Microbial Succession on Honey Bee Body Surfaces Reflects Behavioral Maturation

Abstract

Behavioral maturation is essential for the proper functioning of honey bee societies and is regulated by multiple factors such as juvenile hormone (JH) and nutritional deficiency. Although recent studies have shown that surface-associated microbiota in insects can modulate host behavior, the relationship between body surface microbiota and behavioral maturation in honey bees remains largely unexplored. This study aimed to determine whether the surface microbial communities of honey bees shift with behavioral maturation. By using 16S rRNA gene amplicon sequencing, we analyzed the surface microbiota of worker bees at different behavioral stages (newly emerged bees, nurses, and foragers) in both Eastern honey bee Apis cerana and Western honey bee Apis mellifera. The results showed that in both honey bee species, nurse bees exhibited the lowest microbial diversity, while forager bees showed the highest, and newly emerged bees had an intermediate level of microbial diversity. Moreover, beta diversity analyses revealed that the body surface microbiota of worker bees significantly varied across behavioral stages in both bee species and differed between the two bee species at the same behavioral stage. Additionally, in both bee species, at the phylum level, Pseudomonadota, Bacillota, and Actinobacteriota dominated the worker bee body surface microbiota; at the genus level, foragers had more Gilliamella, while nurses harbored more Lactobacillus. Together, our findings reveal the emergence of distinct microbial signatures on honey bee body surfaces during behavioral maturation.

1. Introduction

As a highly social insect, the honey bee undergoes a strictly socially regulated and biologically controlled process of behavioral maturation [1]. In recent years, significant progress has been made in the study of the mechanisms underlying honey bee behavioral maturation thanks to the rapid development of fields such as molecular biology, neurobiology, and ecology. Gene expression affects bee behavioral maturation, which is closely related to the expression changes of thousands of genes in the brain. These changes occur with age and are largely completed at specific time points, such as eight days old [2,3]. Hormonal regulation plays a crucial role in bee behavioral maturation, with hormones such as juvenile hormone (JH) and queen mandibular pheromone (QMP) playing significant parts in this process [4,5,6]. JH primarily regulates labor division among worker bees, while QMP can inhibit ovarian development in worker bees [7]. Additionally, the insulin/insulin-like growth factor signaling (IIS) pathway also influences the division of labor, and its variations causally affect the timing of behavioral maturation [8]. Chemical pheromones such as ethyl oleate (EO) play a pivotal role in bee behavioral maturation. They are transmitted and perceived through close-range olfaction among individuals, influencing the behavioral development of bees [9,10]. Furthermore, cuticular hydrocarbons (CHCs) also serve as communication cues, playing a significant role in the life of social insects and varying according to the division of labor [11,12,13,14]. Changes in an individual’s nutritional status also have an impact on bee behavioral maturation. In cases of nutritional deficiency, the behavioral maturation of worker bees can be accelerated [15,16]. The absence of pollen in the diet of young worker bees leads to the accumulation of oxidative stress markers in their fat body tissues and alterations in their cuticular hydrocarbon profiles, which resemble those of older bees [17].

Increasing evidence reveals that the insect surface microbiota is complex and dynamic, providing new insights into insect health and environmental interactions. The surfaces of insects host diverse microorganisms, including bacteria, fungi, and viruses, that form intricate symbiotic relationships with their hosts [18]. These microbes significantly influence host physiology, immunity, and behavior [19]. Several studies indicate that epidermal microorganisms produce antibacterial and antifungal compounds, effectively protecting hosts from pathogens [20,21]. For example, surface Actinomycetes on leaf-cutting ants produce antimicrobial compounds, significantly reducing fungal pathogen infections [22,23]. Epidermal microorganisms also help resist the invasion of entomopathogenic fungi [24]. Dominant surface bacteria on Grapholita molesta larvae enhance survival against Beauveria bassiana infections [25]. Moreover, adding Lactobacillus plantarum to axenic Drosophila melanogaster delays fungal infections [26]. Additionally, epidermal microorganisms regulate insect nutrient metabolism [27]. Certain symbiotic bacteria break down complex organic substances into nutrients readily absorbed by hosts, enhancing their survival and adaptability [28]. This relationship is especially evident in fungus-farming ants, which effectively obtain nutrition through microbial assistance [29,30,31]. Microbial communities also influence host behavior [32]. For instance, during egg-laying, female D. melanogaster transfer gut bacteria onto eggshell surfaces, causing newly hatched larvae to exhibit increased locomotor attraction and preference towards these beneficial bacteria, compared to the larvae hatched in sterile conditions [33]. Symbiotic microorganisms are also involved in the reproductive and developmental processes of insects [34,35]. For example, certain beetle species require the chemical protection provided by symbiotic bacteria on their surface during molting to ensure successful molting and continued development [36].

Recent studies have also unveiled the complexity and diversity of microbial communities on bee surfaces and within their nests, as well as the interplay with bee health, behavior, and environmental factors. Research has shown that the microbial community on the surface of the solitary red mason bee Osmia bicornis is more intricate compared to that of A. mellifera, exhibiting higher functional diversity. This includes a greater abundance of bacteria associated with nitrogen fixation and nitrate reduction, along with a higher proportion of animal–parasitic fungi, reflecting adaptive changes in microbial community functions corresponding to different bee lifestyles. Meanwhile, variations in the surface microbial communities of different A. mellifera subspecies, influenced by climate adaptation, display marked differences in fungal abundance and diversity [37]. Furthermore, the microbial community on bee exoskeletons, encompassing potential novel Actinomycetota species, holds potential implications for bee health, behavior, and hive dynamics [38]. The discovery of Actinomycetota on the cuticle of stingless bee workers further supports the role of bees as carriers of beneficial microorganisms [39]. Paenibacillus, a common bacterium found in various wild bees and their nests, albeit in relatively low abundance, is abundant on the surface of certain species like O. caerulescens. It possesses antimicrobial properties and can produce a range of secondary metabolites with antibacterial activity, aiding in the prevention of fungal penetration through bee cuticles and inhibiting fungal threats within the moist and nutrient-rich environment of wild bee nests [40]. Another study focusing on honey bees from Brazil and the United States revealed notable differences in the yeast and bacterial communities associated with these bees, thereby highlighting the crucial role of environmental factors in shaping bee microbial communities [41]. The diversity and functions of microbial communities on bee surfaces and within their nests, and their interactions with environmental factors, offer new perspectives for bee conservation strategies. This underscores the importance of future research in exploring the complete diversity of these microbial communities and potential microbial exchanges between pollinating crops. Despite these insights, comparative studies on surface microbial communities of worker bees from A. cerana and A. mellifera remain limited.

Given the growing body of evidence that insect surface microbiota can regulate host behavior [33], we hypothesize that the body surface microbiota of honey bees may be involved in the process of behavioral maturation. In this study, we investigated the composition and diversity of microbial communities on the body surfaces of worker bees from different behavioral maturation stages (newly emerged bees, nurse bees, and foragers) and from two bee species, A. cerana and A. mellifera. We observed that honey bee body surface microbiota changed dynamically with behavioral maturation in both bee species, suggesting that the body surface microbiota may play a role in the process of behavioral maturation. These findings will enhance our understanding of bee-associated microbiota and thus offer valuable insights for the regulation of honey bee behaviors.

2. Materials and Methods

2.1. Honey Bee Sample Collection

Three healthy colonies of A. mellifera and three of A. cerana were maintained at the experimental apiary of the Institute of Zoology, Guangdong Academy of Sciences, Guangzhou, China (23°50′44″ N, 113°17′17″ E). Sample collection for the experiment was completed from August to September 2024. Only scattered nectar sources were available during this period.

These colonies were chosen based on their similar colony strength and absence of visible diseases or pests. The queens in all experimental colonies were within one year old. No antiparasitic treatments were administered during the course of this study, as the bee colonies under investigation were confirmed to be healthy and free from parasitic infestations.

One brood frame was selected from each experimental colony and kept in an incubator at 34.5 °C with 65% relative humidity. After 24 h, 30 newly emerged bees (Eb) were collected from each brood frame. Nurse bees (Nb) were collected using sterile tweezers while they were feeding larvae in the brood area, with 30 individuals being sampled from each colony. Foraging bees (Fb) with pollen loads were caught at the hive entrance, and 30 were sampled from each colony. As a result, six groups were formed, namely Am_Eb, Am_Nb, Am_Fb, Ac_Eb, Ac_Nb, and Ac_Fb (Am: A. mellifera; Ac: A. cerana; Eb: newly emerged bee; Nb: Nurse bee; Fb: Foraging bee).

2.2. Enrichment of Microbes on Honey Bee Body Surface

On a laminar flow ultra-clean workbench, the hind legs with pollen loads of Fb were cut off using sterile scissors, while no treatment was applied to the Nb and Eb. Then, all 30 bees from each colony were placed in an autoclaved 50 mL centrifuge tube with 15 mL of autoclaved phosphate-buffered saline (Xinkailai Biochemical, Guangzhou, China) added. The mixture was vortexed for 5 min on a vortex mixer [26]. The bees were then removed using sterile tweezers, and the remaining liquid was dispensed into 1.5 mL sterile centrifuge tubes, with 1 mL in each tube. The tubes were centrifuged at 13,200 rpm for 10 min, and 950 µL of the supernatant was carefully removed, leaving the remaining 50 µL to be combined in one tube. Matrix-matched negative controls (using PBS buffer instead of bee samples) were implemented to eliminate nonspecific interference, with all procedures strictly following standardized experimental protocols. During microbial sample collection from honeybee surfaces, negative controls were implemented in parallel (n = 5 technical replicates). These controls rigorously followed the identical experimental. The concentrated body surface wash solution was immediately placed in a −80 °C freezer until DNA extraction. Three colonies of bees at the same stage were merged, with each colony having three replicates, meaning that every group had nine replicates. (Am_Fb, n = 9, Am_Nb, n = 9, Am_Eb, n = 9, Ac_Fb, n = 9, Ac_Nb, n = 9, Ac_Eb, n = 9).

2.3. 16S rRNA Gene Amplicon Sequencing

Using a bacterial genomic DNA extraction kit (Tiangen Biochemical Technology, Beijing, China), total DNA was extracted following the manufacturer’s instructions. The primers 341F and 805R were selected as specific primers for bacterial V3–V4 16S rRNA amplification due to their high specificity and efficiency in capturing diverse bacterial communities [42]. Catalyzed by Phusion Hot start flex 2× Master Mix (NEB, Beijing, China) polymerase, a reaction system consisting of 2.5 μL of forward and reverse primers and 50 ng of template DNA was used to generate 25 μL of amplicons. The PCR amplification reaction conditions were set as follows: initial denaturation at 98 °C for 30 s; followed by 32 cycles, each consisting of denaturation at 98 °C for 10 s, annealing at 54 °C for 30 s, and extension at 72 °C for 45 s, and a final extension at 72 °C for 10 min to complete the amplification process (Hangzhou Longji Scientific Instrument Co., Ltd., Hangzhou, China). The PCR products were purified using AMPure XT beads (Beckman Coulter Genomics, Danvers, MA, USA), and the purified products were quantified using a Qubit (Invitrogen, Carlsbad, CA, USA) fluorometer.

2.4. Data Analysis

The sequencing was performed using the NovaSeq 6000 sequencer from LC-Bio Technologies (Hangzhou) Co., Ltd. (Hangzhou, China) (https://www.omicstudio.cn/tool/1) (accessed on 9 December 2025) with 2 × 250 bp paired-end reads using the NovaSeq 6000 SP Reagent Kit (500 cycles). During the processing of paired-end sequencing data, the sample data were first split based on barcode information, with adapters and barcode sequences removed. Subsequently, cutadapt software (v1.9) was utilized to eliminate primer sequences and balancing bases from the raw data (with parameters set as ‘-g R1 -G R2 -n 1 -O 17 -m 100’). FLASH software (v1.2.8) was then employed to merge each pair of paired-end reads into a longer tag sequence based on overlapping regions (with parameters set as ‘-m 10 -M 100 -x 0.25 -t 1 -z’). Quality scanning of the sequencing reads was conducted using a sliding window approach with a default window size of 100 bp. Any read with an average quality score below 20 within the window was truncated from the window’s start position to the 3′ end fqtrim software (v0.94) is used to remove sequences shorter than 100 bp after truncation, sequences with more than 5% N content, and chimeric sequences (with parameters set as ‘-P 33 -w 100 -q 20 -l 100 -m 5 -p 1 -V -o trim.fastq.gz’). Following this, the DADA2 algorithm was invoked through Qiime DADA2 (v2019.7) denoise-paired for length filtering and denoising, resulting in ASV (Amplicon Sequence Variants) characteristic sequences and ASV abundance table. Singleton ASVs, which have a total sequence count of only 1 across all samples, were removed. We filtered out ASVs that align to chloroplasts and mitochondria, and filtered out ASVs with a relative abundance of less than 0.01 in all samples (fqtrim software, v0.94). Based on the obtained ASV characteristic sequences and ASV abundance tables, alpha diversity and beta diversity analyses were conducted. Specifically, alpha diversity was assessed using four indices, Ace, Chao1, Shannon and Simpson, to evaluate the diversity within habitats. To evaluate similarities in bacterial communities on the body surfaces of worker bees of different behavioral maturation stages within the same species, as well as among worker bees of the same behavioral maturation stages but from different species, we conducted Principal Coordinates Analysis (PCoA) using R software (v3.4.4) with the ade4 package (v1.7.13) and the vegan package (v2.5.4). To evaluate similarities in bacterial communities on the body surfaces of worker bees of different behavioral maturation stages within the same species, as well as among worker bees of the same behavioral maturation stages but from different species, we conducted Anosim (Analysis of Similarities, qiime2, v2019.7). Beta diversity analysis was performed using the Bray–Curtis method to evaluate the diversity between groups. Finally, based on the ASV sequence files, species annotation was carried out using the SILVA database (Release 138, annotation threshold: --min_confidence 0.7), and the NT-16S database (Release 20230718, annotation threshold: --min_ident 90 --min_cov80 --max_e 1 × 10⁻⁵). The abundance of species at various taxonomic levels in each sample was then statistically analyzed and summarized based on the ASV abundance table. In this study, the core microbiota was defined as microbial taxa which are consistently present in no less than 50% of the samples.

This study implemented a prevalence-based statistical approach for microbial contaminant identification with the following analytical workflow: All samples were categorized into two distinct groups—Experimental group containing target microbial communities and Control group (CK) consisting of blank controls without biological material. Using an R-based microbiome analysis pipeline, data preprocessing was performed by integrating ASV (Amplicon Sequence Variant) feature tables and sample metadata through the phyloseq package (v1.26.1). Contaminant detection was conducted via the prevalence method in the decontam package (v1.2.1), employing chi-square tests to compare ASV occurrence frequencies between true samples and controls with a significance threshold of 0.5 (indicating contaminants must demonstrate significantly higher prevalence in control samples). Finally, data purification involved complete removal of all ASVs identified as contaminants from downstream analyses, ensuring the integrity of subsequent microbiome characterization.

16S amplicon functional prediction analysis was performed as follows: 16S sequences were quality-controlled, denoised, and clustered into Amplicon Sequence Variants (ASVs). Taxonomic annotation was conducted using RDP Classifier or SILVA databases. Functional potential was predicted via PICRUSt2, Tax4Fun, and FAPROTAX by linking 16S sequences to KEGG/COG databases through phylogenetic relationships or literature-derived mapping tables, thereby generating functional pathway abundances and gene family distributions. The results were visualized with stacked bar charts.

To explore the potential interactions and co-occurrence patterns among microbial taxa, we carried out network analysis. In order to guarantee the reliability of the network and minimize the noise interference from rare taxa, amplicon sequence variants (ASVs) with a total abundance lower than 5 were excluded before the analysis. The pairwise Spearman’s rank correlation coefficients among the remaining ASVs were computed using the WGCNA package. When both the absolute value of the correlation coefficient and the significance level satisfied specific criteria, the co-occurrence relationship was considered statistically significant. Network construction was implemented using the igraph package, in which nodes symbolized unique ASVs and edges represented significant correlations. To visually demonstrate the habitat preferences of specific taxa, nodes were colored according to their abundance bias towards particular host species and behavioral stages. Specifically, the mean abundance of each ASV was calculated for each “species-behavioral stage” combination, and the ASV was allocated to the group with the highest mean abundance. A dual-coding approach was adopted for node coloring: hue represented the host species (yellow for A. cerana and blue for A. mellifera), while brightness denoted the developmental stage (ranging from light to dark for Eb, Nb, and Fb, respectively). Network visualization was accomplished using the ggraph package.

2.5. Statistical Analysis

All statistical analyses were performed using GraphPad Prism software (version 10.1.2). To compare alpha diversity between the two groups, independent samples t-tests were applied with Welch’s correction for unequal variances when appropriate. For beta diversity analysis, permutational multivariate analysis of variance (PERMANOVA) based on the Bray–Curtis distance matrix was performed. To examine differential expression of bacterial genera across groups, one-way analysis of variance (ANOVA) was conducted, followed by Tukey’s multiple comparison test.

3. Results

3.1. α and β Diversity of Microbiota on Honey Bee Body Surfaces

In the different behavioral stages of A. cerana, in terms of ASV numbers, the newly emerged bees showed the highest number of ASVs detected on their body surfaces, followed by foragers, while nurse bees had the lowest number of ASVs detected (Figure 1A,B). Regarding species diversity, the highest microbial diversity was detected on the body surfaces of forager bees, followed by newly emerged bees, with nurse bees showing the lowest microbial diversity (Figure 1C,D).

In A. mellifera at different developmental stages, regarding ASV numbers, foragers showed the highest number of ASVs detected on their body surfaces, followed by newly emerged bees, while nurse bees had the lowest number of ASVs detected (Figure 1A,B). In terms of species diversity, the highest microbial diversity was detected on foragers’ body surfaces, followed by newly emerged bees, with nurse bees showing the lowest microbial diversity (Figure 1C,D).

In both bee species, identical results were observed: nurse bees exhibited the lowest number of detected species and the lowest species diversity. The highest species diversity was detected on the body surfaces of foragers.

The Bray–Curtis distance metrics and the ANOSIM (Analysis of Similarities) algorithm were utilized to examine the disparities in ecological or biological community structures among different groups. The “Between-group” is of particular importance, as it represents the rank-transformed distance distribution of cross-group sample pairs constructed by the ANOSIM algorithm for comparison with the “Within-group”. In this analysis, it can be defined as the cross-group background distribution for statistical comparison (Figure 2A).

In the different developmental stages of A. mellifera, the body surface microbiota of the three distinct stages formed three clustering circles. Among them, there was substantial overlap between foragers and nurse bees, while only partial overlap existed between foragers and newly emerged bees (Figure 2B).

In the different developmental stages of A. cerana, the body surface microbiota of the three distinct stages similarly formed three clustering circles. Among them, there was partial overlap between foragers and nurse bees, while foragers and newly emerged bees formed significantly separated clustering circles. Nurse bees and newly emerged bees showed partial overlap (Figure 2C).

At the newly emerged bee stage, the body surface microbiota of A. cerana and A. mellifera formed two significantly separated clustering circles (Figure 2D). At the nurse bee stage, the body surface microbiota of A. cerana and A. mellifera similarly formed two significantly separated clustering circles (Figure 2E). At the forager stage, the body surface microbiota of A. cerana and A. mellifera also formed two significantly separated clustering circles (Figure 2F).

3.2. Composition of Microbiota on Body Surface of Honey Bees

In A. mellifera, the top 5 phyla on the body surfaces of forager bees were Cyanobacteriota, Pseudomonadota, Bacillota, Actinobacteriota, and Bacteroidota, whereas nurse bees harbored Actinomycetota, Pseudomonadota, Cyanobacteriota, Bacillota, and Bacteroidota. The body surfaces of newly emerged bees are dominated by Pseudomonadota, Bacillota, Cyanobacteriota, Actinomycetota, and Bacteroidota (Figure 3A and Figure 4).

In A. mellifera, the top 10 genera on the body surfaces of forager bees were Gilliamella, Erwiniaceae, Escherichia-Shigella, Brevundimonas, Enterobacter, Bifidobacterium, Faecalibacterium, Ligilactobacillus, Streptococcus, and Stenotrophomonas; nurse bees were dominated by Pseudarthrobacter, Lactobacillus, Hydrogenophaga, Enterobacter, Commensalibacter, Caulobacter, Snodgrassella, Mycobacterium, Gilliamella, and Brevundimonas; newly emerged bees were dominated by Acinetobacter, Lachnospiraceae_NK4A136_group, MethyloversatilisPseudomonas, Sphingomonas, Caulobacter, Lachnospiraceae, Muribaculaceae, Enterococcus, and SWB02. Gilliamella was identified as the dominant bacterial genus on the body surfaces of forager bees, whereas Arthrobacter predominated in nurse bees. Newly emerged bees showed a high abundance of Acinetobacter (Figure 3B and Figure 4).

In A. cerana, the top 5 phyla on body surfaces of forager bees consist of Cyanobacteriota, Pseudomonadota, Bacillota, Actinomycetota, and Bacteroidota; nurse bees were dominated by Cyanobacteriota, Pseudomonadota, Bacillota, Bacteroidota, and Actinomycetota; and newly emerged bees were dominated by Cyanobacteriota, Pseudomonadota, Bacillota, Actinomycetota, and Bacteroidota (Figure 3C and Figure 4).

In A. cerana, the top 10 genera on the body surfaces of forager bees consist of Sphingomonas, Gilliamella, Xanthomonas, Pseudomonas, Methylobacterium, Snodgrassella, Lactobacillus, Apibacter, Acinetobacter, and Massilia, while nurse bees were dominaed by Lactobacillus, Gilliamella, Sphingomonas, Enterobacter, Apibacter, Pseudomonas, Snodgrassella, Klebsiella, Apilactobacillus, and Ralstonia, and newly emerged bees were dominated by Sphingomonas, Apilactobacillus, Pseudomonas, Rhodococcus, Bacillus, Chroococcidiopsis_PCC_7203, Chroococcidiopsaceae, Brevundimonas, Ralstonia, and Burkholderia-Caballeronia-Paraburkholderia. Sphingomonas was consistently abundant, while Pseudomonas declined with maturation; Gilliamella appeared on nurse and forager bees, and Lactobacillus was significantly abundant on nurse bees (ANOVA: F(2,24) = 4.719, p = 0.0187) (Figure 3D and Figure 4).

Common patterns were observed between two bee species: At the phylum level, Pseudomonadota, Bacillota, and Actinobacteriota showed relatively high relative abundances. At the genus level, Lactobacillus was predominantly detected on nurse bee body surfaces, while Gilliamella was highly abundant on forager bee body surfaces but rarely detected on newly emerged bees.

From the chart, it is clearly observable that the Metabolism category (Metabolism, denoted by the pink section) prevails in all samples, indicating that Metabolic processes might play a pivotal role in the functional activities of worker bees. Simultaneously, the proportional disparities in other functional categories, including Environmental information processing (Environmental Information Processing, denoted by the blue section), Genetic information processing (Genetic Information Processing, denoted by the purple section), Human diseases (Human Diseases, denoted by the orange section), Cellular processes (Cellular Processes, denoted by the green section), Organismal systems (Organismal Systems, denoted by the light purple section), and others (Others, denoted by the light pink section), offer significant evidence for further exploring the functional characteristics among worker bees with different functions (Figure 5).

The network diagram employs node colors and sizes to depict the characteristics of microbial communities and their interaction relationships in A. cerana and the A. mellifera at different behavioral maturation stages (Figure 6). Nodes with yellow tones represent amplicon sequence variants (ASVs) with higher abundance in A. cerana. Specifically, light yellow, medium yellow, and deep orange correspond to its unique/enriched microbial communities during the Eb, Nb, and Fb stages, respectively. Nodes with blue tones represent ASVs with higher abundance in A. mellifera, where light-blue, medium-blue, and deep-blue correspond to its unique microbial communities during the Eb, Nb, and Fb stages, respectively. The size of nodes reflects the number of connections. The observation of tightly clustered deep-blue nodes indicates that the foraging bees of the A. mellifera possess a stable core microbial community structure. The intermixing of yellow and blue nodes suggests that the two honeybee species share environmental microbial communities at specific stages.

4. Discussion

Regarding α diversity, microbial abundance on nurse bees is significantly lower than on foraging and newly emerged bees. This difference likely relates to the distinct roles that bees play during their lifecycle. Foragers, frequently interacting with the external environment, naturally harbor higher microbial abundance [43]. A micro-ecosystem formed by specific microbial communities exists in nectar and pollen. Despite in-hive activities, forager bees, as worker bees specifically tasked with collecting pollen, nectar, and water, serve a dual function as microbial carriers within this ecosystem [44,45,46]. Firstly, there is active contact transmission. This implies that through their high-frequency flower-visiting behavior, their body surfaces directly interact with the specific microbial communities in the floral parts of different plants [44]. A second factor to consider is passive environmental enrichment, which refers to the fact that during long-distance foraging flights, they are exposed to multi-source environmental microbial reservoirs, such as soil, air, and water bodies [47]. As a result, the microbial diversity detected on the body surfaces of forager bees is the most diverse. Nurse bees, whose activities are restricted to hive interiors, show reduced microbial loads. Surprisingly, newly emerged bees often exceed the microbial abundance of nurses and occasionally surpass that of foragers. Dense hairs on newly emerged bees function as microbial collectors, enriching microbes from comb surfaces. Microbial composition thus dynamically changes as bees age, influenced by seasonal and geographical factors. These findings reveal how bee roles and physical structures subtly influence microbial community composition, providing valuable insights into bee–microbe interactions.

Β diversity analysis demonstrates clear clustering of body surface microorganisms among worker bees of different behavioral stages within the same species, significantly separating these clusters. ANOSIM analysis confirmed significant microbial community differences among behavioral groups (foragers, nurse bees, newly emerged bees) with p = 0.001, which aligns with behavioral patterns: foragers interacting with external pollen or nectar exhibit higher surface microbial diversity than hive-dwelling nurse bees. However, the moderate correlation coefficient (R = 0.6023) indicated limited inter-group differentiation, as supported by overlapping Bray–Curtis distance distributions in boxplots. In combination with the comprehensive analysis of Figure S1, this suggests that the boundaries between groups are not clearly defined, with notable overlap or similarity among the groups. This may result from shared microbial communities among co-hive workers despite behavioral differentiation, methodological limitations, including 16S sequencing overlooking fungi, viruses, and a lack of cross-seasonal and geographical sampling coverage. This result strongly supports differences driven by behavioral stages in microbial communities. Furthermore, between different bee species, even bees of the same behavioral stages form distinctly separate microbial clusters, underscoring species-specific microbial compositions influenced by biological characteristics and ecological niches.

At the genus level, most detected bacteria are common in natural environments, although core gut bacteria such as Gilliamella and Lactobacillus were also identified [48,49]. These genera were detected in the epidermal microbiota of both A. cerana and A. mellifera. Study also confirmed the presence of Gilliamella and Lactobacillus on bee cuticles using 16S rRNA sequencing [37]. Similarly, groups of five forager bees or nurse bees were allowed to walk on the surfaces of YM, MRS, and Eugon agar plates for 5 min before being released, and Lactobacillius as well as Bifidobacteria were successfully isolated in this way [41]. Consequently, the core gut microbiota of honey bees is not restricted solely to the intestinal region; it is also found on the external surfaces of bees, a finding that diversifies the transmission routes of the gut core microbiota within bee populations.

Notably, in our study, Gilliamella exhibits higher abundance on the body surfaces of foraging bees, while Lactobacillus is more prevalent on nurse bees. The detection of Gilliamella, typically found in bee hindguts, on their body surfaces is intriguing. In bee intestinal ecosystems, Gilliamella plays a pivotal role. It facilitates pollen wall degradation, enabling bees to access pollen nutrients efficiently. Additionally, it actively participates in nutrient metabolism to provide energy. Additionally, Gilliamella uniquely metabolizes certain toxic sugars, providing bees with natural protection against external threats [50,51,52]. Lactobacillius are widespread within bees and their environments, inhabiting not only bee intestines but also honey sacs, bee bread, honey, honeycomb structures, and beeswax [53,54,55]. Supplementation with lactic acid bacteria can alleviate gut microbiota imbalances caused by antibiotics and improve immune function [56]. Nurse bees transfer food by extruding honey between their mandibles, where Lactobacillus likely remains, explaining its high abundance on nurse bees’ surfaces and highlighting the relationship between bee feeding behavior and microbial distribution. Detecting Gilliamella and Lactobacillus on bee bodies suggests a complex interaction worthy of further research. Similarly, it is noteworthy that researchers have also discovered the presence of Gilliamella bacteria on the surface of Drosophila fruit flies [26]. This discovery further confirms that Gilliamella is not only a core member of the gut microbiota in bees, but may also have close and complex associations with a variety of other species.

In all worker bee samples, the metabolic function category (Metabolism) predominates, indicating that the microbial community on the bee body surface has highly active metabolic activities, suggesting that it may play a key role in the host’s physiological processes. The remaining functional categories, including Environmental Information Processing, Genetic Information Processing, Human Diseases, Cellular Processes, Organismal Systems, and Others, have relatively low relative abundances in each sample. This distribution pattern suggests that these functions may be in a relatively secondary position in the regular functional activities of the microbial community on the bee body surface.

Network graph analysis with color-coded node clustering revealed dynamic changes in body surface microbiota structure between A. cerana and A. mellifera across behavioral maturation stages. For A. cerana, yellow-series nodes showed stage-specific enrichment: light yellow for newly emerged bees (Eb), medium yellow for nurse bees (Nb) with core microbiota, and dark orange for foraging bees (Fb) with stage-specific microorganisms.

In A. mellifera, blue-series nodes demarcated stage-specific microbiota, with dark-blue nodes forming a high-density connected module in foragers (Fb), indicating a stable species-specific core microbiota. Notably, the mixed distribution of yellow/blue nodes in newly emerged bees (Eb) suggests initial microbiota colonization via shared environmental microorganisms between the two stages.

Honey bees, as highly social insects, exhibit a close relationship between their behavioral maturation and the dynamic changes in their microbial communities [57,58,59,60]. Honey bees display specific behavioral patterns when performing different tasks, such as nursing larvae, guarding the hive, and collecting food [1,45,61]. This study, by comparing the microbial composition and diversity on the body surfaces of honey bees at different behavioral stages (newly emerged bees, nurse bees, and forager bees), has revealed significant differences. Within the same bee species, a notable separation is observed in the microbial makeup on the body surfaces of forager bees and newly emerged bees, suggesting a relationship between the microbiota on worker bees’ exteriors and their functional roles. These differences not only reflect the roles that bees play at various life stages and their level of interaction with the external environment but also suggest that microbial communities may adapt to changes in bee life stages by influencing their behavior.

Combined with the findings of this study, we hypothesize that changes in the microbial communities on bee body surfaces may be intrinsically linked to these behavioral adjustments. For example, forager bees, which frequently venture out to collect pollen and nectar, exhibit a notable increase in microbial diversity on their body surfaces. These microbes may affect the foragers’ olfactory receptor gene expression, food preferences, or social behaviors, thereby influencing their foraging efficiency and success rate. Nurse bees, which have a high abundance of Lactobacillus on their cuticles, may influence the microbial colonization of larvae through food transmission or other mechanisms during the nursing process, subsequently affecting larval development and immune status. Therefore, the microbial communities on bee body surfaces are not solely an adaptation to the external environment; they could also serve as a vital mechanism for regulating bee behavior, facilitating social division of labor, and maintaining colony health. Given that the experimental data from this study failed to provide direct evidence supporting the aforementioned viewpoints, subsequent research efforts will focus on delving deeper into the specific functions and underlying mechanisms of the microbiota on the worker bees’ body surface. It is recommended that future studies systematically integrate and analyze the correlation between gene expression profile data of forager bees and the structural composition of their body surface microbial communities. Additionally, it is essential to conduct a systematic analysis of the microbial community structure within the beehive and perform a comprehensive comparative analysis between the microbial compositions on the worker bees’ body surface.

There exists a complex and subtle interplay between the dynamic changes in the microbial communities on bee body surfaces and their behavioral maturation and patterns. By combining behavioral monitoring technologies, we can further elucidate the underlying mechanisms of this interaction, providing a scientific foundation for bee conservation and sustainable beekeeping.

One notable limitation of this study lies in its restricted scope, focusing solely on the composition of microbial communities on bee surfaces within a single geographic location and a specific season. However, the diversity and functionality of bee surface microbiota are profoundly influenced by a multitude of environmental factors, encompassing but not limited to weather conditions, availability of external nectar and pollen sources, as well as broader geographical variations. Consequently, to gain a more comprehensive understanding of the dynamic changes in bee surface microbial communities and their relationship with host health, subsequent research endeavors should strive for more meticulous and in-depth explorations under diverse environmental conditions.

Firstly, weather conditions, such as variations in temperature, humidity, and precipitation, can exert direct impacts on the composition of bee microbiota [62,63]. Secondly, the diversity and abundance of external nectar and pollen sources also play a crucial role in influencing bee surface microbiota [64]. Furthermore, geographical differences, including altitude, latitude, and ecosystem type, may also lead to significant variations in bee microbial communities [65]. Through cross-regional sampling and comparative analysis, we can reveal how geography serves as a driving force behind the diversity of bee surface microbial communities, and provide a scientific basis for formulating regional bee health management strategies.

Studying honey bee body surface microorganisms, as essential components of the bee microecosystem, holds significant theoretical and practical value. Investigating the functions and mechanisms of these microorganisms can offer innovative strategies for sustainable bee health. Using 16S rRNA gene amplicon sequencing, this study demonstrates that worker bee behavioral maturation and species evolutionary divergence dually regulate microbiota assembly, providing novel insights into microbiota succession mechanisms in social insects from a host–microbe interaction perspective. Advances in high-throughput sequencing technologies have deepened our understanding of insect–microbe interactions, utilizing methods such as metagenomics and transcriptomics.

5. Conclusions

Our results reveal that the surface microbial diversity of forager bees is the highest, while that of nurse bees is the lowest. The same bee species exhibits distinct surface microbiota compositions across different behavioral maturation stages, while different bee species within the same behavioral maturation stage also show variations in microbial composition. In both species, the microbial communities of work bees of different behavioral mutation stages were primarily dominated by three major phyla: Pseudomonadota, Bacillota, and Actinobacteriota. At the genus level, significant differences exist, but a common pattern is observed: a high abundance of Gilliamella on forager bee body surfaces and a high abundance of Lactobacillus on nurse bee body surfaces. These results confirm a link between honey bee surface microbiota and behavioral maturation stages.