Introduction
Gut microbial communities can be important mediators of host health and
fitness (McFall-Ngai et al. 2013). Many bee species (Hymenoptera:
Apoidea: Anthophila) host distinct and functionally important bacterial
communities in the GI tract (Martinson et al. 2011, Lee et al. 2015,
Moran 2015, Engel et al. 2016). As a low-diversity and tractable
experimental system, the bee gut represents an excellent model to
examine metabolic specialization, function and coexistence within
microbial communities (Engel et al. 2016). However, bee species vary in
microbiome composition, including the presence of specialized taxa and
the relative abundance of environmental bacteria (Kwong et al. 2017,
McFrederick et al. 2017). The factors that predict this variation among
species and their functions remain poorly understood (Engel et al.
2016), but sociality has been proposed as an important driver of gut
microbiome evolution for bees, like other macroorganisms (Moeller et al.
2016, Kwong et al. 2017, Moran et al. 2019).
Social corbiculate bees in the subfamily Apinae (‘pollen basket’ bees),
including honey bees and bumble bees, are characterized by distinctive
gut microbial communities that are relatively consistent among
individuals within a species (Kwong et al. 2017). In honey bees and
bumble bees, gut bacterial communities are consistent among individuals
and transmitted by social interactions (Koch and Schmid-Hempel 2011,
Powell et al. 2014, Billiet et al. 2017). By contrast, in non-social bee
species, including those closely related to social corbiculates
(McFrederick and Rehan 2019), individuals host more variable and less
distinctive microbiomes, likely driven by environmental rather than
social acquisition of microbes (McFrederick et al. 2012, McFrederick and
Rehan 2019, Cohen et al. 2020). However, key tests of the sociality
hypothesis using bee species in the genus Halictus (which
contains solitary and social species) found limited influence of
sociality on bacterial composition (McFrederick et al. 2014, Rubin et
al. 2018). These results raise the possibility that other traits instead
of or in addition to sociality may be more important in shaping
microbiome composition and specialization among bees.
Carpenter bee species in the genus Xylocopa (Apidae: Xylocopini)
offer a unique system to study the relative role of sociality in
structuring microbiome composition. Xylocopa are large-bodied
bees and close phylogenetic relatives of social corbiculate apids
(Bossert et al. 2019). Xylocopa are locally common and
economically important pollinators in some systems (Keasar 2010,
Giannini et al. 2015), and nest in timber or dead stalks of plants
(Barrows 1980). Of particular note, several species of carpenter bees
have been characterized as facultatively or incipiently social (Gerling
et al. 1989, Michener 1990). In characterized species, the oldest female
in a nest maintains reproductive dominance and feeds younger nestmates
via trophallaxis (Lucia et al. 2015), but cooperative brood care is
rarely documented. One species in which sociality has been well-studied
is X. sonorina , which lives in small, fluid, dominance-based
societies with reproductive division of labor, where the proportion of
individuals nesting socially is temporally dynamic (Ostwald et al.
2020). Moreover, in all of the best-studied Xylocopa species,
both social and solitary nests are present within the same population
(Gerling et al. 1989). Investigation of the microbiome of carpenter
bees—close relatives of corbiculates with contrasting social
structure– may offer insights into the role of sociality in the
evolution of the bee microbiome.
Here, we examine the composition of bacterial full-length 16S rRNA genes
in Xylocopa sonorina Smith [previously X. varipunctaPatton] and Xylocopa tabaniformis orpifex Smith (Bezark
2013), two carpenter bee species common in the western North America. We
leverage PacBio Sequencing and a sample inference method with
single-nucleotide resolution (Callahan et al. 2019) to examine
strain-level resolution and phylogenetic relationships to previously
characterized bee microbial taxa. These Xylocopa species often
co-occur locally, collect nectar and pollen from the same plant hosts,
and show a similar seasonal phenology in activity and reproduction. We
examined microbiome composition in two tissues, the crop (foregut) and
gut (combined midgut and hindgut), which are disparate in function and
separated by a proventricular valve affecting movement between these
regions. We hypothesized that the crop would be variable in bacterial
composition among individuals, due to frequent intake of food including
pollen, low microbial biomass and predominance of
environmentally-sourced microbes, (Anderson et al. 2013). We
hypothesized that if Xylocopa is similar to social apids, the gut
would host a core microbiome distinctive from the crop that was
consistent among individuals (Moran 2015), or if Xylocopamicrobiomes resemble solitary bees sampled to date, the gut would host a
variable microbiome with high similarity to the crop (Voulgari-Kokota et
al. 2019).
To address these hypotheses, we sampled bees from three geographic
locations and first compared how bacterial alpha and beta-diversity
differ among tissues, species and geographic locations. We also examined
if sex or foraging status was associated with bacterial composition.
Next, we defined the core bacterial taxa for these species and examined
phylogenetic patterns among strains within these core clades. Lastly, to
examine if shorter regions of the 16S rRNA gene could also detect these
patterns, we repeated analyses using the full-length data that had been
trimmed to the V4 region only. Our results suggest that carpenter bees
host distinctive gut bacterial communities including bacterial clades
previously detected in corbiculates. In addition, Xylocopaspecies host phylogenetically and geographically distinct lineages
within core clades which are revealed by the full-length 16S but not by
the V4 region alone.