Comparative analysis of Bacillus cereus group isolates' resistance using disk diffusion and broth microdilution and the correlation between resistance phenotypes and genotypes
A collection of 85 Bacillus cereus group isolates were screened for phenotypic resistance to nine antibiotics using disk diffusion and broth microdilution. The broth microdilution antimicrobial results were interpreted using the CLSI M45 breakpoints for Bacillus spp. Due to the lack of Bacillus spp. disk diffusion breakpoints, the results obtained with the disk diffusion assay were interpreted using the CLSI M100 breakpoints for Staphylococcus spp. We identified significant (p < 0.05) discrepancies in resistance interpretation between the two methods for ampicillin, gentamicin, rifampicin, tetracycline, and trimethoprim/sulfamethoxazole. Antimicrobial resistance genes were detected using unassembled and assembled whole-genome sequences with Ariba and Abricate, respectively, to assess the sensitivity and specificity for predicting phenotypic resistance based on the presence of antimicrobial resistance genes. We found antimicrobial resistance gene presence to be a poor indicator for phenotypic resistance, calling for further investigation of mechanisms underlying antimicrobial resistance in the B. cereusgroup. Genes with poor sensitivity and/or specificity, as determined based on broth microdilution results included rph(rifampicin, 0%, 95%), mphgenes (erythromycin, 0%, 96%), and all vangenes (vancomycin, 100%, 35%). However, Bc(ampicillin, 64%, 100%) andtet genes (tetracycline, 67%, 100%) were highly specific, albeit moderately sensitive indicators of phenotypic resistance based on broth microdilution results. Only beta-lactam resistance genes (Bc, BcII, and blaTEM) were highly sensitive (94%) and specific (100%) markers of resistance to ceftriaxone based on the disk diffusion results, providing further evidence of these beta-lactams' role in nonsusceptibility of Bacillus cereus group isolates to ceftriaxone.