six strains of Bacillus cereus from different sources were determined to be either mesophilic or psychrotrophic by growth at 6 and 42°C. The strains were also screened by two polymerase chain reaction (PCR) methods designed to discriminate between mesophilic and psychrotrophic types. Seventeen of the 26 strains were able to grow at 6°C, but only four conformed to the new psychrotolerant species Bacillus weihenstephanensis. Among the 26 strains were two which caused outbreaks of food poisoning in Norway, and three others that were isolated from food suspected of causing illness. The presence of the gene components encoding production of enterotoxins Nhe, Hbl, EntT and a recently described cytotoxin K was determined by PCR. All the strains possessed genes for at least one of these toxins, and 19 of the 26 strains were cytotoxic in a Vero cell assay. We conclude that there are psychrotrophic B. cereus strains which cannot be classified as B. weihenstephanensis, and that intermediate forms between the two species exist. No correlation between cytotoxicity and the growth temperature of the strains was found.


Psychrotolerant (psychrotrophic) bacteria are widespread in nature. A large proportion of our planet has a cold climate or is subjected to wide temperature fluctuations, and psychrotrophic bacteria are well adapted to exist in such habitats [1]. Refrigerated foods are a new niche for such bacteria and they have indeed become a problem for the food industry both as spoilage flora and as foodborne pathogens [2]. Bacillus cereus is not traditionally considered a psychrotrophic species [3,4]. During the last decade there have been many reports on the occurrence of cold-tolerant strains belonging to the B. cereus group [5–7]. A new species, Bacillus weihenstephanensis, has been proposed to accommodate these psychrotrophic strains of B. cereus[8]. Although B. cereus has been recognized as an organism causing food poisoning since the 1950s, a full understanding of the pathogenicity mechanisms in B. cereus food poisoning has not yet been reached. These bacteria produce a number of putative virulence factors including enterotoxins, of which at least three have been characterized [9,10].

In this study we investigated whether psychrotrophic strains of B. cereus had an equivalent ability to produce enterotoxins as mesophilic strains, and if there was a correlation between the presence of enterotoxin genes, or level of cytotoxicity, and the growth temperature phenotype. Moreover we have characterized the growth temperature range of our strains and examined them using two polymerase chain reaction (PCR)-based methods designed to discriminate between mesophilic B. cereus and psychrotrophic B. weihenstephanensis.

2Materials and methods

2.1Bacterial strains

Twenty-six strains of B. cereus were investigated (Table 1). Eighteen strains were isolates from Norwegian dairy products, 15 of the dairy strains had previously been provided to our laboratory by Norwegian Dairies Association for a study carried out in 1992 [11]. PHLS Food Hygiene Laboratory, London, UK, provided two clinical strains for the same study, while six strains were isolates from foods, submitted to the National Reference Laboratory for B. cereus at the Norwegian School of Veterinary Medicine, Oslo, Norway. These isolates were identified by the reference laboratory as described by the Nordic Committee on Food Analysis No. 67, 4th edn., 1997. In the latter group a reference strain (1230-88) was included. All strains were stored at −80°C in a meat broth with glycerol.

Table 1.  The B. cereus strains used in this study and their growth data
Strain No.SourceGrowth at 42°CGrowth at 6°CPCR cspA (product)aPCR 16S rDNA (product)b
  1. aThe cspA PCR assay amplifies a 160-bp product from psychrotrophic strains (+).

  2. bIn the 16S rDNA PCR assay, a 249-bp product is amplified from mesophilic strains (M), and a 132-bp product is amplified from psychrotrophic strains (P). (For growth and PCR conditions, see Table 2 and text.)

72semi-skimmed milk++M+P
87whipped cream+M+P
95dairy product+M
96dairy productM+P
F4501-83clinical (PHLS)+++M+P
15unknown (PHLS)+M+P
1230-88stew (food poisoning)+M
401-92scrambled eggs++P
0075-95stew (food poisoning)+M+P
674-98scrambled eggs++M+P

2.2Growth profiles

Growth at 6 and 42°C was determined on Plate Count Agar (Difco). Plates were inoculated from B. cereus-selective agar plates that had been incubated at room temperature. Plates were monitored for visible colonies for 7 days. Growth studies were also repeated on blood agar (strains 100, 1230-88, 0075-95 and 674-98 were only tested on blood agar). Isolates were inoculated on blood agar directly from the −80°C stocks. Before incubation at the different temperatures, the blood agar plates were left at room temperature for approximately 2 h.

Some strains gave uncertain results and were additionally tested in liquid culture (Brain Heart Infusion broth, Difco) with agitation (at 42°C: approximately 100 rpm, and at 6°C: 180 rpm). Cultures were monitored for 1 week for visible growth (Table 1).


All PCRs were carried out in an MJ Research Minicycler™ PTC-150 with a heated lid or an MJ Research Minicycler™, with the latter using a layer of liquid paraffin on top of the reaction mixture. DyNAzyme II DNA polymerase (supplied with 10× buffer) and dNTP Mix from Finnzymes were used (Cat. No. F501L and F560L, Finland) as instructed by the manufacturer.

The primers and annealing temperatures used in the reactions are listed in Table 2. The primer pair BcAPF1/BcAPR1 amplifies a 160-bp fragment of the major cold-shock protein gene, cspA, from psychrotrophic strains only. In the 16S rDNA PCR assay, the primer pair MF/UR amplifies a 249-bp fragment from mesophilic strains only, while from psychrotrophic strains a 132-bp fragment is produced using the primer pair PR/UF. These two PCR assays were designed as a rapid way of discriminating mesophilic and psychrotrophic strains of the B. cereus group [14,15].

Table 2.  Primers used in PCR
Primer nameSequence (5′–3′)Gene amplifiedAnnealing temperature (°C)Sequence position/accession number (5′–3′)
  1. aPrimer pair BcAPF1/BcAPR1 amplifies a 160-bp product from psychrotrophic strains only.

  2. bPrimer pair MF/UR amplifies a 249-bp product from mesophilic strains, while primer pair PR/UF amplifies a 132-bp product from psychrotrophic strains.

  3. cFor five strains, test was repeated at 45°C.

  4. dFor one strain, test was repeated at 45°C.

BcAPR1cttyttggccttcttctaa  182–164/X93039
MFataacattttgaaccgcatg16S rDNAb52131–149/Z84576
PRgagaagctctatctctaga  989–970/Z84578
UFcaaggctgaaactcaaagga  861–880/Z84576
URcttcatcactcacgcggc  377–359/Z84578
4091Rccatatgcatttgtaaaatctgc  711–689/Y19005
7173FgctctatgaactagcaggaaacnheA 643–664/Y19005
7174Rgctacttacttgatcttcaacg  1203–1183/Y19005
8368Rgatcccattgtgtaccattgg  2512–2492/Y19005
1141Rcgacttctgcttgtgctcctg  3594–3614/Y19005
Rcgaatacataaataattggtttcc  2451–2429/AJ277962
F3aacagatatcggtcaaaatgc  1858–1878/AJ277962
4Rctccttgtaaatctgtaatccct  2041–2020/L20441
TR1tgtttgtgattgtaattcagg  1781–1761/D17312

Varying cycling programs were used for the different PCRs. The standard program was as follows: 95°C for 1 min, 30 cycles of 95°C for 1 min, annealing temperature for 1 min and 72°C for 1 min, then a final extension step of 72°C for 7 min.

For the cold-shock protein gene (cspA) and 16S rDNA PCR we initially used target DNA from bacterial colonies boiled in distilled water, but because not all strains yielded PCR products, the amplifications were repeated using genomic DNA as template. All toxin PCRs were carried out using genomic DNA. The DNA was isolated using ADVAMAX? beads as described in the supplied protocol, or by following a method described by Pospiech and Neumann [12]. PCR products were visualized by agarose (1–2%) gel electrophoresis using Seakem?, GTG? or Metaphor? agarose from FMC BioProducts USA.

2.4Cytotoxicity test on Vero cells

Cytotoxicity was tested using a Vero cell assay according to Sandvig and Olsnes [13]. The assay monitors inhibition of protein synthesis in the Vero cells after addition of the toxic culture supernatant.

Toxins were produced by growth of the B. cereus strains in Brain Heart Infusion broth (Difco) with 10 g of glucose added l−1 (BHIG). Test tubes containing 5 ml of BHIG were inoculated and incubated at 32°C with agitation (approximately 90 rpm) overnight. Cultures (0.5 ml) were transferred to Erlenmeyer flasks containing approximately 50 ml of BHIG, which were then incubated for 6 h at 32°C with agitation (100 rpm). Crude toxin was harvested by centrifugation (12 000×g at 4°C for 20 min); the supernatants were transferred to Eppendorf tubes and stored at −20°C before measuring cytotoxic activity. All strains were tested in duplicate; two parallels each of 10 μl (five strains) or 30 μl and 100 μl supernatant were tested. Strains with low activity (inhibition of protein synthesis 20–30%), as well as negative strains, were tested again using 10-fold-concentrated toxin. Supernatants were concentrated by precipitation with 80% ammonium sulfate, the precipitates were collected by centrifugation at 12 000×g for 20 min, dissolved in 20 mM phosphate buffer and dialyzed against the same buffer overnight.

2.5DNA sequencing

To confirm the presence of the cytotoxin K (CytK) gene, part of this gene was sequenced from strain 23. Sequencing reactions were performed on PCR products which had been amplified using primers F3 and Rc (Table 2) and the standard PCR program with an annealing temperature of 48°C. PCR products were purified using a QIAquick™ PCR Purification Kit (Cat. No. 28104). Sequencing reactions were done with primer Fc (Table 2) and using the ABI Prism BigDye™ Terminator Cycle Sequencing kit following instructions from the manufacturer. Sequencing was performed on a Perkin Elmer ABI Prism 377 automatic sequencer.

2.6Detection of the L2 component of hemolysin BL (Hbl)

The BCET-RPLA B. cereus Enterotoxin (diarrheal type) Test Kit (Cat. No. TD 950) from Oxoid, UK, was used as instructed by the manufacturer. Isolation of toxin for the test is described in Section 2.4.


3.1Psychrotolerance and growth profiles

All strains were separated into psychrotrophic/psychrotolerant or mesophilic groups by their ability to grow at 6 and 42°C (Table 1). Two PCR-based methods to discriminate mesophilic strains from psychrotrophic (B. weihenstephanensis) have been developed by Francis et al. [14] and von Stetten et al. [15]. One is based on the amplification of a segment of the cold-shock protein A gene (cspA), in which only psychrotrophic strains have the specific signature sequence that makes amplification possible. The other method takes advantage of specific sequence differences between mesophilic and psychrotolerant strains in the 16S rDNA. When screening our strains according to these methods four strains (43-92, 132, 401-92 and 453-92) showed growth and PCR amplification results consistent with the definition of the species B. weihenstephanensis[8] and should be reclassified as such. The definition includes the ability to grow below 7°C but not at 43°C, amplification of the 160-bp product from cspA, amplification of a 132-bp fragment from 16S rDNA using primers PR/UF, while primers MF/UR should not produce a fragment. Additionally, 17 strains were able to grow at 6°C but without showing the other characteristics of B. weihenstephanensis. Two of these strains (17 and 674-98) were not able to grow at 42°C, but yielded both a mesophilic (using primers MF/UR) and a psychrotrophic (using primers PR/UF) PCR product from the 16S rDNA amplifications. The strains 14, 36, 67, 68, 80, F4501-83 and 261-92 could grow at high and low temperatures and were PCR-positive for cspA and both types of 16S rDNA fragments. Only three strains were strictly mesophilic in the sense that just the mesophilic type (249-bp fragment) of 16S rDNA was amplified and that they were negative in the cspA PCR (no fragment amplified using primers BcAPF1/BcAPR1). Of the 26 strains studied, both types of 16S rDNA could be amplified from 19 strains. Interestingly, four strains (55, 59, 61 and 72) of the mesophilic genotype grew at 6°C. Strain 96 did not grow at 6°C or at 42°C but both of the possible PCR fragments were amplified from 16S rDNA.

3.2Cytotoxicity and toxin genes

It is believed that psychrotolerant strains of B. cereus are rarely able to produce enterotoxins at high levels since dairy products, from which most psychrotolerant strains are isolated, are not frequently reported to be involved in food poisoning [16]. In this study we found that seven of the 17 strains that were able to grow at 6°C, also showed more than 90% inhibition of protein synthesis in the Vero cell assay (strains 14, 17, 67, 68, 80, 453-92 and 674-98). This level of cytotoxicity is usually seen for food poisoning strains (Norwegian reference laboratory for B. cereus). Another three psychrotolerant strains (72, F4501-83 and 261-92) produced enterotoxins to a level of 50–90% protein synthesis inhibition and three strains (36, 55 and 132) to a level of 20–50% (Tables 1 and 3).

Table 3.  Toxin genes and toxicity
StrainPCR nheaPCR bceTPCR hblL2 (HblC) OxoidPCR cytKCytotoxicity (Vero cells)
  1. Vero cell assay: +++: >90% inhibition of protein synthesis using 100 μl supernatant in assay; ++: 50–90% inhibition of protein synthesis using 100 μl supernatant in assay; +: 25–50% inhibition of protein synthesis using 100 μl supernatant in assay; −: <20% inhibition of protein synthesis using 100 μl supernatant in assay. See text for annealing temperatures and PCR conditions.

  2. aPCR products from all three gene components coding for the NHE protein could be amplified from all the strains.


All of the strains that were screened in this study possessed the gene components for producing one or several of the B. cereus enterotoxins (Table 3). The hemolysin BL (Hbl) was the earliest described of the enterotoxins involved in B. cereus diarrheal food poisoning [17–20]. Approximately one third of the strains in this study yielded a PCR product using primers from the hblD (transcribing L1 component) and into hblA (transcribing the B component) of the hbl operon. The third Hbl component (L2) was detected in all PCR-positive strains using the Oxoid BCET-RPLA kit, as well as in three strains that did not come out positive on PCR for hblA and hblD.

The genes encoding the three components of the non-hemolytic enterotoxin (Nhe) were present in all of the strains, indicating that these genes are widespread among strains of B. cereus. This enterotoxin was first purified from one of the strains also used in this study (0075-95) after a large outbreak of food poisoning in Norway [20,21].

A PCR product was amplified from four strains using primers constructed from the sequence of a recently discovered toxin, CytK, which was found in a B. cereus strain that caused an incident of food poisoning with three lethality cases in France [22]. To confirm that the amplified product was correct, part of the cytK gene was sequenced from strain 23 (results not shown). Approximately 400 bases of sequence were obtained from the early part of the gene (from residue 58 of the mature protein). By BLAST comparison [23] with the sequence from the French strain (accession number AJ277962) we found a high degree of similarity (89%). In addition, a similarity of 42% with the hemolysin II from B. cereus was found.

The possible significance of the enterotoxin T in B. cereus foodborne illness has not been established. This toxin is a single protein, which was identified by Agata et al. [24]. A recently published study indicates that the bceT gene is widely distributed among B. cereus strains, and that there is strain variation in the gene sequence [25]. In our assay, 14 of 26 strains possessed the bceT gene, detected using one set of primers (Table 3).


In 1998 the species B. weihenstephanensis was suggested as the sixth species within the B. cereus group comprising psychrotolerant, but not mesophilic, B. cereus strains [8]. Isolates of this new species grow at 4–7°C but not at 43°C and can be identified rapidly using rDNA- or cspA-targeted PCR. In this study we have shown that there are problems using these criteria to differentiate between strains of B. cereus and B. weihenstephanensis. The main problem seems to be that the majority of the strains that we have looked at in this study are positive for both a mesophilic and a psychotropic PCR product using the 16S rDNA primers [15], as pointed out before by Prüss et al. [26]. Nineteen of our 26 strains were positive for both PCR products, while only three and four strains, respectively, were positive for the mesophilic and psychrotrophic product only (Table 1). The four strains that gave a psychrotrophic PCR rDNA product were all positive for the psychrotrophic cspA, and grew at 6°C but not at 42°C and should, according to the definition, be B. weihenstephanensis. However, the main problem is that the two strains (674-98 and 17) that grew at 6°C but not at 42°C were positive for the psychrotrophic cspA and positive for both the mesophilic and psychrotrophic rDNA. For the time being we will still call these two strains B. cereus. The last strain (96) did not grow at 42°C or at 6°C, contained both types of rDNA and was negative for the psychrotrophic cspA. However, 17 of the 26 strains grew at 6°C (Table 1) and all of them were PCR-positive for the psychrotrophic rDNA but four of the strains were PCR-negative for the psychrotrophic cspA. We can therefore conclude that B. cereus strains can grow at 6°C, and we have found that there are intermediate forms between B. cereus and B. weihenstephanensis based on the suggested criteria [8].

There is no correlation between cytotoxicity and growth temperatures, although two of the B. weihenstephanensis strains were non-toxic in our tests, one was low in cytotoxicity and the last strain was highly toxic (Table 3). Two of the three strict mesophilic strains were highly cytotoxic while the last was negative in the cytotoxicity test. Distribution of the four different cytotoxin (enterotoxin) genes we have screened for (Table 3) does not give a pattern linked to growth temperature. Of the 26 strains, all were positive for nhe, 14 positive for bceT and eight positive for hbl. For the hbl we have only used one set of primers, and mismatch in these areas is quite possible. We have therefore also used the Oxoid detection kit (detecting L2 or hblC) to show production of the first protein in the Hbl operon. This kit detected the L2 protein from another three strains. The five strains negative in the Vero cell test all contained genes for cytotoxins. The reason for lack of cytotoxicity in these strains may be due to missing components of Hbl and/or Nhe, or mutations in the genes that might have dramatically reduced biological activity. The results can also be explained by changes in the promoter areas or in the plcR gene, as seems to be the case in B. anthracis[27].

The newly described CytK, which might cause necrotic enteritis [22], was detected in four of the strains. PCR results were confirmed by sequencing the PCR product from strain 23, and we found that it was about 90% identical to the original sequence. However, this protein was shown to have much lower activity than the original CytK (results to be reported elsewhere).

In conclusion, strains of B. cereus that can grow at low temperature are common and these organisms may contain the full complement of enterotoxin genes.


We thank The Research Council of Norway (Grant 124097/110 to L.P.S.) for supporting this work. We also thank Marianne Skeie for her excellent technical assistance, and Felix von Stetten at the Forschungszentrum für Milch und Lebensmittel Weihenstephan for testing six of our strains for thermal genotype and phenotype.