1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

An improved McCoy cell cytotoxicity assay for Bacillus cereus diarrhoeal toxin, which includes a staining procedure in addition to visual examination, was developed and the method was compared with two commercially available kits (Oxoid BCET-RPLA and Tecra BDE-VIA). A total of 71 strains of 15 different Bacillus, Brevibacillus and Paenibacillus species, including 16 strains of B. cereus from outbreaks of food-borne illness, were evaluated for toxin production. Eleven of the outbreak strains exhibited cytotoxicity, including all six B. cereus strains associated with diarrhoeal illness. Several other isolates of B. cereus, and its relatives B. anthracis, B. mycoides and B. thuringiensis, exhibited similar cytotoxicity. The other species showed no cytotoxicity, with the exception of certain B. subtilis strains. The cytotoxicity assay was more sensitive than the Oxoid kit and unlike the Tecra kit, did not give false positive results with supernatant fluids heat-treated to destroy the toxin.

In the past, fluid accumulation in ligated rabbit ileal loops (RIL) and the vascular permeability reaction (VPR) in guinea-pigs was used successfully in the assay of Bacillus cereus diarrhoeal enterotoxin. More recently, the purification of what is believed to be the enterotoxin has led to the construction of two commercially available immunological kits: Bacillus Diarrhoeal Enterotoxin Visual Immunoassay (BDE-VIA) from Tecra and the Bacillus cereus Enterotoxin-Reversed Passive Latex Agglutination (BCET-RPLA) from Oxoid. There has, however, been some controversy surrounding these kits and the significance of their results ( Christiansson 1993; Beecher & Wong 1994) because the enterotoxin has only been identified provisionally and it is still unclear whether B. cereus diarrhoeal food poisoning is caused by the activity of a single protein enterotoxin ( Shinagawa et al. 1991 ), a multicomponent enterotoxin ( Thompson et al. 1984 ), or more than one multicomponent enterotoxin ( Lund & Granum 1997). Also, the two kits have been shown to detect different antigens ( Beecher & Wong 1994).

The use of cell culture in cytotoxicity assays has also been widely used to detect the diarrhoeal enterotoxin. Thompson et al. (1984) showed that a fraction containing three detectable antigens purified from the culture fluid from B. cereus caused fluid accumulation in the RIL test, gave a positive VPR and was cytotoxic to Vero cells. From this, they suggested that all three biological activities were due to a single entity. In view of this, Kawano et al. (1987) suggested, on the basis of the relation between cytotoxicity and the VPR, that cytotoxicity assays may be used for the analysis of B. cereus enterotoxins. However, most such assays have been tested with a limited number of Bacillus species and are somewhat subjective, being reliant upon visual examination for the determination of results.

In this study, the McCoy cell-based assay described by Jackson (1993) was examined for specificity and compared with the commercial kits using a range of authenticated Bacillus, Brevibacillus and Paenibacillus strains, with a view to producing an improved cytotoxicity assay (ICA) and removing reliance upon visual examination.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

Bacterial strains

Seventy-one strains of Bacillus, Brevibacillus and Paenibacillus species identified by the methods of Logan & Berkeley (1984) were tested: B. cereus from emetic outbreaks (serovars in parentheses, NT = non-typeable): F4810/73 from vomit (1), F4562/75 pancake (1), F4552/75 vomit (3), F2875/77 rice (12), F2549 A/76 pancake (5); from diarrhoeal outbreaks: F4433/73 from meat loaf (2), F2057/77 meals-on-wheels (17), F289/78 pork pie (17), F284/78 pork pie (18), F2769/77 pâté (NT), F2668/90 lamb (11/15); from individual cases of food-borne illness: F3927/89 from ham (NT), F3053/89 quiche (1); strains associated with diarrhoea and vomiting: F4815/94 from food remnants (18), F352/90 chow mein (2); another outbreak strain F4397/92 from rice remnants (1); B. cereus from food (B0083, F3841/75, B1516, B1574, B1617), from clinical (F2038/78, B0688, B1269, B1281, B1283) and from veterinary cases (B0020, F1820/77, B0542, B0752); and the type strain of B. cereus ATCC 14579T; food and veterinary strains of B. anthracis (ASC 431, ASC 432) cured of their virulence plasmids; B. mycoides (LMG 7128T, B0040, B0087, B0520, B0950, B0952); B. thuringiensis (B0528T, B0015, B0021, B0153, B0157, B0161, B0754); strains of other species associated with food-borne and other illnesses ( Drobniewski 1993; Logan & Turnbull 1999 ; Ljungh, personal communication): B. licheniformis (ATCC 14580T, B0025, B0028, B0244, B1508), B. pumilus (B0701, B0768), B. sphaericus (B0121), B. subtilis (ATCC 6051T, B0009, B0223, B0226, B0263, B0268, B0774), Brevibacillus agri (L2 and L3, water-borne outbreak), Paenibacillus macerans (B0198, B0904), P. polymyxa (B0107, B0264); and B. amyloliquefaciens (B0500), B. megaterium (B0407) and P. thiaminolyticus (B0771). (ASC, Dr P.C.B. Turnbull, CAMR, Salisbury; ATCC, American Type Culture Collection, Maryland, USA; B, Logan Bacillus Collection, Glasgow Caledonian University; F, Food Hygiene Laboratory, PHLS, Colindale, London; L, Dr Åsa Ljungh, Lund University, Sweden.)

Toxin production

Strains were grown overnight in Brain Heart Infusion broth (Oxoid) supplemented with 0·1% glucose (BHIG) in an orbital incubator at 30 °C. Samples were centrifuged (900 g for 10 min) and passaged through a 0·45 μm low protein-binding filter (Nalgene) to remove bacterial cells. Each supernatant fluid was checked to be pH 7–8. Supernatant fluids from two strains isolated in connection with emetic outbreaks were also heated at 121 °C for 20 min to destroy the heat-labile diarrhoeal enterotoxin but not any of the heat-stable emetic toxin which might have been present.

Cytotoxicity assay

McCoy cell monolayers were prepared according to Jackson (1993) but diluted to a density of 6 × 104 cells ml−1 and grown to confluent monolayers at 37 °C under 5% CO2 ( Fig. 1a). The growth medium was Eagles’ minimal essential medium (Sigma) supplemented with 10% (v/v) foetal bovine serum, 1% (w/v) l-glutamine, 1% (v/v) non-essential amino acids and 1% (w/v) of an antibiotic/antimycotic mixture containing penicillin, streptomycin and amphotericin B (Imperial Laboratories). Supernatant fluids were double-diluted in maintenance medium (foetal bovine serum reduced to 2% (v/v)) to a final titre of 1/64 and added to the wells in triplicate, with sterile BHIG as the negative control for toxin production. Wells were observed for cytopathic effect at 2, 4, 6 and 20 h intervals using an Olympus CK2 inverted microscope (Olympus Optical Ltd, London, UK).


Figure 1. Monolayer of McCoy cells viewed at × 400. (a) control; (b) treated with a preparation of enterotoxigenic Bacillus cereus, viewed 30 min after incubation; (c) after 2 h incubation; (d) treated with a preparation of B. subtilis, viewed after 2 h incubation

Staining procedure

After standardization of the visual assay, a staining procedure was applied. After a 2 h incubation, and visual examination, plates were fixed and stained in accordance with Christiansson et al. (1989) but the cells were fixed with 2% (v/v) formalin. Absorbances were measured with a microplate reader (MRX, Dynatech Laboratories, Billingshurst, UK) at 630 nm. The appearance of a microtitre plate after staining is shown in Fig. 2. Four samples (in triplicate) were analysed in each microtitre plate. Row A shows supernatant fluid diluted 1/2 and Row F, 1/64. In Fig. 2, the first sample analysed (in triplicate) was the negative control for toxin production, sterile BHIG broth. The remaining three samples were B. cereus isolates producing cytotoxicity of varying degrees; G and H were control rows containing only maintenance medium. The optical density of the control wells were averaged and used to calculate monolayer damage (cytotoxicity). Test well results were expressed as fractions of the control wells and percentage change was calculated. Strains with results greater than 50% at 1/4 dilution were regarded as positive for this assay.


Figure 2. An example of a microtitre cytotoxicity assay plate after fixing and staining

Download figure to PowerPoint

Commercial assays

Two commercial assay kits, Oxoid BCET-RPLA (Unipath, Basinstoke, UK) and Tecra BDE-VIA (Tecra Diagnostics, Reading, UK), were used according to manufacturers’ instructions using the same culture supernatant fluid preparations as the ICA.

Results and discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

The earliest time at which an unambiguous cytopathic effect could be observed was after 2 h, at the 1/4 dilution. By this time, nine of the 16 B. cereus strains associated with outbreaks exhibited cytotoxicity, and some caused monolayer damage at dilutions up to 1/32, but three strains showed weak reactions. The staining procedure clarified these uncertain results and the results are compared in Table 1. Eleven outbreak strains were positive using the ICA; all six strains associated with diarrhoeal illness, two strains associated with emetic illness, one strain from a single case of vomiting, one strain associated with diarrhoea and vomiting, and one other outbreak isolate. The progress of McCoy cell monolayer destruction is shown in Fig. 1(a–c). Emetic strains positive in this assay may also be capable of producing diarrhoeal enterotoxin, and other workers have noted this ( Turnbull 1986); furthermore, two of the emetic outbreak isolates studied here (F4562/75 and F2875/77) were of serotypes 1 and 12, which have also been associated with diarrhoeal outbreaks ( Drobniewski 1993). When tested again after heat treatment, they caused no monolayer damage, implying that any of the heat-stable emetic toxin present was not cytotoxic in this assay.

Table 1.  Results of McCoy cytotoxicity assay at 1/4 dilution after 2 h of incubation
Visual examination
Species and sourcesNo. strains testedPositiveDoubtfulPositive in staining examination
  • *

    Results for heat-treated supernatant fluids of strains F4562/75 and F4397/92 shown in parentheses.

  • Bacillus amyloliquefaciens, B. firmus, B. licheniformis, B. megaterium, B. sphaericus, Brevibacillus agri, Paenibacillus macerans, P. polymyxa and P. thiaminolyticus gave negative results.

Bacillus cereus
food-borne illness16(2) *9311(0) *
other clinical5313
routine food5404
type strain1101
B. anthracis2021
B. mycoides6324
B. thuringiensis7707
B. subtilis7303

Six clinical and veterinary strains of B. cereus gave positive results with the ICA; strains from these sources were examined as it has been suggested that the toxin involved in food poisoning incidents is also involved in non-gastrointestinal infections ( Turnbull 1986).

Except for its insecticidal parasporal crystals, B. thuringiensis is known to elaborate the same range of extracellular metabolites as B. cereus. Spira & Goepfert (1972) found that B. thuringiensis was also capable of producing fluid accumulation in the RIL test, and noted that fluid accumulation caused by B. thuringiensis culture filtrate was neutralized by the antiserum to culture filtrates of B. cereus. This suggests that B. thuringiensis is capable of producing the same enterotoxin as B. cereus and explains the results observed in the present work ( Table 1). Moreover, work with DNA probes revealed that several outbreak B. cereus were actually B. thuringiensis ( teGiffel et al. 1997 ), and diarrhoeal food poisoning caused by B. thuringiensis has indeed been reported ( Jackson et al. 1995 ).

Four B. mycoides strains were positive in the ICA; this was not unexpected as strains of this species have also been shown to produce fluid accumulation thought to be associated with the diarrhoeal enterotoxin ( Jørgensen 1976).

It is unknown whether B. anthracis expresses the enterotoxin associated with B. cereus diarrhoeal syndrome; virulent strains produce the anthrax toxin which might be cytotoxic to our cell line, two avirulent strains (plasmid cured) were examined, and one strain was found to be weakly positive, indicating that further strains of this species should be analysed.

With the exception of three B. subtilis strains, strains of other species showed no cytotoxic reactions. However, the monolayer disruption pattern shown by B. subtilis was distinctly different in appearance ( Fig. 1d). The cells became rounded and detached from the monolayer but remained circular and not crenated, unlike B. cereus treated monolayers. Strains of B. subtilis have been tested for fluid accumulation and VPR, and it was concluded that these strains produce toxin to a very limited extent or not at all, or exhibit unusual reactions ( Turnbull 1986); nonetheless, there have been reports of food poisoning incidents associated with B. subtilis ( Kramer et al. 1982 ; Turnbull 1997). Occasional reports have also appeared implicating B. licheniformis, B. pumilus, B. sphaericus, P. macerans and P. polymyxa in gastrointestinal and other disease ( Logan & Turnbull 1999), but the few strains of these species studied in the present work caused no monolayer damage.

Fourteen B. cereus outbreak strains (F352/90 and F3927/89 not included) were tested by both commercial kits ( Table 2). All six of the diarrhoeal outbreak B. cereus strains were positive by both kits and the ICA. However, the results differed when outbreak strains not involved or not exclusively involved with diarrhoeal illness were tested. F4815/94 was positive in both kits and the ICA, F3053/89 was positive using the Tecra but not the Oxoid kit or ICA, and F4397/92 was positive in both Tecra and ICA but not the Oxoid kit. Of five strains associated with emetic illness, four were positive using the Tecra kit and none were positive using the Oxoid kit. This could be due to differences in sensitivities between the two commercial kits, as the Tecra kit is expected to be more sensitive than the Oxoid kit ( Day et al. 1994 ). Two of the isolates (F4562/75 and F2875/77) were also positive in the ICA. These isolates were tested again after heat treatment and were found to be negative in the ICA, but both still registered as being weakly positive using the Tecra kit, presumably because some of the toxin’s antigenic structure survives the heat treatment. Comparison of the results from the Oxoid and ICA methods also revealed that the Oxoid kit failed to detect five B. cereus positive for enterotoxin, including one outbreak isolate (F4397/92).

Table 2.  Comparison of Tecra VIA, Oxoid RPLA and ICA
Species and sourcesNo. strains testedPositive ICAPostive Tecra VIAPositive Oxoid RPLA
  1. *See legend to Table 1 for explanation of abbreviations.

  2. Bacillus licheniformis (three strains) gave negative results in all assays.

Bacillus cereus
food-borne illness14(2) *10(0) *13(2) *7
other clinical2110
routine food4342
type strain1111
B. mycoides2122
B. thuringiensis2222
B. subtilis4300

Bacillus mycoides and B. thuringiensis strains tested positive for B. cereus enterotoxin production with both kits, suggesting that some of their extracellular metabolites are very similar to those of B. cereus. None of the B. subtilis strains tested produced positive reactions with the kits and therefore, given the reactions with the ICA, this implies that any toxins produced by the B. subtilis strains are immunologically different to those of B. cereus and relatives.

Beecher & Wong (1994) reported that the Oxoid and Tecra kits were detecting different antigens, with the former detecting the L2 component of the haemolysin BL (HBL) multicomponent enterotoxin and the latter detecting two apparently non-toxic proteins. The value of detecting the L2 component alone is unknown but has resulted in false positive and false negative reactions ( Day et al. 1994 ). Recently, Lund & Granum (1997) discovered that the Tecra kit detects one protein of the newly described non-haemolytic enterotoxin (NHE) complex. These workers noted that B. cereus strains may produce both the HBL and the NHE complexes, while other strains may only produce one. Therefore, caution is needed when using one of these methods in isolation. Given the expense of the commercial kits and the uncertainties about the significance of their results, the ICA provides an effective and apparently specific method for the detection of strains of B. cereus and its close relatives which produce diarrhoeal toxin, provided that the results for other species of aerobic endospore-formers are supported by visual examination of the assay plates.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

The authors would like to thank all donors for providing strains for this work, especially Åsa Ljungh and Peter Turnbull for providing the Brevibacillus agri and Bacillus anthracis strains.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Acknowledgements
  6. References
  • Beecher, D.J. & Wong, A.C.L. 1994 Identification and analysis of the antigens detected by two commercial Bacillus cereus diarrhoeal enterotoxin immunoassay kits. Applied and Environmental Microbiology, 60, 4614 4616.
  • Christiansson, A. 1993 Enterotoxin production in milk by Bacillus cereus diarrhoeal enterotoxin: a comparison of methods for toxin detection. Netherlands Milk and Dairy Journal, 47, 79 87.
  • Christiansson, A., Naidu, A.S., Nilsson, I., Wadström, T., Petterson, H.E. 1989 Toxin production by B. cereus dairy isolates in milk at low-temperatures. Applied and Environmental Microbiology, 55, 2595 2600.
  • Day, T.L., Tatini, S.R., Notermans, S., Bennett, R.W. 1994 A comparison of ELISA and RPLA for detection of Bacillus cereus diarrhoeal enterotoxin. Journal of Applied Bacteriology, 77, 7 13.
  • Drobniewski, F.A. 1993 Bacillus cereus and related species. Clinical Microbiology Review, 6, 324 338.
  • TeGiffel, M.C., Beumer, R.R., Klijn, N., Wagendorp, A., Rombouts, F.M. 1997 Discrimination between Bacillus cereus and Bacillus thuringiensis using specific DNA probes based on variable regions of 16S rRNA. FEMS Microbiology Letters, 146, 47 51.
  • Jackson, S.J. 1993 Rapid screening test for enterotoxin-producing Bacillus cereus. Journal of Clinical Microbiology, 31, 972 974.
  • Jackson, S.J., Goodbrand, R.B., Ahmed, R., Kasatiya, S. 1995 Bacillus cereus and Bacillus thuringiensis isolated in a gastroenteritis outbreak investigation. Letters in Applied Microbiology, 21, 103 105.
  • Jørgensen, K. 1976 Demonstration of Enterotoxin Production within the Genus Bacillus by Use of the Cutan Test and a Taxonomic Determination of Mesophilic Species of the Genus. Thesis. Copenhagen, Denmark: Institute of Hygiene and Microbiology, The Royal Veterinary and Agricultural University.
  • Kawano, J., Shimizu, A., Fugii, K., Takeuchi, S., Kimura, S. 1987 Characterisation of Bacillus cereus haemolysis pattern, cytotoxicity and vascular permeability factor activity. Science Reports, Faculty of Agriculture, Kobe University, 17, 273 281.
  • Kramer, J.M., Turnbull, P.C.B., Munshi, G., Gilbert, R.J. 1982 Identification and characterisation of Bacillus cereus and other Bacillus species associated with foods and food poisoning. In: Isolation and Identification Methods for Food Poisoning Organisms(eds Corry, J.E.L., Roberts, D. & Skinner, F.A.), pp. 261 286. London: Academic Press.
  • Logan, N.A. & Berkeley, R.C.W. 1984 Identification of Bacillus strains using the API system. Journal of General Microbiology, 130, 1871 1882.
  • Logan, N.A. & Turnbull, P.C.B. 1999 Bacillus and recently derived genera. In: Manual of Clinical Microbiology, 7th edn(eds Murray, P.R., Baron, E.J., Pfaller, M.A., Tenover, F.C. & Yolken, R.H.). Washington, D.C., American Society for Microbiology.
  • Lund, T. & Granum, P.E. 1997 Comparison of biological effect of the two different enterotoxin complexes isolated from three different strains of Bacillus cereus. Microbiology, 143, 3329 3336.
  • Shinagawa, K.S., Ueno, H., Konuma, N., Matsusaka, N., Sugii, S. 1991 Purification and characterisation of the vascular permeability factor produced by Bacillus cereus. Journal of Veterinary Medical Science, 53, 281 286.
  • Spira, W.M. & Goepfert, J.M. 1972 Bacillus cereus fluid accumulation in rabbit ileal loops. Applied Microbiology, 24, 341 348.
  • Thompson, N.E., Ketterhagen, M.J., Bergdoll, M.S., Schantz, E.J. 1984 Isolation and some properties of an enterotoxin produced by Bacillus cereus. Infection and Immunity, 42, 887 894.
  • Turnbull, P.C.B. 1986 Bacillus cereus toxins. In: Pharmacology of Bacterial Toxins(eds Donner, F. Drews, J.), pp. 397 448. New York. Pergamon Press.
  • Turnbull, P.C.B. 1997 The role of the Bacillus genus in infection. Culture, 18, 5 8.