Competing dopamine neurons drive oviposition choice for ethanol in Drosophila

Abstract

The neural circuits that mediate behavioral choice evaluate and integrate information from the environment with internal demands and then initiate a behavioral response. Even circuits that support simple decisions remain poorly understood. In Drosophila melanogaster, oviposition on a substrate containing ethanol enhances fitness; however, little is known about the neural mechanisms mediating this important choice behavior. Here, we characterize the neural modulation of this simple choice and show that distinct subsets of dopaminergic neurons compete to either enhance or inhibit egg-laying preference for ethanol-containing food. Moreover, activity in α'β' neurons of the mushroom body and a subset of ellipsoid body ring neurons (R2) is required for this choice. We propose a model where competing dopaminergic systems modulate oviposition preference to adjust to changes in natural oviposition substrates.

Fig. 1.

Flies prefer to lay their eggs on ethanol-containing food. (A) Cartoon of egg-laying assay. (B) Given the choice between food with and without ethanol, flies preferred to lay eggs on ethanol [ANOVA: F_(5,71) = 25.36, P < 0.0001; n = 13–16; Tukey’s post hoc comparisons: for 0% vs. 3%, P = 0.04; for 0% vs. 5%, 7.5%, or 10%, P < 0.0001]. (C) The total number of eggs laid is not significantly affected by the presence of ethanol [ANOVA: F_(5,66) = 1.46, P = 0.22; n = 11–13]. (D) Flies preferred to lay their eggs on 5% ethanol over 1% [Wilcoxon: χ²_(1,8) = 6.05, P = 0.01], 3% [χ²_(1,8) = 6.05, P = 0.01], or 10% [χ²_(1,18) = 16.31, P < 0.0001] ethanol. There was no preference between 5% and 7.5% ethanol [χ²_(1,10) = 0.48, P = 0.49]. (E) The number of eggs laid on ethanol concentrations lower than 5% was significantly greater than those laid on concentrations higher than 5% [ANOVA: F_(3,26) = 249.45]. (F) Oviposition preference depends on the context of the oviposition substrate. Two-choice tests between 1%, 3%, 5%, 7.5%, and 10% ethanol suggest that concentrations near 5% ethanol are preferred. Preference indices for all pairwise comparisons are summarized. Detailed data are shown in Fig. S2. Bars on graphs represent means ± SEM. *P < 0.05; **P < 0.001; ***P < 0.0001.

Fig. 2.

Activity of dopamine and not serotonin neurons affects oviposition preference. (A) Dopaminergic neuron cell body positions in one hemisphere of the adult central brain are marked based on anti-TH immunohistochemistry (27, 28, 31, 32). PAM cell number is underrepresented in the schematic (33). (B) Cells colabeled with anti-TH antibody and Ddc-GAL4 in the central brain (33). Disrupting neurotransmission in Ddc cells decreased oviposition preference [n = 17–21 per strain; ANOVA: F_(3,73) = 47.51, P < 0.0001]. (C) Schematic of the TH-GAL4 expression pattern in the central brain (33). See Fig. S5 for more details. Arrow highlights that most PAM neurons do not express TH-GAL4. Disrupting neurotransmission in TH cells increased oviposition preference [n = 9–16 per strain; ANOVA: F_(3,51) = 9.62, P < 0.0001]. (D) Disruption of transmission in serotonergic neurons with TRH-GAL4 did not disrupt oviposition preference [n = 19–21 per strain; ANOVA: F_(2,76) = 4.23, P = 0.08; all Tukey’s comparisons to TRH/TeTx: P > 0.05]. Bars on graphs represent means ± SEM. *P < 0.05; **P < 0.001; ***P < 0.0001. Clusters of dopaminergic neurons are named based on their location in the brain: PAM, protocerebral anterior median; PPL, protocerebral posterior lateral; PPM, protocerebral posterior median; PAL, protocerebral anterior lateral; Sb, subesophageal ganglion.

Fig. 3.

Dopaminergic neurons compete to affect oviposition preference. (A–D) Schematics represent dopamine-expressing cell bodies (black dots) and a subset expressing the particular GAL4 pattern (red dots) with their respective innervation patterns (red lines) for each GAL4 line; MB is drawn in gray (A, C, and D) and EB in black (B). Nondopaminergic expression is omitted in schematics. (A) Disrupting transmission in the HL7-expressing PAM neurons decreased preference [n = 9–19 per strain; ANOVA: F_(2,42) = 9.33, P = 0.0004]. (B) Disrupting transmission in the PPM3 neurons (33) decreased preference [n = 12–18 per strain; ANOVA: F_(2,49) = 20.65, P < 0.0001]. Eliminating expression of TH within the c346 expression pattern using THGAL80 reverted the behavior to control levels, whereas eliminating MB expression using MB-GAL80 did not. (C and D) Disrupting transmission in subsets of PPL1 neurons (17) increased preference [NP2758: n = 10–12 per strain; ANOVA: F_(2,30) = 5.30, P = 0.01; kra: n = 20–21 per strain; ANOVA: F_(2,59) = 3.98, P = 0.02]. Eliminating expression of TH within the NP2758 and kra expression patterns using TH-GAL80 reverted the behavior to control levels. (E and F) Activating the PPL1 neurons using dTrpA1 decreases oviposition preference at 29°C [n = 8–12 per strain; ANOVA: F_(4,41) = 20.48, P < 0.0001] but not 22 °C [n = 6–15 per strain; ANOVA: F_(4,49) = 1.07, P = 0.38]. Bars on graphs represent means ± SEM. *P < 0.05; **P < 0.001; ***P < 0.0001.

Fig. 4.

The MB and EB are required for oviposition preference for ethanol. (A) The MB consists of three major classes of neurons whose axonal branches occupy distinct lobes elaborated by the αβ, α′β′, and γ neurons. (B) Disrupting neurotransmission in distinct subsets of MB neurons revealed that the α′β′ neurons mediate oviposition preference for ethanol [MB247: n = 10–11 per strain; ANOVA: F_(2,30) = 0.52, P = 0.60; NP65: n = 9–11 per strain; ANOVA: F_(2,28) = 7.61, P = 0.003; 4–59: n = 10–11 per strain; ANOVA: F_(2,30) = 6.00, P = 0.007]. (C) The major contributions to the EB are the projections of R neurons that arborize as a ring. (D) Disrupting neurotransmission in R2 neurons decreased oviposition preference [2–72: n = 7–11 per strain; ANOVA: F_(2,26) = 28.56, P < 0.0001; 11–27: n = 10–12 per strain; ANOVA: F_(2,32) = 52.03, P < 0.0001; 4–67a: n = 12–15 per strain; ANOVA: F_(2,32) = 52.03, P < 0.0001; c232: n = 10–14 per strain; ANOVA: F_(2,35) = 1.62, P = 0.21]. (E) Decreasing DopR2 expression in the MB increases oviposition preference for ethanol [n = 18 per strain; ANOVA: F_(2,51) = 7.63, P = 0.001]. (F) Decreasing DopR1 and DopR2 expression in the EB increases oviposition preference for ethanol [DopR1: n = 18 per strain; ANOVA: F_(2,50) = 9.17, P = 0.0004; DopR2: n = 18 per strain; ANOVA: F_(2,50) = 10.60, P = 0.001]. Bars on graphs represent means ± SEM. *P < 0.05; **P < 0.001; ***P < 0.0001.

Fig. 5.

A model for the neural circuitry mediating oviposition preference for ethanol. (A) A schematic representation of one hemisphere of the Drosophila central brain depicting innervation of the MB (blue) and EB (green) by dopaminergic neurons of the PAM (red), PPL1 (purple), and PPM3 (yellow) clusters. (B) Sensory information about ethanol is gated through dopaminergic neurons. The PAM (red) and PPM3 (yellow) neurons promote oviposition preference, whereas the PPL1 neurons (purple) inhibit preference. Dopaminergic input is relayed to the MB via the PAM and PPL1 neurons and to the central complex EB via the PPM3 neurons. The output response is relayed through the MB α′β′ neurons and EB R2 neurons. Sensory information about 5% ethanol activates the PAM and PPM3 neurons, which activate the MB α′β′ neurons and EB R2 neurons, respectively, leading to an increase in oviposition on ethanol. Sensory information about 5% ethanol simultaneously activates PPL1 neurons, which inhibit the MB α′β′ neurons, resulting in a decrease in oviposition on ethanol. When given choices of low ethanol concentrations (for example 5% vs. 0%) activation of the PAM neurons would have a stronger effect on oviposition behavior than activation of PPL1 neurons, resulting in increased preference for ethanol 5%. (C) When given choices that include detrimental ethanol concentrations (for example 10% vs. 5%), we hypothesize that activation of the PPL1 neurons has a stronger effect on oviposition relative to PAM neurons, resulting in a decrease in oviposition preference for the higher ethanol concentration. See Table S1 for detailed expression data.