Development of pheromonal tools for honeybee breeding

Final Report for GNE14-090

Project Type: Graduate Student
Funds awarded in 2014: $14,998.00
Projected End Date: 12/31/2015
Grant Recipient: Pennsylvania State University
Region: Northeast
State: Pennsylvania
Graduate Student:
Faculty Advisor:
Dr. Thomas Baker
Pennsylvania State University
Faculty Advisor:
Expand All

Project Information


The purpose of this project is to lay the groundwork for the development of natural, large-scale stock improvement programs necessary for the conservation of a principal crop pollinator, the managed honey bee. Currently, the decline of populations of honey bees (Apis mellifera), the primary providers of pollination services in agriculture, is a major threat to food security in the United States, given that a majority of our fruit, vegetable, and nut crops are largely or entirely (i.e. almonds) dependent on honey bee pollination. This decline is attributed to the multiple challenges these pollinators currently face, which include pressure from parasites and pathogens, exposure to pesticides, reduced nutrition, and even climate fluctuations. In addition, honey bee breeding currently requires a high level of technical expertise and expensive equipment to facilitate artificial insemination. Our goal is to ultimately find/develop accessible tools and methods to facilitate controlled natural mating.

Because many honey bee behaviors are mediated by pheromones, we have explored the effects of a major queen pheromone (9-ODA) on the physiology and behavior of males. This pheromone has previously shown to regulate worker behavior, development and physiology, thus potentially similar impacts on drones could provide ways on regulating drone reproductive behavior. Indeed, we have found 9-ODA exposure to impact both drone physiology and the timing and frequency of drone reproductive behavior. A second objective has aimed to identify whether a male-produced pheromone exists and its role in mediating male mating aggregations. We have characterized the contents of the male mandibular gland and have employed field tests to demonstrate its biological activity.



The decline of honey bees (Apis mellifera) is a major threat to food security in the United States, given the magnitude of dependence that our agriculture has on managed pollinators. Because declines are attributed to multiple and diverse threats, finding conservation solutions has been difficult. However, the development of genetic stocks which are more resistant to diseases and parasites, provision the colony with more nutritional resources, and survive the winter more successfully, is gaining traction as a, and possibly the only, sustainable long-term solution.

However, there are significant challenges to implementing large-scale honey bee breeding programs. Currently, controlled bee breeding can only be achieved by instrumental insemination, a process that is restrictive to implement on a large-scale since it is labor intensive and requires sophisticated instrumentation and training. In addition, instrumental insemination has been shown to have multiple detrimental effects on the queen that may themselves impact the survival and productivity of honey bee colonies.

I propose to characterize the factors mediating mating behavior in honey bees, and use this knowledge to develop approaches to control matings in a natural manner, laying the groundwork for the development of large-scale and local stock improvement programs necessary for the conservation of this critical pollinator species.

Controlling natural mating is challenging due to the nature of reproduction in honey bees. First, queens and drones reach sexual maturity at specific times , and thus it is necessary to coordinate production of mature drones and queens. While it is possible to manipulate and time the production of queens with exquisite precision, the mechanisms regulating drone sexual maturation have not been investigated, and thus cannot be controlled. In Objective 1, I will investigate how the pheromonal environment within the colony may determine the pace and onset of sexual maturity in drones. Older, sexually-mature males fly out from colonies and congregate at specific sites to mate. How these congregations form is not well understood, but a drone-produced pheromone has been theorized to facilitate their formation. However, this putative pheromone has not yet been identified or characterized. In Objective 2, I will investigate whether a drone-produced pheromone exists and ascertain its identity and role in facilitating aggregative behavior. These objectives are part of a larger project, which includes a third objective (not funded as part of this proposal) that aims to identify the pheromones produced by virgin queens that act as chemical attractants for drones. Overall, these studies will provide critical information that will allow bee breeders to control the timing of sexually mature drone production, the aggregation of drones outside the colony, and selected drones’ attraction to and copulation with selected queens. In addition to facilitating controlled mating between selected stocks, these new pheromone components may increase the mating number of queens and thus the genetic diversity of the resulting colonies, which has been demonstrated to improve colony productivity, health and survival. Conservation of this pollinator will serve to improve the productivity of the crops they pollinate, reduce the costs of commercial pollination services for farmers and thus improve overall farm income and long-term food security.

Project Objectives:

Objective 1: Characterize the effect of the queen pheromone, 9-ODA, on the rate of drone sexual maturation.

  1. Characterize Vg/Juvenile hormone levels in 9-ODA exposed/unexposed male drones. (Completed)
  2. Characterize the effects of 9-ODA exposure on the reproductive behavior of drones under natural colony conditions. (Completed)

Objective 2: Identify and characterize a putative drone-produced pheromone involved in the formation of drone aggregations.

  1. Dissect putative pheromone glands from male drones and extract the contents. (Completed)
  2. Analyze gland contents using GC-MS. Identify compounds produced in the glands. (Completed)
  3. Ascertain the biological activity of identified gland compounds using behavioral assays. (Completed)

Objective 3: Disseminate the practical outcomes of these objectives to research and stakeholder communities

  1. Present ongoing findings at professional and stakeholder meetings. (Completed)

     2. Publish findings in professional research journals. (1 publication in review; 1 publication slated for submission December 2016)


Click linked name(s) to expand/collapse or show everyone's info
  • Dr. Thomas Baker
  • Dr. Christina Grozinger


Materials and methods:

Material and methods for Objective 1:

General honey bee rearing

Honey bee colonies were maintained according to standard apicultural practices at

apiaries at The Pennsylvania State University (University Park, PA, USA). To minimize genetic diversity among the workers used in the cage studies, we used workers produced by single-drone inseminated (SDI) queens (obtained from Glenn Apiaries, Fallbrook, CA). For the field trials with single cohort colonies, bees were derived from naturally mated (multiple-drone inseminated) queens. Note that since drones are produced from unfertilized eggs, brother drones are equally related regardless of the mating status of the queen. Replication of experiments occurred at the colony level and replicates are referred to as “trials.”

For experiments requiring newly emerged workers (callow), honeycomb frames of emerging brood were collected and stored in a dark incubator at 34°C and 50% relative humidity. Emerging bees (<24 hours old) were collected and placed in cages/colonies. For experiments requiring mature nurses, individuals were identified and collected as they inspected and fed larvae in comb cells within the brood nest of their colony.

Drone brood was obtained by caging a queen overnight on a honeycomb frame of drone comb cells (which are larger than the cell used to rear female worker brood), facilitating the laying of unfertilized, drone-destined eggs by the queen. These frames were maintained in the colony until drone emergence was imminent, at which point the frame was removed from the colony, placed in an incubator, and monitored for emergence as with the callow workers. For the cage experiments requiring both callow workers and drones, worker emergence was synchronized to occur with drone emergence.

In the cage studies, 20 callow or mature nurse (henceforth “nurse”) workers and 10 callow drones from a single colony source were placed in individual Plexiglas cages 42

(10x7x7cm). Workers were included because young drones require feeding by workers to survive (Villar, personal observation). Cages were provided with 50% sucrose and crushed pollen ad libitum. Cages were treated once daily with 20 μg of synthetic 9-ODA (Contech International, Victoria, BC) dissolved in 1% water/ isopropanol or a solvent only control on a glass slide, as in. This amount of 9-ODA corresponds to 0.1 queen equivalents of the pheromone and has previously been shown to result in nurse-like levels of JH and Vg. Cages were maintained in a dark incubator at 34°C and 50% relative humidity throughout the duration of the experiments.

Effect of 9-ODA exposure on Vg expression

In the first set of cage studies, 10 one-day old drones were caged with 20 callow workers. We set up 6 replicates of untreated and 9-ODA treated cages. After 48 hours, the bees from all the cages were collected on dry ice and stored at -80°C. This was replicated across two colony sources, for a total of two trials.

In the second set of cage studies, 10 one-day old drones were caged with either 20 callow workers or 20 nurses. Cage exposure to 9-ODA or a solvent control was increased to 72 hours, for a total of 4 different treatments, with 6 cages per treatment. Bees were collected on dry ice after 72 hours and stored at -80°C. These studies were replicated across three colony sources, for a total of three trials.

To characterize Vg expression, 2-3 frozen individuals per cage were collected, using at least four of the six cages/treatment (see figures for the exact number of replicates in each trial). The digestive system was completely removed from the abdomen, leaving the abdominal cuticle and attached fat body, the primary site of vitellogenin synthesis.

Individual abdominal fat bodies from collected drones were homogenized and RNA extracted using an RNeasy RNA isolation kit (Qiagen). cDNA was synthesized from 150 μg of extracted RNA using Superscript II reverse transcriptase (Invitrogen, Carlsbad CA). Expression levels were quantified using an ABI PRISM 7900 sequence detector using the SYBR Green detector method (Applied Biosystems). Samples were tested in triplicate and averaged. Genomic DNA dilutions were used to construct a standard curve and relative quantities of RNA in each sample were calculated. Primer specificity and lack of genomic DNA contamination was confirmed using a dissociation curve and negative controls. Expression levels reported for Vg (F: AGTTCCGACCGACGACG, R: TTCCCTCCCACGGAGTCC) were normalized to the geometric mean of two housekeeping genes, actin (F: CCTAGCACCATCCACCATGAA, R: GAAGCAAGAATTGACCCACCAA) and GAPDH (F: GCTGGTTTCATCGATGGTTT, R: ACGATTTCGACCACCGTAAC). The Grubbs test for statistical outliers was used to identify outliers and only one sample was removed from the analysis. Significant differences in relative expression levels were assessed using a two-way ANOVA with trial, pheromone treatment, and rearing conditions as factors for multiple trials or a Student’s t test for single trials (when applicable). JMP 9.0.2 (SAS Institute Inc.) was the statistical software used for all analyses.

Effect of 9-ODA exposure on trophallaxis rates and hypopharyngeal gland development

To assess the impacts of 9-ODA exposure on worker's likelihood to feed drones, we measured trophallaxis rates in each of the three trials of the second set of cage studies. Behavior was monitored under red light, two hours after the pheromone or solvent control treatment had been administered. Worker-drone trophallaxis events were recorded within each cage every five minutes for 40 minutes, repeated daily over the 72 h duration of the experiment. Event totals per cage per day served as the raw data for statistical analysis. Daily trophallaxis event means were compared within worker rearing groups using a repeated measures ANOVA using trial and treatment as factors.

To assess the impacts of 9-ODA exposure on the worker's ability to feed drones, we measured hypopharyngeal gland size in callow or nurse bees in trials one and two of the second cage study, using a protocol described in. Six workers were collected across at least four cages per treatment after 72 hours of treatment. For each individual, the left hypopharyngeal gland was dissected in molecular grade water and eight random acini were measured lengthwise using an ocular micrometer at high magnification. Acini measurements were averaged per individual and means were compared across treatment groups using an ANOVA with trial and treatment as factors.

Effect of 9-ODA exposure on drone sexual maturation and reproductive behavior

Two single cohort colonies with equivalent amounts of callow workers, two nectar frames, one pollen frame and two empty frames were established in tandem from a single colony source. 100 newly emerged drones were individually number tagged (Betterbee, Greenwich, NY) on their thorax and placed within each colony, marking the start of the experiment. One colony received a daily application of 200μg, or one queen equivalent, of synthetic 9-ODA while the other colony received a solvent control. The pheromone treatment was hung in a central location within the colony and allowed to diffuse throughout. The experiment was replicated in a second trial with a new colony source.

Each colony was outfitted with long runways to facilitate the observation of drone flights. Colonies were observed daily and continuously during typical drone flight times, with observations ceasing two days after the initiation of drone flights since all the drones had taken flights at this point. Individual drone flights from and returning to the colony were documented, allowing us to discern the age of initial flight and flight rates for all individuals. The cumulative proportion of drones flying across treatments was compared using Kaplan-Meier survival analysis across both trials. The average number of daily flights taken per flying individual was compared using a Kruskal-Wallis Rank Sums test or a Welch’s ANOVA and in accordance with data adherence to test assumptions of normality and variance.

Material and methods for Objective 2:

General Bee Rearing

These studies were performed during the summer of and 2015 and 2016 at the research apiaries of Pennsylvania State University (University Park, PA) and the research laboratory at the I. Meier Segals Garden for Zoological Research at Tel Aviv University (Ramat Aviv, Israel). Drones were reared in strong colonies headed by naturally mated queens according to standard practices. Briefly, eight frames of drone honeycomb were introduced into the brood nest of 8 colonies in advance of planned experiments. This facilitated the laying of large number drone eggs and the rearing of large numbers of drone brood, which were expected to reach sexual maturity in just over a month.

In anticipation of performing the attraction studies and collecting drones for extract formulation, colonies with previously introduced drone comb were checked for the presence of a large adult drone population. The sexual maturity of drones in these colonies was confirmed by visually observing the entrance of these colonies for large numbers of drones leaving and returning from mating flights, but also by waiting until the first generation of drones emerging from introduced drone comb were at least 12 days old, the age at which drones are expected to be fully sexually mature and at which they have already begun to take consistent mating flights.

Extract Preparation

Sexually mature drones were collected from the entrance of colonies with introduced drone comb as they were returning from mating flights. Drones were collected on wet ice and dissected in molecular grade water and at room temperature, within 3 hours of being collected. For each drone dissected, the paired mandibular glands were separated from the mandible at their base and gently extracted in 100μl of diethyl ether. Glands were extracted for 24 hours under dark, ambient conditions after which the solvent was separated from the tissue and stored at 4°C until it was used. The glands from 20 drones were extracted to produce an individual aliquot of natural extract for behavioral assays; the number of extracts tested is provided in the description of the assays below.

The extract for GC-MS (gas chromatography coupled mass spectrometry) analysis was similarly obtained but consisted of the extracted glands of 40 sexually mature drones. The solvent was gently evaporated under nitrogen stream and the residue was silylated for two hours using 5μl of the derivatization reagent BSTFA (N,O bis- trimethylsilyl-trifluoroacetamide) (Sigma Aldrich, St. Louis, MO). After the reaction was allowed to take place, the sample was reconstituted in 100μl of analytical grade hexane and analyzed by GC-MS. Note that an internal standard was excluded in order to prevent obscuring volatile peaks, favoring information on total number of compounds over absolute quantification of the compounds.

Drone Mandibular Gland Extract Characterization

GC-MS analysis was conducted on the gland extract and method control (sample containing only solvents and reagents in gland extracts but no glands) using an Agilent 7890 GC with a 5975C MS equipped with an Agilent HP-5MS column (30m x 0.25mm, 0.25μm phase thickness). The temperature program used was 40°C for 2 minutes followed by a 15°C per minute increase to 305°C, which was held for 5 minutes. Initial compound identification was carried out using the National Institute of Standards and Technology’s mass spectral library and confirmed using comparisons with mass spectra of analytical standards when available and/or deduction using established compound fragmentation rules and retention indices. Peak integration was performed using ChemStation software (Agilent Technologies Inc., Santa Clara, CA) to identify the number of compounds present in the extract and assess relative quantities of each compound based on peak area (calculated as the area of a given peak relative to the total area of all identified peaks x 100). Peaks present in the method control and/or identified as solvents, reagents or contaminants were excluded from this analysis.

Attraction Assays: Outside the Colony

Drone flyways, or corridors which drones consistently use to travel to and from their colony apiary, were visually identified and an adjacent field between two flyways was selected for the attraction assays. Whatman paper discs (42.5mm diameter) were impregnated with 20 drone equivalents of drone mandibular gland extract or 100μl of a solvent control (diethyl ether). The solvent was allowed to evaporate and the discs were placed in a 16 gauge wire mesh cage and immediately hoisted 10 meters in the air using telescoping poles. Lures were set up 20 meters apart and were individually video recorded for 10 minutes. In total, we performed pairwise comparisons of DMG extract and a solvent control. The videos of the extracts and solvent control lures were observed frame by frame for at the tenth minute. The number of drones coming within two meters of a lure (long-range attraction) and reorienting in response to the lure (short range attraction) was documented. Honey bees navigating a biologically relevant odor plume will often reorient in an attempt to re-align themselves to track the odor source, hence we used this behavior to identify detection and short-range attraction to the contents of a lure.

Synthetic blends were similarly tested. We hypothesized one or more of the most abundant compounds in the drone gland might be responsible for a portion of the biological activity previously attributed to the mandibular gland extract of the drone. The blend we composed consisted of a mixture of the six compounds found in highest proportion in the mandibular gland extract (hereafter termed 'primary gland components'). Their relative quantities as found in the gland extract were used to create a blend containing quantities of each compound in similar proportion to each other. Lures were once again impregnated with a solvent control or 300μg of the synthetic blend, a quantity we hypothesize to be high and representative of significantly more than what is present in the glands of 20 drones, the quantity tested using natural extracts. We performed pairwise comparisons of synthetic extract and a solvent control, as with the natural extracts. Outdoor attraction assays with natural extract and the synthetic blend were replicated seven times.

Attraction Assays: Inside the Colony

Attraction of workers to natural drone mandibular gland extracts versus a solvent control was tested in a colony setting. A three-frame queenright observation hive was set up in a dark shed, with workers allowed to exit and forage through an entrance tube connected to a window. All drones were removed from the observation hive. To test worker attraction to DMG extracts, small squares of Whatman paper (2cm x 2xm) were impregnated with a 100μl of a given treatment (constituting 20 drone equivalents or a solvent control), the solvent was allowed to evaporate and they were then pinned onto a central brood frame, where nurses might be expected to spend most of their time. The observation hive was closed and the two lures were video recorded for 10 minutes under red light. The number of workers antennating or licking each treatment lure was documented over the course of a minute, and scored at the tenth minute of recording, as in the field lure experiments. In-colony attraction assays were repeated 15 times using workers from three unrelated colonies.

Statistical analysis

Counts at each observation period for each treatment served as the raw values for our analysis and were treated independently, as in. Attraction count data was analyzed using a student’s t-test or a non-parametric Wilcoxon Rank Sums test (when there were parametric test assumption violations of the data). JMP 9.0.2 (SAS Institute Inc.) was the statistical software used for all analyses.

Research results and discussion:

We have finished completing all outlined objectives since our last update.

In the studies outlined in Objective 1, we have found that 9-ODA exposure impacts Vg expression in the fat body of drones (Figure 1), in the same way that it does workers. As Vg has been implicated in specific pathways that play a role in regulating behavioral maturation in workers, we assessed whether the onset and frequency of drone reproductive behavior was affected by exposure to 9-ODA. We find that indeed, 9-ODA exposure appears to have a subtle but significant effect in terms of delaying the age that adult drones begin taking mating flights (Figure 2). More impressively (and quite unexpectedly) the rate at which exposed vs. unexposed drones take mating flights is robustly different.

In the studies outlined in Objective 2, we have, to our knowledge, identified for the first time, the majority of the individual compounds drones are producing in their mandibular glands (organic mid-length acids and alcohols were found in greatest proportion). We employed field trials to test the biological activity of natural mandibular gland extracts from drones, as well as a synthetic blend consisting of the most highly produced compounds in the gland. We found that the extracts were significantly more effective at eliciting long- and short-range responses from drones than a solvent control (Figure 3).

The aims for Objective 3 include the dissemination of our findings through publication in peer-reviewed journals and at professional and stakeholder meetings. See ‘Publication/Outreach’ section for a detailed summary of the outcomes pertaining to this objective. figures

Research conclusions:

include any outreach or publications.

Our findings for Objective 1 have broad potential impacts for basic and applied research and innovation. By identifying a physiological effect of 9-ODA exposure, we have characterized a new primer pheromone in honey bees, furthermore, one that impacts males. This is an incredibly rare phenomenon in animals and will require further inquiry to clearly identify the extent to which male honey bees are impacted by their chemical environment in the hive. We have also identified a potential pathway through which the impacts of 9-ODA may be mediating its effects on drone reproductive behavior. This is a promising lead that directly addresses one of our original aims to this study: to identify potential pheromonal tools that might allow us to control different aspects of honey bee reproductive behavior.

Outcomes from Objective 2 have very exciting implications for the study of honey bee chemical ecology. We have demonstrated that drones produce a large suite of compounds in their mandibular glands. This now motivates a renewed interest in drone biology in the research community. Questions regarding the biological activity and use of these compounds by drones, other workers, and the queen will be important to address going forward. Our findings here also address our original aim. We have found that within the suite of chemicals in the drone mandibular gland, one or more compounds are effective at recruiting drones over a distance. This suite of chemicals should be considered when developing chemical tools to manipulate the presence of drones at a specific mating site.

Participation Summary

Education & Outreach Activities and Participation Summary

Participation Summary:

Education/outreach description:

Publication and outreach are the specific aims of Objective 3. We are confident in our progress/outcomes for this objective:

The outcomes of Objective 1 have been developed into a manuscript. This manuscript was submitted to the journal Animal Behavior in October 2016 and is currently under review.

The outcomes of Objective 2 have been developed into a manuscript. We are awaiting final comments and approval from our research collaborators and expect to submit this manuscript to the Journal of Chemical Ecology in December 2016.

Outreach/dissemination of the findings from Objectives 1 and 2 (in addition to publication) took place at the following professional and stakeholder meetings/workshops:

  1. “Compounds of the Drone Mandibular Gland”

Pennsylvania State Beekeeper’s Association (PSBA) Annual Meeting – 2016

Audience: Stakeholder

  1. “Behavioral and Chemical Characterization of Drone-produced Mandibular Gland Compounds in the Honey Bee”

International Conference on Pollinator Biology, Health and Policy – 2016

Audience: Professional: 80%, Stakeholder: 15%, Industry: 5%

  1. “Primer Effects of a Queen Pheromone on Drone Physiology and Behavior”

American Bee Research Conference – 2016

Audience: Professional: 60%, Stakeholder: 30%, Industry: 10%

  1. “Honey Bee Drones: New Insights Into the Chemical Ecology of the Male Caste”

Agricultural Progress Days, Pennsylvania State University – 2016

Extension Event

Audience: Stakeholders, Farmers, Local Community

  1. “Consequences of Social Living: Indirect Effects of a Honey Bee Pheromone on Male Physiology and Behavior”

Annual Meeting of the Animal Behavior Society (ABS) – 2014

Audience: Professional




Project Outcomes

Project outcomes:


Farmer Adoption

Beekeepers are the primary stakeholder group that stands to most directly relate to and benefit from our findings. We have been able to engage locally and nationally with this group through our presentations. We expect the publication of our outcomes to significantly bolster our exposure to the stakeholder and professional communities.

Assessment of Project Approach and Areas of Further Study:

Areas needing additional study

In both Objectives 1 and 2 we identified compounds that may affect reproductive development and behavior in male drones. This opens the door to more deeply explore the limits that these and other individual or suites of compounds may have in doing so. The fact the 9-ODA exposure reduces the frequency of mating attempts by young adult drones is particularly promising because its presence in the hive is easily manipulated and every stakeholder is already knowledgeable on how to do so (9-ODA can be easily reduced in the colony by removing the queen or transplanting drones to a queenless colony). It is important to note that we only considered the effect of one queen-produced compound on drones, and actually produces hundreds of compounds. Whether a larger complement of queen compounds can have even more robust effects in this regard is an outstanding question that warrants future attention. Our finding that drones produce a large suite of compounds and that other drones are receptive to these compounds is also important. Future studies should use our identified list of compounds to pursue several major lines of research driven questioning: In what social contexts are drones utilizing these compounds to communicate? What individuals are receptive to and respond to drone-produced compounds? What are the individual compounds that elicit specific and consistent responses in conspecifics (and are these compounds acting alone or as a blend)?

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.