On 15 May and 22 May 2022, Apis mellifera foragers and adult Centris decolorata were collected with an insect net (Departamento Recursos Naturales, permit ID 2022-IC-019). Apis mellifera (honeybees) were collected in three sites from two different towns, to make sure they came from different hives: Coamo (18.036814, −66.374096) and Mayagüez (2 plots, 18.250797, −67.177461 and 18.251412, −67.178063), Puerto Rico, United States. Mayagüez is a coastal town and in these exact coordinates, Centris decolorata specimens were also collected (Figure 1A). Centris decolorata nests were excavated in two Mayagüez plots following the method by Marinho et al. (2018). A total of 46 individuals were collected for this study. These individuals include, 9 A. mellifera foragers—6 collected from Coamo and 3 from Mayagüez–; and 24 C. decolorata bees (12 females, 12 males and 13 brood cell contents), all from Mayaguez (Figure 1B). The adult digestive tract (foregut to hindgut) of each species were dissected using sterilized tools under the stereomicroscope. The brood cells were also dissected to retrieve the whole individual larvae and the associated brood cell provision (Figure 1C). Because some brood cells were empty and solely contained the starting/remaining brood cell provisions and some A. mellifera had very small sizes and had to be pooled for extractions, a selection of 39 samples was done for analyses: 6 A. mellifera workers (female foragers); 10 female (mother bees) and 10 male C. decolorata adults, 4 of their larvae, and 10 of their brood cell provisions. Even though reproducing female solitary bees are further referred as “solitary mother bees,” their sampling has been done independently from their larvae. The female solitary bees we collected were considered as to be mother bees based on their behavior: returning to the nest at the end of the afternoon or carrying plant materials into the nest.

In feral foragers of Puerto Rico honeybees, the simple and recurrent phylotypes of the gut microbiota remain as previously described in other honeybees (Martinson et al., 2011; Thompson et al., 2017). The gut microbiota of both bee species has the same phyla (Figures 2C, 3); however, the families are different. The core taxa of Apis mellifera comprises the families of Lactobacillaceae, Bifidobacteriaceae, Bartonellaceae, Neisseriaceae, Orbaceae, Rhizobiaceae, and Acetobacteraceae (i.e., Commensalibacter spp.; Figure 2D; Supplementary Figure 2). In contrast, the core microbiota of Centris decolorata is composed by bacteria from the Acetobacteraceae (i.e., undescribed Acetobacteraceae) and Moraxellaceae (Figures 2D, 4C,D). Some C. decolorata females displayed trace levels of undescribed species of Lactobacillaceae and Bifidobacteriaceae (Figure 2D; Supplementary Figure 2).

Microbiota composition and structure of brood cell provisions are different from all other samples (Figures 4A,B). Brood cell provisions had significantly higher diversity than any gut microbiota of adult solitary bees, but not strongly different than larvae (alpha-diversity differences using Chao1, adjusted-p = 0.286 and using Shannon, adjusted-p = 0.081; Supplementary Table 1). While no significant differences in diversity metrics were found between larvae and brood cell provisions, both beta and alpha diversities of brood cell provisions are significantly different from that of adult females and males (PERMANOVA, p = 0.001; ANOSIM, p = 0.001; Chao1 and Shannon index, p < 0.05, Supplementary Tables 1, 2; Figures 4A,B). Brood cell provisions are composed by diverse families of bacteria having a higher number of taxa as part of the core microbiome as compared to other sample types. Females and males were mostly composed in Acetobacteraceae and in Moraxellaceae (Figure 4C), in fact only Acetobacteraceae are part of the female core microbiome. Furthermore, brood cell contents and adults did not display the same core microbiota at 50% sample prevalence. Rhizobiaceae, Chitinophagaceae, Lachnospiraceae, Enterobacteriaceae, Xanthomonadaceae, Alcaligenaceae, and Microbacteriaceae constituted the core microbiota of brood cells, while Acetobacteraceae constituted that of adults (Figures 4C,D). Undescribed Acetobacteraceae explained most of the differences between adults and brood cells (mostly females) using MaAsLin statistical analysis. Only males had abundant Staphylococcus sp. (Supplementary Figure 3).

A total of 28 pollen types were identified from A. mellifera and C. decolorata pollen loads (cf. “Plant Pollen and Bee Pollen Grain Catalog” in Supplementary Presentations 1, 2). Of these, 16 pollen types were excluded after filtering for pollen types with less than five grains per slide. Seven of the remaining pollen types were found in C. decolorata. Three pollen types, including Byrsonima sp. (an oil flower) and Combretaceae, were found shared by both bee species (Supplementary Figure 4A). The pollen types associated with C. decolorata did not reflect the plant species near their nests, such as Canavalia rosea, Ipomoea pes-caprae, and Bidens Alba (Martinez-Llaurador, 2021). Apis mellifera had a weaker interaction with Combretaceae sp.1 and a stronger interaction with Byrsonima sp. than expected compared to the null model (Supplementary Figure 4B). The remaining interactions between both bee species and plant resources were described as consistent with the null model (Supplementary Figures 4B,C). Although A. mellifera had a higher plant resource species richness and diversity than C. decolorata, there were no significant differences (Kruskal-Wallis test p = 0.20 and p = 0.28 respectively).

The lower physical contact between solitary bee individuals, compared to social ones (Wittwer et al., 2017) is one of the factors that lead to variability in microbial communities among individuals. With social interactions, including trophallaxis, social bees directly share their gut bacteria, reducing probability of interindividual variation. This participates in the maintenance of a consistent core gut microbiota. Compared to social bees, environmental transmission pathways of solitary bees play a stronger role in the acquisition of bacteria, probably due to differences in nesting habits and materials (Voulgari-Kokota et al., 2019b). Solitary bees such as Centris use pollen, nectar, secretions from mandibular and Dufour’s glands, and floral oils to build their nest. After the brood cell is completed, provisioned, provided with an egg, and sealed, the female has no contact with its brood. Through various strategies, the brood is protected from parasites, microbes, predators, and external environment variation, which is especially important in warm and humid environments (Danforth et al., 2019). As a first-line defense, female solitary bees coat their brood cells with glandular secretions which may be combined with other collected materials. Secretion from their Dufour’s gland is the primary source for lining brood cells. It consists mostly of large polar molecules, providing waxy, hydrophobic coating to the brood cell (Danforth et al., 2019). Their exact composition varies among Centris species (Cane and Brooks, 1983), though Centris from the Antilles have yet to be analyzed. Further studies should evaluate if female Centris use these secretions only to coat the brood cells or also to mix them with provisions, as do some other solitary bees, e.g., Megachilids (Williams et al., 1986).

Some solitary bees also use mandibular gland secretions as antimicrobials (Cane et al., 1983), sometimes to first disinfect the brood cell prior to lining (Cane and Tengö, 1981). For instance, linalool, citral, geraniol, nerol or citronellol, all mandibular secretions, are effective inhibitors of fungal and bacterial growth in multiple species of solitary bees (Cane et al., 1983). These molecules and their specific targets are yet to be described for Centridini. Floral oils may also serve as protective coating materials (mostly stearic acid and elaiophore lipids; Danforth et al., 2019) which females collect from multiple plant families such as Malpighiaceae, Calceolariaceae, Iridaceae, Orchidaceae, Krameriaceae, Plantaginaceae, and Solanaceae (Martins et al., 2015).

In addition to brood cell lining, females may also mix the floral oils with provisions as an energy source (Buchmann, 1987), but the generality of this incorporation is not well understood, due to the paucity of species for which nest provisions and linings have been chemically analyzed (Neff and Simpson, 2017). Whether floral oils replace nectar is not absolute: brood cell provisions may contain trace to appreciable amounts of sugar and oils (Neff and Simpson, 1981a). Alternatively, Neff and Simpson proposed that mixing floral oils with provisions is advantageous to oil-collecting bees nesting in environments susceptible to flooding. This argument is based on the absence of oil-collecting habits for Centris species in xeric habitats. Indeed, incorporating floral oils to the provisioning could inhibit hygroscopic effects of provisions from bees nesting in extremely moist environments, but also control mold or bacterial infection, or act as a deterrent against nest parasites (Neff and Simpson, 1981b). For instance, levulinic acid, an oil collected for brood cells, acts as an anti-fungal agent (Neff and Simpson, 1981a).

Given this wide variety of materials used for brood cell construction and provisioning, it is not surprising to see such diverse microbiota present in the guts of adult solitary bees, and even more in the brood cell provisions. How these bacteria survive in this antimicrobial, yet nutrient-rich brood cell microcosm is unknown, yet. We assume that the presence of these bacteria is important, as they will determine the digestibility of the raw pollen clump by larvae. Further research should describe the chemical composition of the brood cell provisions of Centridini bees, as well as the microbial targets of glandular secretions and floral oils.

While other studies showed that bacteria from the honeybee gut is transferred to their corbicula pollen during the process of pollen packing (Prado et al., 2022), the bacteria isolated from brood cells of C. decolorata clearly have a plant origin. Some of these bacteria are known to induce plant growth, e.g., Rhizobiaceae, Chitinophagaceae, or Lachnospiraceae, or to inhibit it, e.g., Xanthomonadaceae, or Alcaligenaceae (Gnanamanickam, 2006). It would be interesting to test if bees are able to modulate the abundance of plant-inhibiting bacteria in later stages of brood cell provisions, as pollen from brood cells is no longer available for pollination. Another constituent of the core microbiome from C. decolorata’s brood cell, Enterobacteriaceae, has been previously found in pollen (Madmony et al., 2005; Ambika Manirajan et al., 2016; Straumite et al., 2022) and larvae (Parmentier et al., 2018). This latter family also has a plant origin, especially flowers (Gnanamanickam, 2006; Junker et al., 2011; Junker and Keller, 2015). Other studies reported on the presence of some Lactobacillus, namely L. micheneri, L. timberlakei, and L. quenuiae, in the pollen provisions, bee guts and flowers. In addition to being tolerant of osmotic stress, these lactic acid bacteria are able to degrade the outer pollen wall (Lipiński, 2018), which makes pollen digestible for early larval feeding (Gurevitch and Padilla, 2004; Gilliam, 2006; Vuong et al., 2019; Voulgari-Kokota et al., 2019a). Given Lactobacillus presence in trace amounts, future description of the brood cell core microbiome combined to pollen wall degradation analyses at brood cell age should help identifying bacteria involved in pollen pre-digestion process.

All these brood cell constituents (pollen, possibly evaporated nectar and floral oils, and the associated bacteria) are definitely brought by Centris decolorata females to the larvae. But even though females are the ones provisioning the brood cells for the larvae, their gut microbiota is significantly different from the microbiota of the brood cell provisions, at least at early stages. These results contrast with previous findings of another solitary bee from a semi-arid region (although in dense aggregations of millions of bees), where the gut microbiota of females and larvae were similar (Kapheim et al., 2021). In C. decolorata, females provision the brood cell for the larva independently to what she ingests. In our study, the gut microbiota of both male and female solitary bees are more similar to each other than to that of brood cell contents, at least initially. Indeed, the collected brood cells were at early stages of larval development. The provisions were thus essentially composed in flower-specific bacteria, which is coherent considering that provisions are mainly constituted by pollen. Later, these flower-specific bacteria shift to bacteria able to grow on nutrient rich mixture, i.e., the proteins from pollen, and possibly sugar from nectar and floral oils. Microbial composition of the larvae and provisions therefore changes along with larval development (Voulgari-Kokota et al., 2018). To confirm the differences in microbial composition between females and nest provisions, future microbial assessment should consider a larger brood cell sampling, with early and later larval stages (using larval development as a proxy for bacterial shift). Indeed, assessing the brood cell provisions at the middle/end of the C. decolorata season would probably lead to higher probability to encounter brood cells of later larval stages. These brood cells would thus be composed in bacteria able to grow on nutrient rich mixture, probably similar to microbiota of females and larvae.

Gut acidification of solitary and social bees seems to be driven by different phylotypes: Lactic Acid Bacteria in honeybees vs. Acetic Acid Bacteria in solitary C. decolorata. The trace amounts of Lactobacillaceae in C. decolorata guts could represent horizontal acquisition from A. mellifera, or environmental pools of related strains to C. decolorata. Acetic Acid Bacteria (AAB) are strict aerobic bacteria and ubiquitous. They occur in a wide variety of substrates such as in plants and flowers (Crotti et al., 2010). They are widespread in carbohydrate-rich, acidic, and alcoholic niches, such as nectar, which has been proposed as an origin for these bacteria (Morris et al., 2019; Ravenscraft et al., 2019). In addition to being considered environmental and ubiquitous bacteria, AAB are also important insect symbionts, as for food uptake and host survival. Insect associations are stable and follow several transmission routes for their propagation (Crotti et al., 2010). In our samples, two families of AAB were found: Moraxellaceae and Acetobacteraceae. Bacteria from the Moraxellaceae family, especially Acinetobacter have been isolated from the solitary male guts, but not from the brood cell provisions of our study, even though reported in the literature as present in pollen provisions coming from Mediterranean plants (Álvarez-Pérez et al., 2013). Although not found in pollen provision of our study, the bacteria from Acetobacteraceae were present in the guts of adult Centris decolorata (both males and females) and constitute their core microbiome. It was scarcely present in foraging honeybees that typically do not interact with larvae in hives (Winston, 1987), as reported in domestic local honeybees (gentle Africanized Honeybee, gAHB) by Ortiz-Alvarado (2019). Acetobacteraceae has been isolated from the gut of adult honeybees (Sabree et al., 2012; Anderson et al., 2013; Corby-Harris et al., 2014; Thompson et al., 2017; Ortiz-Alvarado, 2019), and different Acetobacteraceae (i.e., Alpha 2.2 Acetobacteraceae) from honeybee larvae (Corby-Harris et al., 2014; Ortiz-Alvarado, 2019). Interestingly, Acetobacteraceae was found in larvae, nymphs, young nest bees, and royal jelly in the same study, but it was almost absent in honeybee foragers. Acetobacteraceae could thus be a family of bacteria related to nursing bees, i.e., larvae, nymphs, young nest bees, and royal jelly in honeybees, and in female solitary bees, who play a nursing role. The absence of Acetobacteraceae in the brood cell of this study could suggest that the transfer of these symbionts could mostly be vertical for Centris decolorata, but this remains to be identified in a larger sample size of solitary bee larvae. If AAB such as Acetobacteraceae are present in females and larvae but not in the brood cell provisions, then vertical transmission of these symbionts would be preferred by Centris decolorata.

Some Megachilids show a significant association between the composition of their foraged pollen and the pollen bacterial communities and larval bacterial communities. In these bees, where bacterial transmission pathways through eusociality are impossible, pollen foraging appears to be very important to obtain their bacterial symbionts (Voulgari-Kokota et al., 2019a). In our limited pollen study, we found that pollen resources foraged by Apis mellifera and Centris decolorata were not significantly different whereas their gut microbiota were composed of different bacterial phylotypes. Therefore, pollen acquisition routes cannot explain differences in the gut microbiota between the studied social and solitary bees. Apis mellifera has been previously shown to be weakly impacted by microbiota from pollen (Donkersley et al., 2018; Jones et al., 2018). Pollen samples of C. decolorata evidence the presence of additional plant resources along their foraging range, suggesting they forage for pollen over long distances from their nest locations, in addition to use pollen resources from plant species found near their nests (Pers. Obs.). Interactions between C. decolorata and Byrsonima sp. reflected a lower pollen abundance, suggesting that individuals visit Byrsonima sp. to primarily collect oils and incidentally collect pollen along their bodies. Although C. decolorata transports fewer pollen grains between individuals of Byrsonima sp., these could be sufficient to pollinate the flowers. However, A. mellifera appears to be an effective pollen forager having a stronger interaction, influenced by a greater abundance of pollen on its corbicula, with Byrsonima sp. For a broader overview of plant species visited by C. decolorata, bee sampling and pollen analyses over the season and on different daily periods should be considered for future studies.