Introduction

Honey bees are among the most important pollinators, playing a crucial role in agroecosystems (Reilly et al. 2020). Semi-domestication of bees allows beekeepers to provide pollination services with hives to farmers at a commercial scale assuring world food supply (Aizen and Harder 2009). However, agricultural crops do not always fulfill the minimal honey bee nutritional demands in terms of protein and carbohydrates (Colwell et al. 2017). Pollen and nectar are the only sources of proteins and lipids for honey bees (Corby-Harris et al. 2018). After pollen collection by forager honey bees, the pollen is mixed with honey, inoculated with microorganisms, and stored. After hatching, honey bees start to consume bee bread, which is known to be important in relation to larval growth, longevity, body size and ovary development (Vanderplanck et al. 2014). Biochemical transformation of pollen into  bee bread is mainly made by yeasts (Haydak 1958) and Lactobacillus spp. (Chevtchik 1950), transforming pollen into a rich and complex food, increasing nutrient bioavailability of carbohydrates, enzymes, vitamins, and lactic acids (Khalifa et al. 2020). Besides pollen pre-digestion, pollen-borne microorganisms play an important role for bee nutrition and deficiencies of microorganisms in their diet have been shown to cause poor development and survival (Dharampal et al. 2019). In hives, microorganisms are transmitted between honey bees via social contact between nestmates and by adult bees feeding recently emerged honey bees. Hence, the bee microbiome is well conserved, but susceptible to environmental changes, diseases, feeding habits, and pesticide exposition (Kwong and Moran 2016; Diaz et al. 2019).

Protein is one of the key elements of honey bee nutrition and is naturally obtained by plant pollen. The nutritional value of pollen is defined by its protein percentage and amino acid composition (Crailsheim 1990), which has been shown to fluctuate from 9 to 37% depending on the origin (Liolios et al. 2015). Lipids are also an important component for honey bee nutrition, which are crucial in the production of fat, glycogen, and cell membranes (Graham 1992). Oleic and palmitic fatty acids are two of the principal fatty acids found in honey bees (Manning 2001) and imbalance of these fatty acids may cause honey bee learning deficiencies (Arien et al. 2018; Bennett et al. 2022) and alterations in immune reactions (Bedick et al. 2001). Fatty acids not only can give us insight of the honey bee nutritional status, but can also provide qualitative and quantitative information of the microbial community through specific microbial biomarkers (Grubbs et al. 2015).

As before mentioned, bee bread fermentation is made principally by microorganism like yeast and bacteria. Pesticides like fungicides, herbicides, and insecticides are known to have strong non-target effects on bee bread microorganisms and in this way indirectly affect honey bee health and nutrition (El Agrebi et al. 2020).

Protein and lipid ratios define the pollen preferences of bees (Vaudo et al. 2024) according to the nutritional requirements of the colony and the amount of essential fatty acids in pollen (Zarchin et al. 2017; Hendriksma et al. 2019). The fatty acid composition of pollen plays a significant role in shaping the normal development of the colony (Arien et al. 2020). Therefore, the availability of adequate resources in the environment is crucial. Diversified floral resources in agroecosystems provide spatial and temporal nutrient supply (Decourtye et al. 2010). In contrast, conventional crop systems with monoculture represent a challenge for honey bee nutrition (Colwell et al. 2017) since pollen and nectar nutritional values differ between crops of which some have low levels of essential fatty acids, amino acids, and minimal protein content to fulfill the honey bee nutritional needs to support colony development. In conventional crop systems, pesticides adversely affect honey bee nutrition through intensive herbicide applications which reduce bee food sources (Bretagnolle and Gaba 2015). In addition, pesticides can lead to alterations in bee bread microbial composition (El Agrebi et al. 2020). Added to this, in general, pesticides negatively affect fatty acid metabolism by suppressing fatty acid synthesis (Erban et al. 2019) and by reducing food consumption and sugar levels in the hemolymph (Tosi et al. 2017).

Avocado (Persea americana Mill 1768) and blueberry (Vaccinium corymbosum L. 1753) are two high value crops that require optimal fertilization and pest management practices to achieve maximum yield. These practices can involve the use of either biological methods (organic) or agrochemicals (conventional), such as pesticides and mineral fertilizers (Aktar et al. 2009). In Mexico, avocado is one of the leading crops produced in large areas as monocultures, while blueberry is commonly produced under closed plastic tunnels limiting natural pollination (Ramírez-Mejía et al. 2022; Wu and Guan 2021; Kendall et al. 2021). The direct effects of agricultural management have been widely studied in terms of honey bee pesticide exposure, but knowledge about the indirect effects of agricultural management on honey bee nutrition is limited (Sánchez-Bayo et al. 2016). Pesticides can reach honey through different routes, including oral intake, inhalation, or contact (Crenna et al. 2020).

Here, we studied the effects of crop system (organic and conventional) and crop type (avocado and blueberry) on honey bee nutritional cascades focusing on fatty acids and proteins in pollen, bee bread, and in honey bees. We hypothesized that conventional management would have negative effects on honey bee nutrition cascades reducing potential honey bee food sources. Our prediction is that honey bee nutrition depends on both crop species and management with conventional blueberry systems representing the lowest nutritional value for honey bees.

Materials and methods

Field site location

In central Mexico, newly created nucleus hives were established in avocado and blueberry agroecosystems both with organic and conventional production systems. Organic systems were certified by third-party organizations according to the federal regulatory program for organic produced agricultural products. Conventional sites followed the guidance of the secretary of agriculture and rural development (Secretariat of Agriculture 2006). Sites with similar size plots and characteristics (climate, age of the crop, weed management, pesticide, and fertilization schemes) were selected in avocado and blueberry (blueberry: organic 19°22′13.3"N 101°28′54.4"W, and conventional 19°22′37.7"N 101°29′21.6"W and avocado: Organic 19°28′16.9"N 102°01′32.3"W, and conventional 19°28′49.1"N 102°01′24.8"W).

45 days before peak blooming, 10 nucleus hives with a newly mated hybrid Italian/Africanized queen obtained from a local supplier and three frames of different age brood and one as reserve (honey and bee bread) were placed in each site. Hives were supplied with empty frames with organic certified bee wax foundation. At peak blooming, and after one month of hive placement, three randomly selected hives were sampled per crop, management, and site (n = 3). Only colonies showing normal development, with a solid brood pattern and a healthy amount of brood, eggs, larvae, pupae, and young adults covering all the frames, were selected for the study. These colonies had varroa levels below 1%, measured by alcohol wash (Gregorc and Sampson 2019) and showed no visible signs of other diseases. In each hive, a capped frame with pupae ready to emerge was caged with a metallic mesh frame with a 3 mm opening to allow trophallaxis, while preventing the bees from escaping. Emerged bees were marked daily on the thorax with a color code, and after 7 days, ten adult honey bees per hive were sampled for subsequent analysis. Also, five grams of randomly collected bee bread was collected from newly introduced foundation frames, using an alcohol disinfected spatula.

Pollen

Pollen from the different sites and crops above mentioned was collected from flower anthers of three randomly selected plants per site, a compound sample was made for each plant. Pollen was cold macerated using 1 ml of saline phosphate buffer, vortexed 10 s, and the supernatant was separated by centrifugation 1 min at 3000 g. Hereafter, pollen protein from the supernatant was measured using a BCA Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions at 595 nm. Final protein concentration was calculated according to the corresponding standard curve (Degrandi-Hoffman et al. 2015). Fatty acid methyl esters (FAMEs) were extracted from 200 mg of hand collected pollen in a three-step analysis: saponification with sodium hydroxide, methanol, and water for 30 min at 100 °C; methylation with chlorohydric acid and methanol for 10 min at 80 °C and extraction with hexane and methyl tert-butyl methyl ether. Fatty acids were detected on a gas chromatograph (Agilent 7890B) at 300 °C equipped with a fused silica capillary column and analyzed by Sherlock software version 3.1 (MIDI Inc. Delaware, USA). Fatty acid identification was made by comparison with an external calibration standard and quantification with the addition of an internal standard 19:0 (nonadecanoic methyl ester, Sigma) of known concentration and comparing peak areas. To determine pollen quality, saturated, unsaturated, and total fatty acid biomarkers were summed and compared in the different treatments.

Bee bread

Floral origin of bee bread was determined by microscopy at 400 ×magnification in an Olympus BX41 microscope, according to the standard procedures adapted for bee bread (Bakour et al. 2019). Identification was made by morphological comparison with pollen collected from the main crop and with the pollen atlas database (pollenatlas.net, accessed on November 22, 2022) of 500 grains per sample. Two hundred milligrams of bee bread per hive was homogenized with a mortar and pestle at 4 °C until no visible intact grains were observed at 400 ×magnification. Hereafter, bee bread pollen diversity and microbial composition were evaluated by FAME analysis as described above. For each treatment, fatty acid biomarkers were grouped according to taxonomy (eukaryotic and prokaryotic) and type (saturated, unsaturated, and total). To compare the relative contribution of the prokaryotic and eukaryotic fatty acids in the bee bread, we summed iso and ante-iso and monounsaturated fatty acids as biomarkers of prokaryotes, while eukaryotic fatty acids were calculated by summing polyunsaturated fatty acids (Table 1). Protein percentage was measured as before mentioned.

Table 1 Fatty acid used as biomarkers
Full size table

Bee head

After thawing, heads of the bees collected earlier were dissected and crushed in a mortar containing a phosphate buffer solution. The protein was separated from debris by centrifugation at 3000 rpm for 10 min and the protein percentage was assessed as before mentioned.

Bee fat body

Fatty acids were extracted from bee recently crushed abdomens according to the analysis as described above. Fat body quality was assessed by comparing the amount of saturated, unsaturated, and total fatty acids.

Honey bee gut yeasts and Lactobacillus

In order to identify the beneficial microorganisms associated with honey bee guts, nine dissected honey bee midguts per site, three for each hive, were added to two different broths. MRS (Man, Rogosa and Sharpe) to promote Lactobacillus growth and YEPD for yeasts (yeast extract, peptone, and dextrose) acidified to a pH of 4.5 adding HCl 1N as needed. Broth enrichment cultures were incubated for 72 h at 37 °C for MRS and 30 °C for YEPD. Hereafter, 40 mg of cells from the exponential growth phase was harvested and characterized by biomarker fatty acid analysis as described above.

Statistics

Fatty acids in pollen, bee bread and abdomens, pollen protein, and bee head protein were analyzed with generalized linear mixed model (GLMM) with gamma distribution, using crop and management as fixed effects and random effect for hive ID was included (Bates 2010). For visualization of the dataset, a principal component analysis (PCA) was performed in R studio (Team, R. C. 2013 version 1.2.5019) with Factoextra package (Kassambara 2017). Structural equation models were used to test the direct and indirect effects of pollen protein and fatty acids on honey bee nutrient assimilation, all the plausible variables and their interactions are tested by d-separation test for the evaluation of possible missing paths among unconnected variables (Shipley and Douma 2021). Overall fit of the network of causality was made by piecewiseSEM and model acceptance based on Fisher´s C statistics (α > 0.05) and selecting the model with the best global goodness of fit with the piecewiseSEM package (Lefcheck 2016) in R studio (Team, R. C. 2013 version 1.2.5019).

Results

Protein in pollen was significantly higher in avocado than in blueberry, regardless management type (Fig. 1A). In the same manner, bee bread protein content showed higher levels in avocado when compared with blueberry (Fig. 1B). Neither in pollen protein nor in bee bread, significant interactions were observed between management and crop.

Fig. 1
figure 1

Protein in the pollen (A) and bee bread (B) in avocado and blueberry under organic and conventional management. Values are mean ± SE and different letters represent significant differences (n = 3)

Full size image

Saturated fatty acids of pollen from both avocado and blueberry plantations were significantly higher with organic management compared to conventional management. In conventional management, pollen from blueberry had even lower levels of saturated fatty acids when compared with avocado pollen (Fig. 2A). In terms of unsaturated fatty acids, pollen from avocado with conventional management showed the highest levels, whereas conventional blueberry showed the lowest. The opposite was observed with organic management, where the amount of unsaturated fatty acids was higher in blueberry than in avocado (Fig. 2B). In terms of total fatty acids, the only difference observed was that blueberry pollen from conventional management had the lowest amount of total fatty acids (Fig. 2C).

Fig. 2
figure 2

Saturated (A), unsaturated (B), and total fatty acids (C) in pollen from avocado and blueberry under organic and conventional management. Values are mean ± SE and different letters represent significant differences (n = 3)

Full size image

Microscopy analysis of bee bread showed that in organic avocado, 78% of identifiable pollen belongs to the main crop, 83% in conventional avocado, 92% in organic blueberry, and 92% in conventional blueberry.

Total fatty acids in bee bread showed significant differences for crop management, with organic management resulting in higher levels of total fatty acids in bee bread than conventional (Fig. 3A). Saturated fatty acids of bee bread showed higher levels in organic blueberry than in organic avocado, and for conventional management, lower levels of saturated fatty acids in bee bread were observed in avocado and blueberry crops (Fig. 3B). In terms of unsaturated fatty acids in bee bread, no significant differences were found, neither for crop management nor for crop (Table 1).

Fig. 3
figure 3

Factor treatment mean of total fatty acids of the factor crop system (A, n = 6) and saturated fatty acids in bee bread from avocado and blueberry under organic and conventional management (B, n = 3). Values are mean ± SE and different letters represent significant differences

Full size image

The ratio between eukaryotic and prokaryotic fatty acid biomarkers in bee bread was higher in organic than in conventional management (Fig. 4A) and higher in blueberry than in avocado (Fig. 4B), while no interaction was observed for these two factors. Yeast biomarkers in bee bread showed no significant differences neither for crop management nor for crop (Table 2). Lactobacillus biomarkers in bee bread were higher in organic management when compared with conventional, whereas no significant differences were found for the factor crop (Fig. 5). Protein percentage of the bee head was significantly higher in avocado in both organic and conventional management, whereas bees from conventional blueberry showed the lowest level of head protein percentage (Fig. 6).

Fig. 4
figure 4

Factor treatment means of crop system (A, n = 6)) and crop species (B, n = 6) of the ratio between eukaryotic and prokaryotic fatty acids in bee bread. Values are mean ± SE and different letters represent significant differences

Full size image
Table 2 P values of ANOVA of the individual factors and their interaction for all variables measured
Full size table
Fig. 5
figure 5

Lactobacillus biomarkers in bee bread culture from avocado and blueberry under conventional and organic management (n = 3). Values are mean ± SE and different letters represent significant differences

Full size image
Fig. 6
figure 6

Protein in the bee head in avocado and blueberry under organic and conventional management (n = 3). Values are mean ± SE and different letters represent significant differences

Full size image

The amount of saturated fatty acids in bee abdomen was highest in organic avocado and lowest in conventional avocado (Fig. 7). No differences were observed between the treatments for unsaturated fatty acids in honey bee abdomen.

Fig. 7
figure 7

Saturated fatty acids in bee abdomen from avocado and blueberry under organic and conventional management (n = 3). Values are mean ± SE and different letters represent significant differences

Full size image

The first two components of the principal component analysis contributed to the 61.6% of the experiment variance. Eigenvalues and factor loadings of the principal component analysis are presented in Table 3. In the first component (Dim1), a clear separation was observed between the conventional avocado and blueberry groups. In the first component (Dim 1), conventional blueberry group is in the bottom followed by conventional avocado, organic blueberry, and in the top, organic avocado. This separation is caused principally by an increase in saturated fatty acids in bee bread, total fatty acids in bee bread, saturated fatty acids in pollen, and unsaturated fatty acids in bee bread (Fig. 8).

Table 3 Eigenvalues and factor loadings of the fatty acids (FA) that mainly contribute to the treatment separation of the principal component analysis
Full size table
Fig. 8
figure 8

Principal component analysis biplot of all variables examined

Full size image

The structural equation model of honey bee nutrition cascades best fitted model p = 0.27, explained 68% of the variation in bee bread protein, 67% in bee bread microorganisms, 76% of bee protein, 80% in linoleic fatty acid, and 92% on oleic fatty acid. Oleic and linoleic acids in fat were significantly positively correlated with total fatty acid in pollen. Besides, bee bread protein, bee protein, linoleic acid in fat, and bee microorganism abundance in pollen were positively correlated with pollen protein. In addition, bee bread microorganisms were positively correlated with total fatty acids in pollen and linoleic fatty acid with unsaturated fatty acids of bee bread. Finally, oleic acid in fat was negatively correlated with bee bread protein and multigroup analysis showed that all paths are constrained to the global models showing no differences depending on the management or crop (Fig. 9).

Fig. 9
figure 9

Structural equation best fitted model (p = 0.27) of nutrition cascades in honey bees. Green and red arrows are significant positive and negative coefficients. The number in the circles represent regression coefficients. Only significant regressions are presented, and d-separation test showed no significant missing paths

Full size image

Discussion

Here, we show that both crop species and crop system determine honey bee nutrition, though crop species to a higher extent than crop system. Overall, our results clearly show that avocado pollen provides higher-quality honey bee nutrition than blueberry pollen. As predicted, crop species and system significantly affected honey bee nutrition with the strongest contrast found in conventional blueberry providing lowest values for honey bee nutrition. Crops, define the availability of alternative pollen sources to the main crop, being blueberry the one with the poorest pollen diversity in bee bread, most likely because bees are kept inside plastic macro tunnels. Also, avocado crops in the area are widely distributed in large extensions in the avocado strip, preventing bees to obtain alternative sources of pollen to the main crop (Saenz-Ceja et al. 2022). In addition, regulatory programs prohibit weeds in crops destined to exportation reducing alternative sources of pollen (De la Federación 2002). Our results showing low pollen protein in blueberry is in accordance with Colwell et al. (2017), which has been further shown to lead to poor colony development (Dufour et al. 2020), since such low protein levels do not fulfill the minimal honey bee nutritional requirements (Haydak 1970). Though a slight increase in protein content in bee bread compared to pollen was seen for blueberry systems, this was still lower than what was found for avocado systems. Assimilation of protein in honey bee head seems to be linked with both crop species and management, where bees from plantations with low pollen protein percentage showed low levels of protein in the head, which was further underlined in conventional crops. Protein is important for nursing, hypopharyngeal, and mandibular glands, which is provided to larvae and young bees by nurse bees storing and disseminating protein (Crailsheim 1990). After emerging, honey bee head protein has been shown to increase by 92% (Haydak 1934), which is crucial for carbohydrate metabolism, antioxidant activities, detoxification, and learning (Zaluski et al. 2020).

The observed increase in saturated fatty acids in pollen from organic agroecosystems could be attributed to several factors associated with organic farming practices. Organic management often involves the use of natural fertilizers and reduced pesticide application, which may contribute to alterations in the fatty acid profile of pollen. Overall, the amount of fatty acids, saturated as well as unsaturated, was higher in bee bread than in pollen, most likely mediated by microbial fermentation and the addition of other pollen sources. However, though avocado pollen from conventional agroecosystems had the highest amount of unsaturated fatty acids, this was not transformed into higher levels of unsaturated fatty acids in bee fat. Unsaturated fatty acids play an important role in honey bee nutrition (Manning 2001) and disease prevention (Feldlaufer et al. 1993).

The ratio between eukaryotic (plants and fungi) and prokaryotic (bacteria) fatty acids can provide information about the general microbial composition in the bee bread during pollen fermentation revealing useful information about the relative abundance of microorganisms and their fatty acid contribution to honey bee nutrition. Interestingly, in the present study, this ratio differed both between crop species and management system. Organic management increased the eukaryotic/prokaryotic fatty acid ratio and in terms of crop species. Bee bread from avocado agroecosystems showed a better balance between eukaryotic and prokaryotic biomarkers compared to blueberry. Bee bread microbial diversity and abundances are related to pollen origin and nutrients (Vásquez and Olofsson 2009) and can be altered by pesticide exposition (Bartlewicz et al. 2016). Myristic, palmitic, and linoleic fatty acids play an important role in honey bee attraction to pollen (Manning 2001), which, in the present study, were more abundant in bee bread from organic agroecosystems than in the conventional ones.

Contrary to other bees, honey bees seem to lack the ability to select pollen with desirable nutritional characteristics (Corby-Harris et al. 2018), which, however, in this study, was quantified from corbicular pollen without considering possible effects of pollen-borne microorganisms. Indeed, it has been shown that pollen-borne microorganisms like yeasts and Lactobacillus represent a decision choice for the bees controlled by volatile emissions (Rering et al. 2018). Pollen microorganisms not only release basic nutrients to bees, but also play a direct role in bee nutrition (Dharampal et al. 2019). Principal component analysis showed clear differences in the crop species and management studied, because of the changes in the fatty acid profiles, protein content of bee bread and pollen, and in bee nutrient acquisition. Although the specific mechanisms underlying these differences require further investigation, it is possible that variations in management practices and pesticide usage, or nutrient availability might be responsible for the observed differences. Several studies have tested the effects of organic farming on the nutritional profiles of the crops and the consequent effects on human nutrition (Capouchová et al. 2021; Maggio et al. 2013; Montgomery and Biklé 2021). However, to our knowledge, this is one of the first studies to evaluate this on bees. As an overall synthesis of our work, the structural equation model showed how fatty acid and protein assimilation is mediated by different drivers in the nutrition cascades, which should be taken into consideration, when studying honey bee nutrition. Despite the negative effects of monocultures in the honey bee nutrition, pollen diversity, protein, and bee bread microorganisms honey bee nutrition cascades seem to be very well conserved, in concordance with the nutrition pathways in bees (Kunieda et al. 2006).

This study presents basic findings regarding fatty acid and protein cascades in honey bees. The main results show that both crop species and management systems alter fatty acid and protein assimilation but not the honey bee nutrition cascades. Site replication is labor intensive due to the difficulty of finding sites with similar characteristics, but in future experiments, this should be considered thoroughly to improve the robustness of the results. Our findings suggest that pollen quality should not only be addressed with protein content, but that also fatty acids should be taken into consideration, which is important in honey bee health and nutrition.