Abstract
Stingless bees are major flower visitors in the tropics, but their foraging preferences and behavior are still poorly understood. Studying stingless bee interactions with angiosperms is methodologically challenging due to the high tropical plant diversity and inaccessibility of upper canopy flowers in forested habitats. Pollen DNA metabarcoding offers an opportunity of assessing floral visitation efficiently and was applied here to understand stingless bee floral resources spectra and foraging behavior. We analyzed pollen and honey from nests of three distantly related stingless bee species, with different body size and social behavior: Melipona rufiventris, Scaptotrigona postica and Tetragonisca angustula. Simultaneously, we evaluate the local floristic components through seventeen rapid botanical surveys conducted at different distances from the nests. We discovered a broad set of explored floral sources, with 46.3 plant species per bee species in honey samples and 53.67 in pollen samples. Plant families Myrtaceae, Asteraceae, Euphorbiaceae, Melastomataceae and Malpighiaceae dominated the records, indicating stingless bee preferences for abundant resources that flowers of these families provide in the region. Results also reinforce the preference of stingless bees for forest trees, even if only available at long distances. Our high-resolution results encourage future bee-plant studies using pollen and honey metabarcoding in hyper-diverse tropical environments.
Similar content being viewed by others
Introduction
Plant-pollinator interactions mediate most flowering plant reproduction, maintaining terrestrial ecosystems and crops1. The current decline in pollinator abundance and diversity worldwide threatens pollination services, with direct consequences to nature conservation and food security2. Therefore, understanding interactions between pollinators and flowering plants became crucial, since it can provide an information framework to subsidize conservation policies and decisions in a changing world. The most important group of animal pollinators—the bees—are completely dependent on floral resources to complete their life cycles3, while 87.5% of the animal-pollinated flowering plants depend on bees to reproduce4. Interaction between bees and angiosperms has been a major research focus in the last two decades, due to the potential threats of massive insect and bee declines5 on ecosystem services and food security6,7. In this context, pollen, and honey DNA metabarcoding emerged as an efficient technique to identify plant taxa visited by pollinators based on samples extracted from bees' bodies or nests8,9,10,11. DNA metabarcoding of pollen and honey has been largely applied to temperate systems, and recently to (sub) tropical species of stingless bees12,13.
The stingless bees (Apidae, tribe Meliponini) comprise c. 500 species of eusocial bees, most of which occur throughout tropical America (more than 400 species), but also in Africa, Asia and Australia3,14. The role of stingless bees as pollinators of the neotropical flora could become even more relevant under climate change scenarios, if the predicted expansion of warmer temperatures pushes the distribution of predominantly temperate bees, such as Apis and Bombus, into cooler regions6,15. Although domestication is still restricted to a few species, stingless bees are also explored commercially for honey production16, which can reduce the use of introduced honeybees and their impact on native species in these regions12.
As all eusocial bees, stingless bees show a predominantly generalist pattern of floral exploitation, i.e. they visit a large number of species in several plant families, supposedly disregarding specific floral traits17. Particularities exist though, since stingless bees exhibits a huge diversity of body size (1.8 to 13.5 mm)3, flight distance (0.3 to 3 km)18, and foraging behavior19. However, the breadth of stingless bees diet—that is, how many different species of flowering plant they can forage on—and the extent of their role as pollinators of tropical plants are still largely open questions. Most studies aiming to answer these questions face methodological difficulties due to inaccessibility of visited flowers, which mostly occupy upper canopy strata especially in rainforests20, and hyper diversity of tropical plants, thus hampering easy identification through bee-pollen morphology21,22.
21,22In this study, we explored the diet breadth of three distantly related species of stingless bees of different body sizes and flight ranges in a hyper-diverse tropical ecosystem, the Cerrado savannas of central South America. The Cerrado is the most species-rich savanna in the world and a hotspot of biodiversity23. The flora encompasses > 13,000 native plant species24 in a highly patchy vegetation with several different physiognomies, ranging from grasslands, marshlands, and typical savanna to closed canopy riverine forests along waterways25. The proportion of pollinator-dependent species in the Cerrado flora is still unknown, although this number is likely to be similar to that of tropical forests4, with some authors estimating c. 60% of angiosperm species being bee dependent26,27.
We analyzed pollen and honey from the pots of the nests (henceforward pot-pollen and pot-honey respectively) from three commonly managed stingless bee species (Melipona rufiventris Lepeletier, 1836, Scaptotrigona postica (Latreille, 1807) and Tetragonisca angustula (Latreille, 1811)) native to the Cerrado to investigate: (i) How broad is the diet breadth of stingless bees in a hyper-diverse flora? (ii) Which plant species and families are the most important sources of pollen and/or nectar for stingless bees in the area? (iii) What can pollen and honey metabarcoding reveal, when combined with floristic surveys of the area, about stingless bees foraging behavior, particularly foraging distances, and floral preferences? We also discuss how efficiently pollen and honey metabarcoding identified plants visited by bees in the area, considering the low DNA sequence coverage of neotropical plant species in public databases28, and the potential role of this technique in improving ecological understanding of bee-plant interactions in the tropics.
Material and methods
Study site
Our study was conducted in the Ecological Reserve of the Brazilian Institute of Geography and Statistics (IBGE) (15°56′41″ S and 47°53′07″ W) that, together with the contiguous Brasilia Botanic Garden and the University of Brasília Experimental Field Station, preserves an area of c. 10,000 ha of native Cerrado in the Distrito Federal, Brazil. The IBGE reserve occupies a central position within the Cerrado domain and was chosen as our study site for being one of the most well-studied areas of Cerrado, with good prospects of building a relatively robust plant DNA reference library, a requirement for our analyses (see below). The climate in the area is typical tropical savanna climate (Aw Köppen classification system) with dry winters and rainy summers with an average annual precipitation of 1453 mm and altitude ranging from 1048 to 1160 m. The IBGE reserve contains the main vegetation types typical of the Cerrado domain: savannas (cerrado sensu stricto), palm swamps (veredas), grasslands (campo limpo and campo sujo) and riverine forests (mata de galeria), surrounded by natural and agricultural areas (Fig. 1). This habitat heterogeneity results in high plant biodiversity. The last published floristic survey in the area recorded 1798 species of angiosperms, of which 1457 are native, distributed in 138 families and 724 genera29.
Stingless bee species and nest material sampling
Three native stingless bee species were chosen for our study based on differences in body size, differences in foraging behavior, and phylogenetic relationships. Melipona rufiventris is the largest with a body length of c. 9.5 mm, Scaptotrigona postica has an intermediate body size varying from 5.7 to 6 mm, and Tetragonisca angustula is amongst the smallest stingless bees with a total body length of c. 4 mm. Phylogenetically, the three genera are not closely related, i.e. they do not belong to sister groups30. Melipona rufiventris is typically found in the more open vegetation types of eastern and central Brazil, S. postica occours in a broader region in Central, Northeast and southeast Brazil, also associated with open vegetation, while T. angustula is widespread in the Neotropics (Mexico to South America)14.
The three species are commonly managed by local beekeepers and were chosen also for their relatively easy management in artificial colonies (Figure S1). The decision to use artificial colonies for sampling in our study relied on four main points: 1. to preserve the natural bee community in the area, by not destroying any nests for sampling; 2. to facilitate sampling, as pollen and honey are stored in accessible compartments in the wooden box; 3. to facilitate the access to the colonies, which in natural conditions would be randomly distributed, depending on availability of cavities, and 4. to make sure nests would have a strong population and enough pot-pollen and pot-honey for sampling. Eighteen pre-established nests were installed: three of M. rufiventris, eight of S. postica and seven of T. angustula. The nests were installed at a distance of about 5 m from each other and c. 150 cm above ground level, in typical savanna or cerrado sensu stricto29 where most plant species are subshrubs, shrubs or small trees.
Nests were moved to the study area eight weeks prior to the first sampling to allow bees time to accumulate pollen and honey from local species in the artificial nests. Pot-pollen and pot-honey samples were collected from the nests (Figure S2) once every 15 days for five months (July 2019–November 2019). This period started at the height of the dry season, moved through the transition between dry and wet seasons and ended at the beginning of the wet season. Samples were always collected from new pots—that is, those built in between two subsequent sampling events. Micropipettes (1000 uL) were used to collect honey from the pots, while pollen was collected with plastic straws, which perforates the pollen mass while collecting it at the same time. Samples were subsequently stored in falcon tubes and stored in a −20 °C freezer until extraction. In total, 191 samples (115 of pollen and 75 of honey) were collected from the three species: 29 of M. rufiventris, 81 of S. postica and 74 of T. angustula.
Metabarcoding protocol
DNA extractions of pollen samples from pot-pollen and pot-honey followed different methodologies, due to the different natures of the samples. For pot-honey, we extracted DNA using the Machery-Nagel (Düren, Germany) NucleoSpin Food Kit; for pot-pollen we used the Machery-Nagel (Düren, Germany) NucleoSpin Plant II.
Pot-pollen DNA extraction
To extract pollen genomic DNA, we added to the pooled samples (weight ranging from 0.1 g to 2 g) 4 mL of deionized and autoclaved water and homogenized it using a vortex. We then placed 200 µL of this emulsion in a 1.5 mL microcentrifuge tube and centrifuged it for 15 min at 8000 rpm. We discarded the supernatant material, froze the pellet obtained in liquid nitrogen, and then used mortar and pestle to break the pollen exine, and the NucleoSpin Plant II Kit to promote cell lysis and to isolate the DNA according to the manufacturer’s instructions.
Pot-honey DNA extraction
To extract pollen genomic DNA from honey, we added deionized and autoclaved water to the samples until the volume of each sample tube reached 1.5 mL. We incubated the tubes at 65 °C for 30 min and, over that period, inverted the tubes slowly to homogenize the material. We then pooled the honey samples collected from the same nest and day by pouring them into falcon tubes, to which deionized and autoclaved water was added until completing 10 mL. Afterwards, we centrifuged these pooled samples for 15 min at 5000 rpm and discarded the supernatant material. Each precipitated pooled honey sample was resuspended in 200 µL deionized and autoclaved water and placed in a 1.5-mL microcentrifuge tube. This procedure was done twice. Finally, we centrifuged the samples for 15 min at 5000 rpm, discarded the supernatant material, dried the pellet in a drying cabinet at 35 °C, and then ground the samples inside the microcentrifuge tube using micro-pestles and liquid nitrogen. We then used the NucleoSpin Food Kit to promote cell lysis and to isolate the DNA according to the manufacturer’s instructions.
The protocol of amplification utilizes a dual-indexing strategy9 to amplify the ITS2 region, using the primers ITS-S2F and ITS4R. Primer sequences, references and other amplification methodological details can be found in Sickel (2015)9 and Campos et al. (2021)31. The triplicate PCR reactions were combined per samples, well mixed and checked on 1% agarose gel using 5 uL of the combined products for quality. PCR products of each sample were then normalized to ensure library concentrations were approximately equivalent across samples using the SequalPrep Normalisation kit (Invitrogen, CA, USA) according to the manufacturer's protocol. After pooling the multiplex-index samples, we quantitated the pools with a dsDNA High Sensitivity Assay on a Qubit Fluorometer and assessed fragment lengths with a Bioanalyzer High Sensitivity DNA Chip (Agilent Technologies, CA, USA). For library dilution, we followed the Illumina Sample Preparation Guide for a 2 nM library where 5% consisted of PhiX control library. In addition, the reagent cassette of the sequencing kit was spiked with the Read1, Read 2 and index primers according to Sickel et al. (2015). Sequencing was then performed on the Illumina MiSeq system at the University of Würzburg. Sequence data are available at NCBI (Bioproject 976708).
Bioinformatic data analyses
We used VSEARCH v2.14.232 to join paired ends of forward and reverse reads and to remove reads shorter than 150 bp, quality filtering (EE < 1)33, de-novo chimera filtering (following UCHIME3)34, and determination of amplicon sequence variants (ASVs)34, as previously done for pollen metabarcoding networks12. Reads were first directly mapped iteratively with global alignments using VSEARCH against five flowering plant ITS2 reference databases (see below for construction details) for the study region and an identity cut-off threshold of at least 97% (higher percentages are prioritized). These references databases were created with BCdatabaser, then automatically curated35 from GenBank entries with default parameters (length between 200 and 2000 bp, maximum nine sequences per species), from the following species lists: 1) all plant species recorded from IBGE. This database was then manually curated to remove voucher-less entries for greater trustworthiness. Remaining unclassified sequences were then tracked by iterative searches against geographically broadening public sequence reference data, i.e., 2) species lists of the flora of the Distrito Federal, then 3) the large, neighboring state of Goiás, and 4) the entire Cerrado domain flora to increase completeness of reads. These reference databases were created with the BCdatabaser36 from GenBank entries given above mentioned species lists and default parameters (length between 200 and 2000 bp, maximum nine sequences per species). For still unclassified reads, we used SINTAX37 to assign taxonomic levels as deep as possible. Finally, we used 5) global reference database36. After classification, we performed plausibility checks according to geolocation and phenology with the results to verify validity. Thirteen species were automatically matched to genus level only, but were manually attributed to species based on these being the only species of the genus to occur in the Distrito Federal.
Floristic surveys and vegetation characterization
To improve our knowledge of the flora surrounding the nests, we conducted Rapid Botanical Surveys (RBS) in small plots that were demarcated in loco as homogeneous to vegetation type. These plots were exhaustively surveyed for all flowering plant species of all life forms, fertile or not, by a team of 3–5 researchers, where one was the booker, i.e. the most experienced person in the group, who identified the plants in the field and discarded duplicated species; other team member collected and pressed the vouchers (for additional methodological details see38).
Eleven RBS plots had been initially chosen to correspond to one plot near the nests (henceforward nest plot) and ten other plots established at the vertices of two pentagons; the inner pentagon was established with its vertices at 700 m from the nests and the outer pentagon with vertices at 1500 m from the nests. These distances were chosen based on the literature of the flight capabilities of other stingless bees18. These eleven RBS plots mostly fell in areas of well-preserved savanna within the IBGE Reserve, ranging from the more open, grass and herb-rich areas with few shrubs and trees (campo sujo), to dense savanna woodland (cerradão); one outer pentagon plot fell in disturbed cerrado and another in heavily degraded secondary vegetation out of the IBGE. Because none of the plots fell in riverine gallery forest, we included six additional RBS plots in this vegetation type: three in the riverine gallery forest nearest to the nests (Nascente do Roncador, c. 630 m from the nests), and three in a more distant gallery forest (Ponte do Corujão, c. 2070 m from the nests), measured as the crow flies, totaling 17 RBS plots. Lastly, we also surveyed the plants and weeds growing in the ornamental gardens associated with the Main Building and Seat of the Reserva Ecológica do IBGE, which is located c. 650 m from the nests. All specimens collected in RBS inventories were identified by the author CEBP and JEQF and were deposited in the UB Herbarium (University of Brasilia) and the records are available online in the Species Link Network (https://specieslink.net/search/) by searching on the collector name "Projeto Barcode Cerrado". From all species collected, three are categorized as endangered at some level: Virola urbaniana Warb. and Cedrela fissilis Vell. (vulnerable) and Anemopaegma arvense (Vell.) Stellfeld ex de Souza (endangered). The vast majority of herbarium collections (all collections of the two vulnerable and one endangered species) did not require uprooting.
Data integration
The 30 most abundant plant species in the pollen and honey samples were classified by ubiquity (i.e., presence in pollen or honey samples of two or all bee species). We then crossed this information with data from the RBS floristic surveys: distance from the nests, i.e., if they were sampled at nest plot, inner pentagon plots, outer pentagon plots, nearest or furthest gallery forest plots, or the gardens. These 30 species were also characterized from the literature in terms of type of resources (e.g. pollen, nectar, oil, resin), their habitat (savanna, forest or cultivated/weedy) and habit (trees, shrubs, subshrubs, hemiparasites) (Table 1).
Statistical analysis of pollen and honey samples
Data was processed for analyses using R 4.2.239 and the packages phyloseq40 vegan41, bipartite42, circlize43 and viridis44 (all code required to replicate the analyses is available at github.com/tncvasconcelos/barcode_cerrado). In R, non-plant sequences were removed from the dataset, as well as the data transformed to relative read abundances (RRAs) per sample. ASVs that were classified as the same plant species were accumulated at the species level. Low abundance taxa that contributed less than 1% to a sample were removed from those samples. The Shannon diversity index was calculated for each sample (pollen and honey) for each bee species. The diversity was tested for significant differences between stingless bee species using the Kruskal-Wallace test, separately for pollen and honey samples. We also performed an NMDS ordination to visualize clustering of samples of pollen and nectar using Bray–Curtis beta-diversity dissimilarities. The ordination represented by proximity of points shows how similar two samples are in terms of composition and abundance of taxa. We tested for differences between species by using a PERMANOVA, separately for honey and pollen samples. We further calculated network indices of the three stingless bee species to account for their overlap and complementarity in the visited plant resources, i.e. the d’ for each bee species and H2’ for the entire network.
Results
Pollen and honey metabarcoding yielded a total of 5,079,123 quality filtered reads, with mean throughput per sample of 27,307.11 reads ±1756.635 (SE). Significant reads (more than 1% of reads in any sampling) accounted for 110 ASVs, in 86 genera and 40 plant families; c. 36% of these reads were only matched to generic level or above. In total, 95 out of the 110 ASVs recovered from the samples had been previously recorded in the IBGE Reserve flora29; 12 of the 15 absent taxa were exotic cultivated or weedy species. A detailed list of all significant plant species present in pot-pollen and pot-honey samples is available in Table S2. Reads below the threshold value (190 ASVs) still showed a high number of matches to species known to occur in IBGE (86 species, c. 45%) of which 41 were also recorded by us in the RBS floristic inventories.
How broad is the floral resource exploitation by stingless bees in Cerrado Savanna?
Overall, the interaction network was highly generalized (H2’ = 0.2895575), and consequently also that of the three species within the network (M. rufiventris d’ = 0.22, S. postica d’ = 0.04, and T. angustula d’ = 0.22) (Fig. 2). More than a half of plant species appeared in the samples of at least two of the bee species. In terms of relative plant species abundances as evaluated by combined honey plus pollen samples, bees showed an opportunistic foraging pattern, with most plant species with low abundance and a few highly abundant.
Differences among pattern of floral sources exploitation of bee species
The comparison between alpha diversity among samples of different bee species showed that the plant species richness in the pot-honey was higher than in the pot-pollen for all species, but the difference was only significant for M. rufiventris (Fig. 3). In a comparison among the three bee species, Shannon diversity of plant species in pollen samples was not significantly different between bee species (Kruskal–Wallis rank sum test, chi-squared = 4.5138, df = 2, p-value > 0.05), neither was plant species richness (Kruskal–Wallis rank sum test, chi-squared = 1.4733, df = 2, p-value > 0.05). The same applied for honey samples with Shannon diversity (Kruskal–Wallis rank sum test, chi-squared = 2.6469, df = 2, p-value > 0.05) and species richness (Kruskal–Wallis rank sum test, chi-squared = 4.9389, df = 2, p-value > 0.05).
Although the most frequent plant species are shared among the three stingless bee species, samples from different bee species have several compositional particularities, as shown by the NMDS (Fig. 4). The NMDS showed the composition of plants collected differed strongly between bee species, both for pollen (PERMANOVA, df = 2, R2 = 0.12516, F = 7.2246, p < 0.001***) and honey (PERMANOVA, df = 2, R2 = 0.10751, F = 3.8548, p < 0.001***). The NMDS also points to different plant species composition between samples of three species, but in the honey samples little ordination is observed (Fig. 4A). Among pollen samples, on the other hand, we can observe different patterns among the three species, with more overlap between M. rufiventris and S. postica (Fig. 4B).
Most frequent plant species and families recovered from pot-pollen and pot-honey samples
The 30 ubiquitously found plant species in pot-honey and pot-pollen samples belong to the following families: Myrtaceae, Loranthaceae, Anacardiaceae, Phyllanthaceae, Sapindaceae, Melastomataceae, Euphorbiaceae, Primulaceae, Nyctaginaceae, Rosaceae, Asteraceae, Malpighiaceae, Cloranthaceae, Piperaceae, Fabaceae, and Clusiaceae (Fig. 5, Table S3). Out of 110 ASVs, some plant taxa stand out as most frequent in samples of all the three bee species: Myrtaceae: Syzygium cumini (L.) Skeels, Myrcia linearifolia Cambess. and Myrcia pinifolia Cambess.; Loranthaceae: Struthanthus/Psittachanthus, Anacardiaceae: Tapirira guianensis Aubl., Phyllanthaceae: Richeria grandis Vahl, Sapindaceae: Matayba guianensis Aubl., and Melastomataceae: Miconia stenostachya DC. Most of them offer pollen and nectar, except the pollen-only Miconia and the two Myrcia species. Thirteen of these ubiquitous species were nectar or oil flowers (i.e., they provide additional resources beyond pollen). Five highly abundant reads were incompletely matched, i.e. could not be identified to species level (Eucalyptus sp., Myrtaceae sp., Myrsine sp., Croton sp, Struthanthus/Psittachanthus) but Croton, Eucalyptus, Psittacanthus and Struthanthus are known to produce floral nectar. Pollen-only flowers were found in honey samples of all three species: Myrsine sp, Blepharocalyx salicifolius (Kunth) O.Berg, Piper aduncum L., Miconia leucocarpa DC. and several Myrcia species, thus indicating some kind of mixing nectar and pollen trips, manipulation or spill-over inside the nests. Pollen records include similar diversity numbers of pollen-only flowers and flowers offering nectar and pollen. Only four out of the 110 ASVs were not recorded in our RBSs: Baccharis dracunculifolia DC. and M. pinifolia Cambess., both native Cerrado species that occur in IBGE, and exotic Eucalyptus sp. and Toxicodendron succedaneum (L.) Kuntze.
These 30 most abundant plant species had the following characteristics: all were woody perennials, and most were trees or large shrubs (except one climber and one hemiparasite). They could be grouped into two dominant groups according to a combination of the habitat and floral resources. Group 1 is composed of riverine forest species that offer pollen and nectar, recorded as very common in the Forest RBS surveys: Syzigium cumini, Tapirira guianensis, Richeria grandis, Matayba guianensis, Rubus urticifolius Poir. Group 2 includes Cerrado shrubs or trees offering only pollen and recorded as common around the nests, in the Cerrado RBS surveys: Myrcia linearifolia, Blepharocalyx salicifolius and Maprounea guianensis Aubl. (Table 1).
Discussion
Pollen and honey metabarcoding of three stingless bee species in the genus Melipona, Scaptotrigona and Tetragonisca revealed a broad generalized set of used floral sources regarding number of species and plant families explored. We recovered 110 plant species in pot-honey and pot-pollen retrieved from nests of the three stingless bee species. This reveals a broader spectrum of food sources than found by previous surveys on neotropical stingless bees that relied on non-DNA based methods such as field observations, field collections, and palynological studies. For instance, non-DNA based studies in another hyper diverse area in the Neotropics, the Amazon, revealed from 80 to 122 pollen types in nests and pollen loads of 10–15 species of stingless bees45. Other similar studies in species-rich areas of the Neotropics show comparatively lower numbers22. While these studies recorded a maximum of five to eight plant species per bee species, we found a mean of 46.3 plant species per bee species in honey samples and 53.67 in pollen samples. The interaction network and high number of species found in honey and pollen of the three analyzed stingless bee species point to a generalist foraging behavior, known to be common in eusocial bees and in stingless bees in particular17,20. It also points to probable scouting investigative trips, followed by heavy recruitment and opportunistic behavior when a high-quality resource is located, with most plant species with low abundance and a few highly abundant. Note that our results may still be an underestimation, since samples were collected during only 6 months, i.e., did not include all seasons.
The power of pollen DNA metabarcoding in revealing broad food sources for stingless bees had only been demonstrated before in Southeast Asia and Australia. In Sumatra, a study of Tetragonula laeviceps (Smith, 1857) using pollen metabarcoding coupled with light microscopy revealed 99 plant species46. Similarly, a study with Tetragonula carbonaria (Smith, 1854) in Queensland retrieved 302 plant species in pollen samples across seven sites at different seasons of the year over a two-year period13. These are promising results, especially when considering expanding this technique to tropical and subtropical forests in the Neotropics. Studies of pollination and floral biology in these habitats is often very difficult because flowers are often in the upper canopy and difficult to reach. Therefore, direct observations of bees on flowers in tropical and subtropical forests are rare20, and records of stingless bee—flower interaction in these environments became almost restricted to palynological analyses of pollen loads or pot pollen analyses22. Although their utility is undeniable47, morphological identification of pollen may become obsolete for pollination biology studies when compared with the efficiency of DNA metabarcoding to identify different plant species in extremely rich floras.
Pollen analyses via DNA metabarcoding also have the advantage of revealing unexpected food sources used by bees that would perhaps be unnoticed in studies using other methodologies. For instance, our analyses revealed that DNA from 13 wind-pollinated plant species were found among the 50 most abundant species in the sample of the three species, including monocots (Poaceae, Cyperaceae), eudicots (Euphorbiaceae: Acalypha, Amaranthaceae: Amaranthus, Urticaceae: Cecropia, Cannabaceae: Trema), and a conifer genus, the introduced Pinus (Table S2). The presence of non-melitophyllous angiosperms and gymnosperms (e.g. Cyperaceae, Poaceae, Taxaceae and Pinaceae) is relatively common in melisso-palynological studies12,22,46. Despite previous studies demonstrating that pollen from anemophilous species might be a contamination in melisso-palynological samples48, bees are regularly reported visiting such taxa49,50. Our results confirm active collection of pollen from anemophilous species, since their abundance in our analyzed samples is relatively high. One of the most abundant plant species in the pollen analysis was Hedyosmum brasiliense Mart. ex Miq (Chloranthaceae), widely cited in the literature as wind-pollinated51. This species was not only recorded in the pollen samples of all three species of bees, but was amongst the 10 most abundant records for T. angustula in our results. These results reinforce the theory that anemophilous plants, which account for 10% of angiosperms and most gymnosperms, produce enough pollen52 to be attractive to social bees, under certain conditions of colony size and food demands. However, the role of bees and other insects as true pollinators of anemophilous plants remains unresolved, in spite of the importance of wind-pollinated crops53 and of the several records showing that anemophilous plant pollen is important for several bee species (see references above).
A surprising and novel observation is the significant amount of Marchanthyophyte DNA from the liverwort Dumortiera hirsuta (Sw.) Nees found in pot pollen from the three studied stingless bee species (Table S2). Future research would need to seek evidence if the DNA results from the collection of spores or perhaps some chemical compounds from liverworts by stingless bees. Bees collecting spores from fungi and plants is not a novelty, as there is evidence of active collecting54 as well as records of spores in samples of pollen and honey55. In lieu of pollen, spores supposedly have nutritional benefits56. Stingless bees might also visit liverworts to collect lipidic compounds, e.g. terpenoids used in communication among individuals57 commonly occurring in liverworts58.
The high degree of overlap between plant profiles found in the honey of the three bee species suggests that bees may be competing for the same nectar resources. Pollen plant profiles on the other hand showed far less overlap between species, corroborating evidence that pollen exploitation and digestion require a higher degree of specialization59, even in generalist bees60, which is often facilitated by each species’ gut microbiome61. Although some plant species appeared in the samples of all three bee species, S. postica and M. rufiventris shared more species while T. angustula differed from both. Considering body size vs. flower matching, the smallest species, T. angustula, visits the highest number of species of the three, potentially due to solitary foraging behavior, in which females forage alone without recruiting other workers.
Melipona species present a unique foraging pattern among stingless bees, not only because they are amongst the largest stingless bees (up to 15 mm)3, but because they show clear preferences towards some groups of plants22,62. Melipona are also the only stingless bees capable of buzzing to harvest pollen63, but pollen-flowers that require buzz-pollination for pollen harvesting were not abundant in the samples, even though species with poricidal anthers were observed flowering around the nests during the months of collection (e.g., Miconia ferruginata DC, Pleroma stenocarpum (Schrank & Mart. ex DC.) Triana, Solanum falciforme Farruggia).
Our botanical surveys also reinforced the patterns of floral exploitation among the three species, such as the apparent preference for trees with mass flowering by stingless bees, even though their exploitation demands a long flight range. Some stingless bees' sophisticated communication abilities allow a massive recruitment of foragers when mass blooming plants are available19. In the case of T. angustula, which is considered a solitary forager, the range of pollen sources is wider and seems less biased towards mass blooming plants. In the Atlantic rainforest, another hyper diverse neotropical ecosystem, stingless bees have a preference for upper canopy stratum with small hermaphroditic or monoecious whitish flowers and abundant resources (pollen and/or nectar)29. Importantly, most of their preferred trees flower in mass, i.e. produce a large number of flowers over a short period of time20. In the Cerrado savannas, where the nests were, we observed the typical high frequency of shrubs and herbaceous species in stingless bees pollen (ca. 38% of samples), which reflects the savanna physiognomy where herbs and shrubs are predominant64. However, despite the high availability of flowers in the savanna surrounding their nests, they still flew up to riverine forests at least 630 m far from the nests to collect resources where mass-flowering species were more common.
Flight distance in bees is usually related to body size (larger bees tend to have wider flight ranges)65 and social behavior (social bees have a larger foraging distance than solitary bees due to the potential communication and recruitment between individuals)66. Given that the closest riverine forest is located at a distance of 630 m to the nests, and that species from this habitat were among the most abundant in the samples, this suggests that all three stingless bee species will forage and probably recruit at least 630 m from their nests, supporting the hypothesis of long-distance foraging when attractive rewards are available20. This distance is well within the known flight range of Melipona whose typical flight distance is about 2 km, but can be extended up to 10 km18, but it is more surprising for Scaptotrigona and Tetragonisca whose reported maximum flight distances are 1.7 to 0.6 km, respectively18.
These estimates of minimum foraging distance of 630 m are considered trustworthy based on the high frequency of pollen from species occurring only in riverine forests, e.g. Syzygium cumini, an introduced species that only occurs in a small portion of the nearest riverine forest to the nests. Other highly abundant species in our samples are common in the Distrito Federal riverine forests (Clusia cruiva Cambess, Hedyosmum brasiliense Mart. ex Miq., Miconia hirtella Cogn., Piper aduncum L., Richeria grandis)67,68,69 and were only found in our surveys of the riverine forests (Table 1).
Some plant families stand out as the most important floral sources for the three stingless bee species, i.e. have one or more species amongst the 30 most frequent ASVs. Amongst them, Myrtaceae, Anacardiaceae, Sapindaceae, Melastomataceae, Euphorbiaceae, and Asteraceae are well-known as common resources for stingless bees globally17, while Loranthaceae and Malpighiaceae are frequent in other studies62. Phyllanthaceae, Primulaceae, Chloranthaceae and Piperaceae, however, have been only rarely reported22. Asteraceae, Myrtaceae, and Melastomataceae are amongst the most speciose plant families in the IBGE reserve, representing at least 300 species with different life forms (from herbs to trees) in the flora29, but it is surprising that other diverse plant families in the IBGE area, i.e. Fabaceae, Lamiaceae and Orchidaceae, which also represent close to 300 species combined29, are less conspicuous or totally absent from our most frequent 30 taxa. This means that, although the important floral sources for stingless bees partially overlap with the most common plants in the area, indicating that abundant sources are preferred, this is not always the case. This could simply mean that species within these families were not flowering at the time of sampling, but it is worth noting that Lamiaceae, papilionoid legumes and orchids share complex floral morphologies that are different from those of the families recorded as most abundant in our samples These three families tend to present flowers with bilateral symmetry, specialized petals and androecia, and deep, hidden resources that often forces floral visitors to approach and handle the flowers in a specific way70. Our results confirm the hypothesis that stingless bees may be specialized in exploiting small, open resource “bowl-type” flowers52, with exposed stamens and nectar, that are produced in large numbers20. They may also favor plant species with a “big bang” flowering phenology i.e., that that undergo mass blooming for short periods. Floral morphology, floral chemistry and phenology of plants exploited by stingless bees deserve further investigation. Investigations of plant resources exploited by stingless bees using metabarcoding over a longer time periods, in other types of vegetation, and of other bee species, would also be desirable to consolidate our knowledge of stingless bee ecology in the Neotropics.
In conclusion, our high-resolution results should encourage future studies on bee-plant interactions in hyper diverse tropical environments using pollen or honey DNA metabarcoding, even considering the currently low DNA sequence coverage of tropical plant species in public databases28. The method has been prooved useful not only for pot-pollen, but also for pollen loads carried by bees in their scopae or corbiculae12. Given the importance of social and solitary bees for pollination of tropical crops and natural ecosystems, this methodology has the power to answer various important ecological questions, regarding the use of floral resources by bees, the dependance of bees on specific floral sources, their role as pollinators in crops or how well they compete with different species.
Data availability
The data that support the findings of this study is available at public repositories: metabarcoding sequence data is available at NCBI (https://www.ncbi.nlm.nih.gov/bioproject/976708); all plant species vouchers were deposited in the Herbarium of the University of Brasilia (UB) and all related information, including deposit number, is available in the Species Link network by searching under the collector name “Projeto Barcode Cerrado” (https://specieslink.net/search/). Analytical R codes used in our analysis are deposited at https://github.com/tncvasconcelos/barcode_cerrado.git and the bioinformatics analysis pipeline used for data analysis is available at https://github.com/chiras/metabarcoding_pipeline.
References
Klein, A. M. et al. Importance of pollinators in changing landscapes for world crops. Proc. Biol. Sci. R. Soc. 274, 303–313 (2007).
Gallai, N., Salles, J., Settele, J. & Vaissiere, B. Economic valuation of the vulnerability of world agriculture confronted with pollinator decline. Ecol. Econ. 68, 810–821 (2009).
Michener, C. D. The Bees of the World (The John Hopkins University Press, 2007).
Ollerton, J., Winfree, R. & Tarrant, S. How many flowering plants are pollinated by animals?. Oikos 120, 321–326 (2011).
Biesmeijer, J. C. et al. Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 1979(313), 351–354 (2006).
Morales, C. L. et al. Does climate change influence the current and future projected distribution of an endangered species? The case of the southernmost bumblebee in the world. J. Insect. Conserv. 26, 257–269 (2022).
Parreño, M. A. et al. Critical links between biodiversity and health in wild bee conservation. Trends Ecol. Evol. 37, 309–321 (2022).
Keller, A. et al. Evaluating multiplexed next-generation sequencing as a method in palynology for mixed pollen samples. Plant Biol. 17, 558–566 (2015).
Sickel, W. et al. Increased efficiency in identifying mixed pollen samples by meta-barcoding with a dual-indexing approach. BMC Ecol. 15, 1–9 (2015).
Baksay, S. et al. Experimental quantification of pollen with DNA metabarcoding using ITS1 and trnL. Sci. Rep. 10, 4202 (2020).
Khansaritoreh, E. et al. Employing DNA metabarcoding to determine the geographical origin of honey. Heliyon 6, e05596 (2020).
Elliott, B. et al. Pollen diets and niche overlap of honey bees and native bees in protected areas. Basic Appl. Ecol. 50, 169–180 (2021).
Wilson, R. S. et al. Landscape simplification modifies trap-nesting bee and wasp communities in the subtropics. Insects 11, 1–15 (2020).
Camargo, J. M. F., Pedro, S. R. M. & Melo, G. A. R. Meliponini Lepeletier, 1836. In Catalogue of Bees (Hymenoptera, Apoidea) in the Neotropical Region - online version. (2013).
Janousek, W. M. et al. Recent and future declines of a historically widespread pollinator linked to climate, land cover, and pesticides. Proc. Natl. Acad. Sci. USA 120, e2211223120 (2023).
Slaa, E. J., Chaves, L. A. S., Malagodi-Braga, K. & Hofstede, F. E. Stingless bees in applied pollination: Practice and perspectives. Apidologie 37, 293–315 (2006).
Bueno, F. G. B. et al. Stingless bee floral visitation in the global tropics and subtropics. Glob Ecol Conserv 43, e02454 (2023).
Nunes-Silva, P. et al. Radiofrequency identification (RFID) reveals long-distance flight and homing abilities of the stingless bee Melipona fasciculata. Apidologie 51, 240–253 (2020).
Biesmeijer, J. C. & Slaa, E. J. Information flow and organization of stingless bee foraging. Apidologie 35, 143–157 (2004).
Ramalho, M. Stingless bees and mass flowering trees in the canopy of Atlantic Forest: a tight relationship. Acta Bot. Bras. 18, 37–47 (2004).
Bell, K. L. et al. Pollen DNA barcoding: Current applications and future. Genome 640, 629–640 (2016).
Vit, P., Pedro, S. R. M. & Roubik, D. W. Pot-Pollen in Stingless Bee Melittology. https://doi.org/10.1007/978-3-319-61839-5 (Springer, 2018).
Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. Biodiversity hotspots for conservation priorities. Nature 403, 853 (2000).
Flora e Funga do Brasil. Flora e Funga do Brasil. Jardim Botanico do Rio de Janeiro https://floradobrasil.jbrj.gov.br/.
Silva-Souza, K. J. P., Pivato, M. G., Silva, V. C., Haidar, R. F. & Souza, A. F. New patterns of the tree beta diversity and its determinants in the largest savanna and wetland biomes of South America. Plant Divers. https://doi.org/10.1016/j.pld.2022.09.006 (2022).
Silberbauer-Gottsberger, I. & Gottsberger, G. A polinização de plantas do Cerrado. Rev. Bras. Biol. 48, 651–663 (1988).
Alves-dos-Santos, I. Bees of the Brazilian Savanna. in Encyclopedia of Life Support Systems (eds. Del-Claro, K., Oliveira, P. S. & Rico-Gray, V.) vol. 10 ((EOLSS)/UNESCO, 2009).
Vasconcelos, T. A trait-based approach to determining principles of plant biogeography. Am. J. Bot. 110, e16127. https://doi.org/10.1002/ajb2.16127 (2023).
Pereira, B. A. S. & Furtado, P. P. Vegetação da bacia do córrego Taquara: coberturas naturais e antrópicas. Reserva Ecológica do IBGE: Biodiversidade Terrestre 1, 89–117 (2011).
Rasmussen, C. & Cameron, S. A. Global stingless bee phylogeny supports ancient divergence, vicariance, and long distance dispersal. Biol. J. Lin. Soc. 99, 206–232 (2010).
Campos, M. G. et al. Standard methods for pollen research. J. Apic. Res. 60, 1–109 (2021).
Rognes, T., Flouri, T., Nichols, B., Quince, C. & Mahé, F. VSEARCH: A versatile open source tool for metagenomics. PeerJ 4, e2584 (2016).
Edgar, R. C. & Flyvbjerg, H. Error filtering, pair assembly and error correction for next-generation sequencing reads. Bioinformatics 31, 3476–3482 (2015).
Edgar, R. C. UCHIME2: improved chimera prediction for amplicon sequencing. bioRxiv https://doi.org/10.1101/074252 (2016).
Quaresma, A. et al. Semi-automated curation and manual addition of sequences to build reliable and extensive reference databases for ITS2 vascular plant DNA (meta-)barcoding. bioRxiv https://doi.org/10.1101/2023.06.12.544582 (2023).
Keller, A. et al. BCdatabaser: On-the-fly reference database creation for (meta-) barcoding. Bioinformatics 36, 2630–2631 (2020).
Edgar, R. C. SINTAX: A simple non-Bayesian taxonomy classifier for 16S and ITS sequences. bioRxiv https://doi.org/10.1101/074161 (2016).
Marshall, C. A. M., Wieringa, J. J. & Hawthorne, W. D. Bioquality hotspots in the tropical African flora. Curr. Biol. 26, 3214–3219. https://doi.org/10.1016/j.cub.2016.09.045 (2016).
R Core Team. R: a language and environment for statistical computing. Preprint at https://www.r-project.org/ (2021).
McMurdie, P. J. & Holmes, S. phyloseq: An R Package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 8, e61217 (2013).
Oksanen, J., Kindt, R., Legendre, P., O’Hara, B. & Stevens, H. The vegan package. Community ecology package 10, 631–637 (2007).
Dormann, C. F., Gruber, B. & Fründ, J. Introducing the bipartite Package: Analysing Ecological Networks. R News 8, 8–11 (2008).
Gu, Z., Gu, L., Eils, R., Schlesner, M. & Brors, B. circlize implements and enhances circular visualization in R. Bioinformatics 30, 2811–2812 (2014).
Garnier, S. et al. viridis - Colorblind-Friendly Color Maps for R. Preprint at (2021).
Absy, M. L., Rech, A. R. & Ferreira, M. G. Pollen collected by stingless bees: A contribution to understanding Amazonian biodiversity. In Pot-Pollen in Stingless Bee Melittology 29–46 (Springer International Publishing, 2018). https://doi.org/10.1007/978-3-319-61839-5_3.
Moura, C. C. M. et al. Biomonitoring via DNA metabarcoding and light microscopy of bee pollen in rainforest transformation landscapes of Sumatra. BMC Ecol. Evol. 22, 1–15 (2022).
Roubik, D. & Patiño, J. E. M. The stingless honey bees (Apidae, Apinae: Meliponini) in Panama and pollination ecology from pollen analysis. In Pot-Pollen in Stingless Bee Melittology (eds Vit, P. et al.) 47–66 (Springer, 2018).
Pound, M. J. et al. Determining if honey bees (Apis mellifera) collect pollen from anemophilous plants in the UK. Palynology https://doi.org/10.1080/01916122.2022.2154867 (2022).
Malerbo-Souza, D. T., Da Silva, T. G., De Andrade, M. O., De Farias, L. R. & Medeiros, N. M. G. Factors affecting the foraging behavior of bees in different maize hybrids. Rev Bras Cienc Agrar 13, 1–8 (2018).
Costa, A. C. G., Albuquerque, I. S., Thomas, W. W. & Machado, I. C. Influence of environmental variation on the pollination of the ambophilous sedge Rhynchospora ciliata (Cyperaceae). Plant Ecol. 219, 241–250 (2018).
Gottsberger, G. Generalist and specialist pollination in basal angiosperms (ANITA grade, basal monocots, magnoliids, Chloranthaceae and Ceratophyllaceae): what we know now. Plant Divers. Evol. 131, 263–362 (2015).
Faegri, K. & van der Pijl, L. Principles of Pollination Ecology (Pergamon Press, 1979).
Saunders, M. E. Insect pollinators collect pollen from wind-pollinated plants: implications for pollination ecology and sustainable agriculture. Insect. Conserv. Divers. 11, 13–31 (2018).
Oliveira, M. L. & Morato, E. F. Stingless bees (Hymenoptera, Meliponini) feeding on stinkhorn spores (Fungi, Phallales): Robbery or dispersal?. Rev. Bras. Zool. 17, 881–884 (2000).
Barth, O. M., de Freitas, A. S. & Rio Branco, C. D. S. Pollen collected by stingless bees in a reforested urban area of Rio de Janeiro city. Bee World 98, 23–26 (2021).
Parish, J. B., Scott, E. S. & Hogendoorn, K. Nutritional benefit of fungal spores for honey bee workers. Sci. Rep. 10, 15671 (2020).
Leonhardt, S. D. Chemical ecology of stingless bees. J. Chem. Ecol. 43, 385–402 (2017).
Asakawa, Y. Highlights in phytochemistry of hepaticae-biologically active terpenoids and aromatic compounds. Pure Appl. Chem. 66, 2193–2196 (1994).
Sedivy, C., Müller, A. & Dorn, S. Closely related pollen generalist bees differ in their ability to develop on the same pollen diet: Evidence for physiological adaptations to digest pollen. Funct. Ecol. 25, 718–725 (2011).
Bryś, M. S., Skowronek, P. & Strachecka, A. Pollen diet—Properties and impact on a bee colony. Insects 12, 798 (2021).
Keller, A. et al. (More than) Hitchhikers through the network: The shared microbiome of bees and flowers. Curr. Opin. Insect. Sci. 44, 8–15 (2021).
Ramalho, M., Kleinert-Giovannini, A. & Imperatriz-Fonseca, V. L. Important bee plants for stingless bees (Melipona and Trigonini) and africanized honeybees (Apis mellifera) in neotropical habitats: a review. Apidologie 21, 469–488 (1990).
Nunes-Silva, P., Hrncir, M., Da Silva, C. I., Roldão, Y. S. & Imperatriz-Fonseca, V. L. Stingless bees, Melipona fasciculata, as efficient pollinators of eggplant (Solanum melongena) in greenhouses. Apidologie 44, 537–546 (2013).
Klink, C. A., Sato, M. N., Cordeiro, G. G. & Ramos, M. I. M. The role of vegetation on the dynamics of water and fire in the cerrado ecosystems: Implications for management and conservation. Plants 9, 1–27 Preprint at https://doi.org/10.3390/plants9121803 (2020).
Greenleaf, S. S., Williams, N. M., Winfree, R. & Kremen, C. Bee foraging ranges and their relationship to body size. Oecologia 153, 589–596 (2007).
Grüter, C. & Hayes, L. Sociality is a key driver of foraging ranges in bees. Curr. Biol. 32, 5390-5397.e3 (2022).
Darosci, A. A. B., Takahashi, F. S. C., Proença, C. E. B., Soares-Silva, L. H. & Munhoz, C. B. R. Does spatial and seasonal variability in fleshy-fruited trees affect fruit availability? A case study in gallery forests of Central Brazil. Acta Bot. Bras. 35, 456–465. https://doi.org/10.1590/0102-33062020ABB0279 (2021).
Mendonça, R. C. et al. Flora vascular do Cerrado. In Cerrado: ambiente e flora (eds. Sano, S. M. & Almeida, S. P.) 289–556 (Embrapa-Cerrados, 1998).
Ratter, J. A., Bridgewater, S. & Ribeiro, F. Analysis of the floristic composition of the Brazilian cerrado vegetation III: Comparison of the woody vegetation of 376 areas. Edinb. J. Bot. 60, 57–109 (2003).
Willmer, P. Pollination and floral ecology. In Pollination and floral ecology (Princeton University Press, 2011).
Acknowledgements
We would like to thank Instituto Serrapilheira for the grant received (Chamada Pública No 2—2018). We thank the IBGE reserve for allowing and supporting our field work; Antonio Leite from IBRAMEL stingless beekeepers association for kindly donating the nests that were used in the experiments. AJCA acknowledges FAPDF for the financial support (00193-00001229/2021-48). ACM thanks CNPq for the postdoctoral fellowship (159694/2018-3) and CEBP for the CNPQ PQ2 fellowship. ATFL, KN & MMC acknowledge their undergraduate scholarships (ProIC UnB). FAC acknowledges the CAPES/PRINT program (Edital No 41/2017 88887.716844/2022-00) for sponsoring the visit of Alexander Keller to UFMG.
Author information
Authors and Affiliations
Contributions
Conceptualization: A.C.M., T.N.C.V., C.E.B.P., F.A.C., A.J.C.A., M.M.C.B., A.K.; Investigation: A.C.M., A.J.C.A., C.E.B.P., H.F., K.N., J.E.Q.F., M.M.C., M.B.R.C.; Formal analysis: A.K., T.N.C.V., R.D.; Writing - original draft: A.C.M., T.N.C.V., C.E.B.P.; Funding acquisition: A.C.M., F.A.C., T.N.C.V., C.E.B.P., A.J.C.A., M.M.C.B., A.K. All authors revised the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Martins, A.C., Proença, C.E.B., Vasconcelos, T.N.C. et al. Contrasting patterns of foraging behavior in neotropical stingless bees using pollen and honey metabarcoding. Sci Rep 13, 14474 (2023). https://doi.org/10.1038/s41598-023-41304-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-41304-0
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.