Figures and data
Co-expression of Dop1R1 and Dop2R genes in adult Drosophila brain.
(A and B) The expression of Dop1R1-T2A-GAL4 and Dop2R-T2A-GAL4 visualized by UAS-mCD8::GFP. Maximum-intensity projections of the whole brain.
(C) Schematic of the KCs and the MB-innervating dopamine neurons from the PAM and PPL1 clusters.
(D-E, G-I) Double labelling of Dop1R1-T2A-LexA and Dop2R-T2A-GAL4 expressions by lexAop-rCD2::GFP (green) and UAS-CD4::tdTomato (red), respectively. Dopamine neurons were immunostained with anti-TH antibody (blue). Single optical sections are shown. Cell bodies of the PAM cluster (D), the PPL1 cluster (E), ring neurons projecting to the ellipsoid body (G and H) and ensheathing glia (I) are shown.
(F) Schematic of the regions shown in (G-I).
Scale bars, 50 µm (A and B), 5 µm (D, E and I), 20 µm (G and H).
Cell type-specific visualization of endogenous proteins with GFP11 tag.
(A) Principle of cell type specific fluorescent labelling of target proteins by GFP11 tag. Seven copies of GFP11 are fused to the C-terminal of endogenous receptors. GFP1-10 and membrane marker CD4::tdTomato are expressed in the target cells by GAL4/UAS system. In the target cells, reconstitution of GFP occurs on the endogenous proteins tagged with GFP11.
(B) As an example, DopEcR::GFP11 is visualized in KCs using MB-Switch, a ligand-inducible GAL4 driver. To activate Gene-Switch, flies were fed with food containing 200 µM RU486 for 12 hours before dissection. A merged image of reconstituted GFP (green) and cellular membrane visualized by CD4::tdTomato (magenta). Maximum intensity projection of the whole left MB.
(C) The workflow for visualizing subcellular protein enrichment by localization index (LI). A single sagittal section of the MB calyx and peduncle is shown. The ratio of reconstituted GFP to membrane signal is calculated and normalized by the mean of all voxels to provide LI. In the middle image, LI is color-coded so that red represents local receptor enrichment. In the right image, the intensity of LI color is adjusted based on the membrane signal.
Subcellular localization of Dop1R1 and Dop2R in the Kenyon cells.
Subcellular localization of Dop1R1 and Dop2R in the KCs is visualized by GFP11-tag. MB-Switch was used to express GFP1-10 and CD4::tdTomato in the KCs. To activate Gene-Switch, flies were fed with food containing 200 µM RU486 for 72 hours before dissection.
(A) Schematic showing the projection pattern of a α/β KC.
(B) Enrichment of Dop1R1 and Dop2R in the MB lobe. Maximum-intensity projections of the lobe (left) and the calyx (middle) are shown in frontal view. The whole left MB are shown in sagittal view (right). Reconstituted GFP signals for both Dop1R1:: and Dop2R::GFP11 distributed throughout the MB lobe and the calyx.
(C) Visualization by LI showed more pronounced enrichment of Dop2R than Dop1R1 in the lobe.
(D) Mean LI of Dop1R1 and Dop2R in the calyx, the peduncle, the α3 and β2 compartment in the lobe. Student’s t test was performed to compare LI of Dop1R1 and Dop2R in each region (N = 10). Error bars; SEM. p> 0.05, ** p< 0.01, *** p< 0.001, **** p< 0.0001, ns: not significant p>0.05.
Scalebars, 20 µm (B and C).
High-resolution imaging revealed the two opposing dopamine receptors existing on the presynaptic and postsynaptic sites of Kenyon cells.
(A) Schematic showing the projection pattern of α/β and γ KCs.
(B and C) Airyscan images of Dop1R1::rGFP and Dop2R::rGFP in KCs (green) co-labeled with the AZ stained with anti-Brp (magenta). MB-Switch was used with 72 hours of RU486 feeding to express GFP1-10 and CD4::tdTomato in the KCs. Brp puncta that overlap with CD4::tdTomato signals (cyan) are identified to be presynaptic sites of in KCs (white arrowheads), and those do not overlap are determined to be presynaptic sites of non-KCs (outlined arrows). The synaptic localization of these receptors is similar in the α 3 (B) and γ5 (C) compartments. In the right panels, white squared regions in the left panels are magnified. Scalebars, 5 µm (left), 1 µm (right).
(D) Illustration of localization of Dop1R1 and Dop2R to presynaptic and postsynaptic sites in the axon terminal of KCs.
Presynaptic localization of Dop1R1 and Dop2R in Kenyon cells and giant neurons.
(A) Double labelling of dopamine receptors (green) and the AZ of the KCs (magenta). MB-Switch was used with 72 hours of RU486 feeding to express GFP1-10 and Brpshort::mStraw in the KCs. Single focal slices at the α3 compartment are shown. White squares in the left panels are magnified in the right panels. The Brp puncta in KCs were either abutted by the dopamine receptor signals (white arrowheads) or had barely detectable signals nearby (outlined arrowheads). Scale bars, 5 μm (left), 1 μm (right).
(B) Punctate Brp expression in a giant neuron culture differentiated from cytokinesis-arrested neuroblasts of OK371-GAL4/UAS-mCD8::GFP embryos. Aggregated Brp condensates (magenta) were observed in the neurite terminals of the cells marked with mCD8::GFP (cyan) in the right panel. Scale bars, 20 µm (left), 10 μm (right).
(C) Double labelling of dopamine receptors (green) and the AZs (magenta). Dop1R1::Venus or Dop2R::Venus was crossed with Brp::SNAP. In the left panels, giant neurons extending their neurites from the cell body on the left to the right. In the right panels, white squared regions in the left panels are magnified. Scale bars, 10 μm (left), 2 μm (right).
Subcellular localization of Dop1R1 and Dop2R in MBON-γ1pedc>αβ.
(A) The projection pattern of MBON-γ1pedc>αβ from the tracing data in FlyWire (Dorkenwald et al., 2024; Schlegel et al., 2024).
(B and C) R83A12-GAL4 was used to express UAS-GFP1-10 and UAS-CD4::tdTomato in the PAM neurons.
(B) Reconstituted GFP signals (left) of Dop1R1::GFP11 and Dop2R::GFP11 in MBON-γ1pedc>αβ. Maximum-intensity projections of the left MB lobe. Visualization of LI (right) revealing that both Dop1R1 and Dop2R are enriched in the dendritic projection of MBON-γ1pedc>αβ in the γ1 compartment as well as in the presynaptic boutons.
(C) Airyscan images of the presynaptic boutons aroundα3 (left) and dendritic projections in the γ1 compartment (right). White squares in the right panels are magnified in the insertion to show the swelling membrane structures with punctate localization of dopamine receptors.
Scale bars, 20 µm (B), 5 µm (C), 1 µm (C, insertion).
Subcellular localization of Dop1R1 and Dop2R in dopamine neurons.
(A) Maximum-intensity projection image showing the distribution of presynaptic sites in the PAM neurons. Left panel: R58E02-GAL4 was used to express mCD8::GFP (magenta) and nSyb::CLIP (magenta). Right panel: Visualization by LI showing enrichment of nSyb signals in the lobe projection of the PAM neurons.
(B) Illustrated projection pattern of the PAM neurons. Red puncta on the dendrites indicate the sparse distribution of presynaptic sites in dendrites.
(C-F) Subcellular localization of GFP11-tagged Dop1R1 and Dop2R in the PAM neurons. R58E02-GAL4 (C and F) or R15A04-GAL4 (D and E) was used to express UAS-GFP1-10 and UAS-CD4::tdTomato in the PAM neurons.
(C) Reconstituted GFP signals of Dop1R1::GFP11 and Dop2R::GFP11 in PAM neurons (left). LI visualization revealed the stronger presynaptic enrichment of Dop2R than that of Dop1R1 (right). Maximum-intensity projections of the left hemisphere including the whole MB lobe and dendritic projections of the PAM neurons around the MB.
(D and E) LI in PAM-β’1 neuron. (D) The presynaptic terminals of PAM-β’1 neurons are shown (dashed line). (E) Mean LI for Dop1R1 and Dop2R in the β’1 (Mann-Whitney U test, N = 9). Error bars; SEM.
(F) A single optical slice of the γ5 compartment in the MB lobe obtained using Airyscan. Merged image of reconstituted GFP (green) and CD4::tdTomato (magenta). Insertions are the magnified images of the presynaptic boutons of PAM-γ5 (white squares).
Scale bars, 20 µm (A, C, D and F), 5 µm (F, insertion).
Bidirectional modification of dopamine receptor expression in dopamine neurons.
(A) Schematic illustration of the MB projection of the PAM and PPL1 dopamine neurons.
(B) Dop1R and Dop2R in the presynaptic terminals of PAM-γ5 after 48 hours of starvation compared with fed state.
(C) Quantification of rGFP signal levels in the presynaptic terminals of PAM-γ5 after 0, 10, 24 and 48 hours of starvation (n = 6-13).
(D) Reconstituted GFP signals of Dop1R1::GFP11 and Dop2R::GFP11 in the PPL1 neurons. In the MB projections of the PPL1 neurons, Dop1R1 was detected in only the α3 compartment. Dop2R was found in all MB projections. Maximum-intensity projections of the MB lobe.
(E) Dop1R and Dop2R in the presynaptic terminals of PPL1-α3 after 24 hours of starvation compared with fed state.
(F) Quantification of rGFP signal levels in the presynaptic terminals of PPL1-α3 after 0, 10 and 24 hours of starvation (n = 7-10).
Scale bar, 10 µm (B and E), 20 µm (D).
Interaction effects between genotypes and starvation time on protein levels were tested by Two-way ANOVA (C and F). Bars and error bars represent mean and SEM, respectively (C and F). ** p< 0.01, *** p< 0.001, ns: not significant p>0.05.
The dual dopaminergic feedback regulating starved-state dependent expression of appetitive behavior.
(A) A working model showing the role of the dual dopaminergic feedback regulation. In the starved state, increased Dop1R1 in PAM neurons and increased Dop2R in PPL1 neurons changes the balance between the synaptic outputs from these DANs to favor appetitive behavior.
(B) According to the model, loss of Dop2R in PPL1 upregulates output from PPL1 to attenuate appetitive behavior in starved flies.
(C) Knockdown of Dop2R in the PPL1 neurons by MB504B-GAL4 reduced 3-hour appetitive memory performance (t-test with Bonferroni correction, n=14-15). Bars and error bars represent mean and SEM, respectively. * p< 0.05.
Co-expression of Dop1R1 and Dop2R genes in Kenyon cells.
Double labelling of Dop1R1-T2A-LexA and Dop2R-T2A-GAL4 expressions by lexAop-rCD2::GFP (green) and UAS-CD4::tdTomato (red), respectively. Cell bodies of Kenyon cells are shown. A single optical section is shown.
Scale bars, 5 µm.
GFP11-tagging on Dop1R1 and Dop2R do not affect aversive olfactory memory.
To test the functionality of the knock-in alleles, memory of flies carrying homozygous 7xGFP11-tagged Dop1R1 and Dop2R were compared with those without GFP11 tags. To induce reconstitution of GFP in Kenyon cells, GFP1-10 was expressed under the control of MB-Switch. To activate Gene-Switch, flies were fed with RU486 for 3 days before training.
Box plots represent the median as the center line, the upper and lower quartiles as the box boundaries, the minimum and maximum values as the whiskers. Dunn’s test was performed (N = 9-12). ns: not significant p>0.05.
Localization index remains stable across different experimental batches.
Two experimental batches showed consistent LI for both Dop1R1::GFP11 and Dop2R::GFP11 measured in any region in the MB. These two batches were conducted on different days and at different times (Zeitgeber time 5 and 9). The same data set was used in Figure 3D.
Student’s t test was performed to compare the two batches (N = 5). Bars and error bars represent mean and SEM, respectively. ns: not significant p>0.05.
Presynaptic localization of Dop1R1 and Dop2R in γ Kenyon cells.
Double labelling of dopamine receptors (green) and the AZ of the KCs (magenta). As in the α3 compartment (Figure 5A), both Dop1R1 and Dop2R were found near the AZ of KCs in the γ5 compartment.
Scale bars, 5 μm (left), 1 μm (right).
Starvation-dependent change of dopamine receptors in PAM and PPL1.
(A and C) Dop1R (A) and Dop2R (C) in the presynaptic terminals of PAM-α1, PAM-β2 and PAM-β’1 after 48 hours of starvation compared with fed state.
(B and D) Quantification of dopamine receptor levels in the presynaptic terminals of the PAM neurons after 0 (fed), 10, 24 and 48 hours of starvation (n = 6-13).
(E) Dop2R in the presynaptic terminals of PPL1-α’2, PPL1-γ1pedc and PPL1-γ2 and after 24 hours of starvation compared with fed state.
(F) Quantification of Dop2R levels in the presynaptic terminals of the PPL1 neurons after 0 (fed), 10 and 24 hours of starvation (n = 7-10).
Scale bar, 10 µm (A, C, E). Bars and error bars represent mean and SEM, respectively (B, D, F). * p< 0.05, ** p< 0.01, *** p< 0.001, ns: not significant p>0.05.
