ORIGINAL RESEARCH

published: 03 September 2021

doi: 10.3389/fmicb.2021.717990

Resistance and Vulnerability of

Honeybee (Apis mellifera) Gut
Bacteria to Commonly Used
Pesticides
Ana Cuesta-Maté 1 , Justinn Renelies-Hamilton 1 , Per Kryger 2 , Annette Bruun Jensen 3 ,
Veronica M. Sinotte 1* and Michael Poulsen 1
1
Section for Ecology and Evolution, Department of Biology, University of Copenhagen, Copenhagen, Denmark,
2
Entomology and Plant Pathology, Department of Agroecology, Aarhus University, Aarhus, Denmark, 3 Section
for Organismal Biology, Department of Plant and Environmental Sciences, University of Copenhagen, Copenhagen, Denmark

Agricultural and apicultural practices expose honeybees to a range of pesticides that

have the potential to negatively affect their physiology, neurobiology, and behavior.
Accumulating evidence suggests that these effects extend to the honeybee gut
Edited by: microbiome, which serves important functions for honeybee health. Here we test the
Joel White, potential effects of the pesticides thiacloprid, acetamiprid, and oxalic acid on the gut
Université Toulouse III Paul Sabatier,
France microbiota of honeybees, first in direct in vitro inhibition assays and secondly in an in vivo
Reviewed by: caged bee experiment to test if exposure leads to gut microbiota community changes.
Erick Motta, We found that thiacloprid did not inhibit the honeybee core gut bacteria in vitro, nor
University of Texas at Austin,
did it affect overall community composition or richness in vivo. Acetamiprid did also
United States
Kasie Raymann, not inhibit bacterial growth in vitro, but it did affect community structure within bees.
University of North Carolina The eight bacterial genera tested showed variable levels of susceptibility to oxalic acid
at Greensboro, United States
in vitro. In vivo, treatment with this pesticide reduced amplicon sequence variant (ASV)
*Correspondence:
Veronica M. Sinotte richness and affected gut microbiome composition, with most marked impact on the
[Link]@[Link] common crop bacteria Lactobacillus kunkeei and the genus Bombella. We conducted
network analyses which captured known associations between bacterial members and
Specialty section:
This article was submitted to illustrated the sensitivity of the microbiome to environmental stressors. Our findings
Microbial Symbioses, point to risks of honeybee exposure to oxalic acid, which has been deemed safe for
a section of the journal
Frontiers in Microbiology
use in treatment against Varroa mites in honeybee colonies, and we advocate for more
Received: 31 May 2021
extensive assessment of the long-term effects that it may have on honeybee health.
Accepted: 30 July 2021
Keywords: microbiome, symbiosis, anthropogenic stressor, social insect, neonicotinoid, acaricide
Published: 03 September 2021
Citation:
Cuesta-Maté A, INTRODUCTION
Renelies-Hamilton J, Kryger P,
Jensen AB, Sinotte VM and The deleterious effects of pesticides on non-target organisms have been a critical area of study in
Poulsen M (2021) Resistance
the last decades (Mancini et al., 2019). The vital ecosystem services pollinators provide prioritize
and Vulnerability of Honeybee (Apis
mellifera) Gut Bacteria to Commonly
assessment of pesticide toxicity in this group, and evidence of pesticide harm to honeybees
Used Pesticides. continues to accumulate (Decourtye et al., 2004a; Gill et al., 2012; Goulson et al., 2015). The western
Front. Microbiol. 12:717990. honeybee, Apis mellifera, maintains a wide distribution, generalist foraging behavior and pollination
doi: 10.3389/fmicb.2021.717990 competences that make it one of the most important species of pollinators across the world

Frontiers in Microbiology | [Link] 1 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

(Hung et al., 2018). Specifically, honeybees play a keystone Until recently, most research on the effect of pesticides
role in pollination of natural ecosystems and agricultural crops on honeybee health focused on the direct effects on the bees,
(Goulson et al., 2015; Hristov et al., 2020). However, honeybee but accumulating evidence suggests that effects extend to
populations have declined globally over the past decades, with the honeybee gut microbiome. The honeybee microbiome is
a continued reduction in colony numbers recorded in the simple and conserved, with 8–10 core bacterial taxa that are
United States, northwestern Europe, and Russia between 1940 omnipresent, regardless of geographical origin (Martinson
and 2010 (Potts et al., 2010; US Department of Agriculture et al., 2012; Moran et al., 2012; Ellegaard and Engel, 2019;
(USDA), 2016, 2021; Brosi et al., 2017), and a striking average Figures 1A,B). The microbiome modulates immunity
annual loss of 30% of colonies reported by northern hemisphere against pathogens, partakes in digestion of pollen and in
beekeepers over the past decade (Brosi et al., 2017; vanEngelsdorp the neutralization of toxins (Engel et al., 2012; Engel and Moran,
et al., 2017). A combination of anthropogenic stressors appears 2013; Kwong and Moran, 2016; Kešnerová et al., 2017; Raymann
to drive honeybee declines (Meeus et al., 2018), including and Moran, 2018), promotes host weight and health, and
reduced habitats (Naug, 2009; Breeze et al., 2014), reduced mediates hormonal signaling (Zheng et al., 2016). Adverse effects
genetic diversity (Espregueira Themudo et al., 2020), use of of certain agricultural compounds on honeybee gut microbes
antimicrobials in apiculture (Tian et al., 2012; Li et al., 2017; have been documented. The herbicide glyphosate can perturb
Raymann et al., 2017), and pesticide application (Boncristiani the absolute and relative abundances of dominant bacterial
et al., 2012; Williamson et al., 2014; Johnson, 2015). community members and increase honeybee susceptibility
Agricultural and apicultural practices expose honeybees to pathogens (Motta et al., 2018; Blot et al., 2019). Chronic
to a range of pesticides that can damage their physiology, exposure to the highly toxic nitroguanidine neonicotinoids
neurobiology, and behavior, and ultimately may result in colony (e.g., imidacloprid and thiamethoxam) in laboratory settings
decline. Two pesticide groups with potentially detrimental effects can induce changes in gut bacteria community composition
to honeybee health are insecticides, which are employed within of healthy honeybees (Rouzé et al., 2019). However, these
their natural environment, and acaricides, which are applied results have not been found when bees are returned to the
within colonies. Commonly used neonicotinoid insecticides hive after exposure (Raymann et al., 2018b). Thiacloprid, a
inhibit nervous system function by antagonizing acetylcholine cyanoguanidine neonicotinoid, potentially invokes dysbiosis
receptors (Casida and Durkin, 2013). In honeybees, sublethal of the gut microbiome (Liu et al., 2020), and although lower
exposure to nitroguanidine neonicotinoids can alter locomotion field-level concentrations may not alter colony performance,
(Williamson et al., 2014; Charreton et al., 2015), memory and they reduce immune expression against pathogens (Siede et al.,
learning (Decourtye et al., 2004b; Dacher et al., 2005; Shi 2017). The effects of other commonly used cyanoguanidine
et al., 2019), consequentially impairing flight ability, navigation neonicotinoids, such as acetamiprid, have yet to be investigated.
(Tosi et al., 2017), and the proboscis extension response to Pesticides and antimicrobials used in apiculture to mitigate
sucrose (El Hassani et al., 2008; Thany et al., 2015). Due to parasite and pathogen infections can also impact the honeybee
the adverse effects, nitroguanidine neonicotinoids have been gut microbiota (Li et al., 2017; Raymann et al., 2017, 2018a).
banned in several countries (European commission, 2013; Acaricides such as coumaphos and tau-fluvalinate can influence
Dewar, 2017), while the less toxic cyanoguanidine neonicotinoids the structure of the bacterial community in the gut (Kakumanu
remain widely used (Iwasa et al., 2004; Shi et al., 2019). et al., 2016). Oxalic acid has antibacterial activity, including
Although cyanoguanidine neonicotinoids, such as thiacloprid against Lactobacillus strains isolated from honeybees (Diaz et al.,
and acetamiprid, are considered safer, they can still affect 2019), yet inhibition of other key core taxa and alteration of the
honeybee behavior, memory, and immune functions, and can gut microbial community have not been evaluated. Thus, our
be lethal at high concentrations (Iwasa et al., 2004; Thany knowledge remains sparse despite the potential adverse effects of
et al., 2015; Brandt et al., 2017; Liu et al., 2020). Honeybees commonly used pesticides on honeybee gut microbes.
are also frequently exposed to acidic acaricides that are used In this study, we address the potential effects of the
to treat parasitic Varroa mite infestations (Sammataro et al., pesticides acetamiprid, thiacloprid, and oxalic acid on
2008). Oxalic acid is the most widely used acidic acaricide the gut microbiota associated with honeybees. These
(Nanetti et al., 2003; Brodschneider et al., 2019) and although compounds are still used on a global scale (Adjlane et al.,
its mode of action remains unknown, it is presumed to be 2016; Amulen et al., 2017; US Environmental Protection
safe for bees as it is naturally found in honey (Bogdanov Agency Office of Pesticide, 2020; Begna and Jung, 2021),
et al., 2002; Moosbeckhofer et al., 2003). Honeybees tolerate although thiacloprid was recently withdrawn from approval
treatment concentrations of 3.5% (Rademacher and Harz, in Europe (European Commission, 2020; European Food
2006), but exposure to higher concentrations can increase Safety Authority (EFSA), 2020). Oxalic acid is a widely used
mortality and induce behavioral changes, such as reduced treatment against Varroa (Rademacher et al., 2017). These
nursing efforts or general inactivity (Schneider et al., 2012; pesticides have been deemed safe for bees, but insight into
Rademacher et al., 2017). Oxalic acid treatment may also alter bee how they may affect honeybee gut bacterial communities is
physiology, reduce pH in the digestive tract and the hemolymph necessary to obtain a more complete risk assessment. We
(Rademacher et al., 2017), and create permanent lesions in the first performed in vitro assays to elucidate the potential direct
digestive tract (Martín-Hernández et al., 2007), like necrotic cells negative effects of these compounds on main core bacterial
(Gregorc and Smodiš Škerl, 2007). members of the honeybee gut microbiome (Figure 1C).

Frontiers in Microbiology | [Link] 2 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 1 | The honeybee gut microbiota and experimental setup. (A) The honeybee gut is compartmentalized and includes the crop (purple), midgut (dark blue),
and hindgut (soft blue), which is further divided into the ileum and rectum (created and reproduced with permission from [Link]). (B) The relative abundance
of core microbial genera in the honeybee gut microbiota (based on our dataset). (C) Overview of the in vitro experiment. Lawns of eight core bacteria were used to
test the effects of three pesticides inoculated on 1 cm filter paper discs (structures drawn in ChemDraw). Inhibition was scored as the presence/absence of a distinct
zone of inhibition forming around the disc. (D) Schematic of the in vivo experiment. For each of the three colonies, 20 sub-colonies were established each with 25
bees exposed to either of the three pesticides dissolved in sugar water or sugar water alone (control). After 7 days of exposure, three bees per sub-colony were
randomly picked, their guts dissected and extracted, and the community composition established using amplicon sequencing.

Subsequently, we performed an in vivo experiment to investigate and Frischella strains were maintained on Columbia Blood
if the indications from the in vitro experiment translated to agar (CBA) +5% blood medium. The former three genera
community effects on the microbiome and changes in honeybee were kept at 5% CO2 , while the latter was maintained
health (Figure 1D). under anaerobic conditions. All strains were kept at 34◦ C (cf.
Kešnerová et al., 2016, 2017).
For standardized growth, five biological replicates of each
MATERIALS AND METHODS strain were grown in liquid media, de Man, Rogosa and
Sharpe (MRS) for Lactobacillus and Bifidobacterium strains,
In vitro Assay of Gut Bacteria Columbia Blood (CB) for Gilliamella, Snodgrassella, Bartonella,
Susceptibility to Pesticides and Frischella strains), and once OD600 reached 0.5 (after 1 day),
To test for effects of the acetamiprid, thiacloprid, and oxalic 100 µl of each culture was plated in solid media in two 90 mm
acid on the growth of honeybee gut microbes, we obtained petri dishes to generate a bacterial lawn. Then, adapting from
15 strains from eight core gut bacteria of the western Balouiri et al., 2016, once the inoculum dried on the plate, four
honeybee (A. mellifera) (Figure 1C; Moran et al., 2012; 5 mm sterilized paper filter discs were placed on the surface
Bonilla-Rosso and Engel, 2018; Ellegaard and Engel, 2019). of the plate. The paper disc received 5 µl of the compound
Specifically, we obtained two strains of Lactobacillus Firm-4 solution, and 5 µl of sterile water in the case of the controls. The
(codes ESL0291 and ESL0292), two strains of Lactobacillus Firm- concentrations chosen for each of these compounds were based
5 (codes ESL0183 and ESL0184), two strains of Lactobacillus on their solubility in water: acetamiprid (4.2 mg/ml), thiacloprid
kunkeei (ESL0216 and ESL0219), two strains of Bifidobacterium (0.1 mg/ml), and oxalic acid (95 mg/ml) to ensure that we tested
asteroides (ESL0199 and ESL0200), two strains of Gilliamella effects of maximum concentrations, although these are likely
apicola (ESL0171 and ESL0172), two strains of Snodgrassella to be far higher than those encountered in the field. This trial
alvi (ESL0145 and ESL0252), two strains of Bartonella apis revealed that the bacteria were only sensitive to oxalic acid,
(ESL0058 and ESL0059) and one strain of Frischella perrara and we consequently tested its effects at concentrations 0.095,
(ESL0215). Lactobacillus and Bifidobacterium strains were 0.95, 9.5 mg/ml, which are conceivably more common doses
maintained on de Man, Rogosa and Sharpe agar (MRSA) for microbes to be exposed to than the very high maximum
and cultured anaerobically; Gilliamella, Snodgrassella, Bartonella, dose (Charrière and Imdorf, 2002; Schneider et al., 2012). Each

Frontiers in Microbiology | [Link] 3 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

strain was grown either without (controls) or with (treatments) to be one of the most common concentration beekeepers use in
a compound present in replicates of five. Presence/absence of beehives (Charrière and Imdorf, 2002). Rademacher et al. (2017)
inhibition was scored 2 days after exposure (Figure 1C). used a concentration of 50 µg/bee, considering this to be more
representative of what was found in bee colonies (acknowledging
the probable accumulation within colonies). Therefore, we used
In vivo Assessment of Pesticide the second value and calculated the concentration in the same
Consumption on Gut Microbiota manner as for acetamiprid. Based on this, we provided bees access
Honeybee Collection and Sub-Colony Set-Up to sugar water with a concentration of 3.5 µg/ml of acetamiprid,
Honeybees (A. mellifera) from three different colonies were 0.05 µg/ml of thiacloprid, or 1750 µg/ml oxalic acid. Although
obtained from the Department of Agroecology, Aarhus concentrations used thus vary greatly between compounds, they
University, Research Centre Flakkebjerg in Slagelse, Denmark on were chosen to best reflect ecological relevance.
the 7th of September, 2020. The colonies were opened, and the After sub-colonies were setup, they were kept at room
queen was located. Once found, bees were sampled from frames temperature with dim light. The sub-colonies were checked daily
without the queen so that the colony would keep producing to count the number of dead bees and to check if sugar water
brood. Drones were avoided. Adult workers were picked from was consumed or not. We did not remove dead bees during
the bottom of the frame in an effort to avoid foragers and older the experiment to minimize disturbance and to reduce the risk
bees. This implies that we most likely predominantly sampled of escapees. The approximate volume of sugar water consumed
younger bees working within the hive, with more stable gut was recorded on days one through four, after which the volume
microbiomes than foragers (Jones et al., 2018). Although it would decreased such that it was not easily discernable and thus could
have been ideal to sample comparably aged bees across the three not be recorded. The Falcon tube was refilled as needed with each
nests, the consistent responses of the three colonies (see section treatment. The experiment lasted for 7 days, after which dead and
“Results”) suggest that differences in age are unlikely to have live bees were collected and frozen separately. The gut dissections
impacted our findings. The bees were then placed in Styrofoam following the experiment were performed only on the live bees.
boxes for transport to the University of Copenhagen. The three
queens from the colonies were half-sisters.
At the University of Copenhagen, bees were anesthetized with Gut Dissection, DNA Extraction, and 16S
CO2 so that they could be separated into experimental groups. rRNA Amplicon Sequencing
The Styrofoam boxes were put in plastic bags that were filled Gut dissections were performed in sterile conditions on up to
with CO2 and, once the bees were inactive, they were sorted three bees per sub-colony, for a total of 146 bees (ncontrol = 36,
into their respective experimental groups. Bees from each colony nacetemiprid = 39, nthiacloprid = 39, and noxalicacid = 32). The
were divided in five replicates (25 bees per replicate) for each dissection area was sterilized with 70% ethanol and the dissection
experimental group and housed in rectangular plastic boxes of materials with 96% ethanol and a flame, and the dissections were
18 × 11 × 6 cm, similar to previous work (Evans et al., 2009; performed in the presence of a Bunsen burner (cf. Carreck et al.,
Huang et al., 2014; Figure 1D). Each of the boxes had sixteen 2013; Engel et al., 2013). The bee cuticle was sterilized prior to
2 mm diameter holes in the lid for ventilation, and a larger dissection by soaking the bee in a 1% aqueous solution of bleach
hole was drilled in the lid to allow placing a 15 ml Falcon for 3 min and then rinsing it in sterile purified water for three
tube for feeding. The Falcon tube contained 12 ml 50% sucrose times 30 s (Binetruy et al., 2019). The bee gut was obtained by
diet without pesticides for controls and with one of the three gently pulling the sting out of the abdomen, and the whole gut
pesticides for treatment sub-colonies. Each Falcon tube had three came attached to it. The gut was placed in a screw top 2 ml
holes in the bottom from which the bees were able to feed (Huang Eppendorf tube with 50 µl sterile PBS and frozen at −20◦ C
et al., 2014; Figure 1D). for later DNA extraction. In addition to the experimental bees,
Compound concentrations in the sugar water provided to we included 12 bees from each colony that had been frozen
treatment sub-colonies were based on previous work. Thany et al. immediately after collection (termed field controls). These field
(2015) tested 0.1 µg acetamiprid/bee and El Hassani et al. (2008) controls turned out to differ in community composition from
tested 0.5 µg/bee. We decided to follow the former, because it our experimental controls, so we excluded them from the main
had the concentration that was closest to what we expect bees manuscript analyses, but include them in a set of Supplementary
to encounter in the environment (134 ppb, cf. Mullin et al., Text, Supplementary Figures 1–3, 5–7, and Supplementary
2010). For the calculations, the onetime dose was multiplied Tables 1, 2, 9.
by the number of bees per sub-colony (25) and the number of For DNA extractions, we used the Qiagen Blood and Tissue
days that the experiment lasted (7). For thiacloprid, Mullin et al. kit (Qiagen, Hilden, Germany), following the manufacturer’s
(2010) detected traces of thiacloprid in 2% of pollen samples, protocol, but modified based on Engel et al. (2013). For the tissue
at levels of 7.8 PPB. Zaworra et al. (2019) registered an IC50 lysis, we used sterile 7 mm stainless beads and added 250 µl of
(inhibitory concentration for half the replicates) of 0.19 µg. Tison 0.1 mm sterile crystal beads, running the bead beater at 30 Hz
et al. (2017) found that concentrations of 0.5–50 µg/ml did not twice for 1 and 2 min, respectively. A total of 180 µl ATL buffer
increase mortality, but the highest did alter memory, we therefore and 20 µl proteinase K were added to each sample, vortexed and
used the lowest concentration. Schneider et al. (2012) used a incubated at 56◦ C for 3 h on a rotor. After incubation, 4 µl RNase
3.5% oxalic acid treatment (3500 µg/ml), which they reported A was added and the sample was incubated for 15–20 min. The

Frontiers in Microbiology | [Link] 4 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

remainder of the protocol was as per the manufacture’s protocol. met assumptions of normality and homoscedasticity. In both
The elution volume was 100 µl of buffer AE, and it was passed models, subsequent model reduction and ANOVA comparisons
through the column twice in a joint elute. were performed to address the significance of the fixed effects on
Diagnostic PCR was conducted to confirm the presence consumption and volume consumed.
of sufficient bacterial DNA in 143 samples. We used primers For the 16S rRNA gene amplicon sequencing analysis, we
for the V4 region of the 16S rRNA gene (0 V4.SA504: 50 -AAT used the dada2 pipeline (v.1.12.1; Callahan et al., 2016) and
GATACGGCGACCACCGAGATCTACACCTGCGTGTTATGG performed downstream analysis with default parameters with
TAATTGTGTGCCAGCMGCCGCGGTAA-30 and 0 V4.SB711: the following adjustments: we set the truncLen parameter in
50 -CAAGCAGAAGACGGCATACGAGATTCAGCGTTAGTCA filterAndTrim to c(240,220), trimleft (6) and maxEE to c(2,3).
GTCAGCCGGACTACHVGGGTWTCTAAT-30 ). PCR condi- We obtained 610 amplicon sequence variants (ASVs) after quality
tions were 94◦ C for 4 min, 35 cycles of 94◦ C for 30 s, 56◦ C filtering and removing chimeric reads. The merged sequences
for 30 s, and 72◦ C for 30 s, and a final extension step of 74◦ C were assigned to taxonomic ranks using the assignTaxonomy
for 4 min. Amplifications were confirmed on a 2% agarose gel. function in the dada2 package in R using the SILVA database
Three negative controls without guts and a positive control with release 138.1 (Quast et al., 2013). Since negative controls included
a cellular mock community standard (Zymobiomics, Nordic core honeybee gut microbiome taxa, an abundance threshold
BioSite ApS, Copenhagen) were included in the DNA extraction, was chosen to include genera in our study, while avoiding
and three additional negative control were added during library contaminants (cf. Kešnerová et al., 2017; Raymann et al., 2017).
preparation and sequencing. DNA was sent to the University Only genera with an abundance >0.08% across the whole dataset
of Michigan’s Microbiome Core1 for paired-end 250 bp 16S were retained for further analysis, resulting in 15 genera and
rRNA Illumina MiSeq amplicon sequencing with 0 V4.SA504 and 203 ASVs. By removing non-abundant genera, we reduce the
0 V4.SB711 primers. presence of putative contaminants, while keeping 98.7% of the
total number of original reads, including genera that are part of
Statistical Analyses the honeybee gut core microbiota and present in low levels in
All statistical analyses were performed in R studio v. 4.0.3 R Core negative controls. The cellular mock community validated the
Team (2020). detection of all eight expected taxa and allowed for an estimation
For the bacterial inhibition assay, a generalized linear mixed- of the random error in our community abundances, averaging
effects model (GLMM) with family binomial was used to address 0.07% of the abundance of any given taxon.
the effect of oxalic acid concentration on bacterial inhibition. Richness was estimated as the number of ASVs using the
Concentration was included as a fixed effect and bacterial function estimate_richness in the R package phyloseq (v.1.34.0;
strains as a random slope. Model reduction with subsequent McMurdie and Holmes, 2013). To test for statistical differences
ANOVA comparison was performed to address the effect of in richness between treatments and colony origin, a LM was used,
concentration on inhibition. where the log of the observed number of ASVs per treatment was
To analyze the mortality data (Supplementary Table 3), we the response variable. Both treatment and colony were included
used Cox proportional hazard regression models and likelihood as fixed effects, and the model included their interaction. Pairwise
ratio (LR) test statistics, employing the R package survival and comparisons between treatments were preformed using Tukey
the function coxph (version 3.2-7; Therneau, 2021). Treatment honestly significant difference (HSD) test, after confirming that
was included as a fixed effect, where treatments were compared the model assumptions were met using [Link] (v.0.9-38;
to the control, and colony was included as a stratified fixed Royston, 1982) and bptest (Breusch and Pagan, 1979) from
effect. Additionally, we created models to assess colony-level the lmtest package (v.0.9-38; Zeileis and Hothorn, 2002). Beta
responses to treatment with treatment a fixed effect. For all diversity metrics were calculated based on Bray–Curtis and
models, the proportional hazard assumptions and the Cox-Snell Jaccard distances, using the vegdist function from the vegan
residuals were tested according to Mills (2012); our models met package in R (v.2.5-7; Oksanen et al., 2018). Additionally, Unifrac
both assumptions. distances were also calculated using the UniFrac function. The
We used a GLMM with family binomial to assess if the differences between controls and each treatment were separately
type of pesticide affected whether or not the bees consumed tested using multivariate PERMANOVAs (adonis function in
sugar water. Treatment, colony, and day were inserted as fixed the vegan package), together with a colony fixed effect and the
effects and sub-colony as a random slope. We then assessed interaction between colony and treatment.
the average volume of pesticide or control solution consumed ALDEx2 (v.1.22.0; Fernandes et al., 2013) was used to
by each bee over days one through four by dividing the daily analyze differentially abundant genera between controls and
consumption of the sub-colony by the number of bees alive in each pesticide treatment group. The same was done to detect
the sub-colony that day. We then used a general linear model differentially abundant ASVs. In each case, we controlled for
(LM) to test the fixed effects of treatment, day, and colony on the colony by adding it as a fixed effect.
volume of control or pesticide solution consumed. The minimal Lastly, as previous work has reported syntrophic relationships
model was determined with backwards model reduction and among honeybee gut symbionts (e.g., Kwong and Moran, 2016;
Kešnerová et al., 2017), we performed network analyses using
1
[Link] Bray–Curtis distance in the plot_net function in phyloseq for
microbiome-core each treatment separately to test if associations in the control

Frontiers in Microbiology | [Link] 5 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

group were maintained with different pesticide treatments. We indicated significant effects of treatment (overall model LR test:
performed the analysis across all colonies to secure a sample LR = 23.6, df = 3, p < 0.0001). The hazard ratios (HR) were
size that allowed for a robust analysis. Distances below 0.5 are calculated for each treatment, a hazard ratio >1 means that the
reported as associations. treatment increased mortality compared to the control, while
values <1 indicate that treatment decreased mortality compared
to controls. Exposure to thiacloprid significantly increased bee
RESULTS mortality (Wald statistic: z = −1.04, p = 0.0007, HR = 1.357;
Supplementary Figure 4), as did oxalic acid (z = 3.36, p = 0.0332,
Only Oxalic Acid Inhibited Bacterial HR = 1.217; Supplementary Figure 4), but this was not the case
Growth in vitro for acetamiprid (z = 2.13, p = 0.2974, HR = 0.905) (Table 1
We evaluated the proportion of plates with honeybee gut bacteria and Supplementary Figure 4). For the models on individual
that were inhibited across the specified concentrations of three colonies, treatment had an effect in colony 1 (LR = 32.58, df = 3,
pesticides and found that all strains were resistant to acetamiprid p < 0.0001; Figure 3A), colony 2 (LR = 31.43, df = 3, p < 0.0001;
and thiacloprid, but sensitive to oxalic acid (Figure 2). B. apis Figure 3B), and colony 3 (LR = 81.45, df = 3, p < 0.0001;
and Lactobacillus Firm-5 were only inhibited at the highest Figure 3C). In colony 1, all three pesticides increased mortality,
concentration (95 mg/ml), which is unlikely to be encountered in colony 2, acetamiprid and thiacloprid had less mortality than
in the field. Bacteria susceptible to concentrations relevant controls, and in colony 3, acetamiprid and oxalic acid reduced
to acaricide application and bioaccumulation included S. alvi, mortality and thiacloprid increased mortality (Figure 3 and
F. perrara and B. asteroides, which were inhibited at 9.5 mg/ml, Table 1).
and Lactobacillus Firm-4, L. kunkeei, and G. apicola, which were Treatment significantly impacted the presence/absence
most sensitive and inhibited already at 0.95 mg/ml; although, of consumption (overall GLMM: F 3 ,415 = 8.07; treatment:
the latter for only one of the five test plates (Figure 2). The χ2 = 38.19, df = 3, p < 0.0001) (Supplementary Table 4), while
concentration of oxalic acid had a significant effect on inhibition day and colony did not (day: χ2 = 0.12, df = 1, p = 0.7307; colony:
(overall GLMM: F 2 ,220 = 25.13; concentration: χ2 = 151.8, df = 3, χ2 = 1.62, df = 1, p = 0.2035). Across all experimental days, the
p < 0.0001). total number of events (consumption or no consumption) was
105. Of these accounts, consumption was lowest for oxalic acid
Pesticide Effects on Honeybee Mortality (78 accounts), followed by acetamiprid (96 accounts), controls
and Liquid Consumption (101 accounts), and thiacloprid (103 accounts). Treatment, day,
Mortality was high across treatments, including in controls and colony all significantly impacted consumption volume on the
(Figure 3). The Cox proportional hazard regression model test first 4 days of treatment (overall LM: F 6 ,177 = 12.51, p < 0.0001).

FIGURE 2 | Honeybee bacteria isolates inhibited by oxalic acid. The percentage of cultures that were inhibited after exposure to each of the four concentrations of
oxalic acid (n = 5 for all cultures, except for Lactobacillus Firm-4, strain 291, where n = 4 for 95 mg/ml and n = 3 for 9.5 and 0.95 mg/ml).

Frontiers in Microbiology | [Link] 6 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 3 | Mortality across colonies and treatment in the in vivo caged bee experiment. (A) For colony 1, acetamiprid, thiacloprid, and oxalic acid reduced survival
when compared to the control. (B) For colony 2, acetamiprid and thiacloprid increased survival when compared to the control. (C) For colony 3, both acetamiprid
and oxalic acid increased survival, while oxalic acid increased mortality. *p < 0.05, ***p < 0.001, and ****p < 0.0001 (Table 1).

Treatment had a significant effect (LM: F 3 ,177 = 5.16, p = 0.0019), the non-significant interaction between treatment and colony
where when compared to the control, thiacloprid significantly (F 6 ,134 = 1.297; p = 0.2629).
increased consumption (p = 0.0464), while there was no Beta diversity differences were only to a small extent driven
significant difference for acetamiprid (p = 0.5785) or oxalic acid by pesticide treatment (Figure 4C) but affected more by colony
(p = 0.0773). Consumption significantly reduced over days one (Figure 4D). PERMANOVAs of Bray–Curtis distances indicated
through four (LM: F 1 ,177 = 38.81, p < 0.0001; Supplementary that 29.0% of the variation was explained by colony, 6.2% by
Tables 5, 6), and colony had a significant effect (F 2 ,177 = 7.65, pesticide treatment, and 9.3% by their interaction. Similarly,
p = 0.0006; Supplementary Tables 5, 6). Jaccard distances indicated that 19.7% of the variation is
explained by colony, 5.3% by treatment, and 8.7% by their
Pesticides Affect Richness and Beta interaction, and Unifrac distances gave that 38.8% of the variation
Diversity of the Honeybee Gut Bacterial is explained by colony, 8.6% by treatment, and 8.6% by their
interaction (Supplementary Figure 6). For all diversity metrics,
Community compared to controls, oxalic acid harbored the most distinct
We obtained a total of 4,710,340 clean MiSeq reads of the
microbiome (Bray–Curtis: F 1 ,66 = 8.444; p < 0.0001, Jaccard:
16S rRNA gene (average 27,546/sample) and rarefaction
F 1 ,66 = 5.557; p < 0.0001, Unifrac: F 1 ,66 = 17.6; p < 0.0001),
curves supported sufficient coverage to pursue diversity
followed by acetamiprid (Bray–Curtis: F 1 ,69 = 2.641; p = 0.0305,
analyses (Supplementary Figure 5). When compared to the
Jaccard: F 1 ,69 = 2.151; p = 0.0278, Unifrac: F 1 ,69 = 0.769;
control group, ASV richness was significantly affected by
p = 0.473) and thiacloprid (Bray–Curtis: F 1 ,70 = 2.065; p = 0.0557,
the consumption of oxalic acid, which reduced richness by
Jaccard: F 1 ,70 = 1.844; p = 0.0370, Unifrac: F 1 , 70 = 0.348;
25.7% (p < 0.0001). Treatment had an effect on ASV richness
p = 0.827).
(Figure 4A and Supplementary Table 7; LM: F 3 ,134 = 34.31), but
acetamiprid (p = 0.0522) and thiacloprid treatments (p = 0.2602)
did not differ from controls. There were also significant Treatment Affects Microbiome
differences in microbial richness between colonies (F 2 ,134 = 45.07;
p < 0.0001; Figure 4B and Supplementary Table 8), but the
Composition, Particularly After Oxalic
effect of pesticides was the same on all colonies, as evident from Acid Exposure
Across the full dataset, the most abundant genera were
Snodgrassella, Lactobacillus, Gilliamella, Commensalibacter,
TABLE 1 | Effect of treatment on mortality within the three colonies compared to Frischella, Bombella, Bifidobacterium, and Bartonella (Figure 5
controls, based on Cox proportional hazards regressions. and Supplementary Table 9), accounting for 87.2% of all clean
Colony Treatment (compared Hazard ratio Wald statistic (z) p
sequence reads, and consistent with previous findings (Engel
to control) (HR) et al., 2012; Bonilla-Rosso and Engel, 2018; Ellegaard and Engel,
1 Acetamiprid 1.699 3.367 0.0007 2019). There was, however, extensive variation both between
Thiacloprid 1.971 4.353 <0.0001 colonies and, to a lesser extent, between bees from the same
Oxalic acid 2.236 5.201 <0.0001 colony (Figure 5). Hafnia–Obesumbacterium and Klebsiella, two
2 Acetamiprid 0.565 −3.465 0.0005 environmental bacteria that are opportunistically associated with
Thiacloprid 0.511 −3.972 <0.0001 honeybees, were abundant in some samples, with an overall
Oxalic acid 1.059 0.383 0.7019
average abundance of 3.9 and 9.7%, respectively. However, the
3 Acetamiprid 0.635 −2.466 0.0136
abundance of these opportunistic bacteria was not associated
Thiacloprid 2.349 5.604 <0.0001
Oxalic acid 0.631 −2.552 0.0107
with pesticide treatment (Figure 5 and Supplementary Table 9).
Their elevated levels, including in controls, may have been due
HR > 1 indicates that the treatment increased mortality compared to
the control, while HR < 1 means that the treatment decreased mortality to the presence of dead bees within the boxes, which could
compared to the control. have acted as reservoirs for the transfer of opportunistic and

Frontiers in Microbiology | [Link] 7 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 4 | Gut microbiome richness and beta diversity estimates of pesticide treated and control honeybees. (A) ASV richness across treatments. (B) ASV richness
across colonies and treatments. Boxplots represent the first and third quartiles of the number of ASVs observed per treatment, the horizontal line represents the
median, whiskers extend 1.5 interquartile ranges and dots represent outliers. In panel (B), the letters represent the statistical dissimilarities across groups according
to Tukey HSD test (p < 0.05). (C) Non-metric multidimensional scaling (NMDS) analysis of the samples based on Bray–Curtis distances; plot ordination based on
treatment. (D) Non-metric multidimensional scaling (NMDS) analysis of the samples based on Bray–Curtis distances; plot ordination based on colony. There is a
clear clustering of the samples in the ordination plot based on colonies, which is supported by the PERMANOVA test.

pathogenic bacteria to live bees. Future work should thus the most abundant Gilliamella ASV (t = 7.379, p < 0.0001), and
consider removal of dead bees to prevent effects on microbiomes three Bifidobacterium ASVs (t = 4.543, p = 0.0157; t = 4.210,
of focal individuals. p = 0.0251; and t = 4.067, p = 0.0348) were all relatively more
To better understand the community changes underlying abundant in oxalic acid treated bees than in controls (Figure 6B).
differences in alpha and beta diversity, we assessed the differential These changes in relative abundance of genera and ASVs reflect
abundance of genera (Figure 6A) and ASVs (Figure 6B) altered community composition. However, differences in the
between respective treatments and controls. We did not find absolute abundances would require quantification of the 16S
any differentially abundant genera nor ASVs between controls rRNA gene to estimate bacterial load, and should be considered
and acetamiprid or thiacloprid treatments. However, six genera in future studies.
and seven ASVs differed in relative abundance between controls
and oxalic acid treated bees. At the genus level, the relative
abundance of Bombella was significantly reduced after oxalic-
Pesticide Treatment Affects Network
acid treatment (ALDEx2 test: t = −6.054, p < 0.0001), while Relationships Between Honeybee Gut
Gilliamella (t = 8.989, p < 0.0001), Frischella (t = 4.030, Microbes
p = 0.0023), Lactobacillus (t = 7.986, p < 0.0001), Bifidobacterium To elucidate potential syntrophic relationships among the
(t = 6.912, p < 0.0001), and Snodgrassella (t = 3.942, p = 0.0033) identified genera in our dataset, and to assess whether these were
all increased in relative abundance in response to the oxalic affected by pesticide treatment, we performed network analyses
acid treatment (Figure 6A). At the ASV level, Bombella intestini for controls and each treatment, pooling data from all colonies
(t = −8.326, p < 0.0001) and L. kunkeei (t = −10.56, p < 0.0001; (Figure 7). Overall, we observed flexibility in ecological networks
Supplementary Figure 7) were negatively affected by oxalic acid. of pesticide-treated microbiomes, but some positive associations
Conversely, a Lactobacillus Firm-5 ASV (t = 7.505, p < 0.0001), were shared with controls (Figure 7). All positive interactions in

Frontiers in Microbiology | [Link] 8 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 5 | Relative abundances of the 15 most abundant bacterial genera across the full dataset. The top panel gives relative abundances by treatment, while the
bottom panel gives the relative abundances per honeybee gut, illustrating the abundant variation across samples within and between colonies.

controls were present in the acetamiprid treatment (Figure 7), Lactobacillus and inconsistent positive correlation with other
but here we also saw associations between Commensalibacter bacteria. Nevertheless, overall changes in interbacterial networks
and Lactobacillus, and between Snodgrassella and Frischella point toward the presence of either a flexible or a fragile ecological
and Bartonella. In oxalic acid-treated honeybees (Figure 7), network in the honeybee gut microbiome.
the association between Bombella and Snodgrassella present in
controls disappeared, likely due to Bombella sensitivity to this
DISCUSSION
pesticide. Additionally, Snodgrassella was linked to Lactobacillus,
which potentially fills Bombella’s niche of aerobic respiration. We examined the effect that three commonly used pesticides
Under thiacloprid treatment (Figure 7), it is noteworthy that have on microbial symbionts of honeybees and found evidence
we could not replicate the Bifidobacterium–Frischella association, for both in vitro and in vivo effects of oxalic acid exposure,
while a Lactobacillus–Snodgrassella association emerged. Finally, but no effects of acetamiprid or thiacloprid exposure in the
associations between Lactobacillus and other core members were in vitro assay and far less marked effects in vivo. It is apparent
surprisingly lacking in the controls, despite ample evidence that in vitro testing alone is insufficient to deduce effects on
of syntrophic relationships (Kwong and Moran, 2016). The honeybee gut microbiomes and ultimately honeybee health.
lack of association may be due to the variable niches and Although we were limited from quantifying the amount of liquid
interbacterial interactions of L. Firm 4, L. Firm 5, and L. kunkeei that bees consumed within sub-colonies late in the experiment,
(Bonilla-Rosso and Engel, 2018) that lead to variable levels of when fewer bees were present, all sub-colonies appeared to

Frontiers in Microbiology | [Link] 9 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 6 | Oxalic acid treatment changes microbial relative abundance in vivo. (A) Relative abundance of core microbes at the genus level between controls and
different pesticide treatments. (B) Relative abundance of differentially abundant ASVs between control and pesticide treatments. Asterisks indicate significant
changes between control and oxalic acid treatments assessed in two independent ALDEx2 bivariate models with pesticide treatment and colony as fixed effects
(*p < 0.05, **p < 0.01, ***p < 0.001). Acetamiprid and thiacloprid are not significantly different from controls. Boxplots represent the first and third quartiles of the
relative abundance, the horizontal line represents the median, whiskers extend 1.5 interquartile ranges and dots represent outliers.

consume sugar water, indicating exposure. Oxalic acid was the known to change as the bees transition to the winter season,
only pesticide for which we observed reduced number of days but the impact of this change on bee health remains unclear
with consumption, consistent with previous suggestions that (Ludvigsen et al., 2015; Bleau et al., 2020; Kešnerová et al., 2020).
palatability may cause oxalic acid avoidance (Nanetti et al., 2015). Furthermore, although we conceivably avoided foragers (older
Its consumption is nevertheless corroborated by the observed bees) by sampling from the bottom of the hive frames, we did
effects on gut microbial communities, with reduced microbial not strictly age control and may thus have included older bees
richness, reduced abundance of key microbiome taxa, and with a higher risk of dying during the experiment. However, two
altered interbacterial relationships. We did not see consistently other factors have conceivably played a more important role in
higher mortality induced by the presence of either of the three governing mortality. First, honeybees are sensitive to stress (Klein
compounds compared to controls, so longer-term exposures may et al., 2017), and transport and the use of CO2 to anesthetize bees
be needed to elucidate any such effects. may have negative impacts. Secondly, to minimize disturbance
It should be noted that we observed very high mortality and the risk of escapees we did not remove dead bees during the
across all experimental groups, suggesting that caution should be experiment, and it is likely that this has increased risks of cross
taken when interpreting beyond laboratory experimental group infections with opportunistic pathogens from dead to live bees
comparisons. This high mortality was likely due to a combination (see results on the elevated levels of opportunistic pathogens in
of factors. The low alpha diversity in field control samples (see control bees). Although such high mortality is not optimal, the
Supplementary Text) may suggest that the colonies were not in consistent patterns observed in the microbiome analyses, and the
optimal health condition at the time of sampling (September). impacts we see on specific core gut microbes, warrant meaningful
However, the alpha diversity and microbiome composition is comparisons and reflect biologically relevant effects.

Frontiers in Microbiology | [Link] 10 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

FIGURE 7 | Positive relationships in honeybee gut bacterial genera by experimental treatment. Lines between bacterial genera indicate positive relationships from
individual network analyses. The genera are represented based on their oxygen niche (y-axis) and their position in the gut (x-axis), with the width of text boxes
indicative of their gut placement. The oxygen niche and position in the gut is based on current literature (Anderson et al., 2013; Kwong and Moran, 2016; Butler
et al., 2013; Kešnerová et al., 2017; Bonilla-Rosso and Engel, 2018).

Do in vitro Observations Predict in vivo in vitro. This, and its comparably wider use as an approved
Effects? pesticide, lead us to focus our discussion on potential impacts of
We did not observe in vitro inhibition by thiacloprid or oxalic acid exposure.
acetamiprid of any of the tested bacteria, and the communities The outcome of pesticide treatment on gut bacterial
treated with these pesticides in vivo were only slightly altered abundances in vivo should be the combined effect of bacterial
compared to controls. Oxalic acid inhibition in vitro was sensitivity, direct exposure and interbacterial dependencies.
consistent with greater alteration of the gut microbiomes in vivo, Bacteria that are sensitive in vitro to 0.95–9.5 mg/ml, conceivably
but the bacteria affected in vivo differed from those affected comparable to potential exposures in the field, appear to increase

Frontiers in Microbiology | [Link] 11 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

in relative abundance in vivo after treatment with oxalic acid Other studies have found oxalic acid inhibition in vitro
compared to controls. This is likely due to exposure differences of both honeybee microbes (Diaz et al., 2019) and plant
in vitro and within bee guts, and illustrates the limitations that pathogens (Kwak et al., 2016), and oxalic acid is present
plating experiments have for evaluating impacts of anthropogenic in honey and has been suggested to be responsible for its
compounds on host-associated microbes. Exposure is inevitable antimicrobial properties (Bogdanov et al., 2002; Nozal et al.,
in vitro, but this is not necessarily so in vivo. For example, biofilm 2003). While most strains appear resistant to oxalic acid in low
formation helps protect from environmental changes, including concentrations, oxalic acid can persist in colonies for multiple
chemical or antibiotic exposure (Singh et al., 2017). Sensitivity weeks (Rademacher et al., 2017). If accumulation occurs after
to oxalic acid in the in vitro experiment would predict effects multiple oxalic acid treatments, the pesticide may have a stronger
on e.g., G. apicola, F. perrara, B. asteroides, and S. alvi in vivo. effect on honeybee microbes. Regular oxalic acid application to
However, within bees, G. apicola interacts in a biofilm with S. alvi colonies may also select for resistance in opportunistic bacteria,
and engages in cross-feeding interactions (Engel et al., 2012; Varroa, and honeybee mutualists and commensals. This may
Kešnerová et al., 2017). Such close syntrophic interactions within impair the function constitutive levels of oxalic acid have in
multispecies biofilms are likely to reduce oxalic acid exposure or honeybee colonies, ultimately requiring treatment with higher
buffer its effects. concentrations with impacts on colony health and production
Another potential explanation for the reduced impact on (Adjlane et al., 2016). As Varroa has a larger impact on honeybee
most core bacteria is their location in the intestinal tract. health on the short-term than oxalic acid treatment, we advise
Apart from L. kunkeei and B. apis, the two microbes depleted exploring genetic resistance to Varroa in honeybees, as it may be
by oxalic acid in vivo, all the tested bacteria reside in the a better long-term solution against infection than the application
hindgut (Anderson et al., 2013; Moran, 2015; Kwong and of oxalic acid (Conlon et al., 2019).
Moran, 2016; Powell et al., 2018). The hindgut receives pre-
digested material from the midgut, and oxalic acid may thus
at least partly have been digested before reaching the hindgut Implications of Pesticide Treatment on
(Engel and Moran, 2013). A combination of decreasing oxalic Honeybee Health
acid concentrations along the bee gut and biofilm formation Oxalic acid impact on the honeybee gut microbiome could be
in the hindgut seems plausible to explain the effects on the direct or indirect. Indirect effects could be mediated by necrotic
honeybee gut microbes, suggesting that the most profound cell death in the midgut (Gregorc and Smodiš Škerl, 2007;
implications of exposure for colony health is conceivably effects Papežíková et al., 2017), lesions (Martín-Hernández et al., 2007),
on crop bacteria. In this context, we should note that increases or pH changes (Rademacher et al., 2017). Since the affected
in relative abundances are likely driven mainly by the loss of microbes are present in the crop, rather than the midgut or
other bacteria, increasing the relative abundance of taxa in the hindgut, and the temporal scope of our experiment is restrictive,
microbiome that are unaffected by pesticide treatment. In order we can with some confidence rule out necrotic cell death and
to determine if treatment affects the absolute abundance of lesions for our experiment. This leaves pH changes as the most
bacteria, and to explore potential relationships between bacterial likely effector, which could have secondary effects on nutrient
abundances and alpha diversity, bacterial load would need to be digestion and absorption.
quantified. Our network analyses replicate previously published
data on interbacterial relationships (Kwong and Moran,
2016; Kešnerová et al., 2017; Ellegaard and Engel, 2019).
Implications of Oxalic Acid Bacterial Encouragingly, they revealed that interbacterial relationships
Inhibition on Colony Dynamics are potentially flexible, at least on the short-term, with the
In vitro inhibition of core honeybee microbes and in vivo caveat that our relationships are inferred by correlation,
depletion of crop members are potentially concerning for and hence may not reflect actual syntrophic or other
honeybee colony health. Bacteria move within colonies by types of relationships. Another interpretation is they are
trophallaxis between bees (Martinson et al., 2012), which includes relatively fragile to environmental stressors. Crop bacteria
the exchange of crop contents, via bees picking up bacteria are responsible for a portion of aerobic metabolism in
from hive material (Kwong and Moran, 2016), and through the bee gut, therefore creating an anaerobic environment
consumption of hindgut exudates (Powell et al., 2014). Oxalic in lower portions of the gut. However, even after severe
acid treatment by trickling depends on the redistribution depletion of crop bacteria with oxalic acid, it seems that
of the pesticide by the bees themselves (Schneider et al., interbacterial relationships reassemble in such a way that
2012; Rademacher et al., 2017). While bacteria in most gut Snodgrassella is now responsible for oxygen metabolism and
compartments may be somewhat protected from oxalic acid, we depletion, leading to an association between Snodgrassella and
do observe a significant overall negative effect on microbiome fermentative Lactobacilli (Kešnerová et al., 2017). Snodgrassella’s
richness, estimated as the number of ASVs bee guts contain. association with the facultative aerobe Gilliamella remains
The regeneration of the microbiome after local extinctions intact under this pesticide treatment, as do Gilliamella–
in individual bees, or during inoculation of young bees, also Bifidobacterium and Bifidobacterium–Frischella associations,
depends on bacteria moving from hive to bee and between bees, all of which exist in the lower portions of the honeybee gut.
implying that colony-level impacts from treatment with oxalic The main impacts on interbacterial relationships are therefore
acid are likely. captured by the disappearance of Bombella and L. kunkeei,

Frontiers in Microbiology | [Link] 12 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

which affect other relationships, such as those involving after one week of exposure; however, our experiment tested lower
Bartonella. concentrations (0.05 mg/L), so this may explain the lower effect of
Lactobacillus kunkeei is one of the most abundant bacteria thiacloprid. The bacterial communities recovered by day thirteen
in the honeybee crop, and is also found in nectar, pollen, in Liu et al. (2020), likely due to anal trophallaxis between colony
and honey (Anderson et al., 2013; Corby-Harris et al., 2014a; members compensating for the loss of honeybee gut members
Bonilla-Rosso and Engel, 2018; Kwong and Moran, 2016). (Kwong and Moran, 2016) and could also be buffering the effect
In vitro, we see that this species is relatively more resistant, of thiacloprid in our experiment.
potentially due to exposure of oxalic acid in honey (Bogdanov
et al., 2002; Nozal et al., 2003). Lactobacillus presence increases
brood and honey production (Alberoni et al., 2017), improves DATA AVAILABILITY STATEMENT
the honeybee immune response (Maruščáková et al., 2020),
protects against foulbrood diseases (Vásquez et al., 2012) The 16S rRNA datasets generated in this study can be found in the
and some insect-associated strains can metabolize insecticides SRA archive in GenBank under the BioProject PRJNA732842.
(De Almeida et al., 2017; Daisley et al., 2018). The strain-
level turnover between Lactobacillus strains we see in our
experiment may have implications for colony health, as it is an AUTHOR CONTRIBUTIONS
important factor for honeybee microbiome function (Ellegaard
et al., 2020). For example, L. kunkeei presence decreases larval AC-M, AJ, and MP conceived and designed the experiment.
mortality during Paenibacillus larvae infection and decreases AC-M and PK collected the bees. AC-M performed the laboratory
Nosema infection prevalence in adults (Arredondo et al., 2018). work with assistance from VS. JR-H, AC-M, and VS analyzed
Encouragingly, L. kunkeei can regrow surprisingly quickly after the community data while AC-M and VS analyzed in vitro,
oxalic acid has been removed from the colony environment mortality, and consumption data. AC-M created the first draft
(Supplementary Figure 6). The other affected genus in our of the manuscript and figures with VS, MP, JR-H, and AJ
study, Bombella, has very specific niches within bees, including further editing and providing feedback. All authors agreed on the
in nurse hypopharyngeal glands from which larvae are feed, the final manuscript.
crop of nurse bees, and in royal jelly. Its location in the crop
of nurses and in royal jelly implies potentially relevant roles in
larvae and queen development (Corby-Harris et al., 2014b; Tarpy FUNDING
et al., 2015). Consistent with this assertion, it comprises a large
fraction of the queen gut microbiome, possibly playing a role in This research was supported by a Ph.D. fellowship and
queen nutrition and in modulating queen fertility, fecundity, and research stipend from the Department of Biology, University
longevity (Anderson et al., 2018). This may involve protection of Copenhagen to VS and a European Research Council
from pathogens, as indicated by Bombella apis supplementation Consolidator Grant (771349) to MP.
in colonies significantly reducing Nosema prevalence (Corby-
Harris et al., 2016). Colony-level effects of oxalic acid treatment
on Bombella must thus be assessed, as our results could have ACKNOWLEDGMENTS
implications for honeybee immunity, longevity and fecundity.
Acetamiprid and thiacloprid treatment did not inhibit We thank Philipp Engel for providing bacterial strains, Kasun H.
bacterial growth in vitro or impact the richness of honeybee Bodawatta for assistance in preparing samples for sequencing,
gut microbiota compared to controls in vivo. Acetamiprid and Suzanne Schmidt for ChemDraw structures included
and thiacloprid treatments generate little change in the gut in Figure 1. Further, we thank the Social and Symbiotic
communities of treated bees, maintaining richness levels Evolution Group for discussion of the project and for comments
and core member relative diversity observed in the lab on the manuscript.
controls. Acetamiprid has not been investigated for its potential
detrimental impact on honeybee gut microbiome, but chronic
exposure of its fellow neonicotinoids, like thiamethoxam or SUPPLEMENTARY MATERIAL
imidacloprid, greatly alter intestinal communities of bees (Rouzé
et al., 2019). Our in vivo findings contrast recent work by The Supplementary Material for this article can be found
Liu et al. (2020) who described gut bacterial dysbiosis in online at: [Link]
honeybees exposed to thiacloprid in a dose dependent manner 2021.717990/full#supplementary-material

REFERENCES Alberoni, D., Baffoni, L., Gaggìa, F., Ryan, P. M., Murphy, K., Ross, P. R., et al.
(2017). Impact of beneficial bacteria supplementation on the gut microbiota,
Adjlane, N., Tarek, E. O., and Haddad, N. (2016). Evaluation of oxalic colony development and productivity of Apis mellifera L. Benef. Microbes 9,
acid treatments against the mite Varroa destructor and secondary 269–278. doi: 10.3920/BM2017.0061
effects on honeybees Apis mellifera. J. Arthropod Borne Dis. 10, Amulen, D. R., Spanoghe, P., Houbraken, M., Tamale, A., De Graaf, D. C., Cross,
501–509. P., et al. (2017). Environmental contaminants of honeybee products in Uganda

Frontiers in Microbiology | [Link] 13 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

detected using LC-MS/MS and GC-ECD. PLoS One 12:e0178546. doi: 10.1371/ Charreton, M., Decourtye, A., Henry, M., Rodet, G., Sandoz, J. C., Charnet, P.,
[Link].0178546 et al. (2015). A locomotor deficit induced by sublethal doses of pyrethroid
Anderson, K. E., Ricigliano, V. A., Mott, B. M., Copeland, D. C., Floyd, A. S., and and neonicotinoid insecticides in the honeybee Apis mellifera. PLoS One
Maes, P. (2018). The queen’s gut refines with age: longevity phenotypes in a 10:e0144879. doi: 10.1371/[Link].0144879
social insect model. Microbiome 6:108. doi: 10.1186/s40168-018-0489-1 Charrière, J. D., and Imdorf, A. (2002). Oxalic acid treatment by trickling against
Anderson, K. E., Sheehan, T. H., Mott, B. M., Maes, P., Snyder, L., Schwan, M. R., Varroa destructor: recommendations for use in central Europe and under
et al. (2013). Microbial ecology of the hive and pollination landscape: bacterial temperate climate conditions. Bee World 83, 51–60. doi: 10.1080/0005772X.
associates from floral nectar, the alimentary tract and stored food of honey bees 2002.11099541
(Apis mellifera). PLoS One 8:e83125. doi: 10.1371/[Link].0083125 Conlon, B. H., Aurori, A., Giurgiu, A. I., Kefuss, J., Dezmirean, D. S., Moritz,
Arredondo, D., Castelli, L., Porrini, M. P., Garrido, P. M., Eguaras, M. J., Zunino, P., R. F. A., et al. (2019). A gene for resistance to the Varroa mite (Acari) in honey
et al. (2018). Lactobacillus kunkeei strains decreased the infection by honey bee bee (Apis mellifera) pupae. Mol. Ecol. 28, 2958–2966. doi: 10.1111/mec.15080
pathogens Paenibacillus larvae and Nosema ceranae. Benef. Microb. 9, 279–290. Corby-Harris, V., Maes, P., and Anderson, K. E. (2014a). The bacterial
doi: 10.3920/BM2017.0075 communities associated with honey bee (Apis mellifera) foragers. PLoS One
Balouiri, M., Sadiki, M., and Ibnsouda, S. K. (2016). Methods for in vitro evaluating 9:e95056. doi: 10.1371/[Link].0095056
antimicrobial activity: a review. J. Pharm. Anal. 6, 71–79. doi: 10.1016/[Link]. Corby-Harris, V., Snyder, L., Meador, C. A. D., Naldo, R., Mott, B., and Anderson,
2015.11.005 K. E. (2016). Parasaccharibacter apium, gen. Nov., sp. Nov., Improves Honey
Begna, T., and Jung, C. (2021). Effects of sequential exposures of sub-lethal doses Bee (Hymenoptera: Apidae) resistance to Nosema. J. Econ. Entomol. 109, 537–
of amitraz and thiacloprid on learning and memory of honey bee foragers, Apis 543. doi: 10.1093/jee/tow012
mellifera. J. Asia Pac. Entomol. 24, 77–83. doi: 10.1016/[Link].2021.03.012 Corby-Harris, V., Snyder, L. A., Schwan, M. R., Maes, P., Mcfrederick, Q. S., and
Binetruy, F., Dupraz, M., Buysse, M., and Duron, O. (2019). Surface sterilization Anderson, E. (2014b). Origin and Effect of Alpha 22 Acetobacteraceae in Honey
methods impact measures of internal microbial diversity in ticks. Parasit. Bee Larvae. Appl. Environ. Microbiol. 80, 7460–7472. doi: 10.1128/AEM.02043-
Vectors 12, 1–10. doi: 10.1186/s13071-019-3517-5 14
Bleau, N., Bouslama, S., Giovenazzo, P., and Derome, N. (2020). Dynamics of the Dacher, M., Lagarrigue, A., and Gauthier, M. (2005). Antennal tactile learning
honeybee (Apis mellifera) gut microbiota throughout the overwintering period in the honeybee: effect of nicotinic antagonists on memory dynamics.
in Canada. Microorganisms 8, 1–11. doi: 10.3390/microorganisms8081146 Neuroscience 130, 37–50. doi: 10.1016/[Link].2004.09.006
Blot, N., Veillat, L., Rouzé, R., and Delatte, H. (2019). Glyphosate, but not its Daisley, B. A., Trinder, M., McDowell, T. W., Collins, S. L., Sumarah, M. W.,
metabolite AMPA, alters the honeybee gut microbiota. PLoS One 14:e0215466. and Reid, G. (2018). Microbiota-mediated modulation of organophosphate
doi: 10.1371/[Link].0215466 insecticide toxicity by species-dependent interactions with lactobacilli in a
Bogdanov, S., Charrière, J. D., Imdorf, A., Kilchenmann, V., and Fluri, P. (2002). Drosophila melanogaster insect model. Appl. Environ. Microbiol. 84, 1–13. doi:
Determination of residues in honey after treatments with formic and oxalic acid 10.1128/AEM.02820-17
under field conditions. Apidologie 33, 399–409. doi: 10.1051/apido:2002029 De Almeida, L. G., De Moraes, L. A. B., Trigo, J. R., Omoto, C., and Cônsoli, F. L.
Boncristiani, H., Underwood, R., Schwarz, R., Evans, J. D., Pettis, J., and (2017). The gut microbiota of insecticide-resistant insects houses insecticide-
vanEngelsdorp, D. (2012). Direct effect of acaricides on pathogen loads and degrading bacteria: a potential source for biotechnological exploitation. PLoS
gene expression levels in honey bees Apis mellifera. J. Insect Physiol. 58, 613–620. One 12:e0174754. doi: 10.1371/[Link].0174754
doi: 10.1016/[Link].2011.12.011 Decourtye, A., Armengaud, C., Renou, M., Devillers, J., Cluzeau, S., Gauthier,
Bonilla-Rosso, G., and Engel, P. (2018). Functional roles and metabolic niches in M., et al. (2004a). Imidacloprid impairs memory and brain metabolism in the
the honey bee gut microbiota. Curr. Opin. Microbiol. 43, 69–76. doi: 10.1016/j. honeybee (Apis mellifera L.). Pestic. Biochem. Physiol. 78, 83–92. doi: 10.1016/j.
mib.2017.12.009 pestbp.2003.10.001
Brandt, A., Grikscheit, K., Siede, R., Grosse, R., Meixner, M. D., and Büchler, Decourtye, A., Devillers, J., Cluzeau, S., Charreton, M., and Pham-Delègue, M. H.
R. (2017). Immunosuppression in Honeybee Queens by the Neonicotinoids (2004b). Effects of imidacloprid and deltamethrin on associative learning in
Thiacloprid and Clothianidin. Sci. Rep. 7:4673 . doi: 10.1038/s41598-017- honeybees under semi-field and laboratory conditions. Ecotoxicol. Environ. Saf.
04734-1 57, 410–419. doi: 10.1016/[Link].2003.08.001
Breeze, T. D., Vaissière, B. E., Bommarco, R., Petanidou, T., Seraphides, N., Kozák, Dewar, A. M. (2017). The adverse impact of the neonicotinoid seed treatment ban
L., et al. (2014). Agricultural policies exacerbate honeybee pollination service on crop protection in oilseed rape in the United Kingdom. Pest Manag. Sci. 73,
supply-demand mismatches across Europe. PLoS One 9:e82996. doi: 10.1371/ 1305–1309. doi: 10.1002/ps.4511
[Link].0082996 Diaz, T., del-Val, E., Ayala, R., and Larsen, J. (2019). Alterations in honey bee gut
Breusch, T. S., and Pagan, A. R. (1979). A Simple Test for Heteroscedasticity microorganisms caused by Nosema spp. and pest control methods. Pest Manag.
and Random Coefficient Variation. Econometrica 47, 1287–1294. doi: 10.2307/ Sci. 75, 835–843. doi: 10.1002/ps.5188
1911963 El Hassani, A. K., Dacher, M., Gary, V., Lambin, M., Gauthier, M., and Armengaud,
Brodschneider, R., Brus, J., and Danihlík, J. (2019). Comparison of apiculture and C. (2008). Effects of sublethal doses of acetamiprid and thiamethoxam on the
winter mortality of honey bee colonies (Apis mellifera) in Austria and Czechia. behavior of the honeybee (Apis mellifera). Arch. Environ. Contam. Toxicol. 54,
Agric. Ecosyst. Environ. 274, 24–32. doi: 10.1016/[Link].2019.01.002 653–661. doi: 10.1007/s00244-007-9071-8
Brosi, B. J., Delaplane, K. S., Boots, M., and De Roode, J. C. (2017). Ecological Ellegaard, K. M., and Engel, P. (2019). Genomic diversity landscape of the honey
and evolutionary approaches to managing honeybee disease. Nat. Ecol. Evol. 1, bee gut microbiota. Nat. Commun. 10:446. doi: 10.1038/s41467-019-08303-0
1250–1262. doi: 10.1038/s41559-017-0246-z Ellegaard, K. M., Suenami, S., Miyazaki, R., and Engel, P. (2020). Vast Differences
Butler, È, Alsterfjord, M., Olofsson, T. C., Karlsson, C., Malmström, J., and Vásquez, in Strain-Level Diversity in the Gut Microbiota of Two Closely Related Honey
A. (2013). Proteins of novel lactic acid bacteria from Apis mellifera mellifera: an Bee Species. Curr. Biol. 30, 2520–2531.e7. doi: 10.1016/[Link].2020.04.070
insight into the production of known extra-cellular proteins during microbial Engel, P., and Moran, N. A. (2013). The gut microbiota of insects - diversity in
stress. BMC Microbiol. 13:235. doi: 10.1186/1471-2180-13-235 structure and function. FEMS Microbiol. Rev. 37, 699–735. doi: 10.1111/1574-
Callahan, B., McMurdie, P., Rosen, M., Han, A. W., Johnson, A. J., and Holmes, S. 6976.12025
(2016). P. DADA2: high-resolution sample inference from Illumina amplicon Engel, P., Martinson, V. G., and Moran, N. A. (2012). Functional diversity within
data. Nat Methods 13, 581–583. doi: 10.1038/nmeth.3869 the simple gut microbiota of the honey bee. Proc. Natl. Acad. Sci. U.S.A. 109,
Carreck, N. L., Andree, M., Brent, C. S., Cox-Foster, D., Dade, H. A., Ellis, J. D., et al. 11002–11007. doi: 10.1073/pnas.1202970109
(2013). Standard methods for Apis mellifera anatomy and dissection. J. Apic. Engel, P., James, R. R., Koga, R., Kwong, W. K., McFrederick, Q. S., and Moran,
Res. 52:4. doi: 10.3896/IBRA.[Link] N. A. (2013). Standard methods for research on Apis mellifera gut symbionts.
Casida, J. E., and Durkin, K. A. (2013). Neuroactive insecticides: targets, selectivity, J. Apic. Res. 52, 1–24. doi: 10.3896/IBRA.[Link]
resistance, and secondary effects. Ann. Rev. Entomol. 58, 99–117. doi: 10.1146/ Espregueira Themudo, G., Rey-Iglesia, A., Robles Tascón, L., Bruun Jensen, A.,
annurev-ento-120811-153645 da Fonseca, R. R., and Campos, P. F. (2020). Declining genetic diversity of

Frontiers in Microbiology | [Link] 14 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

European honeybees along the twentieth century. Sci. Rep. 10:10520 . doi: by antibiotic treatment could increase the honey bee’s vulnerability to Nosema
10.1038/s41598-020-67370-2 infection. PLoS One 12:e0187505. doi: 10.1371/[Link].0187505
European commission (2013). European commission. Available online at: Liu, Y. J., Qiao, N. H., Diao, Q. Y., Jing, Z., Vukanti, R., Dai, P. L., et al. (2020).
[Link] Thiacloprid exposure perturbs the gut microbiota and reduces the survival
approval_renewal/neonicotinoids_en (accessed May 23, 2021). status in honeybees. J. Hazard. Mater. 389:121818. doi: 10.1016/[Link].2019.
European Commission (2020). European Commission. Available online at: https: 121818
//[Link]/newsroom/sante/items/667420/en (accessed August 10, 2021). Ludvigsen, J., Rangberg, A., Avershina, E., Sekelja, M., Kreibich, C., Amdam, G.,
European Food Safety Authority (EFSA) (2020). European Food Safety et al. (2015). Shifts in the midgut/pyloric microbiota composition within a
Authority. Available online at: [Link] honey bee apiary throughout a season. Microbes Environ. 30, 235–244. doi:
pesticides-efsa-examine-emergency-use-neonicotinoids (accessed 10.1264/jsme2.ME15019
May 23, 2021). Mancini, F., Woodcock, B. A., and Isaac, N. J. B. (2019). Agrochemicals in the wild:
Evans, J. D., Chen, Y. P., di Prisco, G., Pettis, J., and Williams, V. (2009). Bee identifying links between pesticide use and declines of nontarget organisms.
cups: single-use cages for honeybee experiments. J. Apic. Res. 48, 300–302. Curr. Opin. Environ. Sci. Health 11, 53–58. doi: 10.1016/[Link].2019.07.003
doi: 10.1080/00218839.2009.11101548 Martín-Hernández, R., Higes, M., Pérez, J. L., Nozal, M. J., Gómez, L., and Meana,
Fernandes, A. D., Macklaim, J. M., Linn, T. G., Reid, G., and Gloor, G. B. (2013). A. (2007). Short term negative effect of oxalic acid in Apis mellifera iberiensis.
ANOVA-Like Differential Expression (ALDEx) Analysis for Mixed Population Span. J. Agric. Res. 5, 474–480. doi: 10.5424/sjar/2007054-270
RNA-Seq. PLoS One 8:e67019. doi: 10.1371/[Link].0067019 Martinson, V. G., Moy, J., and Moran, N. A. (2012). Establishment of Characteristic
Gill, R. J., Ramos-Rodríguez, O., and Raine, N. E. (2012). Combined pesticide Gut Bacteria during Development of the Honeybee Worker. Appl. Environ.
exposure severely affects individual- and colony-level traits in bees. Nature 490, Microbiol. 78, 2830–2840. doi: 10.1128/AEM.07810-11
105–108. doi: 10.1038/nature11585 Maruščáková, I. C., Schusterová, P., Bielik, B., Toporèák, J., Bíliková, K., and
Goulson, D., Nicholls, E., Botías, C., and Rotheray, E. L. (2015). Bee declines Mudroòová, D. (2020). Effect of Application of Probiotic Pollen Suspension
driven by combined Stress from parasites, pesticides, and lack of flowers. Science on Immune Response and Gut Microbiota of Honey Bees (Apis mellifera).
347:1255957. doi: 10.1126/science.1255957 Probiotics Antimicrob. Proteins 12, 929–936. doi: 10.1007/s12602-019-09626-6
Gregorc, A., and Smodiš Škerl, M. I. (2007). Toxicological and McMurdie, P. J., and Holmes, S. (2013). Phyloseq: an R Package for Reproducible
immunohistochemical testing of honeybees after oxalic acid and rotenone Interactive Analysis and Graphics of Microbiome Census Data. PLoS One
treatments. Apidologie 38, 296–305. doi: 10.1051/apido:2007014 8:e61217. doi: 10.1371/[Link].0061217
Hristov, P., Neov, B., Shumkova, R., and Palova, N. (2020). Significance of apoidea Meeus, I., Pisman, M., Smagghe, G., and Piot, N. (2018). Interaction effects of
as main pollinators. ecological and economic impact and implications for different drivers of wild bee decline and their influence on host–pathogen
human nutrition. Diversity 12:280. doi: 10.3390/d12070280 dynamics. Curr. Opin. Insect Sci. 26, 136–141. doi: 10.1016/[Link].2018.02.007
Huang, S. K., Csaki, T., Doublet, V., Dussaubat, C., Evans, J. D., Gajda, A. M., Mills, M. (2012). Introducing Survival and Event History Analysis. California: Sage
et al. (2014). Evaluation of cage designs and feeding regimes for honeybee publicactions INC.
(Hymenoptera: Apidae) laboratory experiments. J. Econ. Entomol. 107, 54–62. Moosbeckhofer, R., Pechhacker, H., Unterweger, H., Bandion, F., and Heinrich-
doi: 10.1603/EC13213 Lenz, A. (2003). Investigations on the oxalic acid content of honey from oxalic
Hung, K. L. J., Kingston, J. M., Albrecht, M., Holway, D. A., and Kohn, J. R. (2018). acid treated and untreated bee colonies. Eur. Food Res. Technol. 217, 49–52.
The worldwide importance of honeybees as pollinators in natural habitats. Proc. doi: 10.1007/s00217-003-0698-z
R. Soc. B Biol. Sci. 285:20172140. doi: 10.1098/rspb.2017.2140 Moran, N. A. (2015). Genomics of the honeybee microbiome. Curr. Opin. Insect
Iwasa, T., Motoyama, N., Ambrose, J. T., and Roe, R. M. (2004). Mechanism for the Sci. 10, 22–28. doi: 10.1016/[Link].2015.04.003
differential toxicity of neonicotinoid insecticides in the honeybee, Apis mellifera. Moran, N. A., Hansen, A. K., Powell, J. E., and Sabree, Z. L. (2012). Distinctive gut
Crop Prot. 23, 371–378. doi: 10.1016/[Link].2003.08.018 microbiota of honeybees assessed using deep sampling from individual worker
Johnson, R. M. (2015). Honeybee toxicology. Annu. Rev. Entomol. 60, 415–434. bees. PLoS One 7:e36393. doi: 10.1371/[Link].0036393
doi: 10.1146/annurev-ento-011613-162005 Motta, E. V. S., Raymann, K., and Moran, N. A. (2018). Glyphosate perturbs the
Jones, J. C., Fruciano, C., Marchant, J., Hildebrand, F., Forslund, S., Bork, P., et al. gut microbiota of honeybees. Proc. Natl. Acad. Sci. U. S. A. 115, 10305–10310.
(2018). The gut microbiome is associated with behavioural task in honey bees. doi: 10.1073/pnas.1803880115
Insectes Soc. 65, 419–429. doi: 10.1007/s00040-018-0624-9 Mullin, C. A., Frazier, M., Frazier, J. L., Ashcraft, S., Simonds, R., vanEngelsdorp,
Kakumanu, M. L., Reeves, A. M., Anderson, T. D., Rodrigues, R. R., and Williams, D., et al. (2010). High Levels of Miticides and Agrochemicals in North American
M. A. (2016). Honeybee gut microbiome is altered by in-hive pesticide Apiaries: implications for Honey Bee Health. PLoS One 5:e9754. doi: 10.1371/
exposures. Front. Microbiol. 7:1255. doi: 10.3389/fmicb.2016.01255 [Link].0009754
Kešnerová, L., Emery, O., Troilo, M., Liberti, J., Erkosar, B., and Engel, P. (2020). Nanetti, A., Büchler, R., Charrière, J. D., Fries, I., Helland, S., Imdorf, A., et al.
Gut microbiota structure differs between honeybees in winter and summer. (2003). Oxalic acid treatments for Varroa control (Review). Apiacta 38, 81–87.
ISME J. 14, 801–814. doi: 10.1038/s41396-019-0568-8 doi: 10.1016/[Link].2015.08.003
Kešnerová, L., Mars, R. A. T., Ellegaard, K. M., Troilo, M., Sauer, U., and Engel, P. Nanetti, A., Rodriguez-García, C., Meana, A., Martín-Hernández, R., and Higes,
(2017). Disentangling metabolic functions of bacteria in the honeybee gut. PLoS M. (2015). Effect of oxalic acid on Nosema ceranae infection. Res. Vet. Sci. 102,
Biol. 15:e2003467. doi: 10.1371/[Link].2003467 167–172. doi: 10.1016/[Link].2015.08.003
Kešnerová, L., Moritz, R., and Engel, P. (2016). Bartonella apis sp. nov., a honeybee Naug, D. (2009). Nutritional stress due to habitat loss may explain recent honeybee
gut symbiont of the class Alphaproteobacteria. Int. J. Syst. Evol. Microbiol. 66, colony collapses. Biol. Conserv. 142, 2369–2372. doi: 10.1016/[Link].2009.04.
414–421. doi: 10.1099/ijsem.0.000736 007
Klein, S., Cabirol, A., Devaud, J. M., Barron, A. B., and Lihoreau, M. (2017). Nozal, M. J., Bernal, J. L., Gómez, L. A., Higes, M., and Meana, A. (2003).
Why Bees Are So Vulnerable to Environmental Stressors. Trends Ecol. Evol. 32, Determination of Oxalic acid and other organic acids in honey and in
268–278. doi: 10.1016/[Link].2016.12.009 some anatomic structures of bees. Apidologie 34, 181–188. doi: 10.1051/apido:
Kwak, A. M., Lee, I. K., Lee, S. Y., Yun, B. S., and Kang, H. W. (2016). 2003001
Oxalic acid from Lentinula edodes culture filtrate: antimicrobial activity Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn,
on phytopathogenic bacteria and qualitative and quantitative analyses. D., et al. (2018). Vegan: community ecology package. R package version
Mycobiology 44, 338–342. doi: 10.5941/MYCO.2016.44.4.338 2.5-2.
Kwong, W. K., and Moran, N. A. (2016). Gut Microbial Communities of Social Papežíková, I., Palíková, M., Kremserová, S., Zachová, A., Peterová, H., Babák, V.,
Bees. Nat. Rev. Microbiol. 14, 59–73. doi: 10.1038/[Link] et al. (2017). Effect of oxalic acid on the mite Varroa destructor and its host the
Li, J. H., Evans, J. D., Li, W. F., Zhao, Y. Z., DeGrandi-Hoffman, G., Huang, honeybee Apis mellifera. J. Apic. Res. 56, 400–408. doi: 10.1080/00218839.2017.
S. K., et al. (2017). New evidence showing that the destruction of gut bacteria 1327937

Frontiers in Microbiology | [Link] 15 September 2021 | Volume 12 | Article 717990

Cuesta-Maté et al. Pesticide Effects on Honeybee Microbiome

Potts, S. G., Biesmeijer, J. C., Kremen, C., Neumann, P., Schweiger, O., and Kunin, Therneau, T. (2021). A Package for Survival Analysis in R. R package version 3.2-11.
W. E. (2010). Global pollinator declines: trends, impacts and drivers. Trends Available online at: [Link]
Ecol. Evol. 25, 345–353. doi: 10.1016/[Link].2010.01.007 Tian, B., Fadhil, N. H., Powell, J. E., Kwong, W. K., and Moran, N. A. (2012).
Powell, J. E., Eiri, D., Moran, N. A., and Rangel, J. (2018). Modulation of Long-term exposure to antibiotics has caused accumulation of resistance
the honeybee queen microbiota: effects of early social contact. PLoS One determinants in the gut microbiota of honeybees. mBio 3, 4–5. doi: 10.1128/
13:e0200527. doi: 10.1371/[Link].0200527 mBio.00377-12
Powell, J. E., Martinson, V. G., Urban-Mead, K., and Moran, N. A. (2014). Routes of Tison, L., Holtz, S., Adeoye, A., Kalkan, Ö, Irmisch, N. S., Lehmann, N., et al.
acquisition of the gut microbiota of the honeybee Apis mellifera. Appl. Environ. (2017). Effects of sublethal doses of thiacloprid and its formulation Calypso R
Microbiol. 80, 7378–7387. doi: 10.1128/AEM.01861-14 on the learning and memory performance of honey bees. J. Exp. Biol. 220,
Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013). 3695–3705. doi: 10.1242/jeb.154518
The SILVA ribosomal RNA gene database project: improved data processing Tosi, S., Burgio, G., and Nieh, J. C. (2017). A common neonicotinoid pesticide,
and web-based tools. Nucleic Acids Res. 41, 590–596. doi: 10.1093/nar/gks1219 thiamethoxam, impairs honey bee flight ability. Sci. Rep. 7:1201 . doi: 10.1038/
Rademacher, E., and Harz, M. (2006). Oxalic acid for the control of varroosis s41598-017-01361-8
in honeybee colonies - a review. Apidologie 37, 98–120. doi: 10.1051/apido: US Department of Agriculture (USDA) (2016). US Department of Agriculture.
2005063 Available online at: [Link]
Rademacher, E., Harz, M., and Schneider, S. (2017). Effects of oxalic acid on Apis Thiacloprid#section=USDA-Pesticide-Data-Program&fullscreen=true
mellifera (Hymenoptera: Apidae). Insects 8:84. doi: 10.3390/insects8030084 (accessed May 23, 2021).
Raymann, K., Bobay, L. M., and Moran, N. A. (2018a). Antibiotics reduce genetic US Department of Agriculture (USDA) (2021). US Department of Agriculture.
diversity of core species in the honeybee gut microbiome. Mol. Ecol. 27, Available online at: [Link]
2057–2066. doi: 10.1111/mec.14434 text=Logan%2C%\hbox{20Utah-,}U.S.%20Honey%20Bee%20Losses,
Raymann, K., and Moran, N. A. (2018). The role of the gut microbiome in health Statistics%20Service%20(NASS)%20survey (accessed May 23, 2021).
and disease of adult honey bee workers. Curr. Opin. Insect Sci. 26, 97–104. US Environmental Protection Agency Office of Pesticide (2020). Proposed Interim
doi: 10.1016/[Link].2018.02.012 Registration Review Decision Case Number 7605 January 2020. Washington:
Raymann, K., Motta, E. V. S., Girard, C., Riddington, I. M., Dinser, J. A., and United States Environmental Protection Agency, 1–77.
Moran, N. A. (2018b). Imidacloprid decreases honey bee survival rates but vanEngelsdorp, D., Traynor, K. S., Andree, M., Lichtenberg, E. M., Chen, Y.,
does not affect the gut microbiome. Appl. Environ. Microbiol. 84, e00545–18. Saegerman, C., et al. (2017). Colony Collapse Disorder (CCD) and bee age
doi: 10.1128/AEM.00545-18 impact honey bee pathophysiology. PLoS One 12:e0179535. doi: 10.1371/
Raymann, K., Shaffer, Z., and Moran, N. A. (2017). Antibiotic exposure perturbs [Link].0179535
the gut microbiota and elevates mortality in honeybees. PLoS Biol. 15:e2001861. Vásquez, A., Forsgren, E., Fries, I., Paxton, R. J., Flaberg, E., Szekely, L., et al.
doi: 10.1371/[Link].2001861 (2012). Symbionts as major modulators of insect health: lactic acid bacteria and
Rouzé, R., Moné, A., Delbac, F., Belzunces, L., and Blot, N. (2019). The honeybee honeybees. PLoS One 7:e33188. doi: 10.1371/[Link].0033188
gut microbiota is altered after chronic exposure to different families of Williamson, S. M., Willis, S. J., and Wright, G. A. (2014). Exposure to
insecticides and infection by Nosema ceranae. Microbes Environ. 34, 226–233. neonicotinoids influences the motor function of adult worker honeybees.
doi: 10.1264/jsme2.ME18169 Ecotoxicology 23, 1409–1418. doi: 10.1007/s10646-014-1283-x
Royston, J. P. (1982). An Extension of Shapiro and Wilk’s W Test for Normality to Zaworra, M., Koehler, H., Schneider, J., Lagojda, A., and Nauen, R. (2019).
Large Samples. Appl. Stat. 31:115. doi: 10.2307/2347973 Pharmacokinetics of Three Neonicotinoid Insecticides upon Contact Exposure
Sammataro, D., Finley, J., and Underwood, R. (2008). Comparing oxalic acid and in the Western Honey Bee, Apis mellifera [Rapid-communication]. Chem. Res.
sucrocide treatments for Varroa destructor (Acari: Varroidae) control under Toxicol. 32, 35–37. doi: 10.1021/[Link].8b00315
desert conditions. J. Econ. Entomol. 101, 1057–1061. Zeileis, A., and Hothorn, T. (2002). Diagnostic Checking in Regression
Schneider, S., Eisenhardt, D., and Rademacher, E. (2012). Sublethal effects of Relationships. R News 2, 7–10.
oxalic acid on Apis mellifera (Hymenoptera: Apidae): changes in behaviour and Zheng, H., Nishida, A., Kwong, W. K., Koch, H., Engel, P., Steele, M. I., et al.
longevity. Apidologie 43, 218–225. doi: 10.1007/s13592-011-0102-0 (2016). Metabolism of toxic sugars by strains of the bee gut symbiont Gilliamella
Shi, J., Liao, C., Wang, Z., Zeng, Z., and Wu, X. (2019). Effects of sublethal apicola. mBio 7, 1–9. doi: 10.1128/mBio.01326-16
acetamiprid doses on the lifespan and memory-related characteristics of honey
bee (Apis mellifera) workers. Apidologie 50, 553–563. doi: 10.1007/s13592-019- Conflict of Interest: The authors declare that the research was conducted in the
00669-w absence of any commercial or financial relationships that could be construed as a
Siede, R., Faust, L., Meixner, M. D., Maus, C., Grünewald, B., and Büchler, R. (2017). potential conflict of interest.
Performance of honey bee colonies under a long-lasting dietary exposure
to sublethal concentrations of the neonicotinoid insecticide thiacloprid. Pest Publisher’s Note: All claims expressed in this article are solely those of the authors
Manag. Sci. 73, 1334–1344. doi: 10.1002/ps.4547 and do not necessarily represent those of their affiliated organizations, or those of
Singh, S., Singh, S. K., Chowdhury, I., and Singh, R. (2017). Understanding the the publisher, the editors and the reviewers. Any product that may be evaluated in
Mechanism of Bacterial Biofilms Resistance to Antimicrobial Agents. Open this article, or claim that may be made by its manufacturer, is not guaranteed or
Microbiol. J. 11, 53–62. doi: 10.2174/1874285801711010053 endorsed by the publisher.
Tarpy, D. R., Mattila, H. R., and Newton, I. L. G. (2015). Development of the honey
bee gut microbiome throughout the queen-rearing process. Appl. Environ. Copyright © 2021 Cuesta-Maté, Renelies-Hamilton, Kryger, Jensen, Sinotte and
Microbiol. 81, 3182–3191. doi: 10.1128/AEM.00307-15 Poulsen. This is an open-access article distributed under the terms of the Creative
Thany, S. H., Bourdin, C. M., Graton, J., Laurent, A. D., Mathé-Allainmat, Commons Attribution License (CC BY). The use, distribution or reproduction in
M., Lebreton, J., et al. (2015). Similar comparative low and high doses of other forums is permitted, provided the original author(s) and the copyright owner(s)
deltamethrin and acetamiprid differently impair the retrieval of the proboscis are credited and that the original publication in this journal is cited, in accordance
extension reflex in the forager honeybee (Apis mellifera). Insects 6, 805–814. with accepted academic practice. No use, distribution or reproduction is permitted
doi: 10.3390/insects6040805 which does not comply with these terms.

Frontiers in Microbiology | [Link] 16 September 2021 | Volume 12 | Article 717990