D1 dopamine receptor dDA1 is required in the mushroom body neurons for aversive and appetitive learning in Drosophila

Abstract

Drosophila has robust behavioral plasticity to avoid or prefer the odor that predicts punishment or food reward, respectively. Both types of plasticity are mediated by the mushroom body (MB) neurons in the brain, in which various signaling molecules play crucial roles. However, important yet unresolved molecules are the receptors that initiate aversive or appetitive learning cascades in the MB. We have shown previously that D1 dopamine receptor dDA1 is highly enriched in the MB neuropil. Here, we demonstrate that dDA1 is a key receptor that mediates both aversive and appetitive learning in pavlovian olfactory conditioning. We identified two mutants, dumb1 and dumb2, with abnormal dDA1 expression. When trained with the same conditioned stimuli, both dumb alleles showed negligible learning in electric shock-mediated conditioning while they exhibited moderately impaired learning in sugar-mediated conditioning. These phenotypes were not attributable to anomalous sensory modalities of dumb mutants because their olfactory acuity, shock reactivity, and sugar preference were comparable to those of control lines. Remarkably, the dumb mutant's impaired performance in both paradigms was fully rescued by reinstating dDA1 expression in the same subset of MB neurons, indicating the critical roles of the MB dDA1 in aversive as well as appetitive learning. Previous studies using dopamine receptor antagonists implicate the involvement of D1/D5 receptors in various pavlovian conditioning tasks in mammals; however, these have not been supported by the studies of D1- or D5-deficient animals. The findings described here unambiguously clarify the critical roles of D1 dopamine receptor in aversive and appetitive pavlovian conditioning.

Figure 1.

dDA1 IR in the adult head sections of Canton-S, dumb¹, and dumb² flies. The frontal sections at the levels of the MB lobes (top) and pedunculi along with the central complex (bottom) are shown. dDA1 IR is visualized by red florescence. Canton-S has prominent dDA1 IR in the MB lobes and pedunculi as well as the central complex; however, no dDA1 IR is visible in those structures of dumb¹ and dumb². p, Pedunculus; cc, central complex. All images are at the same magnification. Scale bar, 100 μm.

Figure 2.

The learning phenotype of dumb mutants in aversive olfactory conditioning. A, B, Flies were trained with BA and OCT as CS and tested immediately after (acquisition) or 1 h after training (1 h memory). A, dumb¹ homozygous and dumb¹/Df(3R)su(Hw)7 trans-heterozygous mutants exhibited severely impaired learning, whereas dumb¹/+, Df(3R)su(Hw)7/+, and dumb¹/Df(3L)AC1, which have one copy of the dDA1 gene, showed performance similar to that of Canton-S (AVOVA; F_(5,35) = 35.9; p < 0.0001; n = 6 for all groups; asterisks indicate significant difference by post hoc Tukey–Kramer tests). B, At 1 h after training, dumb¹, dumb¹/Df(3R)su(Hw)7, Df(3R)su(Hw)7/+, and dumb¹/Df(3L)AC1 showed defective performance compared with Canton-S and dumb1/+ (ANOVA; F_(5,35) = 26.04; p < 0.0001; n = 6; asterisks indicate significant difference compared with Canton-S by post hoc Student's t test). C, dumb² homozygous and dumb¹/dumb² transheterozygous mutants showed no trace of learning and 1 h memory, whereas learning or 1 h memory performance of dumb² heterozygous flies (dumb²/+) was similar to that of the genetic control line w¹¹¹⁸ (w) (acquisition ANOVA: F_(3,23) = 55.3, p < 0.0001; 1 h memory ANOVA: F_(3,23) = 21.6, p < 0.0001; n = 6; asterisks indicate significant difference by Tukey–Kramer tests). Error bars indicate SEM.

Figure 3.

Restored dDA1 expression in dumb transgenic mutants. A–C, dumb¹/dumb² trans-heterozygous mutants carrying Elav-GAL4 and GAL80^ts reared at room temperature (uninduced) had no detectable dDA1 expression (A); however, when they were incubated at 30°C for 3 d (induced), dDA1 IRs were visible in the MB lobes (B) and pedunculi, the central complex, and other brain structures (C). D, GFP driven by MB247-GAL4 in the wild-type genetic background was visible in most, if not all, dDA1-positive MB neurons. E, F, dumb¹/dumb² carrying MB247-GAL4 had conspicuous dDA1 expression in the MB lobes (E) and pedunculi (F) but not in the central complex (F). p, Pedunculus; cc, central complex. All images are at the same magnification. Scale bar, 100 μm.

Figure 4.

Rescue of the dumb mutant's phenotype in aversive learning. A, The restricted dDA1 expression in the adult nervous system rescued the learning phenotype of dumb mutants. dumb¹/dumb² carrying Elav-GAL4 and GAL80^ts reared at room temperature (Elav,GAL80^ts/+;dumb¹/dumb² uninduced) showed poor performance immediately after training; however, performance of the same genotype reared at 30°C for 3 d (Elav,GAL80^ts/+;dumb¹/dumb² induced) was not significantly different from that of Canton-S (Kruskal–Wallis test, p = 0.0009; n = 6; the asterisk indicates significant difference by Mann–Whitney tests). B, When subjected to brief (submaximal) training with two pulses of electric shock (2 shocks), dumb² heterozygous flies carrying Elav-GAL4 and GAL80^ts that were reared at 30°C for 3 d to induce ectopic dDA1 expression (Elav,GAL80^ts/+;dumb¹/dumb² induced) had the learning score comparable with that of Canton-S, whereas dumb² homozygous mutants showed impaired learning (ANOVA; F_(2,17) = 27.2; p < 0.0001; n = 6; the asterisk indicates significant difference by Tukey–Kramer tests). Likewise, the dumb² heterozygous flies with ectopic dDA1 expression (Elav,GAL80^ts/+;dumb¹/dumb² induced) had performance comparable with that of Canton-S and the same genotype without heat treatment (Elav,GAL80^ts/+;dumb¹/dumb² uninduced) when subjected to regular training with 12 pulses of electric shock (regular) (ANOVA; F_(2,17) = 0.04; p = 0.96; n = 6). C, The dumb transgenic mutants expressing dDA1 in the MB lobes (MB247/+;dumb² and MB247/+;dumb¹/dumb²) had the learning scores similar to those of Canton-S and MB247/+;dumb¹/+, whereas all three lines with deficient dDA1 expression (dumb², dumb¹/dumb², and MB247/+;dumb¹) had significantly low learning scores (ANOVA; F_(6,41) = 43.6; p < 0.0001; n = 6; asterisks indicate significant difference by Tukey–Kramer tests). D, The temporally induced dDA1 expression only in the adult MB rescued the learning phenotype of dumb mutants. dumb¹/dumb² carrying MB247-GAL4 and GAL80^ts that were reared at 30°C for 3 d (MB247/+;GAL80^ts,dumb²/dumb¹ induced) had the learning score comparable with that of Canton-S, whereas the same genotype reared at room temperature (MB247/+;GAL80^ts,dumb²/dumb¹ uninduced) and dumb¹ homozygous mutants had significantly low learning scores (ANOVA; F_(3,23) = 77.6; p < 0.0001; n = 6; asterisks indicate significant difference by Tukey–Kramer tests). Error bars indicate SEM.

Figure 5.

Impaired learning of dumb mutants in appetitive conditioning and rescue by reinstated dDA1 expression in the MB. A, Both dumb¹ and dumb² mutants were moderately impaired in acquisition (ANOVA; F_(2,17) = 14.2; p < 0.0001; n = 6) and 1 h memory (ANOVA; F_(2,17) = 11.5; p < 0.005; n = 6) of sugar-mediated olfactory conditioning. B, The dDA1 expression in the MB driven by MB247-GAL4 was sufficient to rescue the learning phenotype of dumb² homozygous (MB247/+;dumb²) and dumb¹/dumb² transheterozygous (MB247/+;dumb¹/dumb²) mutants (ANOVA; F_(3,50) = 10.2; p < 0.0001; n = 6 for MB247/+;dumb¹/dumb²; n = 15 for other groups). Asterisks denote significant difference by Tukey–Kramer tests. Error bars indicate SEM.