Bees, ants, and Lactobacillus
I’ve just published a paper (link here)(pdf) looking at host-specificity between Hymenoptera and Lactobacillus in Applied and Environmental Microbiology. This paper is a follow up to my earlier paper where I found that sweat bees associate with Lactobacillus species that have also been isolated from flowers (link here). Other researchers (mostly in Nancy Moran’s lab) have found that honey- and bumble-bees, which have larger colonies than my sweat bees, associate with host specific lactobacilli. So I wanted to determine if other Hymenoptera that form large colonies also have host specific lactobacilli. To answer that question, I collaborated with ant researchers that have found lactobacilli in their ants and built a phylogeny representing all publicly available Lactobacillus sequences along with our ant and bee sequences. To make the phylogeny as accurate as possible, I collaborated with Robin Gutell and Jamie Cannone, who study the secondary structure of 16S rRNA (check out their work here). They built models of the 16S rRNA secondary structure, and used those models to create the most accurate alignment possible (thanks for the awesome work Robin and Jamie!). I used this alignment to reconstruct the phylogenetic history of the genus Lactobacillus, and found that host specificity is only found in corbiculate apids and not ants that live in colonies with up to hundreds of thousands of individuals. Host specificity is therefore not related to large colony size, but instead is either a relic of a shared evolutionary history of bumble bees and honey bees or related to something special about corbiculate apid ecology. This was a really fun project because I collaborated with people from several different fields. Here is one of the phylogenies from our paper:
Hey Quinn – I’m wondering what you walk away from when looking at this tree. The phylogenies generated based on the 16S for the firmicutes seem to be much more highly branching than those for other parts of the bacterial tree — could you comment on this? Do you think this reflects the evolution in this group or sampling in the databases?
Hi Irene,
Thanks for stopping by! This is a good question. For sure Lactobacillus is an important genus for humans, and therefore gets a lot of research attention. Lactobacilli could therefore be over represented in the databases and falsely appear to evolve more quickly, at least in the 16S rRNA gene, than other bacteria. There are 3,913,617 bacterial 16S sequences on GenBank, and 248,282 of those are Firmicute sequences, so roughly 1/16th of the 16S sequences are Firmicutes sequences. There are 30 bacterial phyla, so maybe the Firmicutes are overrepresented? But not grossly so. The greatest genetic distance in the tree is definitely in the delbrueckii/acidophilus clade, which is again an important clade for humans and probably the most studied Lactobacillus clade. So sorting out evolution from sampling bias is again hard to do. But Lactobacilli are found in diverse environments, and Lactobacillus is a huge genus, so perhaps the amount of evolution is real. I could see Lactobacillus being split into several genera, it is such a big genus and the taxonomy is such a mess (several genera such as Oenococcus and Pediococcus fall within Lactobacillus).
I am mostly interested in the hymenoptera associates, and host-specificity. So that is the main take away for me. But my collaborators built what is probably the best Lactobacillus 16S rRNA alignment out there, so systematists might take away an entirely different message than I do. I certainly hope so!