Performance Testing of Bacillus cereus Chromogenic Agar Media for Improved Detection in Milk and Other Food Samples

Abstract

In this study, the performance of four alternative selective chromogenic B. cereus agar was compared to the reference mannitol-yolk polymyxin (MYP) agar (ISO 7932) using inclusion and exclusion test strains (n = 110) and by analyzing naturally contaminated milk and other food samples (n = 64). Subsequently, the panC group affiliation and toxin gene profile of Bacillus cereus senso lato (s.l.) isolates were determined. Our results corroborate that the overall best performing media CHROMagar™ B. cereus (93.6% inclusivity; 82.7% exclusivity) and BACARA® (98.2% inclusivity, 62.7% exclusivity) are more sensitive and specific compared to Brilliance™ B. cereus, MYP and ChromoSelect Bacillus Agar. Both media allow unequivocal detection of B. cereus with low risks of misidentification. Media containing ß-D-glucosidase for the detection of presumptive B. cereus may form atypical colony morphologies resulting in a false negative evaluation of the sample. Naturally contaminated samples presented high numbers of background flora, while numbers of presumptive B. cereus were below the detection limit (<10 CFU g−1 or mL−1). Recovery after freezing resulted in the highest detection of B. cereus s.l. on BACARA® (57.8%), CHROMagar™ B. cereus (56.3%) and MYP agar (54.7%). The panC/toxin profile combination IV/A was the most abundant (33.0%), followed by III/F (21.7%) and VI/C (10.4%). More panC and toxin combinations were present in 15.6% of samples when reanalyzed after freezing. In order to improve detection and confirmation of B. cereus s.l. in food samples, we recommend the parallel use of two complementary selective media followed by molecular characterization (e.g., panC typing combined with toxin gene profiling). When determining psychrotolerant or thermophilic members of the B. cereus group, the selective agar media should additionally be incubated at appropriate temperatures (5 °C, ≥45 °C). If high-risk toxin genes (e.g., ces or cytK-1) are detected, the strain-specific ability to produce toxin should be examined to decisively assess risk.