Influence of Gene Action Across Different Time Scales on Behavior

See allHide authors and affiliations

Science  26 Apr 2002:
Vol. 296, Issue 5568, pp. 741-744
DOI: 10.1126/science.1069911


Genes can affect natural behavioral variation in different ways. Allelic variation causes alternative behavioral phenotypes, whereas changes in gene expression can influence the initiation of behavior at different ages. We show that the age-related transition by honey bees from hive work to foraging is associated with an increase in the expression of the foraging (for) gene, which encodes a guanosine 3′,5′-monophosphate (cGMP)–dependent protein kinase (PKG). cGMP treatment elevated PKG activity and caused foraging behavior. Previous research showed that allelic differences in PKG expression result in two Drosophila foraging variants. The same gene can thus exert different types of influence on a behavior.

Some genes influence behavior via genetic polymorphisms, whereas other genes influence behavior via developmental polymorphisms. But little is known about whether the same gene, or orthologs of a gene, can influence behavior in both ways. This knowledge is necessary to develop a comprehensive understanding of how genes and the environment influence behavior, because both involve genomic responsiveness, albeit over vastly different scales of time.

The foraging gene (for) affects naturally occurring variation in insect behavior (1). Twofor alleles are commonly present in populations ofDrosophila melanogaster: forR (rover) flies have higher levels of for mRNA and PKG activity and collect food over a larger area than dofors (sitter) flies. Patchy food and high population densities provide a selective advantage for rovers; more uniformly distributed food and low population densities favor sitters (2). These results suggest that behavioral evolution in flies has involved selection for alternative for alleles under different ecological conditions.

We used the honey bee (Apis mellifera) to study the possibility that for also is involved in developmentally regulated behavioral variation. Unlike in flies, foraging in honey bees unfolds as part of a complex process of behavioral maturation, and in a social context. Honey bee colonies exhibit an age-related division of labor; adult worker bees perform tasks in the hive such as brood care (“nursing”) when they are young, and then shift to foraging for nectar and pollen outside the hive. The transition to foraging typically occurs at about 2 to 3 weeks of age, is preceded by a series of orientation flights, and involves changes in brain chemistry, brain structure, endocrine activity, and gene expression (3). The age at onset of foraging is not rigid; it depends on the needs of the colony, mediated in part by inhibitory social interactions with older individuals and pheromones from the brood and queen (3). Foraging in honey bees is also different from flies because foragers collect food for their colony, and not necessarily when they themselves are hungry (4).

We hypothesized that foraging in honey bees is associated with an up-regulation of the for transcript in the brain, with foragers having higher levels than nurses. This hypothesis was based on the notion that nurse bees loosely resemble sitter flies because they obtain food only in the more restricted confines of the beehive, whereas forager bees display rover-like behavior by ranging widely throughout the environment. Specifically, we investigated whether the same gene that results in alternative allelic-based phenotypes (sitters and rovers in Drosophila) is also involved in developmentally regulated alternative phenotypes, nursing, and foraging in honey bees.

To test this hypothesis, we cloned a honey beefor ortholog (Amfor) (5). The predicted protein sequence of Amfor contains all regulatory, cGMP binding, and kinase domains typical of a PKG and is >80% similar to PKGs from other organisms (5). Northern blot analysis (6) indicated the presence of a single transcript in the head and suggested higher expression in forager heads relative to nurse heads (Fig. 1A). Real-time quantitative reverse-transcription polymerase chain reaction (qRT-PCR) (7) demonstrated that foragers had significantly higher brain levels of Amfor mRNA (by a factor of 2 to 8) than did nurses in all three colonies studied (Fig. 1, A and B). Foragers also exhibited about four times as much PKG activity as did nurses [34.9 ± 3.1 versus 12.5 ± 1.2 pmol min–1mg–1 protein (±SE), foragers and nurses, respectively;N = 8 heads per group, analysis of variance (ANOVA),P < 0.001; assayed as in (1)].

Figure 1

Differential expression of Amforduring honey bee behavioral maturation. (A) Northern blot analysis of nurses and foragers from a typical colony (5 μg of mRNA from five heads per lane). The same blot was probed withAmfor-specific probe, stripped, and reprobed withEf1α probe as RNA loading control. Lane 1, nurse head; lane 2, forager head. (B) qRT-PCR analysis ofAmfor expression in individual brains of nurses and foragers from three unrelated typical colonies (colony 1, bees of unknown age; colonies 2 and 3, nurses 7 days old and foragers >21 days old;N = 8 brains per group). Data are means ± SE (converted to the same arbitrary scale as the mean). Results of ANOVA for each trial are shown (**P < 0.01). Two-way ANOVA showed significant (P < 0.001) differences between nurses and foragers (task) overall, significant differences (P < 0.001) between colonies, and a significant task × colony interaction (P < 0.01). (C) qRT-PCR analysis ofAmfor expression in individual brains of nurses and precocious foragers (7 to 9 days old) from four unrelated single-cohort colonies. Sample sizes and analyses were as in (B). Two-way ANOVA showed significant (P < 0.001) differences between nurses and foragers (task) overall, significant differences between colonies (P < 0.001), and a significant task × colony interaction (P < 0.05). The data for each colony are normalized relative to a control gene (7) and hence cannot be compared in absolute terms. However, PKG activity data (see text) indicate similar levels for nurses in both typical and single-cohort colonies. Bee colonies can differ as a result of both genotypic and environmental factors, and these factors may have influenced the magnitude of the relative difference between nurses and foragers; differences were in the same direction in all seven colonies studied, and they were significantly different in six of the colonies.

These results are consistent with our hypothesis; however, foragers typically are also older than nurses. To resolve whether for up-regulation is associated primarily with foraging behavior or with the foragers' advanced age, we manipulated colony social structure to obtain precocious foragers. We established “single-cohort colonies” initially composed only of 1-day-old bees (8); the absence of foragers results in some colony members initiating foraging as much as 2 weeks earlier than usual (9). In support of our hypothesis, 7- to 9-day-old precocious foragers had significantly higher levels of AmformRNA (by a factor of 2 to 4) than did same-age nurses in three of four colonies (Fig. 1C).

We used a pharmacological approach to test the hypothesis that increased PKG activation causes an increase in the likelihood of precocious foraging. Bees were chronically treated with 8-Br-cGMP (10), a membrane-permeable analog that is relatively resistant to degradative phosphodiesterases. As expected, the treatment significantly elevated PKG activity (Fig. 2A); treated bees had forager-like levels of PKG activity, whereas control bees had levels similar to nurse bees. This treatment significantly increased the likelihood of precocious foraging in a dose-dependent manner (Fig. 2B). In contrast, 8-Br-cAMP treatment, which elevated cAMP (adenosine 3′,5′-monophosphate)–dependent protein kinase activity (11), did not elevate PKG activity and did not affect the likelihood of precocious foraging (Fig. 2C). These results demonstrate a specific treatment effect and suggest that PKG activation can influence the initiation of foraging behavior.

Figure 2

Effects of treatment with cGMP or cAMP on honey bee foraging behavior. (A) Treatment with 8-Br-cGMP (500 μM), but not 8-Br-cAMP (1000 μM), significantly increased PKG activity [ANOVA, N = 8 heads per group; PKG activity measured as in (1)]. (B) Treatment with 8-Br-cGMP induced precocious foraging. Two trials were performed with no significant differences between them (P = 0.62), allowing the data to be pooled. P value on the graph is based on a survival analysis for dose dependence [Cox proportional hazards test, see (28)]; N = 35 to 45 bees per trial for each treatment. One-day-old bees were treated for 4 days in the laboratory and then introduced to single-cohort colonies. Observations at the hive entrance were made (9) to ensure that the onset of foraging was identified; the graph indicates the cumulative percentage of bees initiating foraging on each of the first 4 days of observation, when they were 4 to 10 days old. (C) Treatment with 8-Br-cAMP did not induce precocious foraging. Design and analysis were as in (B). One trial was performed: N = 44 to 49 bees per group. No significant (P = 0.22) effect on precocious foraging was found in a second trial with 3000 mM 8-Br-cAMP (11). •, 0 μM; ○, 100 μM; ▾, 250 μM; ▿, 500 μM; ▪, 1000 μM.

In situ hybridization analysis was performed (12) to explore where Amfor might exert its effects in the brain (Fig. 3). Amfor is highly expressed in the lamina of the optic lobes and in the mushroom bodies. The mushroom bodies constitute the main center for multimodal sensory processing in the insect brain (13). In the mushroom bodies,Amfor is preferentially expressed in a central column of intrinsic (Kenyon) cells that receive mainly visual input (14). On the basis of these results, we speculate thatAmfor is involved in higher order integration of visual information associated with orientation and foraging behavior; involvement in other neural functions related to division of labor is also possible.

Figure 3

Amfor expression in the honey bee brain. OL, optic lobes; KC, Kenyon cells. (A) Coronal section. Squares delineate regions shown magnified in (B) and (C). No labeling was seen in sections probed with a sense control (11,12). There were no obvious spatial differences between nurses and foragers in expression patterns (N = 5 brains per group); these images are from a forager brain. Results suggest that the differences in Amfor mRNA levels between nurses and foragers detected with qRT-PCR may represent increased expression in the same cells. Brains were sectioned from anterior (antennal lobes) to posterior (subesophageal ganglion).

Division of labor in honey bees involves intricate processes that integrate the effects of age, social interactions, colony needs, and resource availability on the likelihood of engaging in foraging behavior. Other genes show changes in brain expression in association with the transition from hive work to foraging (3), and quantitative trait loci for pollen versus nectar foraging also have been identified (15). Our results suggest that the up-regulation of Amfor in the brain and the resultant increase in PKG activity is causally related to the transition from hive work to foraging outside. Hence, Amforapparently influences the division of labor in honey bees and is one of only a few genes implicated in the organization of an animal society (16, 17).

Both fly (1) and bee foraging involve for, and PKG plays a role in the control of feeding arousal in some other invertebrates and vertebrates (18–20). This suggests that the responsiveness of for expression over evolutionary (flies) and ontogenetic (bees) time scales reflects aspects of a phylogenetically conserved process of regulation of feeding. We propose that evolutionary changes in food-related behaviors, including complex social foraging, are based in part on changes in the regulation offor and other related genes. Given the importance of gene regulation in generating biological complexity, further studies offor and other genes that are both under selection and subject to regulation by extrinsic factors (21, 22) should provide important insights into the influences of genes on behavior.

  • * To whom correspondence should be addressed. E-mail: generobi{at}


View Abstract

Navigate This Article