We describe wide-ranging endogenous bacterial taxa that are capable of inhibiting an important honey bee pathogen and show considerable variation within and across colonies in the distribution of these taxa. The results have general implications for the expression of bacterial virulence in insects and for the maintenance of both beneficial and disease-associated bacteria in social insects. They also point to new avenues for the prophylactic or therapeutic treatment of honey bee diseases. None of the genera represented in this survey matched genera found in a previous 16S survey of bacteria from adult honey bees , although they do mimic, broadly, the microbial biome measured in bee colonies to date (as reviewed by Gilliam, ).
Most of the bacteria cultivated in this study belonged to the genus Bacillus, a result that is consistent with the high frequency of isolates placed in this genus by Gilliam and colleagues . Among the Bacillus species, the majority fell into the Bacillus cereus group. Both 16S rRNA sequencing and multi-locus sequencing (GlyP, PyC) indicate that these isolates represent several distinct taxa from this group, although interference with P. larvae did not fall out with species identification. The high frequency of bees harboring bacteria from the B. cereus group suggests a stable symbiosis between bees and this taxon, perhaps helping to explain the fact that bees are more tolerant than many other insects toward B. thuringiensis . Curiously, isolates from this group which shared both 16S haplotypes and sequence identity at the two protein-coding genes differ substantially in their ability to inhibit P. larvae. This variation could result from undetected genetic variation within subspecies, conditional activation of inhibitory substances, or a role for plasmids or other mobile elements in inhibition. Future experiments will help resolve the causes of conditional inhibition by Bacillus cereus subspecies.
There is a growing appreciation for the potentially beneficial roles of bacteria in honey bee colonies. Evans and Lopez  recently showed that non-pathogenic bacteria can stimulate the innate immune response of honey bee larvae, perhaps helping bees survive exposure to pathogens. Further, Reynaldi et al.  recently showed that bacteria isolated from bees in Argentina are inhibitory of the important bee fungal pathogen, Ascosphaera apis. It will be interesting to determine whether these species, in addition, are also inhibitory toward P. larvae, and to contrast the microbes associated with bees across different continents.
Bacterial symbionts likely play roles in individual and colony fitness across the social insects. Sharing of symbiotic bacteria is notoriously important for termite nutrition, and it is increasingly clear that both obligate and facultative symbioses are widespread in social insects. Recent evidence for a socially communicable defense against pathogens in termites  might indeed reflect sharing of bacteria among termite colony members, rather than the proposed induction of host-specific physiological changes.
Perhaps, as is apparent in the termites and ants [24, 25], honey bees have evolved behavioral or physiological mechanisms to enhance the transmission of beneficial microbes, while battling those species which are pathogenic. This would indicate a delicate balancing act for bees and other social insects, allowing for the encouragement of beneficial species while maintaining barriers against exploitation by pathogens. If so, discrimination at the levels of behavior and individual immune responses might be used to bias the microbial biome within insect colonies toward mutualists and against parasites and pathogens. Beneficial symbionts can potentially be fed to developing bees as a prophylactic against disease , and can regardless be used to better understand the complexity of interactions between the microbial biota of bees. It will, in this vein, also be important to look more closely at transmission mechanisms of microbes within and between bee colonies.