Pollinator Visitation Alters Cranberry Flower Fungal Communities in Wisconsin Cranberry Agroecosystems
Celeste C. Mezera, Shawn Steffan, Leslie A. Holland
TL;DR
This study shows that pollinators like bees and flies transfer fungi to cranberry flowers, increasing fungal diversity and possibly affecting crop health.
Contribution
The first evidence of shared fungal communities between pollinators and cranberry flowers in agroecosystems.
Findings
Cranberry flowers with pollinators had higher fungal richness compared to those without.
Fungi linked to cranberry fruit rot were found in several pollinator groups, especially Toxomerus flies.
Shared fungal communities were detected between cranberry flowers and insect visitors.
Abstract
Pollinators are known dispersal agents of microbial communities between flowering plants, although the role of insect‐mediated microbial assembly in flowering agricultural crops is not well understood. In cranberry ( Vaccinium macrocarpon Ait.) agroecosystems, the blossom period is a vulnerable time for infection from pathogens within the cranberry fruit rot fungal disease complex, and understanding the components and assembly dynamics in cranberry flower fungal communities may provide important insights to the relationship between the cranberry microbiome and crop health. This 2‐year study uses a combination of culture‐dependent and next‐generation sequencing approaches to compare the community structure of cranberry flowers, honey bees ( Apis mellifera ), bumble bees (Bombus sp.), wild solitary bees, hover flies (Syrphidae), and nearby wildflowers to identify shared fungal…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5| Fungal genera | Cranberry flowers ( | Honey bees ( | Wild bees ( | Wildflowers ( | Analysis of variance |
|---|---|---|---|---|---|
|
| 27.86% | 49.78% | 32.50% | 54.17% |
|
|
| 18.66% | 22.20% | 15.00% | 21.35% |
|
|
| 11.02% | 8.84% | 4.38% | 19.27% |
|
|
| 8.94% | 14.01% | 8.13% | 15.10% |
|
|
| 7.29% | 13.15% | 6.25% | 12.76% |
|
|
| 10.59% | 7.97% | 14.38% | 5.99% |
|
|
| 8.51% | 7.11% | 5.63% | 9.77% |
|
|
| 2.52% | 8.84% | 6.25% | 8.85% |
|
|
| 3.21% | 3.02% | 1.88% | 7.03% |
|
|
| 2.26% | 6.25% | 3.75% | 5.73% | 0.184 |
|
| 5.38% | 0.86% | 0.63% | 2.34% |
|
|
| 3.65% | 1.72% | 0.63% | 3.26% |
|
|
| 2.86% | 3.02% | 0.00% | 2.73% |
|
|
| 1.30% | 1.08% | 3.13% | 2.60% | 0.0855 |
|
| 1.74% | 0.43% | 0.63% | 2.34% |
|
|
| 0.61% | 0.43% | 0.00% | 0.13% | 0.0992 |
- —U.S. Department of Agriculture10.13039/100000199
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPlant and animal studies · Insect and Pesticide Research · Plant and fungal interactions
Introduction
1
The flower microbiome plays an important role in the ecology and evolution of angiosperms and their pollinators alike (Aizenberg‐Gershtein et al. 2013; Brysch‐Herzberg 2004; Lachance et al. 2001; Steffan et al. 2024). Despite the ephemeral nature of flowers, the floral microbiome contains a rich host of bacteria and fungi. These microbial communities include a mixture of generalist fungi and bacteria, specialists that can thrive in floral structures, plant pathogens, and the associates of pollinating visitors (Vannette 2020). The microbial associates of flowers may be present on all floral structures and at all stages of the flower's life, including before anthesis (Shade et al. 2013; von Arx et al. 2019); however, the diversity and abundance of bacteria and fungi on flower surfaces increase with the age of the flower, which may be attributed in part to longer exposure to environmental and animal‐mediated microbial dispersal factors (Morris et al. 2020; Nicolson and Thornburg 2007; Shade et al. 2013; Vannette and Fukami 2018). Microbes have been known to ‘hitchhike’ on bees to visit locations with ideal resources for nutrition and secure protection from antagonistic environmental conditions (Keller et al. 2021; McFrederick et al. 2017, 2018; Steffan et al. 2024; Zemenick et al. 2021). Flowers that have been visited by pollinators hold increased yeast abundance in nectar (Belisle et al. 2012; Herrera et al. 2008, 2010), and some yeast species rely on bumble bees for dispersal to new flowers (Brysch‐Herzberg 2004).
The fungal associates of flowers and insect visitors are thought to assemble and interact differently in agricultural systems, where the environment has been heavily manipulated through decreased plant diversity, agrichemical inputs, and cultivation methods (Fernandez De Landa et al. 2023; Morris et al. 2020). In cranberry agroecosystems, the fungal inhabitants of mass‐flowering cranberry blooms are of particular relevance to the health of the cranberry fruit. The most prevalent disease to the North American cranberry industry is cranberry fruit rot, a disease in which all commercially available cranberry cultivars are susceptible (Johnson‐Cicalese et al. 2015), and crop losses due to cranberry fruit rot may reach up to 100% in some regions (Oudemans et al. 1998; Polashock et al. 2009). Cranberry fruit rot is a disease complex containing between 10 and 15 fungal genera, including Colletotrichum, Phyllosticta, Physalospora, Monilinia, Godronia, Coleophoma, Botrytis and Allantophomopsis (Oudemans et al. 1998). Fungi within the cranberry fruit rot complex are categorised into field rot and storage rot, although many of the pathogens within the complex are shared in both categories (McManus 2001; Oudemans et al. 1998). The field rot pathogens within the complex are monocyclic, and several members have a sporulation and infection period coinciding with cranberry bloom, between June and July. Infection of cranberry fruits may occur during the flowering period, and fungicide application during cranberry bloom is documented to decrease instances of cranberry fruit rot compared to fungicide applications during other phenological periods (McManus 2001; Oudemans et al. 1998; Zuckerman 1958). Cranberry fruit rot fungi are often detected in cranberry flowers in managed cranberry marshes, although detection of these pathogens may vary between regions and between farms (Wood et al. 2023). However, the disease cycle of pathogens within the cranberry fruit rot complex has not been fully identified for cranberry plant hosts.
Understanding the fungal community identities and dynamics between pathogens, flower specialists, and environmental fungi may help further our understanding of the uneven spatial distribution of disease in Wisconsin. Although fruit rot poses a consistent and imposing threat to cranberry growers in the northeastern United States, Wisconsin cranberry marshes have uneven, spotty distribution of this disease, which may vary within marshes, between farms, and across years (McManus et al. 2003; Wells‐Hansen and McManus 2017). In addition to the fungi within the CFR disease complex, several fungal associates of cranberries are considered with uncertain pathogenicity, including the fungi in the genera Alternaria, Aureobasidium, Cladosporium, Curvularia, Epicoccum, Penicillium, Pestalotia, Pseudotracylla, Rhabadospora, Septoria, Sphaeropsis and Trichoderma (Oudemans et al. 1998). Additionally, several yeasts have been cultured from rotten cranberry fruits; some of which have been shown to contribute to the cranberry fruit rot disease complex (Zalewski et al. 2021), however, the floral‐inhabiting yeasts of the cranberry flower have not yet been examined in a community context.
Although pollinators are known culprits of vectoring plant pathogens in flowering fruit crops such as apples and pears (Alexandrova et al. 2002; Johnson et al. 1993), bee‐vectored biological antagonists of pathogens have also been intentionally incorporated into commercial bee pollination in apples, blueberries, tomatoes and canola (Park et al. 2013). Additionally, pollinators have been identified as important in assembling the floral microbiome in flowering crops (Schaeffer et al. 2023). During cranberry bloom, cranberry producers rely on insect pollinators to pollinate their crop (Gaines Day 2013; Loose et al. 2005). Rentals of honey bees ( Apis mellifera ) and purchased bumble bee colonies ( Bombus impatiens ) are the main sources of pollination services during cranberry bloom (Evans and Spivak 2006; Gaines‐Day and Gratton 2015), although some cranberry producers also use habitat conservation as a method to maintain wild bee populations on the marsh (Amon et al. 2023; Gaines‐Day et al. 2017). Additionally, hoverfly species are common visitors of cranberry flowers, and the significance of their contribution to the pollination of cranberries has yet to be fully examined (Gervais et al. 2018). The microbial associates of social bees such as honey bees and bumble bees include widely‐shared microbiomes (Anderson et al. 2011; Gilliam 1997; Kwong et al. 2017), which are likely to differ from solitary bee and fly pollinators, with less conclusive and more transient fungal associates (Fernandez De Landa et al. 2023; Graystock et al. 2017; Voulgari‐Kokota et al. 2019). Pollinators may be important agents for introducing environmental, flower‐specialising, and even pathogenic microbes to the floral microbiome. The contributions of pollinator visitation to cranberry flower microbial communities have not yet been evaluated in cranberry systems, although pollinators may play an important role in the assembly of the cranberry microbiome.
Examining the fungal community dynamics occurring during the cranberry bloom period is an important step in understanding the cranberry fruit rot disease cycle, and we recognise that pollinator visitation may influence fungal communities in cranberry flower mycobiomes. We hypothesise that flower visitation from bees and flies increases the fungal richness and abundance of cranberry flowers. First, we use culture‐dependent and culture‐independent methods to compare the community structure of cranberry flowers, pollinators of cranberry and vegetation with synchronous flowering phenology during the cranberry bloom period to identify shared fungal associates between hosts. Next, we compare the fungal richness and abundance between cranberry flowers with and without a tenting treatment preventing insect visitation to demonstrate the impact of pollinator visitation on the cranberry flower community assembly. Finally, we compare the detection of cranberry fruit rot pathogens between cranberry flower treatments, insect groups and proximate wildflowers to advance our understanding of the movement of cranberry fruit rot pathogens within and between potential inoculum sources in cranberry systems.
Materials and Methods
2
Site Selection and Flower Sampling
2.1
During the cranberry bloom period of 2023 and 2024, samples of cranberry flowers, wildflowers, and insect pollinators were collected at eight cranberry marshes located in Central Wisconsin. Each marsh was visited twice per year for sample collections between June 15 and July 10. In 2023, four cranberry beds and four patches of wildflowers located along the marsh margins 500 m apart from each other were selected for sampling within each marsh (Figure S1). Sampling in cranberry took place on the ‘Stevens’ cultivar. In 2024, sample sites were reduced to two cranberry beds and two wildflower patches per marsh. In 2024, wildflowers were collected within 100 m of the cranberry beds at each site, and samples of all unique wildflower species on site were collected. Pictures of the flowering plants were taken for species‐level identification prior to collection for identification validation. Collections of cranberry and wildflowers were performed with forceps sterilised with 95% ethanol into 50‐mL Falcon tubes and transported on ice to the lab where they were stored at −80°C until processed.
At each sampling site within cranberry beds, two treatments were established prior to cranberry bloom: (1) ‘tented’, which blocked pollinator access to flowers within 1 m^2^ using a white polypropylene garden cover (Summerweight Fabric, Gardener's Supply Company), and (2) ‘open’, allowing pollinator access to flowers within a flagged 1 m^2^ area (Figure S1). Tent frames were constructed using two perpendicular 2‐m segments of black irrigation tubing secured with rebar in four corners of a 1 m^2^ area. The white polypropylene garden cover was secured over the tent frame using four 1‐m segments of 3/4″ PVC tubing at the tent's base with 4″ landscaping staples secured to the ground. If any flowers were present prior to tent construction, they were removed to ensure no flowers collected in future sampling had been visited by pollinators. Flowers were collected from ‘tented’ and ‘open’ treatments twice during the cranberry blossom period using the same procedure as described above.
Insect Pollinator Observations and Sampling
2.2
In 2023, bees were sampled passively at each cranberry and wildflower sampling site using blue, yellow and white sticky traps (Great Lakes IPM, Vestaburg, MI). This method of bee collection was used to prevent contamination between bees and between samples and collection equipment. Traps were left for 2 days, then wrapped in plastic wrap and transported to the lab on ice. Bees were counted within genera groupings for each site, and one individual within each unique genera was extracted using ethanol‐sterilised forceps to be placed in sterile 15‐ml Falcon tubes. In 2024, pollinators were actively sampled. When possible, insects were collected directly from flowers into the 15‐mL tubes; otherwise they were collected with a net, which was ethanol‐sterilised between collections. Active sampling of bees and flies visiting cranberry flowers took place for 10 min at each site, or until the collection threshold of three insects for each pollinator category (honey bees, wild bees and hover flies) was reached. Each of the marshes visited in this study has rental contracts with beekeepers, and in the instance of low honey bee counts in the beds, samples were supplemented with honey bees collected near the closest hives on site. In the instance of a cranberry marsh using commercial bumble bee colonies on their marshes, Bombus impatiens were excluded from the ‘wild bee’ samples, although other Bombus species were collected.
Pollinator visitation on cranberry flowers was documented in 2024 at each collection site within 1 m^2^ of ‘Stevens’ cranberry flowers within a 5‐min period. Each site was observed by two observers over two collection periods for a total of 4 observations per site. Visitation observations were recorded generally between 8:00 AM Central Daylight Time (CDT) and 4:00 PM CDT, and temperature and cloud cover information was recorded. Using the pollinator morpho‐groups including honey bees, bumble bees, large dark‐bodied bees (about 12 mm long or larger), small dark‐bodied bees (about 6 mm long or smaller), green metallic bees, flies, and other non‐bee pollinators, pollinator visits to cranberry flowers were recorded within a 5‐min period. Visits were defined as instances in which a pollinator contacts the flower within the 1 m^2^ area. Observations were recorded using protocols established by the WiBee app (Gratton Lab, UW Madison, https://pollinators.wisc.edu/wibee/).
Insect Identification
2.3
All insects collected for fungal culturing were returned to the laboratory on ice and stored in a −20°C freezer to prevent decay and associated alterations to the fungal community. Flowers were stored at 4°C for a maximum of 2 days until processing for fungal culturing. Wild bees and flies from active netting collections in 2024 were imaged under a dissecting microscope from multiple angles for identification, then placed in 2 mL microcentrifuge tubes, and stored at −80°C until processing for DNA extraction. All bees were identified to genus‐ and several to species‐level identification from morphological traits (Ascher and Pickering 2020) using an M125 C Encoded stereo microscope (Leica Microsystems), which were verified by an additional entomologist familiar with the pollinators in cranberry systems. Similarly, all flies were identified to genus and further classified to species if possible using the Field Guide to the Flower Flies of Northeastern North America (Skevington et al. 2019).
Isolation of Fungi From Cranberry Flowers, Wildflowers and Bees
2.4
Potato dextrose agar (PDA) and two selective media were prepared to target a wide range of fungi for environmental samples. Acidified 1/4 strength PDA (0.8% PDA, 0.5% agar, 0.38% of 25% lactic acid) was prepared to target filamentous fungi and exclude bacterial contaminants. Yeast peptone mannitol (YPM, 0.3% peptone, 0.5% yeast extract, 2.5% mannitol, 1.5% agar) was prepared to target fungal yeast species (Yarrow 1998).
To increase the surface area of the samples, each sample of 20 cranberry flowers and approximately equal volume of wildflowers collected were flattened using an ethanol‐sterilised rolling pin between sheets of sterilised aluminium foil under a laminar flow hood. In order to culture the fungi present on the surface of flowers, as well as material in floral tissue and structures such as the nectaries, two different culturing methods were used for each flower sample. Half of each crushed flower sample was spread evenly across the three different media types in aseptic conditions. The other half of the sample was placed in a sterile 15‐mL Falcon tube, containing 3 mL sterile phosphate buffer solution (PBS) with 0.01% Tween 80. This solution was vortexed for 20 s, then 100 μL of the resulting aqueous material was spread onto the three media types described under a laminar flow hood. Similarly, to capture surface fungi on bee samples, each bee was washed in 1 mL sterile PBS‐T (1× PBS + 0.01% Tween 80) within a sterile 15‐mL Falcon tube, then vortexed gently for 20 s. The bee was then extracted from the tube and embedded in a 10‐cm diameter petri dish of PDA. An aliquot of 100 μL of the resulting aqueous material was spread onto each of the three media types described.
All cultures were incubated at 24°C for 3 days, or until distinct colonies developed. Unique morphotypes, characterised by the colour, texture, and aerial growth of the organism, were distinguished, and morphotype presence was documented for each isolation plate. The 15 most frequently observed morphotypes were subcultured on PDA to create a pure culture of the microorganism for molecular genetic characterisation.
Molecular Identification of Cultured Fungi
2.5
DNA was extracted from subcultures using a modified cetrimonium bromide (CTAB) DNA extraction protocol adapted from Doyle and Doyle (Doyle and Doyle 1987). Tissue was collected from pure subcultures, combined with garnet sand and 500 μL of pre‐heated (55°C) CTAB buffer (100 mM Tris–HCL [pH 8], 20 mM EDTA, 1.4 M NaCl, 2% [wt/vol] CTAB, 2% [wt/vol] polyvinylpyrrolidone) with 2.5 μL of 2‐beta‐mercaptoethanol (BME) for homogenisation at 3200 rpm for 30 s for two cycles, with 10 s of rest between cycles. Then, the samples were incubated at 55°C in a heat block for 1 h. After incubation, 500 μL chloroform was added, and the tubes were centrifuged for 7 min at 16,000 rcf (13,053 rpm). The aqueous phase of this product was transferred to a new tube, and 0.08 volumes of cold 7.5 M ammonium acetate and 0.54 volume of cold isopropanol were added, gently mixed by repeated inversion, and incubated on ice for at least 30 min. After incubation, samples were centrifuged for 3 min at 16,000 rcf to form a pellet. Pellets were rinsed in 700 μL of 70% ethanol and centrifuged for 1 min at 16,000 rcf, then repeated with 700 μL of 95% ethanol. Pellets were air dried at room temperature, then rehydrated with 50 μL TE buffer. The quality and quantity of DNA from each sample were assessed using a Nanodrop One (Thermo Scientific, Waltham, MA). Higher concentrations were diluted prior to end‐point PCR amplification.
For all fungal samples, PCR was performed using a mixture of 3 μL of AccuPower PCR Premix (Bioneers, Alameda, CA), following the provider's guidelines. In filamentous fungi, the ITS spacer region was amplified using universal primers ITS1 (5′TCCGTAGGTGAACCTGCGG ′3) and ITS4 (5′ TCCTCCGCTTATTGATATGC ′3) (White et al. 1990) with a temperature cycle of an initial heating of 95°C for 5 min, then 31 cycles of 94°C for 20 s, and 60°C for 30 s, 72°C for 40 s, followed by a final heating at 72°C for 5 min. For yeast species, the D1/D2 region was amplified using the universal primers NL1 (50‐GCA TAT CAA TAA GCG GAG GAA AAG‐30) and NL4 (50‐GGT CCG TGT TTC AAG GAC GG‐30) (O'Donnell 1992). The temperature cycle for this primer pair is an initial heating of 94°C for 4 min, followed by 35 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min, followed by a final heating of 72°C for 5 min. SuccessfμL PCR amplification of target regions were verified using gel electrophoresis using a 1.5% agarose gel at 110 V for 20 min. PCR products with suitable quality were cleaned using an ExoSap‐IT protocol (PCR Product Pre‐Sequencing Kit, Applied Biosystems, Thermo Fisher Scientific) by combining 15 μL of PCR product, 3 μL exonuclease I, and 3 μL Shrimp Alkaline Phosphatase with a temperature cycle of 37°C for 15 min, followed by 15 min at 80°C. Samples were then submitted for Sanger sequencing to Functional Biosciences (Madison, WI).
DNA Extraction From Flowers and Insects
2.6
DNA was extracted from cranberry flowers, wildflowers, honey bees, wild bees, and hover flies collected in 2024 using the cetrimonium bromide (CTAB) DNA extraction protocol described previously with a few modifications. Each flower sample included 0.25 g of tissue, and each insect sample included one individual insect. Tissue from all samples was combined with lysing beads in 2‐ml microcentrifuge tubes, then homogenised at 3200 rpm for 30 s for two cycles with 10 s rest between cycles. Each sample received 500 μL of pre‐heated (55°C) CTAB buffer (100 mM Tris–HCL [pH 8], 20 mM EDTA, 1.4 M NaCl, 2% [wt/vol] CTAB, 2% [wt/vol] polyvinylpyrrolidone) with 2.5 μL of 2‐mercaptoethanol (BME) for a second round of homogenisation. Due to the higher concentration of phenols in the wildflower tissue, the CTAB buffer solution for these samples included 4% [wt/vol] polyvinylpyrrolidone. Then, the samples were incubated at 55°C in a heat block for 1 h. After incubation, 500 μL chloroform was added, and the tubes were centrifuged for 7 min at 16,000 rcf (13,053 rpm). The aqueous phase of this product was transferred to a new tube, and 0.08 of the transferred product's volume of cold 7.5 M ammonium acetate and 0.54 of the product's volume of cold isopropanol were added, gently mixed by repeated inversion, and incubated on ice for at least 30 min. After incubation, samples were centrifuged for 3 min at 1600 rcf to form a pellet. Pellets were rinsed in 700 μL of 70% ethanol and centrifuged for 1 min at 16,000 rcf, then repeated with 700 μL of 95% ethanol. Pellets were air dried at room temperature, then rehydrated with 50 μL TE buffer. The quality and quantity of DNA from each sample were assessed using a Nanodrop (Thermo Scientific, Waltham, MA).
Illumina Next‐Generation Sequencing
2.7
Samples with adequate quality (OD260/280 between 1.9 and 2.15) and quantity (> 20 ng/μl) of DNA were submitted for Illumina Next Generation sequencing at a coverage of 0.05 M raw tags per sample using the ITS2 spacer region using primers ITS3‐2024F (GCATCGATGAAGAACGCAGC) and ITS4‐2409R (TCCTCCGCTTATTGATATGC) (Novogene Corporation Inc., Sacramento, CA). This depth was selected to broadly understand the fungal communities within each sample type to make comparisons between communities, rather than providing an in‐depth characterisation of the fungal community within one sample type. All PCR reactions were carried out with 15 μL of Phusion High‐Fidelity PCR Master Mix; 0.2 μM of forward and reverse primers, and about 10 ng template DNA. Thermal cycling consisted of initial denaturation at 98°C for 1 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and elongation at 72°C for 30 s and 72°C for 5 min. The PCR products were purified using magnetic bead purification. Samples were mixed in equidensity ratios based on the concentration of PCR products. After thorough mixing, the PCR products were detected and target bands were recovered. Sequencing libraries were generated and indexes were added. The library was checked with Qubit model 3.0 fluorometer (Life Technologies, Carlsbad, CA) and real‐time PCR for quantification and bioanalyzer for size distribution detection. Quantified libraries were pooled and sequenced on Illumina platforms, according to effective library concentration and data amount required.
Bioinformatic Analysis
2.8
Paired‐end reads were assigned to samples based on their unique barcode and truncated by cutting off the barcode and primer sequence. Paired‐end reads were merged using FLASH (V1.2.1 1, http://ccb.jhu.edu/software/FLASH/) (Magoč and Salzberg 2011). Quality filtering on the raw tags were performed using the fastp (Version 0.23.1) software to obtain high‐quality Clean Tags (Bokulich et al. 2018). The tags were compared with the reference database (Unite Database version 10.0, https://unite.ut.ee/acknow.php) to detect chimera sequences, and the effective tags were obtained by removing the chimera sequences with the vsearch package (V2.16.0, https://github.com/torognes/vsearch) (Rognes et al. 2016). For these effective tags, a denoising step was performed with the deblur module in the QIIME2 software to obtain initial ASVs (Amplicon Sequence Variants). Species annotation was performed using QIIME2 software with the Unite Database.
Statistical Analyses and Data Visualisation
2.9
The absolute abundance of ASVs was normalised using a rarefaction normalisation in order to account for varying sequencing depths and achieve a proportionate representation of all observed sequences (Cameron et al. 2021). The rarefaction depth was calculated based on the 15th percentile of all sample depths, and samples at this depth were randomly subsampled using the rarefy_even_depth function in the phyloseq package in R (McMurdie and Holmes 2013). Subsequent analysis of alpha diversity and beta diversity was all performed based on this normalised data. The relative abundance of the ASVs was calculated as the ASV abundance divided by the total abundance within the sample. For comparisons within groups, the relative abundance was calculated as the ASV abundance for all samples within a group divided by the total abundance of all ASVs within a group.
The alpha diversity, richness, and evenness of fungal communities were determined using the metrics of Shannon diversity, dominance index (1—Simpson index), and Pielou's evenness using the R packages ‘phyloseq’ and ‘vegan’ (Dixon 2003). These metrics were calculated for each group and for each sample within a group. Differences between groupings for the alpha diversity metrics were tested by fitting an Analysis of Variance model to the data using the R package ‘stats’ and the model statement aov(x ~ y) with a TukeyHSD post hoc test to determine differences between groupings using the model statement TukeyHSD(x, conf.level = 0.95). Normality of the data was assessed using a quantile‐quantile normality plot of residuals of the Analysis of Variance model using the ‘qqnorm’ function, as well as the Shapiro–Wilk Normality Test using the ‘shapiro.test’ function in the R ‘stats’ package. In the case of non‐normal data, differences in alpha diversity between sample groups were determined using a non‐parametric Kruskal‐Wallis Rank Sum Test (p < 0.03) using the R package ‘stats’ with the function ‘kruskal.test’ using the model statement kruskal.test (x ~ y), followed by a Dunn test with a Bonferroni adjustment for post hoc pairwise comparisons using the package ‘dunn. test’ with the model statement: dunn. test (x, y, method = ‘bonferroni’). This procedure of testing for normality, performing the appropriate Analysis of Variance or Kruskal‐Wallis test, and performing the appropriate post hoc test was performed to determine differences between detection of fungi in culture‐dependent and culture‐independent methods.
To determine taxa which appear most frequently in each group, a ‘core community’ was determined for each sample group by selecting taxa which were present in 75% or more of the samples within the group (Jumbam et al. 2024). Comparisons of fungal community composition were quantitatively compared using a PerMANOVA test using the package ‘adonis2’ function in the ‘vegan’ package in R with the model statement: adonis2(formula = bray_curtis ~ Treatment, data = metadata, permutations = 999, method = ‘bray’). The data used for these tests included the relative abundance of fungal genera identified through next‐generation sequencing. To test whether the assumptions for the PerMANOVA analysis were met, we tested for homogeneity of multivariate dispersion among groups using PERMDISP with the ‘betadisper’ function in the vegan package. Significance was evaluated using 999 permutations with the function ‘permutest.’ An NMDS plot was created to visually compare fungal community overlap using bray–curtis dissimilarity calculated with the ‘vegdist’ function in the ‘vegan’ package in R. Data was visualised using the R packages ggplot2, upset and pheatmap (Conway et al. 2017; Kolde 2015; Wickham 2011).
Results
3
Identification of Wildflowers, Bees and Flies
3.1
Wildflower and bee collections included 32 plant species and 16 bee genera for culture‐based identification of fungi in 2023 (Table S1 and Table S2). The majority of bees passively collected in 2023 were honey bees ( Apis mellifera ), representing 97.5% of the bees collected, with 1256 counts of honey bees from passive blue, white, and yellow sticky traps. This dominance of honey bees to wild bees collected was consistent across sites, and is likely due to the high numbers of rented honey bee hives on cranberry marshes during the cranberry bloom period. Of the 33 wild bees collected in 2023, Lasioglossum (11 counts), Eucera (5), Agapostemon (3) and Augochlorella (3) were the most common, although the abundance of these wild, solitary bee genera is still quite low. Bombus species were under‐represented in this collection, likely due to the sticky trap collection method which was not sufficient for stronger, larger bees such as bumble bees. Targeted collections of wild bees using aerial nets in 2024 resulted in fewer overall specimens, but included many of the same genera as in 2023: Bombus (29), Ceratina (1), Agapostemon (1), Andrena (3), Lasioglossum (2); and one specimen within the genus Megachile (1), which is only represented in 2024. Although flies were not sampled in 2023, a total of 64 flies were collected in 2024, and all of the flies collected for fungal identification were in the genus Toxomerus (Table S2). In 2023, 25 plant species were collected, 17 of which were also represented in the samples in 2024. The most commonly collected wildflowers include white clover (Trifolium repens, present at 100% of sites), common cinquefoil (Potentilla simplex, 58%), yellow loosestrife (Lysmachia vulgaris, 35%), dewberry (Rubus flagellaris, 35%), and orange hawkweed (Pilosella aurantiaca, 31%, Table S1). The richness of wildflowers increased significantly between the first collection and second collection for both years (t‐test, p < 0.01), although there were no significant differences in wildflower species richness between sites.
Pollinator Visitation on Cranberry Flowers
3.2
A total of 63 pollinator visitation observations were recorded in cranberry flowers within a temperature range of 57°F–84°F (Table S3). There were no significant differences in temperature between the June and July collection periods in 2023 or 2024. The pollinator visits observed ranged from 0 to 73 visits per square meter patch of cranberry flowers within a 5‐min period. Almost all insect visitation on cranberry flowers was from honey bees (428 total visits) or flies (304), with no recorded instances of visits from bumble bees or other groups of wild bees. Twenty times more honey bee visits occurred during the second collection in July (408 total visits) compared to the first in June (20), which can be partially attributed to the colder temperatures in the June collection period, as well as less time since colony installation. Fly visits remained approximately the same for each collection period, regardless of temperature (Table S2). Within a 5‐min period, an average of 6.79 honey bee visits and 4.83 fly visits occurred within a 1‐m^2^ patch of cranberry flowers.
Culture‐Dependent Characterisation of the Flower and Insect Mycobiomes
3.3
A total of 2544 cultures of cranberry flowers (1152 cultures), wildflowers (768), honey bees (464), and wild bees (160) produced 186 unique morphotypes. The 15 overall most frequently observed fungal morphotypes were identified to genera (Table 1), which include, in descending order: Cladosporium (detected 1020 times), Scheffersomyces (506), Fusarium (323), Irpex (297), Rhodotorula (253), Penicillium (228), Curvularia (215), Rhizopus (148), Dothiora (108), Epicoccum (105), Sporobolomyces (85), Streptotinia (76), Alternaria (68), Trichoderma (45) and Cryptococcus (41). These 15 genera include 5 yeasts: Scheffersomyces, Rhodoturula, Dothiora, Sporobolomyces, and Cryptococcus. Allantophomopsis, which is a fungus within the cranberry fruit rot complex, was detected 10 times in different cultures, with 7 observations in cranberry flowers, 2 observations in honey bees, and one occurrence in wildflowers. There are significant differences between cranberry flowers, honey bees, wild bees, and wildflowers in the frequency of detection for many of the top 15 fungal genera from cultures (Table 1). These differences in the frequency of detection are observed in all of the identified yeast genera and all identified filamentous fungi except for Epicoccum, Trichoderma and Allantophomopsis.
Next‐Generation Sequencing Characterisation of the Flower and Insect Mycobiomes
3.4
The bioinformatic analysis of the next‐generation sequencing data with the Unite database (version 10.0, https://unite.ut.ee/) for ITS region has identified 542 fungal genera and an additional 88 groups which could not be classified beyond the levels of class, phylum or subphylum, or the fungal kingdom (Figure 1A). Although a sequencing depth of 50 k paired‐end reads was selected for sequencing, the depth achieved averaged 94.6 k reads with chimera sequences removed, with a median depth of 102.45 k reads. The fungal genus with the highest relative abundance in cranberry flowers is Setomelanomma (31.2%), followed by Rachicladosporium (17.8%) and Taphrina (7.8%). The highest proportion of samples in wildflowers were identified as fungi with no further classification available in the database (76.5%), and the grouping of samples including honey bees, bumble bees, and wild bees, also had a high proportion of unclassified fungi (10.7%). Aureobasidium is represented with the highest relative abundance in fly samples (25.6%), followed by bee samples (7.7%). Claviceps (10%), Starmerella (7.6%), and Taphrina (7.5%) were also common in fly samples.
Stacked bar plot of relative abundance of fungal genera and heat map of fungal genera represented in wildflowers, cranberry flowers and insect pollinators. (A) The relative abundance for fungal genera are displayed from left to right from cranberry, fly, bee, and wildflower samples. Bees include Apis mellifera, Bombus spp., and solitary bees. Flies include hoverflies in the Toxomerus genus. Fungal genera represented with less than 10% of relative abundance within each sample type are grouped in the category ‘< 1% abundance, grouped’. (B) The heat map of the top 35 genera detected in sequencing visualises abundance of each genus compared within sample types, with red indicating that a genus is highly abundant within a sample group, and blue highlighting genera with lower abundance. Branching on the y‐axis corresponds to the similarity of fungal genera, and branching on the x‐axis indicates the similarities between sample groups, with smaller branches indicating closer relationships.
Many of the overall top 35 genera identified through next generation sequencing in cranberry flowers, insect groups, and wildflowers are abundant distinctly for a particular sample type (Figure 1B). Bumble bees and honey bees have more similarity than the other groups, as indicated by the branching of groups (Figure 1B). Bumble bees and honey bees share many of their most abundant fungal genera, including Dothiora, Cryptocline, Botrutis, and Aspergillus. Bumble bees samples include the highest abundance of Candida, Penicillium, and Pseudogymnoascus, and honey bees include the highest abundance of Fonsecazyma, Petunia, Lachancea, and Papiliotrema, as well as high abundance of taxa which were classified to the family Filobasidiaceae. Cranberry flowers which were open to insect visitation have high abundances of Physalospora, Zymoseptoria, Rashicladosporium, Tilletiopsis, and Solicoccozyma. ‘Tented’ cranberry flower samples included high abundances of Rhizospaera and Aureobasidium.
Significant differences were detected between cranberry flowers, insect groups and wildflowers for all of the alpha diversity metrics calculated, including Shannon Diversity Index, Dominance Index, and Pielou's Evenness Index (Table S4). Honey bee samples have the highest diversity for the Shannon Index (3.45) and Pielou's Evenness Index (0.59). Wildflowers include the lowest diversity with a Shannon Index of 0.02 and Pielou's Evenness Index of 0.01. The Dominance Index, calculated as 1‐Simpson Evenness, is 0.99 for overall wildflower samples. No significant differences were detected between ‘open’ and ‘tented’ cranberry flowers for the Shannon Index (Tukey HSD, p = 0.9649330), Pielou's Evenness Index (Dunn test, p = 1.00), or Dominance Index (Tukey HSD, p = 0.9394935), although the ‘open’ cranberry flowers have average values indicating slightly higher diversity for each of these metrics (Table S4).
The Bray–Curtis dissimilarity permutation test (PerMANOVA) across treatments indicates differences in fungal community composition between sample types (p = 0.001). Post hoc pairwise PerMANOVA tests between each sample group indicate significant differences between cranberry flowers and all other sample types (p < 0.01), although no significant difference in community composition between open and tented cranberry flowers (Table S5). Additionally, no significant differences were detected between wild bees with bumble bees or honey bees, and no significant differences were detected between flies and other insect groups. An NMDS plot visualises the significant overlap between fungal communities of the different sample types, notably between both treatments of cranberry flowers, honey bees, bumble bees, and wild bees (Figure S2). However, a test of the PerMANOVA assumption of homogenous multivariate dispersion (PERMDISP) was significant (p < 0.01), indicating that group dispersions differed. Therefore, the significance of the PerMANOVA tests should be interpreted with caution, as dispersion heterogeneity contributes to the observed group differences and may inflate PerMANOVA significance.
Identification of Core Community Components in Insects and Flowers
3.5
A ‘core community’ of all sample types was defined as taxa which appear in at least 75% of all the samples within a group (Figure 2). In cranberry flowers, the genus Setomelanomma appears in 100% of both ‘open’ and ‘tented’ samples, identified to species as Setomelanomma holmii, a fungus associated with Spruce Needle Drop of spruce species. Two other genera were determined to be core to cranberry flowers, including Rachicladosporium and Taphrina (Figure 2). These three core cranberry genera are also shared with the social bee groups, in that the honey bee core includes Setomelanomma and Taphrina, and the bumble bee core includes Setomelanomma and Rachicladosporium. Additionally, Filobasidium is core to honey bees, and three genera are core to bumble bees alone: Fusarium, Aspergillus, and Diplodia. Aspergillus was detected in 76.9% of bumble bees and 56.2% of honey bees. Aspergillus flavus, a common pathogen of several crops and bees, was detected in 7.7% (1) of bumble bee samples and 18.8% (3) of honey bee samples. Shared between honey bees and bumble bees are two core genera: Aureobasidium and Pseudopithomyces. Aureobasidium is also represented highly, but not considered core, in cranberry flowers, with detection of Aureobasidium pullulans in 92.3% of bumble bees, 84.6% of honey bees, and 48.2% of cranberry flowers. In flies, two genera are found to be core to this group: Malassezia and Vishniacozyma. In wild solitary bees, no core genera were identified.
Core community overlap venn diagram between sample types. The ‘core community’ is defined as fungal taxa which are represented in at least 75% of all samples within one group. Genera shared between taxa are centered in the area of overlap between sample types. No fungal genera were identified as ‘core’ components of wild solitary bees nor wildflowers, apart from taxa identified broadly as ‘fungi (unknown)’.
Comparison of Fungal Communities Between Cranberry Tenting Treatments
3.6
All of the genera identified through culturing techniques were found in both ‘tented’ and ‘open’ cranberry flowers (Figure 3), although the presence of these genera in culture is higher in ‘open’ cranberry flowers for Cladosporium, Fusarium, Irpex, Penicillium, Curvularia, Rhodoturula, Streptotinia, Dothiora, Epicoccum, Alternaria and Trichoderma. Cultures of ‘tented’ cranberry flowers included higher detection of Scheffersomyces, Sporobolomyces and Rhizopus. The average difference between tenting treatments is 12.4 more observations of each genus in ‘open’ cranberry flowers than ‘tented’ cranberry flowers. This difference in detection frequency is significant for the filamentous fungi identified, but not in yeasts as determined by a Kruskal‐Wallis test (p < 0.01) (Figure 3). Filamentous fungi were detected more frequently than yeasts in cultures of cranberry flowers, with 923 detections of filamentous fungi and 418 detections of yeasts. Filamentous fungi were detected in 191 more cultures of ‘open’ cranberry flowers than in those with a tenting treatment, compared to 8 additional detections of yeast genera in ‘open’ cranberry flowers than those which were ‘tented’.
Appearance of top 15 filamentous and yeast fungal genera in cultures of cranberry flowers with and without tenting treatment. This graph displays the frequency of detection of identified filamentous fungi and yeasts for cranberry flowers which are tented (‘tented’) or open to pollinator visitation (‘open’). There is a significant difference in the detection of the top 15 filamentous fungal genera between the ‘open’ and ‘tented’ cranberry flowers as indicated by the Kruskal‐Wallis test (p < 0.01). However, there are no significant differences in the frequency of detection of yeasts between cultures of ‘open’ and ‘tented’ cranberry flowers.
More genera were identified from next‐generation sequencing in cranberry flowers which were open to insect visitation (119 genera) than in those which were ‘tented’ (82 genera) (Figure 4). The number of genera identified in each sample is significantly different between ‘open’ and ‘tented’ cranberry flowers, as signified by Welch's t‐test (p = 0.00153). Each genus is also detected more frequently in ‘open’ cranberry flowers than those which are ‘tented’ (Kruskal‐Wallis test, p = 0.004433) (Figure 4). Yeasts in the class Saccharomycetes were only detected in ‘open’ cranberry flowers, including Pichia and Metschnikowia, but were only detected in three samples. Two of the yeast genera detected in cultures of cranberry flowers were also identified using next‐generation sequencing of cranberry flowers, including Dothiora and Sporobolomyces, which were both identified in ‘open’ cranberry flowers, and one instance of Sporobolomyces was detected in ‘tented’ flowers.
Detection of fungal genera in cranberry flowers with and without tenting treatment using next‐generation sequencing fungal identification methods. This stacked bar chart indicates the frequency of detections for each genus, meaning the number of samples in which a genus is identified through next‐generation sequencing. Genera that were detected in fewer than 10% of samples were grouped for this chart. There is a significant difference in the detection of fungal genera between cranberry flowers that are open to insect visitation (‘open’) and ‘tented’, as indicated by Welch's t‐test (p = 0.00153). Additionally, more genera were detected in ‘open’ cranberry flowers than ‘tented’ flowers, as indicated by a Kruskal Wallis test (p‐value = 0.004433).
Detection of Cranberry Fruit Rot Pathogens and Associated Yeasts
3.7
Several cranberry fruit rot (CFR) pathogens and associated yeasts were detected in cranberry flower and insect samples, but not in wildflower samples (Figure 5). The CFR fungal pathogens detected in these samples include Coleophoma cylindrospora, Phomopsis vaccinii, Phyllosticta vaccinii, and Physolospora vaccinii (Figure 5). The associated yeasts of the CFR complex which were detected include Hanseniaspora uvarum , Pichia madshurica, and Pichia membranifaciens. The highest detection of CFR fungal pathogens was found in flies, with 50% of samples containing at least one CFR fungus and the highest instances of detection of C. cylindrospora (16.7%) and Phyllosticta vaccinii (25%). A large portion of bumble bee samples contained at least one of the CFR fungi (30.8%), and the only detection of H. uvarum occurs within this group. Phomopsis vaccinii is detected in all insect groups: flies 8.3%, bumble bees 7.7%, honey bees 6.3%, and wild solitary bees 14.3%. CFR fungi were detected in more cranberry flowers which were open to insect visitation than those which were not. In cranberry flowers, more detections of CFR fungi occurred in the second collection period between July 1–4, 2024 than the first (June 18–26, 2024), and the associated yeasts were only detected during this second collection period.
Proportion of samples containing cranberry fruit rot (CFR) fungi and associated yeasts by sample type. This stacked bar plot displays the proportion (%) of samples in which members of the cranberry fruit rot complex and associated yeasts are detected in flies, bumble bees, honey bees, wild solitary bees, and cranberry flowers which are open to insect visitation (‘open’) or ‘tented’, and wildflowers.
Discussion
4
This study identified overlaps in the fungal communities between cranberry flowers and pollinators, and the data from this study suggests that visitation by pollinators may increase the fungal abundance, and possibly enhance fungal richness, in flowers. Using a combination of culture‐dependent and next‐generation sequencing methods, we highlight a strong fungal community overlap between cranberry flowers and bees (Table 1 and Figure 1). The ‘core’ community of cranberry flowers, defined as fungal genera detected in over 75% of samples, was also a part of the honey bee and bumble bee core communities (Figure 2). Additionally, the tenting treatment preventing pollinator visitation on cranberry flowers reduced the presence of fungi identified through culture (Figure 3) and community sequencing (Figure 4). Fewer genera were consistently identified through next‐generation sequencing methods in tented cranberry flowers compared to those with access to insect visitation (Figure 4). Finally, with a lens focused specifically on the cranberry fruit rot complex, the pathogens and associated yeasts of this complex were most commonly detected in pollinators and flowers open to pollinator visitation, with low detection in tented cranberry flowers and wildflowers. However, we must note that the open and tented cranberry flowers had no statistical differences in their Bray–Curtis dissimilarity analysis of variance or Shannon diversity. Pollinator visitation appears to increase the abundance of common fungi in cranberry flowers (Figures 3 and 4), which includes members of the cranberry fruit rot complex (Figure 5), and increases the richness of the community through the increased detection of fungal genera that appear in low abundance (Figure 4).
The fungal communities of cranberry flowers characterised through culture and next‐generation sequencing are consistent with those of other studies in this system. The top 15 fungal genera identified through culture included fungi represented in other culture‐based studies of cranberry flowers, including Cladosporium, Fusarium, Penicillium, Epicoccum, Alternaria and Trichoderma (Oudemans et al. 1998; Polashock et al. 2009; Waller et al. 2020). Additionally, the cranberry fruit rot pathogen Allantophomopsis was identified through culture during the flowering stage, which has been documented in other studies using culture‐based methods (Tadych et al. 2012; Waller et al. 2020). The next‐generation sequencing results of cranberry flowers identified additional known associates of cranberry flowers, including Aureobasidium and Pichia, and the additional cranberry fruit rot fungi Phyllosticta and Physalospora (McManus 2001; Wells‐Hansen and McManus 2017; Wood et al. 2023; Zalewski et al. 2021).
Common associates of bees were identified in cultures of cranberry flowers, bees, and wildflowers, including Alternaria, Fusarium, Penicillium and Rhizopus, which other researchers have identified in honey bee and bumble bee provisions and guts (Gilliam et al. 1988), (Rutkowski et al. 2023). The core fungal associates shared between cranberry flowers, bumble bees, and honey bees included Setomelanomma and Rachicladisporium, two genera associated with leaf and needle debris of forests (Crous et al. 2009; Plewa et al. 2012). Additionally, the dead hardwood‐inhabiting crust fungus Irpex was consistently detected in culture in all insect and flower sample types. The collection sites for this study were all located in the ‘central sands’ of central Wisconsin, which includes a mixture of wetland marshes, sandy plains, and oak and pine dominated forests (Curtis 1959). The high occurrence of these wood, leaf, and needle debris‐associated fungal genera may be partially due to transmission by bumble bees, which often nest in forested areas in leaf‐litter or below ground (Liczner and Colla 2019). However, Setomelanomma and Rachicladisporium were also detected as core components of ‘tented’ cranberry flowers as well, so bumble bees alone are likely not the sole hosts for these fungi.
Although tenting treatments prevented the visitation of insect pollinators to cranberry flowers, the dispersal of fungi in these systems still may have reasonably been facilitated by wind, water dispersal or extremely small‐bodied insects such as thrips (Blake 1988), (Agrios 2005). Although the observed differences in fungal communities between these treatments of cranberry flowers are considered to be influenced largely by pollinator visitation, we cannot exclude other characteristics of the tenting treatment which may have influenced the fungal communities of cranberry flowers. For example, the tenting structures may have created a microclimate less conducive to wind dispersal of fungal spores. These tents could also offer increased humidity and slight protection from UV light exposure. However, a noticeable increase, rather than the decrease observed in this study, in the detection of existing genera would be expected due to the suitability of the tented environment to fungal growth (Agrios 2005).
The methods used to sample pollinators and understand pollinator visitation were insufficient to fully characterise the wild pollinator community diversity or their microbiomes. The sticky‐trap sampling method used to prevent contamination between bees and collection materials introduced bias in the bees collected, and bumble bees were under‐represented in pollinator sampling from 2023. In order to robustly document the pollinator communities in these cranberry systems, a combination of pan‐trapping and possibly hand‐netting would be necessary, as has been conducted in other studies in cranberry systems (Averill et al. 2018; Broussard et al. 2011; Gaines Day 2013; Gervais et al. 2018; Mackenzie and Winston 1984). The Bombus community represented was further biassed in 2024, as wild Bombus impatiens were not collected from sites which had these commercial bumble bees. The small sample size of pollinators collected by hand‐netting and pollinator visitation observations resulted in a very small wild bee community used to characterise fungal diversity. The lack of a ‘core’ wild bee community is likely in part due to making inferences on fungal communities across a large diversity of bees using a small subset of samples. Additionally, several insect species documented to pollinate in managed cranberry systems were not captured in this study, including the many species within the genera Andrena, Bombus, Halictus, Lasioglossum, Megachile, Melissodes, and Osmia which were documented as abundant and potentially important pollinators in cranberry systems (Averill et al. 2018; Broussard et al. 2011; Gaines Day 2013; Mackenzie and Winston 1984). However, several important genera from these studies have representatives included in this work, including Apis, Agopostemon, Andrena, Bombus, Ceratina Dufourea, Eucera, Hylaeus and Megachile (Table S2). Further, this study only included Toxomerus flies for fungal community characterisation, but a study in Quebec documented 33 species in cranberry systems, although Toxomerus species accounted for 46.41% of all specimens (Gervais et al. 2018). Future research including a robust sample size of wild bees and flies may elucidate core mycobiome components and other fungal associates which were not able to be identified in this study.
Other researchers have also aimed to determine pollinators' contributions to floral microbial communities. Aizenberg‐Gershtein et al. (2013) used a mesh‐bag treatment to prevent pollinator visitation on the flowers of two plant species, finding altered nectar bacterial communities of these flowers with honey bee visitation. However, McFrederick et al. (2017) also used a mesh‐bag treatment to prevent pollinator visitation on wildflower communities and found similar bacterial communities among bagged and open wildflowers, concluding that bees may increase the populations of these microbes, but may not necessarily increase the bacterial richness on these flowers. Although the contributions of pollinators to floral fungal communities has received less research attention, several studies have determined that shared flower visitation is a common transmission route for the pathogens of bees (Evison et al. 2012; Figueroa et al. 2019; Graystock et al. 2015), and that insect visitation changes the yeast communities in the nectar of wildflowers (Belisle et al. 2012; Herrera et al. 2008, 2010).
This study differs from these previous investigations in several ways. First, the setting of this study is within a managed agricultural system. The manipulation of the study agroecosystem adds a layer of complexity due to the chemical and cultural control measures in place in cranberry systems, including fungicides. Additionally, this study documents the contributions of groups of pollinators, including the understudied pollinating hover fly Toxomerus, to the overall fungal communities present in cranberry and wildflowers, rather than targeting bacteria or specific subsets of fungi such as pathogens of bees or yeast species. The understudied nature of fungal communities in the environment is most clear in the limitations presented in the database used to identify ASVs in wildflower samples, in which the majority of reads were determined to be fungi, with no further classification in the database available. Through this study's investigation of whole‐fungal communities, this research adds a foundational understanding of not only pollinator‐angiosperm microbial sharing, but importantly documents the rich diversity of environmental fungi yet to be reliably characterised in genetic databases.
Our research has important implications for the health of managed cranberry plants, which is highlighted by the highest detection of cranberry fruit rot pathogens in Toxomerus hover flies, social bees, and ‘open’ cranberry flowers without the tenting treatment. The cranberry fruit rot pathogen Phyllosticta vaccinii was detected in hover flies, honey bees, and cranberry flowers which were open to insect visitation. The storage rot pathogen Allantophomopsis was detected in culture in cranberry flowers, honey bees, and wildflowers, but not in wild bees. This information, along with no detection of other cranberry fruit rot fungi in wildflowers, suggests that wild bees and wildflowers may not be important reservoirs of CFR pathogens. Hover flies and honey bees had the highest observed cranberry flower visitation, indicating more contact incidences than in wild bee groups. Hover flies had the highest detection of cranberry fruit rot fungi of all insect groups, which may have been accumulated partially during post‐pupation emergence from the soil (Coffey 2023; Gullan and Cranston 2014), a potential overwintering source for CFR pathogens. Additionally, the decaying organic matter available for fly feeding in cranberry marshes may include leaf litter and rotten fruits from previous seasons, which have been identified as important secondary sources of inoculum for CFR fungi (Oudemans et al. 1998). We must be clear that the detection within insect samples does not indicate vectoring capabilities on insects; however, future studies on the vectoring capabilities of hover flies and honey bees could highlight the movement of cranberry fruit rot pathogens within cranberry agroecosystems.
Although research efforts have worked to determine the effects of different fungicide groups on pathogens within the cranberry fruit rot disease cycle (McManus 2001; Oudemans et al. 1998; Wood et al. 2023; Zuckerman 1958), chemical and non‐chemical pest management practices have not yet been evaluated for their effects on the fungal communities of cranberry flowers. The majority of the cranberry marshes visited during this study use two classes of fungicides for the control of cranberry fruit rot diseases, as is standard in the industry. Many non‐target fungi in these systems are likely reduced during these chemical applications, and the fungal communities of cranberry flowers may be significantly altered. Future research on the impact of management programs for disease control on the fungal communities present on cranberry flowers, and the implications of these potentially altered communities on cranberry disease, would improve the research and practice of management for cranberries.
From a pollinator‐health perspective, whereas many of the documented commensal fungal associates of bees have also been identified in this study, the use of honey bee rentals and chemical controls is likely to alter the microbial communities of wild, native bees in cranberry agroecosystems (Fernandez De Landa et al. 2023; Morris et al. 2020). Given the known dispersal of bee pathogen transmission through the use of shared floral resources, the detection of Aspergillus flavus, a pathogen known to cause stonebrood disease in honey bees, provokes some concern for honey bee and wild bee health (Becchimanzi and Nicoletti 2022; Jensen et al. 2013). Additionally, the use of fungicides during the cranberry bloom period may alter the microbiome of social bees and wild pollinators (Porras et al. 2024; Reiß et al. 2023), although the extent of this microbiome alteration has not been explored in cranberry systems.
A known biological antagonist of pathogens in other crop systems, Aureobasidium pullulans, was detected frequently in honey bees, bumble bees, and cranberry flowers. This yeast‐like fungus has been studied for its effectiveness as a biocontrol agent for grey mould disease in strawberries (Adikaram et al. 2002) and grapes (Galli et al. 2021), and the postharvest diseases of apples (Castoria et al. 2001; Mari et al. 2012). Aureobasidium pullulans has been studied in one instance for its use as a biological antagonist of Allantophomopsis cytisporea (formerly Apostrasseria lunata) and Strasseria oxycocci, the black rot pathogens of cranberry (Stretch 1989). The use of A. pullulans as a biological antagonist for other cranberry fruit rot (CFR) fungi could be a promising direction for future studies, as limited biocontrol products are currently registered for the control of cranberry fruit rot.
Conclusions
5
This study provides evidence that pollinators in cranberry systems share fungal associates with cranberry flowers, and pollinator visitation to cranberry flowers may increase the richness and abundance of fungi in cranberry flowers. The most abundant fungi considered ‘core’ to cranberry flowers are also core components of honey bee and bumble bee visitors, indicating shared mycobiome components within this unique spatiotemporal landscape. Additionally, cranberry fruit rot fungi are detected in insect pollinators, but importantly there was no detection of CFR pathogens in adjacent wildflower communities. Future studies should address the implications of increased floral‐fungal diversity during the cranberry bloom period on cranberry fruit health.
Author Contributions
Shawn Steffan: conceptualization, funding acquisition, methodology, supervision. Celeste C. Mezera: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, data curation. Leslie A. Holland: conceptualization, investigation, funding acquisition, methodology, validation, writing – review and editing, project administration, supervision.
Funding
This work was supported by the United States Department of Agriculture (USDA) Agriculture Research Service (ARS) (grant number AWD00000457) and under agreement number 58‐5090‐2‐042.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Sampling design illustration. Samples of cranberry flowers, wildflowers, and pollinating insects were collected at 8 cranberry marshes in central Wisconsin within the area highlighted in red on the map of Wisconsin. Sampling for 2023 included 4 sites, at least 500 m apart, for open and tented cranberry flowers, wildflowers, and insect collection which were collected using sticky traps. Fungi were identified from these samples using culturing techniques. Sampling for 2024 differed in that two sites were used to collect cranberry flowers, wild flowers, and insects, and insects were actively targeted using hand‐nets. Fungal communities of samples from 2024 were identified using Illumina sequencing techniques.
Figure S2: NMDS plot of fungal communities identified using next‐generation sequencing by sample type. Non‐metric multidimensional scaling plot of the fungal genera detected through Illumina sequencing of open and tented cranberry flowers, wildflowers, bumble bees, honey bees, wild bees and flies. Overlap indicates similarity between fungal communities.
Table S1: Wildflowers collected for fungal identification. Wildflowers, consisting of naturally occurring unmanaged flowering plants and plants in pollinator gardens aimed to attract bees, were collected within 100 m of cranberry beds. The common name, scientific name and year collected are expressed in this table.
Table S2: Bees and flies collected from cranberry marshes in 2023 and 2024. Bees were collected differently between years: 2023 collections were from sticky traps, and 2024 collections were from targeted hand‐netting. All organisms were identified to genus using morphological traits, and to species when possible. Species identification and species counts are reflected in the table.
Table S3: Observations of pollinator visitation in cranberry flowers. The number of visits by pollinators within 1‐m^2^ of cranberry flowers within 5 min are recorded with corresponding date, time, temperature, cloud cover, wind speed, and percent of cranberry bloom information. All visits recorded within one instance are displayed in the ‘All visits’ column. Two observers maximum recorded at one time, and the observer's initials are noted in the ‘Observer’ column.
Table S4: Alpha Diversity Metrics by sample type. The Shannon diversity, Simpson evenness, Dominance index, and Pielou's evenness score are displayed by sample type. Significant differences between sample types for each of these diversity metrics are indicated by the p‐value of their respective statistical tests. ∂ = Analysis of Variance test, followed by Tukey's HSD, ß = Kruskal‐Wallis test, followed by Dunn's test.
Table S5: Pairwise perMANOVA analysis tests. Post hoc pairwise PerMANOVA tests results between all sample pairs are listed in this table. Groups compared include both open and tented cranberry flowers, wildflowers, bumble bees, honey bees, wild solitary bees, and flies. For each comparison, the F‐value (top), R ^2^ value (middle), and p‐value (bottom) are listed in the corresponding cell. p‐values in bold text indicate significant differences fungal community composition (p‐value ≤ 0.01). PERMDISP analysis revealed significant differences in multivariate dispersion among treatment groups; therefore, PERMANOVA results may reflect a combination of differences in group centroids and differences in within‐group variability.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Adikaram, N. K. B. , D. C. Joyce , and L. A. Terry . 2002. “Biocontrol Activity and Induced Resistance as a Possible Mode of Action for Aureobasidium Pullulans Against Grey Mould of Strawberry Fruit.” Australasian Plant Pathology 31, no. 3: 223. 10.1071/AP 02017. · doi ↗
- 2Agrios, G. 2005. Plant Pathology. Environmental Effects on the Development of Infectious Plant Disease. Elsevier.
- 3Aizenberg‐Gershtein, Y. , I. Izhaki , and M. Halpern . 2013. “Do Honeybees Shape the Bacterial Community Composition in Floral Nectar?” P Lo S One 8, no. 7: e 67556. 10.1371/JOURNAL.PONE.0067556.23844027 PMC 3701072 · doi ↗ · pubmed ↗
- 4Alexandrova, M. , B. Cimini , C. Bazzi , E. Carpana , S. Massi , and A. G. Sabatini . 2002. “The Role of Honeybees in Spreading Erwinia Amylovora.” Acta Horticulturae 590: 55–60. 10.17660/ACTAHORTIC.2002.590.5. · doi ↗
- 5Amon, N. D. , M. Quezada , D. Labarre , and C. Guédot . 2023. “Pollination Practices and Grower Perceptions of Managed Bumble Bees (Bombus spp.) as Pollinators of Cranberry in Quebec and Wisconsin.” Renewable Agriculture and Food Systems 38: e 43. 10.1017/S 1742170523000352. · doi ↗
- 6Anderson, K. E. , T. H. Sheehan , B. J. Eckholm , B. M. Mott , and G. De Grandi‐Hoffman . 2011. “An Emerging Paradigm of Colony Health: Microbial Balance of the Honey Bee and Hive (Apis mellifera).” Insectes Sociaux 58, no. 4: 431–444. 10.1007/S 00040-011-0194-6/TABLES/1. · doi ↗
- 7Ascher, J. S. , and J. Pickering . 2020. “Discover Life Bee Species Guide and World Checklist (Hymenoptera: Apoidea: Anthophila).”
- 8Averill, A. L. , M. M. Sylvia , N. Hahn , and A. V. Couto . 2018. “Bees (Hymenoptera: Apoidea) Foraging on American Cranberry in Massachusetts.” 25, no. 3: 502–512. 10.1656/045.025.0314. · doi ↗