Identification and characterization of mushroom body neurons that regulate fat storage in Drosophila

Abstract

Background: In an earlier study, we identified two neuronal populations, c673a and Fru-GAL4, that regulate fat storage in fruit flies. Both populations partially overlap with a structure in the insect brain known as the mushroom body (MB), which plays a critical role in memory formation. This overlap prompted us to examine whether the MB is also involved in fat storage homeostasis.

Methods: Using a variety of transgenic agents, we selectively manipulated the neural activity of different portions of the MB and associated neurons to decipher their roles in fat storage regulation.

Results: Our data show that silencing of MB neurons that project into the α'β' lobes decreases de novo fatty acid synthesis and causes leanness, while sustained hyperactivation of the same neurons causes overfeeding and produces obesity. The α'β' neurons oppose and dominate the fat regulating functions of the c673a and Fru-GAL4 neurons. We also show that MB neurons that project into the γ lobe also regulate fat storage, probably because they are a subset of the Fru neurons. We were able to identify input and output neurons whose activity affects fat storage, feeding, and metabolism. The activity of cholinergic output neurons that innervating the β'2 compartment (MBON-β'2mp and MBON-γ5β'2a) regulates food consumption, while glutamatergic output neurons innervating α' compartments (MBON-γ2α'1 and MBON-α'2) control fat metabolism.

Conclusions: We identified a new fat storage regulating center, the α'β' lobes of the MB. We also delineated the neuronal circuits involved in the actions of the α'β' lobes, and showed that food intake and fat metabolism are controlled by separate sets of postsynaptic neurons that are segregated into different output pathways.

Fig. 1

The mushroom body regulates fat storage. a Top, cartoon of mushroom body lobes. Left, TLC assays, with 3 replicates for each genotype. For each TLC plate, the left 3 sample sets show Kir2.1-mediated silencing vs. GAL4- and UAS-constructs alone. The right 3 sample sets show dTrpA1-mediated hyperactivation vs. GAL4- and UAS-constructs alone. Right, Quantification of the TLC data. Control bars (Cont.) in the histogram cases are the averages of measurements of four replicas from three different controls (driver/WT, UAS-Kir2.1/WT, and UAS-dTrpA1/WT), all of which have similar fat levels. WT indicates a wild-type chromosome (+). Top row, silencing or hyperactivation of αβ KCs using C739-GAL4 does not affect fat content. Second row, silencing of α’β’ KCs using VT30604-GAL4 produces leanness, while hyperactivation produces obesity. Third row, silencing of γ KCs using MB009-split-GAL4 produces moderate obesity, while hyperactivation produces moderate leanness. Bottom row, the intensity of fat droplet staining by Nile Red is reduced relative to controls when α’β’ KCs are silenced (middle), and increased when they are hyperactivated (right). Arrows indicate fat droplets. b Top, schematic illustrations of the fly central brain showing the c673a-GAL4 (left) or Fru-GAL4 (right) neurons in green, MB neurons in red, and overlapping neurons in yellow. Most or all γ KCs express Fru-GAL4. Middle row left, TLC plate showing that hyperactivation of c673a-Gal4-positive neurons using NaChBac1 causes leanness, and silencing GAL4 in all KCs using MB247-GAL80 does not alter the phenotype. Middle row right, TLC plate showing that hyperactivation of Fru-GAL4-positive neurons causes leanness, but silencing GAL4 in all KCs using MB247-GAL80 partially suppresses the effect. Cont. samples for both TLC plates are UAS-NaChBac/+. Bottom row, quantitation of TLC data. S, triglyceride standards. Bars indicate means ± SEM, n = 12 samples for pooled controls and n = 4 for other genotypes. Asterisks denote t-test statistical significance: **p < 0.005, ***, p < 0.0005

Fig. 2

Mushroom body input and output neurons involved in fat storage regulation. a Input neurons. Quantitation of fat levels, as measured by TLC, in flies with Kir2.1-mediated silencing (left) or dTrpA1-mediated hyperactivation (right) of DPM neurons using C316-GAL4, or of PAM-γ5 neurons using MB315-GAL4. Far right, Image of PAM-γ5 neurons (orange). In all images in this figure, the brain is a translucent grey skeleton, and the MB lobes are in translucent pink and blue. There are 8–12 PAM-γ5s per brain hemisphere. b MBONs innervating α’ compartments. Left bar graph shows that Kir2.1-mediated silencing of MBON-γ2α’1 using MB077B-GAL4 and MBON-α’2 using MB082C-GAL4 causes leanness. Right bar graph shows that dTrpA1-mediated hyperactivation of these neurons has no effect. Far right, images of MBON-γ2α’1 (blue-green) and MBON-α’2 (light green). There are 2 MBON-γ2α’1 s and 1 MBON-α’2, per brain hemisphere. c MBONs innervating β’2 compartments. Left bar graph shows that Kir2.1-mediated silencing of MBON-γ5β’2a and MBON-β’2mp neurons using MB011B-GAL4 has no effect on fat content. Right bar graph shows that dTrpA1-mediated hyperactivation of the same neurons causes obesity. Far right, images of MBON-γ5β’2a (red) and MBON-β’2mp (blue). There is 1 MBON-γ5β’2a and 1 MBON-β’2mp per brain hemisphere. d Combined images at the bottom show superimpositions of all α’-innervating MBONs (left), all β’-innervating MBONs (middle), all MBONs (right), and MBONs plus PAM-γ5s (bottom). Bars indicate means ± SEM, n = 12 samples for pooled controls as in Fig. 2 and n = 4 for other genotypes, each composed of 10 flies homogenate. Asterisks denote t-test statistical significance: ***, p < 0.0005

Fig. 3

Behavioral and metabolic phenotypes associated with silenced and hyperactivation of the different mushroom body circuits involved in fat storage regulation. a Food intake in flies with silenced or hyperactivated MB. Silencing of any type of mushroom body-associated neuron does not significantly affect food intake (upper row). Hyperactivation of α’β’ KCs, using VT30604-GAL4, (lower left) of MBON-γ5β2’a, MBON-β2’mp (MB011B-GAL4 neurons; lower right) cause increases in food intake; all other genotypes are not significantly different from controls. b Conversion of ingested ¹⁴C aspartic acid to protein (magenta), carbohydrate (light blue), and lipid (yellow) in flies with silenced or hyperactivated MB neurons. Silencing of α’β’ KCs (VT30604), MBON-γ2α’1 (MB077B), MBON-α’2 (MB082C), and PAM-γ5 (MB315C) produces decreases in labeled lipids, and silencing γ KCs (MB009B) produces an increase (upper panels). Hyperactivating γ KCs, using MB009-GAL4, and DPM neurons, using C316-GAL4, produces decreases in labeled lipids (lower panels). Bars indicate means ± SEM, n = 60 single flies in pooled controls (see Fig. 2) and n = 20 for all other genotypes for Café and climbing assay, and n = 12 for pooled controls and n = 4 for other genotypes for the fat store degradation and ¹⁴C incorporation experiments (each sample is a homogenate from 10 flies). Asterisks denote t-test statistical significance: *p < 0.05, **p < 0.005, ***, p < 0.0005

Fig. 4

Paired hyperactivation of neurons with opposing effects on fat storage. a Schematic illustration of parts of the MB circuitry, including neurons, examined. Solid rectangles are dendritic arbors, and solid triangles are axon terminals. Selected MB compartments in α’β’ and γ lobes are indicated. PAMs and MBONs labeled by the drivers used in this paper are indicated. c673a and non-KC Fru neurons are indicated by unconnected yellow boxes. In the diagrams in the other panels, the method of neural hyperactivation is indicated by red color when dTrpA1 is used and orange color when NaChBac1, highlighted on a grey schematic of the MB circuit. b and c Combining hyperactivation of α’β’ KCs with VT30604-GAL4 with hyperactivation of Fru-GAL4 neurons (b) or of c673a-GAL4 neurons (c) results in fat levels that are the same as those in flies in which only α’β’ KCs are hyperactivated, showing the dominant role of α’β’ KC neurons in determining fat content. d and e Combining hyperactivation of MBON-γ5β’2a and MBON-β’2mp with MB011B (d) or of MBON-γ2α’1 with MB077B (e) with hyperactivation of Fru-GAL4 neurons results in suppression of the leanness produced by hyperactivation of Fru-GAL4 neurons alone. Bars indicate means ± SEM, n = 12 samples for pooled controls and n = 4 for other genotypes, as in Fig. 1. Asterisks denote t-test statistical significance: *p < 0.05, **p < 0.005, ***, p < 0.0005. In all experiments, the combined column refers to a single copy of both drivers (GAL4, and split-GAL4) present with a single copy of UAS-NaChBac1 f. A model for fat-regulating MB circuitry. α’β’ KCs regulate both food consumption (via MBONs innervating β’ compartments) and fatty acid synthesis (via MBONs innervating α’ compartments). MBON-γ5β’2a and MBON-γ2α’1 also innervate the γ lobes. γ KC activity affects fatty acid synthesis but not food consumption. The phenotypes shown in Fig. 3 and the epistatic relationships shown in Fig. 4 suggest that the output of the γ KCs that is relevant to fat storage is opposite in sign to that of the α’β’ KCs. However, we do not know the circuits through which γ KCs regulate fat content, so connections from the γ lobes are shown as dotted lines with inhibition bars