Enterotoxin profiles for Bacillus cereus group cells
As described earlier, B. cereus group cells, especially B. cereus and B. thuringiensis, share many phenotypical properties and relatively high levels of chromosomal base sequence similarity ( Ash et al. 1991a,b ), thus it is interesting to know the enterotoxin profiles for these organisms. Table 3 shows the results for the PCR detection of different enterotoxin genes, the results of immunoassay using enterotoxin specific kits, the patterns of haemolysis on the blood agar plates and the cytotoxicity to the CHO cells for the B. cereus group cells including 58 B. cereus strains isolated from foods and samples associated with food-borne outbreaks.
Table 3. Detection of the presence of enterotoxins and their genes as well as cytotoxicity for Bacillus cereus group cells
|Profile and strains||Total strains ||Pattern of haemolysis †||HblA1/ HblA2 ‡||BCET- RPLA ||ETF/ ETR ‡||ENTA/ ENTB ‡||CHO cell cytotoxicity §||Ph1/Ph2 ‡|
| CCRC strains, BC1, BC3, BC5, BC9, BC10, BC14 and BC24 ||7|| || || || || || || |
| Outbreak-associated source, BCN29(h) *, BCN30(h), BCN34(h), BCN35(k), BCN41(h) and BCN42(h) ||6|| || || || || || || |
| CCRC strains, BC8, BC19, BC20, BC25 and BC26 ||5|| || || || || || || |
| Food isolates, BCN12||1|| || || || || || || |
| Food isolates, BCN27 and BCN28||2|| || || || || || || |
| Outbreak-associated source, BCN53(c)||1|| || || || || || || |
| CCRC strain, BC11||1|| || || || || || || |
| Food isolates, BCN1||1|| || || || || || || |
| CCRC strain, BC2||1|| || || || || || || |
| CCRC strain, BC13||1|| || || || || || || |
| CCRC strain, BC12||1|| || || || || || || |
| Food isolate, BCN21||1|| || || || || || || |
| CCRC strains, BC4, BC6, BC7, BC16, BC17, BC21, BC22 and BC23 ||8|| || || || || || || |
| Food isolates, BCN2, BCN5, BCN6, BCN7, BCN9, BCN10, BCN14, BCN15, BCN20, BCN22, BCN23, BCN24 and BCN25 ||13|| || || || || || || |
| Outbreak-associated sources, BCN31(h), BCN33(h), BCN38(c), BCN43(h), BCN45(h), BCN46(h), BCN47(h), BCN48(h), BCN49(h), BCN50(h), BCN51(h), BCN52(w), BCN54(k), BCN55(d), BCN56(c) and BCN57(h) ||16|| || || || || || || |
| CCRC strain, BC18||1|| || || || || || || |
| Food isolates, BCN3, BCN4, BCN8, BCN11, BCN13, BCN16, BCN17, BCN18, BCN19 and BCN26 ||10|| || || || || || || |
| Outbreak-associated sources, BCN32(h), BCN36(h), BCN37(k), BCN39(h), BCN40(h), BCN44(h) and BCN58(h) ||7|| || || || || || || |
| CCRC strain, BC15|
| B. cereus group cells except B. cereus|
| BT3, BT4, BT5, BT6, BT7, BT8 and BT9 |
| BA1 and BA2|
Primers HblA1/HblA2 were designed for the specific detection of the hblA gene encoding the B component of HBL of B. cereus group cells ( Mäntynen & Lindström 1998). Only 26 of the 84 B. cereus cells collected in our laboratory generated the expected PCR product with molecular weight equal to 876 bp. Of the 28 B. cereus strains isolated from food samples, four strains were hblA PCR-positive while, of the 30 B. cereus strains isolated from samples associated with outbreaks, seven strains were hblA PCR-positive. Thus, no significant difference was found for strains from these two sources. Seven B. thuringiensis strains and two B. mycoides strains also generated the expected PCR products. However, two B. anthracis strains tested were found to be PCR-negative.
Beecher & Wong (1994a) reported that the discontinuous haemolytic pattern produced on sheep and calf blood agar was a distinct characteristic for HBL. In this study, the haemolytic patterns for all B. cereus group strains collected in our laboratory were analysed. Twenty-three of the 26 hblA PCR-positive B. cereus strains showed the discontinuous haemolytic patterns. The remaining three strains showed continuous patterns. All the hblA PCR-negative B. cereus strains showed continuous haemolytic patterns. However, seven of the eight hblA PCR-positive B. thuringiensis strains and two hblA PCR-positive B. mycoides strains also showed the discontinuous haemolytic patterns. In contrast, a haemolytic zone on sheep blood agar was not found for B. anthracis strains ( Table 3). Thus, the presence of the hblA gene seems to be highly correlated to the discontinuous haemolytic patterns as reported by Beecher & Wong (1994a).
When the B. cereus and B. cereus group strains shown in Table 3 were assayed with the BCET-RPLA enterotoxin kit (Denka Seiken), it was found that all the 36 hblA PCR-positive strains, including 26 B. cereus strains, eight B. thuringiensis strains and two B. mycoides strains, were also enterotoxin-producing strains, except for strain BC13 which was hblA PCR-positive but BCET-RPLA-negative. All the hblA PCR-negative strains were negative as assayed with this enterotoxin kit. The above observations indicate that the results from the BCET-RPLA assay agree with the results of the hblA PCR assay except for strain BC13. In addition, several B. cereus and other B. cereus group strains, such as strains BC2, BC11 and BT2, although they were hblA PCR- and BCET-RPLA-positive, showed the continuous haemolytic pattern on the blood agar plate. Although it has been reported that the discontinuous haemolytic pattern was only transient for B. cereus cells ( Beecher & Wong 1994a), discontinuous patterns were not found for strains BC2, BC11 and BT2 throughout the whole assay process.
Beecher & Wong (1994b) reported that the BCET-RPLA kit detects the L2 component of HBL, but Mäntynen & Lindström (1998) reported that the primers HblA1/HblA2 used here detected the B component of the hblA gene. Bacillus cereus strains deficient in part of the HBL components have also been reported ( Mäntynen & Lindström 1998). Strain BC13 might be L2 component-deficient and thus was PCR-positive but BCET-RPLA negative. It may also be possible that the expression level of the hblA gene from strain BC13 was below the detection limit of the BCET-RPLA kit. For strains BC2 and BC11, although both the PCR and RPLA assays showed positive results, the concentrations of each HBL component produced might not be enough or the ratio of each component might not be correct to generate the discontinuous haemolytic patterns. Formation of the discontinuous pattern might need a minimum concentration and defined ratios for each component of the HBL enterotoxin ( Beecher & Wong 1994b). The HBL is a tripartite enterotoxin that requires all three components for maximum activity ( Beecher et al. 1995 ). Furthermore, polymorphism of the haemolysin B-component gene within the amplified PCR products has been found for the B. cereus group cells ( Mäntynen & Lindström 1998). Therefore, it could also be possible that the HBL genes for strains BC2, BC11, BT2 and BC13 were heterogeneous and thus showed different haemolytic activities on the blood agar plate. For the haemolysis study, we have found that the time required to form the haemolytic zone was not the same for different B. cereus strains.
When the ETF/ETR primers reported by Agata et al. (1995) were used to amplify the bceT genes of B. cereus group strains, 41 of the 84 B. cereus strains, two of the three B. mycoides strains and eight of the nine B. thuringiensis strains generated the expected PCR products with molecular weight equal to 428 bp. Granum et al. (1996) have reported that bceT is not present in five of the seven food-poisoning B. cereus strains. Mäntynen & Lindström (1998) tested 80 strains of Bacillus spp., including 58 B. cereus group cells, and found that only a bceT gene model strain, i.e. strain B-4ac, was positive in amplification of the bceT gene. None of the B. cereus strains they collected showed the expected bceT PCR product. In contrast, our study shows that the ratio for B. cereus strains carrying the bceT gene was much higher, i.e. 49% of the total of 84 B. cereus strains. In addition, 14 (50%) of the 28 food isolates and 17 (57%) of the 30 outbreak-associated strains were bceT PCR-positive.
ENTA and ENTB primers have been used to amplify the entire entFM gene ( Asano et al. 1997 ). The 45 kD enterotoxin (EntFM) isolated from B. cereus strain FM-1 has been shown to be cytotoxic to the Vero cells, although none of the haemolytic activity has been found ( Shinagawa et al. 1991 ). As shown in Table 3, the 1269 bp PCR products amplified from the entFM gene were found for 78 of the 84 B. cereus strains, one of the three B. mycoides strains and seven of the nine B. thuringiensis strains. In addition, 27 of the 28 food isolates and all 30 outbreak-associated strains were entFM PCR-positive. Thus, the entFM gene was found to be the most prevalent enterotoxin gene for B. cereus group cells.
Although it is stated that EntFM is cytotoxic but not haemolytic, some strains, such as strains of enterotoxin profile VIII, show continuous haemolytic patterns ( Table 3). Since haemolytic activity could be due to other factors, such as haemolysins H-I and H-II ( Coolbaugh & Williams 1978), lecithinase or phospholipase C ( Titball 1993), haemolysin II, cereolysin or sphingomyelinase ( Granum 1994), continuous haemolytic patterns for strains of enterotoxin profile VIII could be expected. Since investigation regarding the distribution of the entFM gene in B. cereus group strains has not been reported by other laboratories, our data could not be compared with those of other laboratories.
When CHO cells were used to analyse the cytotoxicity of B. cereus group cells, all the B. cereus group strains, except one B. cereus strain, i.e. BC15, and two B. anthracis strains, showed cytotoxicity to the CHO cells. None of the enterotoxin genes was found for the B. cereus BC15 and B. anthracis strains which showed no cytotoxicity. Our results implied that the CHO cell cytotoxicity assay showed a positive result if any of the three types of enterotoxin genes was present in the B. cereus group cells. Finally, we have to point out that a similar toxigenic profile could be found amongst strains of B. cereus, B. mycoides and B. thuringiensis. For example, the toxigenic properties for B. cereus strains BC8, 19, 20, 25, 26 and BCN12 and B. mycoides strain BMY2 are the same. The toxigenic profile for B. thuringiensis strains BT3–BT9 and B. cereus strains BC1, 3, 5, 9, 10, 14, 24, BCN29, 30, 34, 35, 41 and 42 are the same. Thus, from the viewpoint of food safety, a method for the detection of all B. cereus group cells rather than only the B. cereus spp. is important. As for the virulence profiles for B. cereus strains isolated from food samples and samples associated with food-poisoning outbreaks, a clear difference was not found between these two sources.
Detection of Bacillus cereus group cells in food samples
Food samples selected for PCR detection were those subjected to a high risk of contamination by B. cereus group cells. These samples were whole milk, cooked rice, egg and pork purchased from local food markets. When such food samples, e.g. milk, were artificially contaminated with B. cereus cells (strain BC3) and incubated in BHIG broth for 8 h, as few as 100 cells g−1 sample could be detected ( Fig. 2). Similar results could be obtained for other food samples, such as cooked rice, eggs and pork ( Table 4). Samples without target cell inoculation showed negative results as assayed by the PCR method and the conventional method using mannitol-egg yolk-polymyxin agar as the culture medium. These food samples, however, were contaminated by endogenous microflora as evidenced by the PCA counting. For these samples, primers Ph1/Ph2 allowed a detection limit of 100 cfu g−1 sample ( Fig. 2).
Figure 2. Sensitivity of the detection of Bacillus cereus cells in milk and cooked rice samples. Lanes: 1 and 9, polymerase chain reaction (PCR) results for blank without inoculation of target cells; 8, 100 bp ladder marker; 2–7, PCR results amplified from 100−105 cfu target cells ml−1 whole milk; 10–15, PCR results amplified from 100−105 cfu target cells ml−1 cooked rice
Download figure to PowerPoint
Table 4. Sensitivity for the polymerase chain reaction (PCR) detection of the sphingomyelinase gene of Bacillus cereus in food samples after pre-enrichment using Ph1/Ph2 primers
|Sample||lContamination evel (cfu g−1) *||MYP †||0||Inoculation level (cfu g−1) N§ × 100||N × 101||N × 102||N × 103||N × 104||N × 105|
|Cooked egg||N§ × 100||−||−||+||+||+||+||+||+|
|Cooked rice||N × 102||−||−||+||+||+||+||+||+|
|Cooked pork||N × 103||−||−||+||+||+||+||+||+|
In conclusion, not only the B. cereus spp. but also the B. cereus group cells are potentially toxigenic. Thus, a method for the detection of all B. cereus group cells in food samples is important. In this report, we have designed a set of novel PCR primers which allows the specific detection of all the B. cereus group cells in food samples.