Extracellular proteins produced by Bacillus cereus B-4ac were separated by chromatography on Amberlite CG-400, QAE-Sephadex, Sephadex G-75, and hydroxylapatite. A fraction, containing three detectable antigens, obtained from chromatography on hydroxylapatite caused fluid accumulation in ligated rabbit ileal loops, was dermonecrotic to rabbit skin, was cytotoxic to cultured cells, and was lethal to mice after intravenous injection. Two other fractions obtained from chromatography on hydroxylapatite showed essentially no toxic activity when tested individually. Each nontoxic fraction contained two of the three proteins present in the toxic material. When the two nontoxic fractions were combined, activity in all of the biological assays was observed. Antiserum against either of the nontoxic fractions neutralized the dermonecrotic response of the combined material. These results suggest that all of these biological activities probably are due to a single entity and that more than one component probably comprise the toxic entity. The association of Bacillus cereus with food-borne illness is now well established and has been reviewed recently (10, 11, 23). Two types of food-borne illness are recognized, one causing diarrhea and the other causing emesis. The ability of B. cereus to cause diarrhea has been attributed to the production of a proteinaceous enterotoxin (12, 13, 20, 22). This study is concerned with a toxin produced by B. cereus B-4ac that is believed to be the agent responsible for the induction of diarrhea. Like culture filtrates of Vibrio cholerae (4, 5), enterotoxigenic Escherichia coli (7, 17), and cell extracts of Clostridium perfringens (6, 18), culture filtrates of B. cereus cause fluid accumulation in rabbit ileal loop (RIL) assays (21) and increase vascular permeability (VP) in rabbit skin (12). The VP activity of B. cereus culture filtrates correlates well with RIL activity (12, 24). Other toxic effects of B. cereus culture filtrates include cytotoxicity to cultured cells (1) and lethality to mice after intravenous injection (3). However, attempts to isolate the factor(s) responsible for these biological activities and to relate them to diarrheal activity have resulted in inconclusive findings. Attempts by Spira and Goepfert (21) to isolate the factor responsible for the VP activity resulted in separation of VP activity from hemolytic and phospholipase activities by ammonium sulfate precipitation of crude culture supernatant fluids and subsequent chromatography on Sephadex G-75. Unable to duplicate this separation procedure, Turnbull et al. (24) used preparative isoelectric focusing to separate proteins in crude culture supernatant fluids. They were able to isolate a fraction enriched for VP and RIL activity and devoid of phospholipase activity, but this fraction contained residual hemolytic activity. Ezepchuk et al. (8) reported the isolation of a permeability factor of molecular weight 100,000; this factor was lethal to mice, but did not give an RIL response.

Murine monoclonal antibodies reactive with staphylococcal enterotoxins B (SEB) and C1 (SEC1) were isolated by hybridoma techniques. Of the nine antibodies, three reacted only with SEB, two reacted with SEB and SEC1, three reacted with all subtypes of SEC, and one reacted only with SEC2 and SEC3. All of the antibodies reacted with protein blotted onto nitroceliulose from electrophoresis gels which corresponded to the enterotoxin band. Fragments of SEB and SEC1 were generated by limited digestion of the toxin with trypsin. With the immunoblot technique, four of the five antibodies reactive with SEB reacted with the tryptic fragment of molecular weight 17,000, and the five antibodies reactive with SEC, reacted with the tryptic fragment of molecular weight 14,000.