From Society to Genes with the Honey Bee

Gene E Robinson. American Scientist. Volume 86, Issue 5. Sep/Oct 1998.

On September 9,1997, an article in The New York Times announced the discovery of the “first gene for social behavior.” Anthony Wynshaw-Boris, of the National Human Genome Research Institute, and his colleagues had discovered odd behavior in laboratory mice lacking a gene called disheveled-1. These mice interacted and huddled with others less than normal, and they failed to perform an important social duty, trimming the whiskers of fellow mice. Whether or not this is really the first gene discovered “for social behavior,” no one should lean toward the notion that genes play an exclusive role in regulating behavior. Biologists long ago came to realize that behavior is influenced by genes, the environment and interactions between the two. To better understand this regulatory combination, scientists can turn to an organism, such as the honey bee, whose behavior can be studied in the field under natural conditions.

A discussion of “genes for behavior” might raise anxiety over the implications of attributing so much control to strings of nucleic acids, or DNA. In particular, some people fear that the concept of biological determinism-the notion that genes play a dominant, if not exclusive, role in regulating behavior-might creep in and diminish our appreciation for the role of the environment in shaping behavior. Nevertheless, genes never act alone. They must operate in an environment, where they code for proteins that participate in many systems in an organism. In fact, genes themselves depend on many of those proteins for replicating DNA and linking together amino acids, which are the fundamental units of proteins. Consequently, biologists need to take a broad approach in assessing the impact of any gene.

To properly appreciate the influence of genes on behavior, we need behavioral studies that demonstrate-at the molecular level-the influences of genes, the environment and their interactions. Social behavior is ideally suited for this challenge because it is especially sensitive to environmental influence. Moreover, these influences are in many cases mediated by specific social signals communicated from individual to individual, which can make them easier to study experimentally. Molecular-genetic studies of social behavior will show how an animal’s phenotype, which indudes social behavior, arises from both its genotype and environment. Making that connection, however, requires identifying genes that influence social behavior, revealing how those genes regulate the neural and endocrine mechanisms through the production of proteins, and, finally, exploring how specific manipulations of an animal’s social environment affect gene expression.

My research group uses the Western honey bee, Apis mellifera, to understand how genes and the environment govern social behavior. As I shall show, we study the development of naturally occurring social behavior, from society to gene. Honey bees are particularly useful for studying social behavior because, like humans, they experience behavioral development. In other words, honey bees pass through different life stages as they age, and their genetically determined behavioral responses to environmental and social stimuli change in predictable ways. Often these responses increase in complexity and involve learning. We hope to explain the function and evolution of behavioral mechanisms that integrate the activity of individuals in a society, neural and neuroendocrine mechanisms that regulate behavior within the brain of an individual, and genes that influence behavior by encoding these mechanisms.

Basics of Bee Behavior

The so-called social insects, including honey bees, live in societies that rival our own in complexity and internal cohesion. For instance, honey bees always follow three rules: They live in colonies with overlapping generations, they care cooperatively for offspring other than their own and they maintain a reproductive division of labor. A colony arises from a queen that performs one task, laying lots of eggs, sometimes as many as 2,000 in one day. Her daughters, called workers, basically take care of the colony-doing everything from foraging for food to building the hive-but they generally do not reproduce. As one might expect, it takes many workers to run a hive, and some honey bee colonies consist of as many as 60,000 workers. Finally, a colony’s males, called drones, can usually be found in the hive, where they do essentially nothing. The drones specialize in reproduction, which takes just a couple of hours on a sunny day when they fly to mating areas away from the hive. Once a drone mates, he dies.

A further division of labor exists among the workers. Although a worker’s adult life span is just four to seven weeks, it undergoes a series of transitions. A worker usually spends its first few weeks tending to duties in the hive and its last few weeks foraging for food outside the hive. During the hive phase, a worker starts out with a couple of days of cell cleaning, literally removing debris from cells in the hive that are used to raise other bees or to store food. Next, a worker serves as a nurse, caring for and feeding larval bees. Toward the end of the hive phase, a worker spends its time processing and storing food and maintaining the nest, including building new sections of hive. Some workers also perform a few other tasks along the way, including guarding the hive or removing corpses. Finally, a worker switches to foraging, which is probably the most challenging task of all. To be a successful forager, a bee must learn how to navigate in the environment and obtain nectar and pollen from flowers. Foragers also communicate the location of new food sources by means of the famous “dance language.” These transitions in occupation typically do not arise abruptly. For example, a worker might slowly decrease its nursing duties and become gradually more involved in maintaining the hive.

Behavioral development in honey bees is a powerful system for integrated analysis. Although it occurs naturally in the field, some underlying mechanisms can be analyzed in the laboratory. Moreover, honey bees have been closely associated with humans for millennia because of their special status as prolific producers of honey and wax and as premier pollinators for our food and fiber crops. As a result, we know more about honey bees than just about any other animal on earth. One consequence of this wealth of knowledge is that the natural social life of honey bees can be extensively manipulated with unparalleled precision.

Colony Adjustments

Although worker bees go through a rather consistent path of behavioral development, it is not rigidly determined. Bees can accelerate, retard or even reverse their behavioral development in response to changing environmental and colony conditions. For example, favorable environmental conditions in the late spring might cause a surge in worker birth rates, and that could resuit in a colony with a reduced percentage of foragers. Under these circumstances, young bees compress their period of hive work from three weeks to one week and become “precocious foragers.” Conversely, a new colony founded by a swarm-a fragment of an old colony that leaves to establish a new colony-soon reaches a point at which it contains predominantly older individuals. In that case, some colony members retard their development and serve as overaged nurses. In those bees, hypopharyngeal glands that produce food for larvae continue with this function rather than producing other substances.

How does the behavior of thousands of individual bees generate a smoothly functioning colony? It seems unlikely that individual bees could monitor the state of their entire colony and then perform the tasks that are needed. Although some workers play special roles in organizing specific tasks, such as leading other bees to a new nest site during swarming, there is no evidence for real leaders or individuals-not even the queen-that perceive all or most of a colony’s requirements and direct the activities of other colony members from one task to another. The challenge is understanding the mechanisms of integration that enable individual bees to respond to fragmentary information with actions that are appropriate to the state of the whole colony.

My colleagues and I and others have discovered that juvenile hormone-one of the most important hormones influencing insect development-helps to time the pace of behavioral maturation in honey bees. This hormone comes from the corpora allata, a gland that lies near a honey bee’s brain. Indirect evidence for this hormone’s role exists in the fact that young bees working in a hive have low levels of this hormone and older foragers have higher levels. Direct proof has also been obtained: Young bees given juvenile-hormone treatments become precocious foragers. Recently, my University of Illinois colleague Susan Fahrbach, graduate student Joseph Sullivan, undergraduate Omar Jassim and I found that removing the corpora allata does not prevent a bee from developing into a forager but does delay it for a few days on average. Juvenile-hormone treatments, however, eliminate that delay.

Manipulating hormone levels on a bee-by-bee basis is one thing, but demonstrating that bees alter hormone levels themselves in response to changing conditions is another. To show how the environment can modulate hormone levels, Robert Page of the University of California at Davis, Colette and Alain Strambi of the Centre National de la Recherche Scientifique in Marseille, France, and I induced precocious foraging by establishing colonies that consisted of only very young bees. Then we tested their blood levels of juvenile hormone and found that one-week-old precocious foragers had approximately 100 nanograms of juvenile hormone per milliliter of blood, which is about the same as that in three-week-old foragers and higher than the 5-20 nanograms usually found in one-week-old nurses. Two weeks later, we obtained overaged nurses from these colonies by preventing new adults from emerging, and these old nurses had levels of juvenile hormone that resembled young nurses, rather than foragers. From work with other experimental colonies, we found that bees that reverted from foraging to nursing were also “young” in terms of their levels of juvenile hormone.

Inhibitory Interactions

How do bees perceive changes in colony needs and adjust their behavioral development to perform the tasks most in demand? Postdoctoral associate Zachary Huang and I found that the rate of endocrine-mediated behavioral development is influenced by inhibitory social interactions. That is, older bees inhibit the behavioral development of younger bees. Bees reared in isolation in a laboratory for seven days have forager-like levels of juvenile hormone and forage precociously when placed in colonies. By carefully manipulating a colony’s age demography but keeping other characteristics unchanged, we found that the rate of behavioral development is negatively correlated with the proportion of older bees in a colony So depleting a colony’s foragers stimulates younger bees to forage earlier than normal. Conversely, younger bees forage later than normal if a colony’s foragers stay in the hive for several days because a sprinkler aimed at the hive entrance makes them think it’s raining.

Someone might imagine that bees could learn about their colony’s condition by monitoring the combs in their hive. For instance, a young bee might notice a food shortage in the combs, which might result in a neuorendocrine response that triggers precocious behavioral development. To explore the possibility that bees pay attention to the combs in this way, Huang, graduate student David Schulz and I recently tested the effects of starvation on the rate of behavioral development. Young bees from starved colonies do start foraging a few days earlier than bees from well-fed colonies. This starvation effect, however, is not mediated by perceiving a shortage of food in the honeycomb. We showed this by keeping a colony well fed from a sugar feeder while we constantly-but discretely-vacuumed any stored food out of their honeycomb. This was accomplished by drilling small holes at the base of each honeycomb cell. Well-fed bees in an empty hive started to forage at ages similar to bees in colonies with ample food stores and not nearly as early in life as did bees in truly starved colonies.

Inhibitory social interactions that influence the rate of behavioral development involve chemical communication between colony members. This is strikingly similar to pheromone regulation of sexual maturation in rodent societies. For example, a queen’s mandibular glands produce a pheromone that inhibits behavioral development. (See “The Essence of Royalty: Honey Bee Queen Pheromone” by Mark Winston and Keith Slessor in the July-August 1992 issue of American Scientist.) Queen mandibular pheromone has been known for some time to exert long-lasting effects on worker physiology and behavior by inhibiting the rearing of new queens. More recently, Mark Winston and Tanya Pankiw of Simon Fraser University, Huang and I demonstrated that queen mandibular pheromone depresses blood levels of juvenile hormone and, more important, delays the onset of foraging.

The primary modulator of behavioral development, however, appears to come from the workers themselves. The mandibular glands of workers contain compounds similar to those found in queen mandibular glands. Huang, Erika Plettner, a graduate student at Simon Fraser University, and I recently found that there must be direct social contact between bees for older ones to inhibit the development of younger ones. Moreover, older bees with their mandibular glands removed do not inhibit behavioral development. The mandibular glands of workers contain compounds similar to those found in queen mandibular glands. The inhibition that results from worker-worker interactions might come from exchanging a pheromone, which might be in the mandibular glands or somewhere else. When we removed the glands, that could have eliminated the inhibition because it removed the source of the pheromone or it simply blocked the pheromone’s flow from another location. Clearly, more work must be done here.

Brain Remodeling

How does a bee’s brain support the striking changes in behavior that take place during maturation? A small part of the answer lies in the mushroom bodies, a brain region thought to be the center of learning and memory in insects. Graduate student Ginger Withers, Fahrbach and I discovered about a 20 percent increase in the volume of a specific area of the mushroom bodies as worker honey bees mature. This volume increase occurs in a mushroom body subregion where synapses, or connections, are made between neurons from other brain regions that are devoted to sensory input. This was the first report of such brain plasticity in an invertebrate, and it was particularly exciting because volume increases in brain regions in vertebrates reflect increases in certain cognitive abilities.

It seemed that the increase in the mushroom bodies might be learningrelated. Young workers take orientation flights prior to the onset of foraging to learn their way around outside the hive, and the increase in volume in the mushroom bodies begins at that time. To test flying’s effect on mushroombody volume, Withers, Fahrbach and I made what we called “big-back bees.” By attaching a large tag to each bee’s back and placing a screen at the hive’s entrance, we prevented some workers from flying out of the hive but allowed them to interact with other bees. Bigback bees showed normal increases in mushroom-body volume despite their deprivation from orientation flights. So far, the volume increase is unstoppable. Fahrbach, Darrell Moore of East Tennessee State University, graduate student Sarah Farris, postdoctoral associate Elizabeth Capaldi and I showed that it takes place even in bees reared in social isolation and complete darkness in a laboratory.

Still, it might be premature to exclude the idea of a connection between the plasticity of the mushroom bodies and orientation flights in honey bees. Our results indicate that a bee’s mushroom bodies need not increase because of taking orientation flights, but we have not ruled out a volume increase that prepares a bee for those flights. In other words, the mushroom bodies might need to increase in volume to provide the necessary brain space for a bee to learn how to get around outside its hive, and how to get back.

After learning to orient outside the hive, a bee learns to forage, and that might also involve an increase in the mushroom bodies. Withers, Fahrbach and I showed that the mushroom bodies increase in volume more rapidly in precocious foragers than in nurse bees of the same age. This result has been confirmed in the laboratory of Randolf Menzel in Berlin, using a somewhat different neuroanatomical analysis. These results suggest that the structure of the mushroom bodies might be sensitive to changes in social context that are associated with the onset of foraging.

While we continue our efforts to unravel the significance of a volume increase in the mushroom bodies, we also wonder how the region gets bigger. The number of cells in the mushroom bodies is highly stable in adult life. The production of new neurons is not detectable, and there is no evidence for cell death, according to research with Fahrbach that was performed by undergraduates Jennifer Strande and Jennifer Mehren. Accordingly, the volume increase in the mushroom bodies probably represents an increased arborization of some subpopulation of brain cells that already exists. This increased proliferation of neuronal branches would likely result in an increase in the number of synapses per neuron, which would impact the processing of information in the mushroom bodies.

Beyond structural changes in a worker bee’s brain, neurochemical analyses have revealed striking changes in levels of biogenic amines, which are well known as modulators of nervous-system function and organismal behavior in animals, including humans. Alison Mercer, her colleagues from the University of Otago in New Zealand and I found changes in brain levels of two biogenic amines-dopamine and serotonin-during behavioral development. Jeffrey Harris and Joseph Woodring at Louisiana State University reported similar findings. Recently, graduate students Christine Wagener-Hulme and David Schulz, research technician Jack Kuehn and I showed that another biogenic amine, octopamine, appears to be most important in honey bee behavioral development. When a bee receives treatments of juvenile hormone, levels of octopamine increase, but dopamine and serotonin do not. Looking specifically at the antennal lobes, a brain region that receives sensory information from a bee’s antennae, we found high levels of octopamine in the antennal lobes of foragers as compared with nurse bees, regardless of worker age. In contrast, levels of all three amines in the mushroom bodies are intimately associated with worker age, but not behavioral status.

These results suggest that octopamine might influence behavioral development by modulating a bee’s sensitivity to the stimuli that elicit the performance of age-specific tasks. We presume that these stimuli are mostly chemical, because bees live in a dark hive and possess modest auditory acuity, at least relative to their renowned chemosensory prowess. That is why we are so encouraged to find behaviorally related changes specifically in the antennal lobes. The hypothesis that octopamine is playing a causal role in behavioral development is currently being tested by chronic administration of octopamine to the brain, followed by behavioral assays. Studies of biogenic amines might also provide some of the missing links between endocrine regulation and behavioral development in honey bees.


Molecular-genetic research in my laboratory has only recently begun, and it currently involves selecting candidate genes and exploring their possible involvement in social behavior. This is done by studying whether differences in social behavior-within and between individuals-are correlated with variation in gene-transcription regulation, gene structure or both. Graduate student Daniel Toma and I are exploring the role of the period (per) gene in honey bee behavioral development. In the fruit fly Drosophila melanogaster, per is a principal component of the fly’s circadian clock. The protein that per encodes is thought to help create circadian rhythms of activity by orchestrating the transcription of other genes according to a precisely timed schedule. We chose per because we have found intriguing links between division of labor and circadian behavioral rhythms in honey bees.

Moore, Fahrbach, undergraduates Iain Cheeseman and Jennifer Angel and I discovered that foragers have pronounced circadian rhythms of activity-including being more active during the day than the night-but workers in the early part of their hive phase do not show such rhythms. This difference is obvious both in beehives and in assays of individually isolated bees in the laboratory For example, consider the behavior of nurse bees that need to feed bee larvae around the clock. How is this accomplished? If nurse bees have a circadian rhythm of brood care, one might expect to find evidence of “shift work,” or groups of nurse bees on different schedules. Alternatively, if nurse bees perform brood care with no circadian rhythm, one might expect to find them performing it randomly with respect to time. Monitoring individually tagged bees every three hours around the clock in glass-walled observation hives, we found no evidence of rhythmicity or shift work for nursing. Nurse bees were arrhythmic in the performance of this task.

Moreover, genetic factors appear to influence both the plasticity in a honey bee’s behavioral development and its circadian rhythm. Using colonies composed of workers with identifiable genotypes, graduate student Tugrul Giray and I learned that workers of some genotypes are more likely to consistently mature rapidly and forage precociously, even in different social environments, and workers of other genotypes are more apt to develop into overaged nurses. Working with such “fast” and “slow” genotypes, Moore, Giray and I found that fast-genotype bees developed a circadian rhythm of locomotor behavior in the laboratory at younger ages than did slow-genotype bees. Fast-genotype bees also had a periodicity to their rhythm of locomotor activity that was faster than that of slow-genotype bees. These results are reminiscent of the strikingly diverse effects of the per gene in fruit flies. This gene governs not only circadian rhythms in flies but some elements of their mating song as well.

Toma and I have cloned a putative bee homologue of the per gene. We hope to determine whether variation in gene expression and gene structure is correlated with variation in the ontogeny of behavioral circadian rhythmicity and the rate of behavioral development. The next step is to manipulate a hive’s age demography to determine the association between gene expression and behavioral status.

Another approach to discovering genes involved in a honey bee’s behavioral development is to build on our neuroanatomical work by identifying molecular mechanisms that contribute to the increase in the volume of the mushroom bodies. We hope to determine whether variation in the expression or structure of genes preferentially expressed in a honey bee’s mushroom bodies affects the functioning of this brain structure, which in turn may affect behavioral development in honey bees.

Luckily, it is relatively easy to find behavioral variation that is correlated with genotypic variation in honey bees. Several laboratories, including mine, have demonstrated that variation in a worker’s genotype influences many aspects of the division of labor in honey bee colonies, including the tendency to specialize in rare tasks such as removing corpses and guarding a nest’s entrance. In addition, controlled mating via instrumental insemination facilitates research on honey bee behavioral genetics.

The molecular-genetic analysis of social behavior in general is a fertile new field. Results of classical quantitative genetic studies have indicated that there are strong correlations between genetic variation and variation in social behavior among individuals, but none of the genes has been identified. (Much more work must be done on disheveled-1 before it could meet these criteria.) The idea that gene expression in the brain is sensitive to social context is supported by recent findings from the laboratories of Fernando Nottebohm of Rockefeller University and David Clayton of the University of Illinois on bird song, Donald Pfaff of Rockefeller University and Thomas Insel of Emory University on rodents, Russell Fernald of Stanford University on cichlid fish and Edward Kravitz of Harvard University and Donald Edwards of Georgia State University on lobsters. I propose that two-way interactions between the nervous system and the genome contribute fundamentally to the control of social behavior. Information about social conditions that is acquired by the nervous system is likely to induce changes in genomic function that in turn adaptively modify the structure and function of the nervous system.

With the presence of abundant genetic variation in behavior and a growing selection of tools needed to exploit it, the prospects are good that honey bees can be used as a new model for molecular genetic analyses of social behavior. Nevertheless, I believe that the difficulty in studying the genetic basis of social behavior demands a bold, new initiative, which I call sociogenomics. In essence, this means taking a wide-ranging approach to identify genes that influence social behavior, determining the influence of these genes on underlying neural and endocrine mechanisms and exploring the effects of the environment-particularly the social environment-on gene action. Implicit in the name sociogenomics is the realization that many genes must be studied simultaneously to decipher the complexity behind social behavior. Such an approach could be based on the revolutionary advances that are emerging from the Human Genome Project. For example, there are new techniques for sequencevariation analysis and simultaneously screening large numbers of genes for differences in expression that are correlated with differences in behavioral state that can contribute significantly to gene discovery in bees.

In continued studies of honey bees, investigators will probably find common mechanisms that govern life in both invertebrate and vertebrate societies. If so, the identification of genes influencing social behavior in honey bees-guided by our emerging understanding of the underlying neural and endocrine mechanisms-will likely yield insights that go well beyond a beehive.