Assessing potential impact of gut microbiome disruptions on the environmental stress resilience of indoor-reared Bombus terrestris

Bumblebees are crucial for both natural ecosystems and agriculture, but their decline in distribution and abundance over the past decade is alarming. The global importance of bumblebees in natural ecosystems and agricultural food production cannot be overstated. However, the reported decline over the past decade has led to a surge of interest in understanding and addressing bumblebee population decline. Hence, we aimed to detect disruptions in the gut microbiome of male and worker bumblebees reared indoor and outdoor to assess potential resilience to environmental stress. Using the Illumina MiSeq platform for 16s rRNA amplicon sequencing, we analyzed the gut microbiome of male and worker bees that were raised indoors (designated as the IM and IW group) and those that were raised outdoors (also designated as the OM and OW group). Our results show presence of core bacteria Neisseriaceae, Orbaceae, Lactobacillaceae and Bifidobacteriaceae from indoor reared worker bees. However, a higher abundance of Bifidobacterium and absence of Fructobacillus from indoor reared worker bees was also observed. Indoor-reared male bees had lower diversity and fewer observed OTUs compared to outdoor-reared male bees. Additionally, the relative abundance of Actinobacteriota, Bacteroidota, and Firmicutes was significantly lower in indoor-reared males, while Proteobacteria was significantly increased. Despite this, we did not observe any dysbiosis in the gut microbiota of indoor-reared bumblebees when comparing the role of the gut symbionts among the groups. These results suggest that indoor-reared Bombus terrestris may be resilient to environmental stress when used as outdoor pollinators.


Introduction
Eusocial corbiculate bees are managed valuable resources which provide crucial pollination services and are hence important for food production [1].The population of these important pollinators has been reported to suffer a global decline [2].Population decline may be attributed to several factors, such as pathogen spillover from managed pollinators or commercially bred colonies [3][4][5], toxins and poor nutrition [6].Research by Hammer et al. (2021) [7], and Kwong and Moran (2016) [1] has shown functional role of bacterial symbionts in bee health.Gut microbiome is of particular interest because of their potential to affect the health and development of the host [8,9].The role of gut microbiome has also been elucidated against parasite defense [10,11].While the significance of gut microbiota has been acknowledged, the dynamics of its composition remains unclear.Changes in the composition of microbial communities serves as basis of understanding the host development and life processes [12,13].Bumblebees are crucial pollinators of wild plants in natural habitats [14], but the structure and composition of gut microbiome is not as well-established as that of honeybees.Hence, focus on examination of bumblebee gut microbiota is required [15].In bumblebee transmission of symbionts occurs through a single queen [7] but the influence of resource acquisition from different habitats has been reported to change microbial composition affecting health and resistance to pathogens [16].Parmentier et al. (2018) [17] showed microbial gut community differed between the larvae and adults of wild Bombus pascuorum.Variation in the bumblebee gut microbiome community during different stages of development [15,[18][19][20] and temporal pattern of community stability during old age has been reported [21].The link between social caste and gut microbiome in B. terrestris from agriculture landscapes and forest meadows showed that microbial community differed only slightly [22].The microbial gut community of worker bees from agricultural and semi-natural sites showed little effect of habitat on community composition [10].Koch and Hemple, 2011 [23] examined microbial flora of worker bumblebee and reported no significant effect of location on the bacterial communities.However, Choi et al. (2023) [24] reported that gut microbiota of indoor reared adult worker bees from two environments differed.
To further investigate the differences in the gut microbiome of bumblebee castes from different habitats, we took advantage of the ability of bumblebees to rear full colonies indoors.This enabled us to compare the gut microbiomes of bees from different habitats.The aim of our study was to determine whether there are differences in the gut microbiomes of indoor and outdoor-reared male and worker bumblebees, and to identify any disruptions in the gut microbiota that may suggest plausible resilience against environmental stress.

Bumblebee sampling and DNA extraction
Bumblebees (Bombus terrestris) used in this study were supplied from the Department of Agricultural Biology, National Academy of Agricultural Science, Republic of Korea.The bees were reared in the insectary at 28˚C, 65% relative humidity and continuous darkness [25].Bees raised in insectary were shifted to tomato-planted plastic house 15-days prior to sampling and divided into four groups: indoor-reared male bees (IM), indoor-reared worker bees (IW), outdoor-reared male bees (OM), and outdoor-reared worker bees (OW).Until processed for DNA extraction, samples were stored at -20˚C in 75% EtOH.DNA extraction was performed using Wizard genomic DNA purification kit (Promega) by following the manufacturer's protocol.

16s rRNA gene amplification and sequencing
Hypervariable regions V3 & V4 were targeted for 16s rRNA gene amplification using universal primers 341F (5 0 TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3 0 ) and 805R (5 0 GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3 0 ) with overhang adaptor sequence.Amplification profile used the following conditions: initial denaturation at 95˚C for 3 min, followed by 25 cycles of denaturation at 95˚C for 30 s, annealing at 55˚C for 30 s, and extension at 72˚C for 30 s, with a final extension step at 72˚C for 5 min.The amplicons were subjected to clean-up by Agencourt AMPure XP PCR Purification system (Beckman Coulter, Brea, USA) for removing PCR primers and primer dimers.The second library preparation and sequencing were performed at NICEM (National Instrumentation Center for Environmental Management), Seoul National University in Seoul, Korea.The libraries were sequenced on Illumina MiSeq platform (Illumina, Inc.).DADA2 (Divisive Amplicon Denoising Algorithm) plug in the QIIME2 (version 2020.08)pipeline was used for low-quality sequence removal, chimeric sequence removal and merging of forward and reverse sequence [26].Using the de novo VSEARCH algorithm (vsearch cluster-features-de-novo) [27] 97% pairwise nucleotide sequence threshold was used for clustering merged sequences into operational taxonomic units (OTUs) (S1 Fig) .The taxonomy of OTUs was assigned using the Naïve Bayes algorithm implemented in the q2-feature-classifier prefitted to the SILVA database (version 138.1gg_13_5) for the V3-V4 region of 16S rRNA gene [28] (S1 Fig).

Statistical analyses
Alpha diversity analysis was performed in R using the Shapiro-Wilk normality test, followed by Sudent's t-test.For beta-diversity analysis PERMANOVA (permutational multivariate analysis of variance) was used with 999 permutations.Pairwise comparisons were done using the Wilcoxon rank-sum test for nonparametric data.For predictive modelling randomforest function in the randomForest R package was used to determine OTUs discriminating the four groups.Differentially abundant taxa were identified by linear discriminant analysis effect size (LEfSE) [29].

Results
To assess the difference between the gut microbiome of indoor and outdoor reared male and worker bees, we analyzed the bacterial communities from the hind guts of the IM, IW, OM and OW groups (n = 10).A total of 5,022,303 raw reads were generated by amplification of V3-V4 regions.Reads were quality filtered (Q score > 25), denoised and truncated at 290 and 250 bp.After removing chimera and low abundance OTU (20% threshold), 28 OTUs at 97% similarity were obtained (Table 1).

Diversity measures
We measured alpha diversity indices (Pielou evenness, Shannon entropy, Faith PD and Observed OTUs) of the four groups (Fig 1).The Pielou evenness index value showed significantly low evenness score for IM group compared to IW, OM and OW groups (p < 0.001), which showed higher evenness.Similarly, the Shannon entropy values indicated significantly lower diversity in IM in comparison to the IW, OM, and OW groups (p < 0.001), which showed higher diversity.The Faith PD values indicated significantly low phylogenetic diversity in IM group in contrast to the IW, OM and OW groups (p < 0.001).The observed feature showed significantly lower diversity in IM group compared to the IW, OM and OW groups (p < 0.001).Ordination of Bray-Curtis dissimilarity (Fig 2) between gut microbiota of IM, IW, OM and OW groups showed significant difference (PERMANOVA, p = 0.001).

Gut microbiome structure of IM, IW, OM, and OW groups
The gut microbiome analysis of IM, IW, OM, and OW groups revealed a total of 28 OTUs.An overview of the OTUs is presented in Table 2.The distribution of the six identified phyla (Fig 3A ) among the four groups showed that except Deniococcota and the unidentified phyla, the abundance of all other phyla varied among the groups.The exclusive presence and overlap of the OTU observed among the four groups is shown in Fig 4 .The abundance of Actinobacteriota, represented by the genera Bifidobacterium and Bombiscardovia, exhibited a caste-dependent variance in bumblebees reared indoors, but not in those reared outdoor.We did not observe caste or habitat dependence of Bombiscardovia or Bifidobacterium for gut colonization, except for OTU22 identified as B. bombi which was exclusively detected in the IM group, and OTU12, B. commune, which exclusively colonized worker bees irrespective of the habitat.Two OTUs, 19 and 14, of genus Bifidobacterium were exclusively detected in IW group.Only one Operational Taxonomic Unit (OTU2) identified as Apibacter mensalis represented Bacteroidota and was detected in all samples, indicating an association irrespective of caste and habitat.The abundance of A. mensalis (OTU2) was significantly higher in worker bees compared to male bees (IW vs IM p <0.001, OW vs OM p < 0.01) (S1 Fig) .Phylum Firmicutes was represented by Lactobacillus, Enterococcus, and Fructobacillus.The relative abundance of Firmicutes was low in the IM group compared to the IW, OM, and OW groups which did not exhibit any significant difference (Fig 3a).Phylum Proteobacteria was the most dominant bacteria detected in all four groups.

RandomForest and LEfSe based identification of caste and habitat associated OTUs
We used random forest (RF) analysis (

Discussion
We document the fitness of gut microbiome from indoor and outdoor reared male and worker bumblebee and infer plausible resilience against environmental stress.
Alpha diversity indexes showed significantly lower diversity in the IM group compared to OM, but we did not observe any significant difference between the OW and IW population (Pielou evenness p = 0.999, Shannon entropy p = 0.984, Faith PD p = 0.961, Observed OUT p = 0.704).Significantly lower alpha diversity in male bee was also reported by Krams et al. (2022) [22] across caste in bumblebee.The lower diversity in indoor reared males may have been due to their limited exposure to environmental bacteria but more evidence is required to substantiate this speculation.Beta-diversity analysis showed significant difference in gut flora which is indicative of the higher microbial plasticity associated with the communities harbored by B. terrestris [23].
Of the three common important operational taxonomic units (OTUs) (7, 14 and 19) identified by both RF and LEfSe (Fig 7) which belonged to Actinobacteriota, OTU7 was detected from all four groups but 2 OTUs (14 and 19) assigned to Bifidobacteriaceae (Table 2) were detected only from the IW group (Fig 4).Remarkably, among the five OTUs that were identified as Actinobacteriota, four of them (OTUs 12, 14, 19, and 22) were found to be affiliated with the genus Bifidobacterium (Table 2).We observed exclusive association of Biofidobacterium with indoor reared bumblebees except B. commune (OTU 12) which was also detected from OW group (Fig 4).Meeus et al. (2015) [30] noted that core-bacteria of the indoor-reared bumblebees is mainly composed of Bifidobacterium.Exclusive detection of B. bombi (OTU 22) from indoor reared adult worker bumblebee was reported by Choi et al. (2023) [24] who argued that the detection of Bifidobacteria was a result of decrease in abundance of core bacteria which allowed better detection based on study by Meeus et al. (2015) [30].
We propose that the gut microbiome of indoor reared bees harboring higher abundance of bifidobacteria compared to the outdoor reared bees may not affect resilience to environmental stresses as bifidobacteria are mainly involved in the carbohydrate metabolism.Zheng et al. (2019) [31] demonstrated in the honeybee gut; genome of Bifidobacterium contains repertoires of glycoside hydrolase (GH) genes which were also detected in Lactobacillus present.Specifically, their investigation revealed that B. bombi strain possessed four GH43 genes, while B. commune did not contain any.Meta-transcriptome of honeybee showed participation of γ-Proteobacteria, Bacilli and Actinobacteria in carbohydrate metabolism [32].However, there is scarce data which explains the higher abundance of bifidobacteria associated with the indoor reared bees and factors like feed, genetics, and other environmental factors e.g., hygienic conditions may be explored to gain better understanding.
Our results showed that Fructobacillus (OTU26) was identified exclusively from OW group (Fig 4).Fructobacillus has been reported to dominate gut microbiome of adult worker B. terrestris from environments with least anthropogenic impact [22].We propose Fructobacillus may have caste dependent colonization irrespective of the habitat.The IW group of our study consumed commercial feed rich in sucrose which could not have sustained Fructobacillus.Fructose dependence is attributed to impaired alcohol fermentation in these bacteria, which results from the lack of the adhE gene encoding alcohol/acetaldehyde dehydrogenase [33].Temperature decrease has been reported to increase the fructose concentration in nectar [34] which allows Fructobacillus abundance in guts of B. terrestris.The absence of Fructobacillus from IW shows a disruption of gut microbiome which may result in loss of fructose metabolization in cold weather.But we argue that this might only be detrimental to B. terrestris inhabiting regions with temperatures dropping below freezing point.However, indoor reared B. terrestris harbor higher abundance of Bifidobacterium spp could still thrive and be suitable for survival.Use of Fructobacillus as a probiotic prior to onset of freezing temperatures may improve fitness of indoor-reared B. terrestris by minimizing the disruption of gut microbiome.
Five important OTUs (1, 3, 5, 6 and 16) were identified by RF and LEfSe (Fig 7) which were assigned to Proteobacteria (Table 2).Two (OTUs 1 and 3) of the five were found associated with all four groups.
In our study, S. floricola (OTU 5) was observed to be exclusively associated with male bees (Fig 4).We observed a significantly high abundance of Saccharibacter in OM group compared to IM group (p < 0.01) (S1 Fig) .The lower abundance might have been a result of closed breeding system which induced a bottleneck in microbiota of indoor-reared B. terrestris [30].Or the high abundance in the OM group may have been a result of immune-mediated apparent competition [35] between a fungal pathogen and S. floricola which favored growth of S. floricola in gut of OM group.S. floricola has been reported to possess type 1 polyketide synthase gene cluster which produces antifungal metabolite responsible for protective effect [36].The identification of S. floricola from indoor reared bees implies that the bees possess a degree of adaptation for enduring biotic stress.Studies of bee gut microbiota are limited in their detection of S. floricola in various bee species.Further research may reveal S. floricola role in the health and performance of worker bees and other pollinators.
Our results indicated absence of Candidatus_Schmidhempelia (OTU6) from indoor-reared bumblebees which may suggest a reduction of the wild microbiota (Fig 4).In comparison to indoor-reared B. terrestris, Choi et al. (2023) [24] also reported presence of Candidatus_Schmidhempelia in adult worker B. terrestris raised outdoors.We presume that Candidatus_Schmidhempelia may rely on a horizontal transmission route, as is the case for most non-core bacteria.Candidatus_Schmidhempelia is phylogenetically closely related to Gilliamella apicola [37], a core bacterium of gut microbiota of wild and indoor-reared B. terrestris [30].The loss of Candidatus_Schmidhempelia represents microbiome dysbiosis in indoor-reared bumblebees, which may result in a lack of benefits provided to the host.
Representation of phylum Bacteroidota exclusively by A. mensalis among all groups was observed in our data (Table 2).Interestingly, our data showed (Table 2) no detection of Snodgrassella from any of the four groups.A. mensalis may utilize the encoded type IX secretion system and gliding motility apparatus to establish biofilms [38].This mechanism is comparable to that of S. alvi, which is also an aerobic organism known to create biofilms on the gut wall [39,40].The co-occurrence of Apibacter alongside S. alvi has been reported in Apis cerana, while A. mellifera generally does not harbor S. alvi [41].Some evidence suggests that there may be competition for a limited ecological niche between Apibacter and S. alvi in bumblebees, as there may be an inverse relationship between their abundance, but the available data on this topic are limited [11,42,43].The co-occurrence and engagement of S. alvi and Gilliamella has been reported due to carboxylate and carbohydrate dependence, respectively [44].There is limited information available on the metabolic requirements of A. mensalis, as it is a relatively newly described bacterial species.However, based on the available literature, there is currently no evidence to suggest that A. mensalis has a carboxylate dependence like Snodgrassella.Association of Apibacter with B. terrestris has been reported to decrease infection by Crithidia bombi, a trypanosomatid gut parasite [43].In our study, we found A. mensalis in both indoor and outdoor bees.As A. mensalis has been linked to disease suppression, we conclude that the gut symbionts of indoor bees may confer resilience against such biotic stresses.
Although the bumblebee gut communities of the IW, IM, OW and OM groups are different from each other, we did not observe a dysbiosis in indoor reared bumblebees when functions of members of the gut symbionts among the group were compared.We evaluated the comparative fitness potential of the gut communities of indoor-reared bees versus outdoor-reared bees.To achieve this, we focused on the functions of the symbionts provided to the host.However, we acknowledge that our taxonomic classification of symbionts based on partial sequencing of the V3-V4 region had limited resolution.As a result, our assessment of fitness was based on limited knowledge regarding the functional capabilities of bumblebee gut symbionts.Further research is required to investigate the functional roles of various unclassified Bifidobacterium spp and Lactobacillus spp associated with indoor and outdoor-reared bees with focus on benefits associated with the resilience provided against environmental stress.