The metabolic state affects the gene profiles of the Gustatory Second Order Neurons (G2Ns) populations analyzed.

(A) The scheme depicts the region of the brain dissected and the sorting of the G2Ns according to the input. Two populations of trans-TANGO neurons were labeled using Gr64f-Gal4 and Gr66a-Gal4 transgenic lines. Flies were either fed or food-starved for 24 hours. (B) Principal Component Analysis (PCA) of the genes. (C) Volcano plots show the genes up- and down-regulated in fed and starved conditions in the two populations of G2Ns.

Number of brains and cells dissected

Transcripts per Million (TPM) for Neuropeptides and Neurotransmitters.

TPMs for Neurotransmitters (A), Neurotransmitter Receptors (B), Neuropeptides (C) and Neuropeptide Receptors (D) for the two G2N populations in the two metabolic conditions studied (fed and 24h starved).

Leucokinin expression is increased in starvation.

(A) Immunofluorescence with anti-GFP (green) and anti-nc82 (blue) on a whole-mount brain of a lk-Gal4>UAS-mCD8:GFP. (B) Scheme showing the approach used to select the Lk GFP+ neurons by FACS. (C) qPCR results for Lk neurons labeled with GFP and sorted by FACS. The central brain includes all four Lk neurons, while the SEZ region includes only the two neurons located in this region. (D) Intensity fluorescence quantification in whole Oregon-R brains stained with anti-Lk antibody in fed and 24-hour starved flies. SELK Fed = 20 brains; SELK Starved = 21 brains; LHLK Fed = 12 brains; LHLK Starved = 16 brains. Scale bar = 50μm. Statistical test: Wilcoxon rank sum test. *p<0.05, **p<0,01, ****p < 0,0001. Scale bar = 50μm.

Lk neurons co-localize with the trans-Tango signal.

Immunofluorescence with anti-Leucokinin (blue) and anti-RFP (magenta) on whole-mount brains of (A) Gr64f- Gal4>trans-TANGO, and (B) Gr66a-Gal4>trans-TANGO. Images correspond to the adult SEZ. Arrowheads point to SELK neurons. Scale bar = 50μm.

Lk neurons are synaptically connected to Gr64fGRN and Gr66aGRN.

(A) Immunofluorescence with anti-GFP (green) and anti-nc82 (blue) on whole-mount brains of Gr64f-LexA>LexAop-CD4spGFP11 and Lk-Gal4>UAS-CD4spGFP1–10 (left column) and Gr66a- LexA>LexAop-CD4spGFP11 and Lk-Gal4>UAS-CD4spGFP1–10 (right column). Arrowheads point to GFP positive signal. (B) Immunofluorescence with anti-GFP (green), anti-RFP (magenta), and anti-nc82 (blue) on whole-mount brains of Gr64f-LexA, Lk-Gal4>BacTrace (left column) and Gr66a-LexA, Lk-Gal4>BacTrace (right column). Arrowheads point to the BacTrace positive signal. Scale bar = 50μm.

Lk neurons are pre- and postsynaptic to a significant number of neurons.

(A) The anatomy of the SELK neurons (Right in cyan and Left in magenta) was reconstructed using the FAFB brainmesh template in Flywire. IDs from Flywire are indicated: 720575940632407826 ID for the Left SELK in cyan (GNG.246) and 720575940623529610 for the Right SELK in magenta (GNG.276). Zoom in to the SEZ from (A), showing the details of the neuronal arborizations. (B) SELK postsynaptic candidate neurons reconstructed in the FAFB brainmesh template in Flywire. Only postsynaptic candidate neurons with ≥10 synaptic points with SELK neurons are shown (174 postsynaptic candidate neurons). (C) Immunofluorescence with anti-GFP (green), anti-RFP (magenta), and anti-nc82 (blue) on a whole-mount brain on a Lk-Gal4>trans-TANGO flies. (D) SELK presynaptic candidate neurons reconstructed in the FAFB brainmesh template in Flywire. Only presynaptic candidate neurons with ≥10 synaptic points with SELK neurons are shown (84 presynaptic candidate neurons). (E) Immunofluorescence with anti-GFP (green), anti-RFP (magenta), and anti-nc82 (blue) on a whole-mount brain on a Lk-Gal4>retro-TANGO flies. Scale bars: 50 µm.

Leucokinin neurons integrate sweet and bitter gustatory information.

(A) Scheme depicting the experimental procedure for Proboscis Extension Reflex (PER). Flies are given a series of increasing concentrations of sucrose (6.25, 12.5, 25, 50, 100, 200, 400 and 800 mM) for Experiment 1; and a fixed concentration of sucrose (50mM) mixed with increasing concentrations of caffeine (1, 5, 10 and 20mM) for Experiment 2. All flies starved 24 hours. (B) PER Dose response curve of Lk-Gal4>UAS-TNTimp (n=30) and Lk-Gal4>UAS-TNT (n=30) flies for Experiment 1. (C) PER of Lk-Gal4>UAS-TNTimp (n=30) and Lk-Gal4>UAS- TNT (n=30) flies for Experiment 2. (D) Scheme depicting the experimental procedure for flyPAD (Adapted from Itskov et al., 2014). Experiment 3: 20mM Sucrose vs. 100mM Sucrose and Experiment 4: 20mM Sucrose vs. 100mM Sucrose + 50mM Caffeine. All flies starved 24 hours. (E) Preference Index (Left) for the highest concentration of Sucrose (100mM) with or without 50mM Caffeine and total number of sips (Right) for Experiment 3 and 4. Experiment 3: Lk-Gal4>UAS-TNTimp (n=48); Lk-Gal4>UAS-TNT (n=59); Experiment 4: Lk-Gal4>UAS- TNTimp (n=20); Lk-Gal4>UAS-TNT (n=25). Only flies that performed a minimum of 25 sips per fly and 2 bouts were considered for the analysis. ns: non-significant, *p<0.05. PER analysis: Binomial regression model; Error bars represent the standard error of the proportion. flyPAD statistics Wilcoxon Rank Sum Test.

Model for the integration of gustatory information by SELK neurons.

(A) SELK neurons receive direct input from Gr64fGRNs (sweet), Gr66aGRNs (bitter), and metabolic information. When flies are fed, the expression of the neuropeptide Leucokinin is reduced. The probability of a fed fly initiating feeding is reduced when facing any food. (B) SELK neurons in starved flies express large amounts of Leucokinin. In this situation, flies are more prone to accept food containing bitter compounds. (C) Starved flies with silenced SELK neurons are less prone to accept sweet foods laced with bitter compounds.