ENTEROTOXIN PRODUCTION IN MILK BY BACILLUS-CEREUS - A COMPARISON OF METHODS FOR TOXIN DETECTION

The enterotoxin production by seven dairy isolates and two reference s trains of Bacillus cereus, as detected by a cytotoxicity test with hum an embryonic lung cells (HEL) and two commercially available immunolog ical kits, a reversed passive latex agglutination and an enzyme-linked immunosorbent assay (ELISA), were compared after growth in milk. Ther e was no correlation between results with the two immunological method s. The positive-control enterotoxins provided with either one kit did not react with the antibodies of the other kit. A qualitatively good a greement was found between the results with the cytotoxicity test and the ELISA test. There was a linear relationship between the HEL cytoto xin titres and bacterial counts for a strongly toxigenic strain grown in milk at 8-degrees-C (detection limit of B. cereus 10. 10(6) mL-1) a s well as between ELISA titres and bacterial counts (detection limit 1 -10(6) mL-1). High ELISA titres (1600) were obtained with B. cereus in milk at 200.10(6) mL-1. Low titres (10-25) were found with B. cereus at 10. 10(6) mL-1. When 26 cardboard milk packages from 18 dairy plant s were incubated at 8-degrees-C under stress (agitation at 200 min-1 f or 7 to 12 d), sweet curdling or unclean bitter taste were found for a lmost all samples. No cytotoxin was detectable (range of counts of B. cereus 1.10(6)-47-10(6) mL-1) in any package. Low ELISA titres of 5 we re found in two out of eight samples containing 16 and 42 million mL-1 . Although toxigenic strains of B. cereus are found in milk, productio n of enterotoxin in milk packages at 8-degrees-C with normal storage s hould be very low through limited growth of B. cereus.