Fungicide ingestion reduces net energy gain and microbiome diversity of the solitary mason bee

Fungicides are frequently used during tree fruit bloom and can threaten insect pollinators. However, little is known about how non-honey bee pollinators such as the solitary bee, Osmia cornifrons, respond to contact and systemic fungicides commonly used in apple production during bloom. This knowledge gap limits regulatory decisions that determine safe concentrations and timing for fungicide spraying. We evaluated the effects of two contact fungicides (captan and mancozeb) and four translaminar/plant systemic fungicides (cyprodinil, myclobutanil, penthiopyrad, and trifloxystrobin) on larval weight gain, survival, sex ratio, and bacterial diversity. This assessment was carried out using chronic oral ingestion bioassays where pollen provisions were treated with three doses based on the currently recommended field use dose (1X), half dose (0.5X), and low dose (0.1X). Mancozeb and penthiopyrad significantly reduced larval weight and survival at all doses. We then sequenced the 16S gene to characterize the larvae bacteriome of mancozeb, the fungicide that caused the highest mortality. We found that larvae fed on mancozeb-treated pollen carried significantly lower bacterial diversity and abundance. Our laboratory results suggest that some of these fungicides can be particularly harmful to the health of O. cornifrons when sprayed during bloom. This information is relevant for future management decisions about the sustainable use of fruit tree crop protection products and informing regulatory processes that aim to protect pollinators.

Contact and ingestion exposure of these fungicides to honey bees are relatively well known since it is part of the pesticide registration process by the U. S. Environmental Protection Agency and many other countries' regulatory agencies [12][13][14] .However, the effects of fungicides on non-honey bees are less well known as they are not required as part of the registration protocols in the US 15 .In addition, there is a general lack of standardized testing protocols for solitary bees 16,17 , and it is challenging to maintain colonies that provide bees for testing 18 .Various species of managed Osmia in Europe and the US are increasingly being tested to examine pesticide effects on wild bees, and recently, a standardized protocol was developed for O. cornifrons 19 .
Osmia cornifrons is univoltine and has been commercially used in tree fruit crops to supplement or replace honey bees.These bees emerge between March and April, with protandrous males emerging three to four days ahead of females.After mating, females actively collect pollen and nectar for provisioning a series of brood cells within tubular nest cavities (natural or artificial) 1,20 .Eggs are laid on the pollen provisions within a cell; the female then builds a mud partition before provisioning the next cell 21 .The first larval instar is enclosed within the chorion, feeding on embryonic fluids.From the second to fifth instar (prepupa), the larva feeds on pollen provisions 22 .Once the pollen provision is completely consumed, the larva forms a cocoon, pupates and becomes an adult in the same brood cell, usually by the end of summer 20,23 .The adult emerges in the following spring 22 .Adult survival is correlated with net energy gain (weight gain) based on the provisions consumed.Therefore, the nutritional quality of the pollen, as well as other factors such as weather or pesticide exposure, are determinants of survival and fitness 24 .
Pre-bloom applications of insecticides and fungicides with the ability to move within the plant vascular system to varying degrees from translaminar (e.g., able to move from the top surface of a leaf to the bottom surface as with some fungicides) 25 to genuinely systemic neonicotinoid insecticides that can move from root applications up into the canopy have been previously shown to move into the nectar of apple bloom 26 where they can kill adult O. cornifrons 27 .Some pesticides can also move into the pollen, affecting O. cornifrons larval development and causing mortality 19 .Other studies have shown that some fungicides can dramatically change nesting behavior in a congener, O. lignaria 28 .Furthermore, laboratory and field studies simulating pesticide (including fungicides) exposure scenarios demonstrated adverse effects on physiology 22 , morphology 29 , and survival in honey bees and some solitary bees 12 .The impact of various fungicide sprays applied directly to open flowers during bloom, which would contaminate the pollen collected by adult O. cornifrons for larval development, has yet to be explored 30 .
It is increasingly recognized that larval development is affected by the microbial community present in the pollen and digestive system.The bee microbiome influences parameters such as body mass 31 , metabolism alterations 22 , and susceptibility to pathogens 32 .Prior research has investigated the effects of developmental stages, nutrients, and environment on solitary bee microbiome.These studies revealed similarities in structure and abundance of the microbiome of both larvae and pollen 33 and the most abundant bacteria genera, Pseudomonas and Delftia, in solitary bee species 34,35 .However, the impact of fungicides on the larval microbiome through direct oral exposure remains unexplored despite its relevance for strategies aimed at preserving bee health.
This study tested the effects of field-realistic doses of six commonly used fungicides registered for use on tree fruit throughout the US, which included contact and systemic fungicides via oral exposure to O. cornifrons larvae from contaminated provisions.We found contact and systemic fungicides reduced bee weight gain and increased mortality, with the most severe impact associated with mancozeb and penthiopyrad.Then, we compared the microbial diversity of larvae fed with pollen provisions treated with mancozeb against those fed with control provisions.We discuss potential mechanisms underpinning the lethality as well as implications for integrated pest and pollinator management (IPPM) programs 36 .

Rearing method and egg collection
Overwintering adults of O. cornifrons in cocoons were obtained from the Fruit Research Center in Biglerville, PA and stored at − 3 to 2 °C (± 0.3 °C).before the experiment (a total of 600 cocoons).In May 2022, groups of 100 cocoons of O. cornifrons were transferred daily to plastic cups (50 cocoons/cup, 5 cm DI × 15 cm length) with a napkin introduced inside the cup to aid emergence and provide a substrate for chewing, which reduces mason bee stress 37 .Two plastic cups with cocoons were placed inside an insect cage (30 × 30 × 30 cm, BugDorm MegaView Science Co. Ltd., Taiwan) with a 10 ml feeder containing 50% sucrose solution and held for four days to ensure emergence and mating at 23 °C with 60% RH and photoperiod 10 L (low intensity): 14 D. One hundred individual mated females, and males were released every morning for six days at the peak of apple bloom (100 individuals/ day) in two artificial nest boxes (trap nests: width 33.66 × height 30.48 × length 46.99 cm; Supplementary Fig. 1 Both nest boxes were faced east, covered with a green plastic garden fence (Everbilt Model# 889250EB12, hole size 5 × 5 cm, 0.95 m × 100 m) to keep rodents and birds out, and provided with mud on the soil surface near the nest boxes (Supplementary Fig. 1a).Osmia cornifrons eggs were harvested daily from the nest boxes by collecting 30 tubes and bringing them to the laboratory.An incision was made on the end of the tube using scissors; then, we unraveled the spiral tubing, revealing the brood cells.Individual eggs and their pollen provisions were

Fungicide exposure
In laboratory we evaluated the oral toxicity of six fungicides applied before and during apple bloom, at three concentrations (0.1X, 0.5X, and 1X, where 1X is the labeled high field dose applied in 100 gal water/acre = field concentration, Table 1).Each concentration was replicated 16 times (n = 16).Toxicity of two contact fungicides (Table S1: mancozeb at 2696.14 ppm and captan at 2875.88 ppm) and four systemic fungicides (Table S1: Penthiopyrad at 250.14 ppm; Trifloxystrobin at 110.06 ppm; Myclobutanil at 75.12 ppm; Cyprodinil at 280.845 ppm) which are widely used in fruits, vegetables, and ornamental crops.We homogenated pollen provisions using a grinder and transferred 0.20 g into a well (24--Well Falcon Plate), added and mixed 1 µL of fungicide solution, and formed a pyramid-shaped pollen provision with a 1 mm deep hole where the egg was placed using a minispatula (Supplementary Fig. 1c,d).Falcon plates were held at room temperature (25 °C) and 70% RH 19 .We compared them against control larvae fed on homogenized pollen provisions treated with pure water.We recorded mortality and measured larval weight every other day until the prepupal larval instar using an analytical scale (Fisher Scientific, accuracy = 0.0001 g).Finally, the sex ratio was evaluated by dissecting cocoons after 2.5 months.

Metagenome sample preparation and sequencing
DNA was extracted from whole O. cornifrons larvae (n = 3 per treatment condition, mancozeb-treated and untreated pollen provisions), we conducted a microbial diversity analysis on these samples, particularly because the highest mortality rates were observed in larvae fed on mancozeb-treated pollen provisions.. Using the DNAZymoBIOMICS ® -96 MagBead DNA Kit (Zymo Research, Irvine, CA), DNA was amplified, purified, and sequenced on Illumina ® MiSeq™ with a v3 reagent kit (600 cycles).Bacterial 16S ribosomal RNA gene-targeted sequencing was performed using the Quick-16S™ NGS Library Prep Kit (Zymo Research, Irvine, CA), employing primers that target the V3-V4 region of the 16S rRNA gene.Additionally, 18S sequencing was conducted with 10% PhiX spike-in, and amplification was performed using the primer pair 18S001 and NS4.

Statistical analysis
All experiments' data were tested for normality (Shapiro-Wilks) and homogeneity of variance (Levene's test) assumptions.Since data sets did not meet the assumptions for parametric analysis and transformation failed to normalize the residuals, we employed nonparametric two-way ANOVAs (Kruskal-Wallis) with two factors [time (three time points 2, 5 and 8 days) and fungicide] to assess the effect of treatments on larval fresh weight, then posthoc nonparametric pairwise comparisons were conducted using Wilcoxon test.We used a generalized linear model (GLM) with a Poisson distribution to compare the impacts of fungicides on survival at three fungicide concentrations 41,42 .For differential abundance analysis, amplicon sequence variant counts (ASVs) were collapsed at the genus level.Differential abundance comparisons between groups using 16S (genus level) and 18S relative abundances were performed using a Generalized Additive Model for Location, Scale, and Shape (GAMLSS) with zero-inflated beta (BEZI) family distribution, implemented in the metamicrobiomeR 43 (v1.1).Mitochondria and Chloroplast genera were removed before differential analysis.Due to the differing levels of the 18S classification, only the lowest level of each taxon was used for differential analysis.All statistical analyses were conducted using R (v. 3.4.3., CRAN project) (Team 2013).

Mortality
The highest mortality was observed in larvae fed on pollen treated with mancozeb, followed by penthiopyrad and trifloxystrobin.Mortality increased with the dose of mancozeb and penthiopyrad (Fig. 2; Table 2).However, O. cornifrons mortality only slightly increased as the concentration of trifloxystrobin rose; cyprodinil and captan did not significantly increase mortality compared to the control treatment.

Microbiome
Analysis of 16S sequences showed that the bacteriome differed between larvae fed on mancozeb-treated pollen and those fed on untreated pollen (Fig. 4a).The microbial index was higher in untreated pollen-fed larvae than larvae fed on mancozeb-treated pollen (Fig. 4b).While the difference in observed richness between the groups was not statistically significant, it was markedly lower than in larvae fed on untreated pollen (Fig. 4c).The relative abundance showed that the microbiome of larvae fed on control pollen was more diverse than in those fed on mancozeb-treated (Fig. 5a).Descriptive analysis indicated the presence of 28 genera across the control and mancozeb-treated samples (Fig. 5b).c Analysis using 18S sequencing did not exhibit significant differences (Supplementary Fig. 2).

Discussion
Our results suggest that oral exposure to contact (mancozeb) and systemic (penthiopyrad and trifloxystrobin) fungicides which are widely applied during bloom significantly reduced weight gain and increased mortality in O. cornifrons larvae.Additionally, mancozeb notably reduced the microbiome's diversity and richness at the  www.nature.com/scientificreports/pre-pupal stage.Another systemic fungicide, myclobutanil, significantly reduced larval weight gain at all three doses.This effect was notable at the second (5th day) and third (8th day) time points.In contrast, cyprodinil and captan did not significantly reduced weight gain or survival compared to the control group.To our knowledge, this work is the first to determine the effects of field doses of a wide range of fungicides used in crop protection on O. cornifrons using direct exposure through pollen provisions.
All fungicide treatments significantly reduced weight gain compared to control treatments.Mancozeb had the most substantial impact on larval weight gain, declining 51% on average, followed by penthiopyrad.However, other studies did not report the adverse effects of a fungicide field dose on the larval stage 44 .Although dithiocarbamate fungicides have been shown to have low acute toxicity 45 , Ethylenebisdithiocarbamate (EBDCS), such as mancozeb, can degrade into ethylene thiourea.Given its reported mutagenic impacts on other animals, this  degradation product may be responsible for the observed effects 46,47 .Previous research indicated that the formation of ethylene thiourea is influenced by factors such as increasing temperature 48 , moisture levels 49 , and length of product storage period 50 .Proper storage conditions for the fungicide could mitigate these adverse effects.In addition, concern about the toxicity of penthiopyrad has been raised by the European Food Safety Authority, which found carcinogenic effects in the digestive system of other animals 51 .
Oral ingestion of mancozeb, penthiopyrad, and trifloxystrobin increased the mortality of O. cornifrons larvae.In contrast, myclobutanil, cyprodinil, and captan did not affect mortality.These results diverge from Ladurner et al. 52 , which showed captan significantly reduced survival in O. lignaria and Apis mellifera L. (Hymenoptera, Apidae) adults.Furthermore, fungicides, such as captan and boscalid, were found to induce larval mortality [52][53][54] or altered foraging 55 .These changes could, in turn, affect the nutritional quality of pollen, ultimately impacting the energy gain during the larval stage.The mortality observed in control groups was consistent with those reported in other studies 56,57 .
The male-biased sex ratio observed in our work can be attributed to factors such as insufficient mating 58 and adverse weather conditions during the blooming period 59 , as previously suggested by Vicens & Bosch 60 for O. cornuta.Although females and males in our study were given four days for mating-a period generally considered adequate for successful mating 38 -we intentionally reduced the light intensity to minimize stress.The modification, however, may have inadvertently hindered the mating process 61 .Furthermore, the bees were exposed to several days of unfavorable weather, including rain and cold temperatures (< 5 °C), which could have also negatively impacted mating success 4,23 .
Although our study focuses on the microbiomes of whole larvae, our findings provide insights into potential relationships among bacterial communities that may be crucial for bee nutrition and fungicide exposure.For instance, larvae fed on pollen treated with mancozeb exhibited a significant reduction in both the structure and abundance of their microbial communities compared to those fed on untreated pollen.In larvae that consumed untreated pollen, the dominant bacterial groups were Proteobacteria and Actinobacteria, which are primarily aerobic or facultatively aerobic.Delftia bacteria, often associated with solitary bee 62 species, are known to exhibit antibiotic activity, suggesting their potential protective role against pathogens [62][63][64] .Another bacterial genus that was abundant in larvae fed on untreated pollen but significantly reduced in mancozeb-treated larvae was Pseudomonas.Our results corroborate previous studies that identified Pseudomonas as one of the most abundant genera in O. bicornis 35 and other solitary bees 34 .Although experimental evidence of the Pseudomonas' role in O. cornifrons health has not been studied, this bacterium has been shown to contribute to the synthesis of a defensive toxin in the beetle, Paederus fuscipes, and promotes arginine metabolism under in vitro conditions 35,65 .These observations suggest its potential role in viral and bacterial defense during O. cornifrons larval development.Microbacterium was another genus detected in our study, which has been reported in high abundance in black soldier fly larvae under starvation field conditions 66 .In O. cornifrons larvae, Microbacterium could contribute to the balance and resiliency of the gut microbiome under stress conditions.In addition, Rhodococcus was found in O. cornifrons larvae and is known for its detoxification capabilities 67 .This genus has also been found in the guts of A. florea but at very low abundance 68 .Our results indicate a diverse range of genetic variations across numerous microbial taxa that may alter metabolic processes in larvae.However, the functional diversity of bacteria in O. cornifrons needs to be better understood.
To conclude, the results suggest that mancozeb, penthiopyrad, and trifloxystrobin reduced weight gain and increased mortality of O. cornifrons larvae.Although there is increasing interest in the impacts of fungicides on pollinators, the effects of residual metabolites of these compounds need to be better understood.These results may be incorporated into recommendations for integrated pollinator management programs, which can instruct farmers to avoid certain fungicides during the prebloom and bloom period of fruit trees by fungicide selection and application timing change or promot use of less detrimental alternatives 36 .This information has critical implications for establishing pesticide application recommedations such as adjusting the existing spray program in fungicide selection and spray timing change or promote use of less detrimental alternatives.Additional studies are needed on the side effects of fungicides on sex ratio, foraging behavior, gut microbiome, and the molecular mechanisms underlying O. cornifrons weight loss and mortality.

Figure 4 .
Figure 4. Comparisons of the Shannon richness and observed richness at the phylum level based on SAV profile for 16S sequences.(a) Principal coordinate analysis (PCoA) based on the overall structure of microbial communities in untreated pollen-fed larvae or control (blue) and mancozeb-fed larvae (orange).Each data point represents an individual sample.PCoA was calculated using Bray-Curtis distances with a multivariate t-distribution.Ellipses represent an 80% confidence level.(b) Boxplots, raw data (points) of Shannon richness, and c. observed richness.Box plots display the median line, interquartile range (IQR) boxes, and 1.5 × IQR (n = 3).