Abstract
American foulbrood (AFB) is the most serious bacterial disease of honey bee brood. Spores of the causative agent Paenibacillus larvae are ingested by bee larvae via brood foods and germinated cells proliferate in the larval midgut. In Japan, a macrolide antibiotic, tylosin, is used as the approved prophylactic for AFB. Although tylosin-resistant P. larvae has yet to be found in Japan, it may emerge in the future through the acquisition of macrolide resistance genes from other bacteria, and bacteria latent in brood foods, such as honey, may serve as a source of resistance genes. In this study, to investigate macrolide resistance genes in honey, we attempted to isolate tylosin-resistant bacteria from 53 Japanese honey samples and obtained 209 isolates from 48 samples in the presence of 1 μg/ml of tylosin. All isolates were Gram-positive spore-forming bacteria mainly belonging to genera Bacillus and Paenibacillus, and 94.3% exhibited lower susceptibility to tylosin than Japanese P. larvae isolates. Genome analysis of 50 representative isolates revealed the presence of putative macrolide resistance genes in the isolates, and some of them were located on mobile genetic elements (MGEs). Among the genes on MGEs, ermC on the putative mobilizable plasmid pJ18TS1mac of Oceanobacillus strain J18TS1 conferred tylosin and lincomycin resistance to P. larvae after introducing the cloned gene using the expression vector. Moreover, pJ18TS1mac was retained in the P. larvae population for a long period even under non-selective conditions. This suggests that bacteria in honey is a source of genes for conferring tylosin resistance to P. larvae; therefore, monitoring of bacteria in honey may be helpful to predict the emergence of tylosin-resistant P. larvae and prevent the selection of resistant strains.