Report

Queen Pheromone Blocks Aversive Learning in Young Worker Bees

See allHide authors and affiliations

Science  20 Jul 2007:
Vol. 317, Issue 5836, pp. 384-386
DOI: 10.1126/science.1142448

Abstract

Queen mandibular pheromone (QMP) has profound effects on dopamine signaling in the brain of young worker honey bees. As dopamine in insects has been strongly implicated in aversive learning, we examined QMP's impact on associative olfactory learning in bees. We found that QMP blocks aversive learning in young workers, but leaves appetitive learning intact. We postulate that QMP's effects on aversive learning enhance the likelihood that young workers remain in close contact with their queen by preventing them from forming an aversion to their mother's pheromone bouquet. The results provide an interesting twist to a story of success and survival.

To advertise her presence in the colony and to exert influence over its members, a honey bee queen produces a complex blend of substances known as queen mandibular pheromone (QMP) (1). Young workers, attracted to the queen by QMP are enticed not only to feed her, but also to lick and to antennate her body. As they do so, they gather samples of QMP, which they distribute to other members of the colony (2, 3). At the colony level, QMP inhibits the rearing of new queens (4), influences combbuilding activities (5), helps prevent the development of worker ovaries (6), and modulates the behavioral development of workers (7, 8).

We recently showed that brain dopamine levels, levels of dopamine receptor gene expression, and brain tissue responses to this amine are altered significantly in young workers exposed to QMP (9). Among its many functions, dopamine contributes to cellular events underlying learning and memory. In insects, dopamine signaling is essential for aversion learning (1012), which raises an interesting question: Does QMP, through its actions on brain dopamine function in young workers, alter their ability to establish aversive olfactory memories?

We examined QMP's impact on associative olfactory learning in bees exposed to QMP from the time of adult emergence. Bees of the same age maintained under identical conditions but without exposure to QMP were used as controls. To begin, we examined QMP's effects on aversive learning in young (6-day-old) bees. A differential conditioning paradigm was used to train bees to extend their sting to an odorant paired with an aversive stimulus (mild electric shock) and not to respond to a nonreinforced odorant (12, 13). Over successive conditioning trials, the percentage of control (untreated) bees responding with sting extension to the reinforced odorant (CS+) increased significantly (Fig. 1A, i), whereas the percentage of bees exhibiting sting extension in response to the nonreinforced odorant (CS–) declined [Fig. 1A, i, and supporting online material (SOM) text]. Comparison of the response curves shows that bees in the control group clearly differentiate between the two odorants (Fig. 1A, i). In a retention test performed 1 hour after the last conditioning trial, the bees responded significantly more to CS+ than to CS–(Fig. 1A, ii), which demonstrated that an aversive memory was established. In contrast, bees exposed to QMP failed to show aversive learning. There was no significant change in responsiveness to the reinforced odorant over successive conditioning trials (SOM text) and no difference between the response curves obtained for the two odorants (Fig. 1B, i). There was also no significant difference in QMP-treated bees between the levels of responses to the two odorants 1 hour after the last conditioning trial (Fig. 1B, ii).

Fig. 1.

Effects of QMP on aversive learning in 6-day-old workers. Associative olfactory conditioning of the sting extension response (SER) was used to compare aversive learning in (A) control bees and (B) bees treated with QMP. Bees were trained to discriminate between an odorant paired with electric shock (CS+, filled circles) and an odorant that was not reinforced (CS–, open circles). (A) (i) After 12 conditioning trials (6 with CS+ and 6 with CS–), control bees clearly discriminated between the two odorants [F(1,78) = 31.86, P < 0.0001]. (A) (ii) One hour after the last conditioning trial, a significantly higher percentage of controls responded with sting extension to the odorant associated with electric shock (black bar) than to the nonreinforced odorant (white bar) (χ2 = 18.05; P < 0.0001). (B) (i) QMP-treated bees did not learn to discriminate between the two odorants as a result of conditioning [F(1,78) < 0.0001, not significant (n.s.)]. (B) (ii) One hour after the last conditioning trial, the percentage of bees responding to the two odorants was similar (χ2 = 3.2, n.s.).

How selective are these effects? In Drosophila, synaptic output from dopamine neurons is important for aversive learning, but is not required for appetitive learning (10). On the basis of these findings, we predicted that, if QMP targets dopamine pathways selectively (9), appetitive learning in young bees should not be affected by QMP. We thus compared the ability of young QMP-treated bees and controls to associate an odorant stimulus with a sucrose reward. Appetitive learning was examined using the proboscis extension reflex (13, 14). QMP-treated bees and control (untreated) bees were trained to differentiate between an odorant stimulus paired with sucrose (CS+) and an odorant stimulus that was not reinforced (CS–). Over successive conditioning trials, the percentage of bees exhibiting proboscis extension in response to the reinforced odorant (CS+) increased significantly (controls, Fig. 2A, i; QMP-treated, Fig. 2B, i), whereas responses to the nonreinforced odorant (CS–) showed no significant change (controls, Fig. 2A, i; QMP-treated, Fig. 2B, i, and SOM text). The response curves show that bees in both groups clearly differentiate CS+ from CS–(Fig. 2). Moreover, in retention tests performed 1 hour after the last conditioning trial, both groups responded significantly more to CS+ than to CS–(controls, Fig. 2A, ii; QMP-treated, Fig. 2B, ii). These results are consistent with evidence that the reinforcing capacity of sucrose in associative olfactory discrimination tasks is mediated not by dopamine, but by octopamine (10, 15, 16), an amine that is not affected by the presence or absence of queen pheromone (17).

Fig. 2.

Effects of QMP on appetitive learning in 6-day-old workers. Associative olfactory conditioning of the proboscis extension response (PER) was used to compare appetitive learning in (A) controls and (B) bees treated with QMP. Bees were trained to discriminate between an odorant paired with sucrose (CS+, filled circles) and a nonreinforced odorant (CS–, open circles). (A) (i) After 12 conditioning trials (6 with CS+ and 6 with CS–), control bees clearly discriminated between the two odorants [F(1,78) = 48.51, P < 0.0001]. (A) (ii) One hour after the last conditioning trial, a significantly higher percentage of controls responded with proboscis extension to the odorant associated with sucrose (black bar) than to the nonreinforced odorant (white bar) (χ2 = 13.47, P = 0.0002). (B) (i) QMP-treated bees learned to discriminate between the two odorants [F(1,78) = 44.6, P < 0.0001]. (B) (ii) One hour after the last conditioning trial, the percentage of bees responding to the reinforced odorant (black bar) was significantly higher than the percentage of bees responding to the nonreinforced odorant (white bar) (χ2 = 7.83, P = 0.005).

What is responsible for mediating the QMP effects on aversive learning in young workers? The aromatic compound, 4-hydroxy-3-methoxyphenylethanol (homovanillyl alcohol, HVA) is a major contributor to QMP's effects on dopamine signaling in the brain (9). To examine HVA's contribution to QMP's effects on aversive learning, we treated newly emerged adults with HVA alone (13). Two control groups were included for comparison: untreated bees and bees treated with the QMP component, methyl p-hydroxybenzoate (HOB) (1). In contrast to HVA, HOB does not modulate brain dopamine function in young worker bees (9).

We found clear evidence of aversive learning in both untreated bees (fig. S1) and bees treated with HOB. As a result of conditioning, the percentage of bees responding with sting extension to the reinforced odorant increased significantly, whereas responses to the nonreinforced odorant decreased (Fig. 3A, i, and SOM text). In both groups, bees clearly differentiated between CS+ and CS–(controls, fig. S1; HOB-treated, Fig. 3A, i), and the level of responses to CS+ was still significantly higher than the level of responses to CS–1 hour after the last conditioning trial (controls, fig. S1; HOB-treated, Fig. 3A, ii).

Fig. 3.

Effects of (A) HOB and (B) HVA on aversive learning in 4-day-old workers. Bees were trained to discriminate between an odorant paired with an electric shock (CS+, filled circles) and a nonreinforced odorant (CS–, open circles). (A) (i) HOB-treated bees clearly discriminate between the two odorants [F(1,78) = 84.61, P < 0.0001]. (A) (ii) One hour after the last conditioning trial, a higher percentage of bees responded with sting extension to the odorant paired with electric shock (black bar) than to the nonreinforced odorant (white bar) (χ2 = 24, P < 0.0001). (B) (i) Comparison of the response curves for CS+ and CS–revealed no significant difference between the levels of responses to the two odorants in HVA-treated bees [F(1,78) = 0.83, n.s.]. (B) (ii) One hour after the last conditioning trial, a higher percentage of HVA-treated bees responded with sting extension to the reinforced odorant (black bar) than to the nonreinforced odorant (white bar) (χ2 = 4, P = 0.045). However, the percentage of bees exhibiting conditioned responses is significantly lower in HVA-treated bees than in HOB-treated bees (χ2 = 9.26, P = 0.002).

In bees treated with HVA, learning was impaired. Over successive conditioning trials, the percentage of HVA-treated bees responding to CS+ declined, mirroring responses to CS–(Fig. 3B, i, and SOM text). Comparison of the response curves for CS+ and CS–revealed no significant difference between the levels of responses to the two odorants (Fig. 3B, i). One hour after the last conditioning trial, the level of responses to CS+ was significantly higher than to CS–(Fig. 3B, ii), which suggested that in some HVA-treated bees an aversive memory was established. However, the percentage of HVA-treated bees exhibiting a conditioned response was significantly lower than that of controls (Fig. 3B, ii).

Despite being exposed to queen pheromone, forager bees show aversion learning (12). How can this be explained? Some QMP effects, including QMP's ability to elicit retinue behavior, are age-dependent (2, 3). We examined aversion learning in bees exposed to QMP for 15 days (Fig. 4) (13) and found that both QMP-treated bees (Fig. 4A) and controls (Fig. 4B) showed robust aversion learning. These results indicate that QMP's effects on associative olfactory learning and, presumably, also those of HVA alone, depend on worker age.

Fig. 4.

Effects of QMP on aversive learning in 15-day-old workers. Associative olfactory conditioning of the SER was used to compare aversive learning in (A) control bees and (B) bees treated for 15 days with QMP. Bees were trained to discriminate between an odorant paired with electric shock (CS+, filled circles) and an odorant that was not reinforced (CS–, open circles). After conditioning, bees in both groups clearly discriminated between the two odorants [(A) (i) F(1,78) = 53.7, P < 0.0001; (B) (i) QMP-treated bees F(1,78) = 47.96, P < 0.0001]. One hour after the last conditioning trial, a significantly higher percentage of bees responded with sting extension to the odorant associated with electric shock (black bar) than to the nonreinforced odorant (white bar); [(A) (ii) χ2 = 18.05, P < 0.0001; (B) (ii) QMP-treated, χ2 = 20.05, P < 0.0001].

HVA has recently been identified as the pheromone most important for increasing queen survival (18). This suggests that HVA blocking of aversive learning in young workers is somehow advantageous to the queen. One possibility is that high dosages of QMP experienced by workers attending the queen have unpleasant side effects. Previous reports have shown, for example, that QMP in high concentrations can be repellent to workers (19, 20) and can elicit aggression (21, 22), behaviors that could jeopardize the survival of the queen. By blocking the establishment of aversive memories, young bees would be prevented from forming an association between QMP and any unpleasant side effects induced by high dosages of this pheromone. This would confer significant benefit, as it would increase the likelihood of workers' remaining in attendance of the queen. Retinue behavior not only ensures that the queen is constantly fed, groomed, and antennated by young workers (13), but also, that the queen's attendants facilitate the distribution of QMP throughout the colony (23). It is noteworthy that, for reasons that remain unclear, QMP's effects on aversive learning are age-dependent. Even in the presence of queen pheromone, older workers exhibit robust aversion learning. This is also significant, as it ensures that this important survival tool can benefit workers and contribute ultimately to the survival of the colony as a whole.

Supporting Online Material

www.sciencemag.org/cgi/content/full/317/5836/384/DC1

Materials and Methods

SOM Text

Fig. S1

References

References and Notes

View Abstract

Navigate This Article