This protocol was developed to enhance the understanding of how agrochemicals affect honey bee (Apis mellifera) reproduction by establishing methods to expose honey bee queens and their worker caretakers to agrochemicals in a controlled, laboratory setting and carefully monitoring their relevant responses.
Current risk assessment strategies for honey bees rely heavily upon laboratory tests performed on adult or immature worker bees, but these methods may not accurately capture the effects of agrochemical exposure on honey bee queens. As the sole producer of fertilized eggs inside a honeybee colony, the queen is arguably the most important single member of a functioning colony unit. Therefore, understanding how agrochemicals affect queen health and productivity should be considered a critical aspect of pesticide risk assessment. Here, an adapted method is presented to expose honey bee queens and worker queen attendants to agrochemical stressors administered through a worker diet, followed by tracking egg production in the laboratory and assessing first instar eclosion using a specialized cage, referred to as a Queen Monitoring Cage. To illustrate the method's intended use, results of an experiment in which worker queen attendants were fed diet containing sublethal doses of imidacloprid and effects on queens were monitored are described.
Due to increased global demand for agricultural products, modern farming practices often require the use of agrochemicals to control numerous pests known to reduce or harm crop yields1. Simultaneously, the growers of many fruit, vegetable, and nut crops rely on the pollination services provided by commercial honey bee colonies to ensure abundant crop yields2. These practices may result in pollinators, including honey bees (Apis mellifera), being exposed to harmful levels of pesticide residues3. At the same time, the widespread presence of parasitic Varroa destructor mite infestations in honey bee colonies frequently require beekeepers to treat their hives with miticides, which may also exert negative effects on the health and longevity of the colony4,5,6. To reduce and mitigate harmful effects of agrochemical products, it is necessary to fully evaluate their safety to honey bees prior to their implementation so that recommendations for their use can be made to protect beneficial insects.
Currently, the Environmental Protection Agency (EPA) relies upon a tiered risk-assessment strategy for honey bee pesticide exposure, which involves laboratory tests on adult bees and sometimes honey bee larvae7. If lower tier laboratory tests fail to alleviate concerns of toxicity, higher tier field and semi-field testing may be recommended. While these laboratory tests provide valuable insight into the potential effects of agrochemicals on worker longevity, they are not necessarily predictive of their effects on queens, which differ significantly from workers biologically8 and behaviorally9. Furthermore, there are numerous potential effects of agrochemicals on insects beyond mortality, which can have considerable consequences for social insects that rely on coordinated behaviors to function as a colony unit10,11.
Although mortality is the most commonly considered effect of agrochemical pesticides12, these products can have a wide range of effects on both target and non-target arthropods including altered behavior13,14,15,16, repellency or attractancy17,18,19, changes in feeding patterns20,21,22, and increased or decreased fecundity20,21,22,23,24,25. For social insects, these effects can systemically disrupt colony interactions and functions11. Of these functions, reproduction, which is heavily reliant on a single egg-laying queen supported by the rest of the of the colony unit9, may be particularly vulnerable to perturbation due to pesticide exposure.
Studies performed on immature queens have demonstrated that developmental exposure to miticides can affect adult queen behavior, physiology, survival26,27. Similarly, studies using full or reduced sized colonies have demonstrated that agrochemicals can affect adult honey bee queens by decreasing mating success28, decreasing oviposition29, and decreasing the viability of the eggs produced25,30,31. These phenomena have previously been difficult to observe without the use of whole colonies, due largely to a lack of available laboratory methods. However, a method to study queen oviposition under tightly controlled laboratory conditions using Queen Monitoring Cages (QMC)32 has recently been adapted to examine the effects of agrochemicals on queen fecundity33. Here, these techniques are described in detail along with additional methods to measure and track worker diet consumption in QMCs.
These methods are more advantageous than experiments requiring full sized colonies because they allow for the administration of precise doses of agrochemicals to a greatly reduced number of workers relative to the tens of thousands typically present inside a colony34, which then provision the queen. This exposure technique mirrors the second-hand exposure that queens would experience in real-world scenarios because, within a colony, queens do not feed themselves and rely upon workers to provision them with diet9. Similarly, queens do not generally leave the hive except during colony reproduction (swarming) for mating flights35. Mated honey bee queens can be purchased from commercial queen breeders and shipped overnight. Typically, queen breeders sell queens directly after confirming that they have started to lay eggs, which is taken as an indication of successful mating. If more precise information on queen age or relatedness is needed, researchers may consult with the queen breeder before placing an order.
QMCs allow for precise observation and quantification of honey bee queen oviposition and egg hatching rates32,33, yielding valuable data related to the effects of agrochemical exposure on queen fecundity. The representative results presented here describe an experiment quantifying oviposition, diet consumption, and embryo viability in QMCs under chronic exposure to field relevant concentrations of the systemic neurotoxicant neonicotinoid pesticide imidacloprid36. Once applied, imidacloprid translocates to plant tissues37, and residues have been detected the pollen and nectar of numerous bee pollinated plants38,39,40. Exposure to imidacloprid can have a broad range of detrimental effects on honey bees including impaired foraging performance16, impaired immune function41, and decreased rates of colony expansion and survival42,43. Here, imidacloprid was selected for use as a test substance because field experiments have shown that it can affect honey bee queen oviposition29
1. QMC assembly
Figure 1. A: Disassembled QMC. B: Partially assembled QMC with 1 ELP inserted. C: Fully assembled QMC with 2 ELPs. Please click here to view a larger version of this figure.
2. Preparing and administering diets laced with agrochemicals
Desired Concentration | Desired Solvent Vehicle Concentration | Desired final volume/mass of sucrose solution | Voume of stock solution | Imidacloprid in stock solution | Suggested stock solution recipe | |
Sucrose solution | 10 ppb (w/w) | 0.05% (v/v) | 81.45 mL/100 g* | 40.7 µL | 0.001 mg/40.7 µL | 0.02 mg/814 µL |
Pollen supplement | 10 ppb (w/w) | 10 g** | 4.07 µL | 0.0001 mg/4.07 µL | 0.02 mg/814 µL | |
Sucrose solution | 50 ppb (w/w) | 0.05% (v/v) | 78.5 mL/100 g* | 40.7 µL | 0.005 mg/40.7 µL | 0.1 mg/814 µL |
Pollen supplement | 50 ppb (w/w) | 10 g** | 4.07 µL | 0.0005 mg/4.07 µL | 0.1 mg/814 µL |
Table 1: Example recipes for treated sucrose solution, pollen supplement, and stock solution. *Volume based on the density of 50% (w/w) sucrose solution (1.228 g/mL). **The density of the pollen supplement will vary depending on what product is used, but if this solvent volume is used, the final solvent concentration in pollen supplement will be within the desired range of ≤ 5% by volume.
3. Monitoring – Egg Production Rate
4. Monitoring – Food Consumption
5. Monitoring – Embryo Viability
6. Worker Sampling
7. Worker Mortality
The production of eggs was monitored in QMCs assembled and maintained as described above with once daily observations of egg production and 15 cages per treatment group. Newly mated queens of primarily Carniolan stock were purchased and shipped overnight from a queen breeder, and honey bee workers were obtained from 3 colonies maintained according to standard commercial methods at The Bee Research Facility at the University of Illinois Urbana-Champaign. Here, 4 dietary treatment groups were used: 1) 50 ppb (g/g) imidacloprid in sucrose solution and pollen supplement (50 ppb – p+s), 2) 10 ppb imidacloprid in sucrose solution and pollen supplement (10 ppb – p+s), 3) 10 ppb imidacloprid in pollen supplement alone (10 ppb – p), and 4) a control group given diet containing an equivalent volume of acetone as the treatment groups (CTRL).
Treatment-related changes in daily egg counts were evaluated as described in Fine et al.32 with minor modifications. Briefly, a Poisson log-linear GEE with an auto-regressive (AR-1) correlation matrix structure was implemented to assess treatment related changes in egg production over time. Here, time (day) was treated as a continuous variable and treatment was categorical. Wald chi-square post hoc tests were used to determine significance. Because no egg laying was observed on day 1 of the experiment, this day was excluded from analysis to conform to the assumptions of the GEE. The results of this analysis are shown in Table S1. Daily egg production was significantly lower in QMCs in the 50 ppb p+s treatment group (χ2=43.99, p<0.001; Figure 2A).
Differences in the total number of eggs produced in QMCs by treatment were analyzed using a one-way ANOVA and Tukey HSD post hoc test (Figure 3). For this analysis, any QMC removed from the experiment before the end of the 14-day monitoring period due to queen or worker death was excluded, resulting in N=13 each for the CTRL and the 50 ppb – p+s groups, N=14 for 10 ppb – p, and N=15 for 10 ppb – p+s. A dose dependent effect was observed for treatments administered in both sucrose and pollen, with the largest reduction in egg production relative to control observed in 50 ppb – p+s followed by 10 ppb – p+s. No difference in total eggs produced was observed between CTRL and 10 ppb – p (F3, 52=17.95, p<0.001, Tukey HSD).
Consumption of pollen supplement and water was recorded every 48 hours for 10 days, and consumption of sucrose solution was recorded every 48 hours for 12 days. Changes in diet consumption rates were evaluated using Gaussian distributed GEEs with the same parameters as described above (Figure 2B-D). Results are summarized in Table S1. Briefly, daily rates of sucrose consumption significantly increased as the experiment progressed (χ2=6.03, p=0.014), but rates of pollen supplement consumption decreased (χ2=174.98, p<0.001). Significantly higher rates of pollen consumption were observed when imidacloprid was administered at 10 ppb in pollen supplement alone (χ2=21.44, p<0.001) and significantly decreased when it was administered at either 10 or 50 ppb in pollen supplement and sucrose solution together (10 ppb – p+s: χ2=6.59, p=0.010; 50 ppb – p+s: χ2=14.47, p=0.0001).
Eggs were collected from QMCs on day 7 of the experiment, and changes in the number of eggs hatching successfully following maternal exposure to agrochemical treatments was assessed using a generalized linear mixed model (GMLR) with a binomial distribution and QMC identity treated as a random effect. Maternal exposure to imidacloprid administered at 10 ppb in pollen alone or in pollen and sucrose solution did not affect egg hatching rates (10 ppb – p+s: Z=-0.139, p=0.290; 10 ppb – p: Z=0.182, p=0.856). Hatching rates could not be assessed for eggs laid by queens in QMCs provisioned with 50 ppb imidacloprid laced diet due to low rates of egg production in this treatment group.
For this work, all statistical analysis was performed in R Studio 1.2.5003 (Boston, MA, USA). Figures were prepared using JMP Pro 15 and Photoshop CC 2019 (Adobe Inc., San Jose, CA). Data are available in Supplementary file S1.
Figure 2. A: Average ± SE eggs per day in QMCs. B: Average ± SE pollen supplement, C: sucrose solution, D: and water (g) consumed during 48-hour periods in QMCs. Significance of treatments (indicated by "*") determined by GEE and Wald chi-square post hoc test. Please click here to view a larger version of this figure.
Figure 3: Average ± SE sum of eggs laid by treatment during experiment. Significance (indicated by letters) determined by ANOVA and Tukey HSD post hoc test. Please click here to view a larger version of this figure.
Table S1:Results of GEEs analyzing changes in egg laying rates and diet consumption in QMCs over time. Please click here to download this Table.
Supplementary file S1: Please click here to download this Supplemental File.
The fecundity of female solitary insects as well as queens in eusocial insect colonies can be influenced by abiotic stressors such as agrochemicals25,28,29,30,33. In honey bees, the effects of agrochemicals on queens may be indirect, as they can occur via changes in their care and feeding by worker bees. Our representative results, which are similar to those reported in a field-based study29, demonstrate that the effects of agrochemicals on queen performance can be measured efficiently in a laboratory environment using QMCs, generating comparable results to field-based approaches. Furthermore, these results shed light on the influence of imidacloprid on worker diet consumption and on egg viability.
Imidacloprid had clear negative effects on egg production when it was administered in sucrose solution and pollen supplement together. This is similar to results reported using observation hives provisioned with imidacloprid laced syrup and permitted to forage freely29. However, here, a dose-dependent response was observed, with the most pronounced effect seen in QMCs provisioned with 50 ppb imidacloprid relative to the lower concentration. Unlike what was reported for field colonies, this group experienced a near cessation of egg production. It should be noted that all concentrations including 50 ppb used in this work are higher than pollen and nectar residues typically observed when imidacloprid is applied as a seed treatment and are more representative of residues found following soil applications40. Examples of relevant plants include cucurbits and ornamentals found in urban landscapes29, and therefore, these results should be interpreted in this context. Additionally, the differences observed between these results and those generated using field colonies, where results were not as pronounced, even in the highest treatment groups, suggest that like other laboratory-based assays, QMCs may be more sensitive than using full sized colonies52, which should be considered when interpreting the data.
Previously reported work examining oviposition with exposure to insect growth regulators (IGR) in QMCs did not find that IGRs cause reductions in queen egg laying rates33, demonstrating that disruption of egg production is not a uniform stress response. Although field level assessments using full-sized colonies may provide a more holistic view of the effects of agrochemicals on colony health, these findings suggests that QMCs have the potential to be used as a tool to identify chemicals like imidacloprid that may affect honey bee queen oviposition. When used in the context of a broad risk assessment strategy accounting for use patterns, exposure patterns, and effects on other metrics of honey bee health, egg production data generated by QMCs may yield a more comprehensive understanding of the potential effects of an agrochemical on honey bee colonies.
In addition to generating quantitative oviposition data, QMCs can be used to assess patterns in worker diet consumption and changes in physiology. Here, it was shown that 10 ppb imidacloprid in pollen diet alone stimulates pollen supplement consumption in workers in the presence of a mated queen. This effect was not observed in other dietary treatments when QMCs were provisioned with imidacloprid in both pollen supplement and sucrose solution, even at the same concentration. It should be noted that more precise estimates of consumption rate can be obtained by tracking mortality and adjusting measures of diet consumption based on the exact number of bees remaining in the QMCs, but if mortality is consistently low across treatments, some comparisons can be made. The discrepancy between treatments in the consumption of pollen diet containing the same concentration of imidacloprid may be related to the difference in the higher total dose administered to bees when imidacloprid is present in both sucrose and pollen supplement compared to when it is present in pollen supplement alone.
At low levels, there is evidence that honey bees prefer food sources containing neonicotinoid pesticides18, and they have been reported to exhibit a similar preference for floral resources containing nicotine53. It has been suggested that these preferences may be due to the neuro-stimulative properties of nicotine and neonicotinoids, which activate nicotinic acetylcholinesterase receptors54 expressed in parts of the honey bee brain involved in learning and memory55. In spider mites, imidacloprid stimulates diet consumption, resulting in increased oviposition and fecundity22. Here, imidacloprid-related increases in pollen supplement consumption were not related to increases in oviposition, and the effects of imidacloprid on worker physiology in this work remain to be explored. However, understanding how much of an agrochemical-laced diet bees inside a colony are likely to consume, particularly workers that require more pollen in their diet to actively provision a laying queen56, can help inform the risk of an agrochemical to various aspects of colony performance.
Imidacloprid did not cause any measurable changes in embryo viability, as measured by hatching rates in eggs collected from QMCs provisioned with 10 ppb imidacloprid in pollen supplement alone or in both pollen supplement and sucrose solution. This differs from the decreases in egg hatching rates reported following IGR exposure in QMCs33, demonstrating again that QMCs can be used to examine specific and diverse aspects of queen fecundity. Imidacloprid is highly water soluble and is likely metabolized and excreted by bees differently than more fat-soluble agrochemicals like IGRs57, which may be transovarially eliminated58,59,60,61 to some extent, resulting in effects on embryo development. Alternatively, imidacloprid, which is a neurotoxin36 may not affect developing embryos in the same manner as IGRs, which target pathways associated with insect development62.
One question commonly asked by researchers seeking to understand the effects of agrochemicals on honey bee reproduction is whether adult queens, who rely on workers to provision her with glandular secretions as food9,63, are directly exposed to agrochemical residues. This was not explored and is not represented in the results reported here. However, agrochemical residues in worker glandular secretions are typically greatly reduced relative to what workers are provisioned with in controlled colony feeding scenarios64. Similarly, when full-sized colonies were exposed to concentrations of imidacloprid that resulted in decreased oviposition, no residues were detected in queens29, suggesting that the changes in oviposition rates observed in the referenced work were due to direct exposure to trace amounts that were readily excreted, or that the observed effects on queens were due to effects of imidacloprid on the workers responsible for caring for and provisioning the queen. The method presented here allows for the sampling of worker bees known to have ingested the treated diet from adult eclosion to the time of sampling. Follow-up work examining the effects of imidacloprid on the physiology of worker bees sampled from the described experiment will help to elucidate this question.
In summary, the methods presented here will allow researchers to better assess the risk of agrochemicals to honey bees by evaluating endpoints related to the fecundity, survival, and development of honey bees. The described technique has the potential to greatly enhance agrochemical risk assessment by generating quantitative data pertaining to queen fecundity that can be difficult and resource intensive to acquire using field and semi-field experiments. Additionally, the presence of a laying queen adds realism to experiments performed on young workers, which are typically the members of the colony responsible for the care and feeding of the queen9. Using this technique, the risks of agrochemicals on honey bee colony health, longevity, and performance can be better predicted and mitigated.
The authors have nothing to disclose.
Thank you to Dr. Amy Cash-Ahmed, Nathanael J. Beach, and Alison L. Sankey for their assistance in carrying out this work. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer. This research was supported by a grant from the Defense Advanced Research Projects Agency # HR0011-16-2-0019 to Gene E. Robinson and Huimin Zhao, USDA project 2030-21000-001-00-D, and the Phenotypic Plasticity Research Experience for Community College Students at the University of Illinois at Urbana Champaign.
Fluon | BioQuip, Rancho Dominguez, CA | 2871A | |
Honey bee queens | Olivarez Honey Bees, Orland, CA | ||
Imidacloprid | Sigma-Aldritch, St. Louis, MO | 37894 | |
MegaBee Powder | MegaBee, San Dieago, CA | ||
Microcentrifuge tubes 2 mL | ThermoFisher Scientific, Waltham, MA | 02-682-004 | |
Needles 20 gauge | W. W. Grainger, Lake Forest, IL | 5FVK4 | |
Potassium Sulfate | Sigma-Aldritch, St. Louis, MO | P0772 | |
Queen Monitoring Cages | University of Illinois Urbana-Champaign | Patent application number: 20190350175 | |
Sucrose | Sigma-Aldritch, St. Louis, MO | S8501 | |
Universal Microplate Lids | ThermoFisher Scientific, Waltham, MA | 5500 |