PK2-R1 is required for larval sleep control

(A) Sleep amounts in PK2-R1 or Oamb knockout mutants. Each dot represents an individual animal; in this and the following panels, “N” indicates the number of biologically independent animals per group, and the thick line and thin error bars indicate the median and interquartile range (Q1–Q3), respectively. *** p < 0.001 (Mann–Whitney U-test with Bonferroni correction). (B) Sleep amounts in larvae in which PK2-R1 neurons were silenced. *** p < 0.001 (Mann–Whitney U-test with Bonferroni correction). (C) Expression pattern of PK2-R1-GAL4 > UAS-Kir::EGFP larvae.

IPCs and Dilps negatively regulate larval sleep

(A) Triple labeling of PK2-R1 neurons expressing nuclear-localized RFP (magenta), Dilp3 neurons expressing nuclear-localized GFP (green), and anti-Dilp2 positive cells (blue). Top panels show signals in the larval brain and the VNC. Bottom panels show magnified images of the white-squared area in each top panel, focusing on the dorsomedial brain region where the cell bodies of IPCs are located. Note that all IPCs labeled with Dilp3-GAL4 overlapped with PK2-R12A-LexA-expressing cells. Similar results were obtained from 5 independent samples of the same genotype. (B) Simultaneous detection of PK2-R1 neurons expressing nuclear-localized RFP (magenta), Dilp5 neurons expressing nuclear-localized GFP (green), and anti-Dilp2 positive cells (blue). Similar results were obtained from 5 independent samples of the same genotype. (C) Effect of IPCs silencing on larval sleep. Each dot represents an individual animal; in this and the following panels, “N” indicates the number of biologically independent animals per group, and the thick line and thin error bars indicate the median and interquartile range (Q1–Q3), respectively. *** p < 0.001, ** p < 0.01, * p < 0.05 (Mann–Whitney U-test with Bonferroni correction). (D) Sleep amounts in Dilp3 or Dilp5 null mutants. *** p < 0.001, ** p < 0.01 (Mann–Whitney U-test with Bonferroni correction).

Hug neurons negatively regulate larval sleep

(A) Sleep amounts in Hug CRISPR-knockout mutant larvae. **** p < 0.0001 (Mann–Whitney U-test). Each dot represents an individual animal; in this and the following panels, “N” indicates the number of biologically independent animals per group, and the thick line and thin error bars indicate the median and interquartile range (Q1–Q3), respectively. (B) Effect of silencing Hug neurons on larval sleep amount. ** p < 0.01, * p < 0.05 (Mann–Whitney U-test with Bonferroni correction). (C) Larval sleep changes induced by optogenetic activation of Hug neurons. For each genotype, sleep duration in the 1-h light-ON period was normalized to that in the 1-h light-OFF phase. * p < 0.05 (Mann–Whitney U-test with Bonferroni correction). (D) Larval sleep during thermogenetic activation of Hug neurons. ** p < 0.01, * p < 0.05 (Mann–Whitney U-test with Bonferroni correction). (E) Visualization of HugPC neurons and IPCs labeled by rCD2::RFP (magenta) and mCD8::GFP (green), respectively. The bottom panels show magnified images of the white-squared dorsomedial region in the top panels, where HugPC neurons project their axons close to the cell bodies of the IPCs. While the top panels are z-stacks of 122 image slices (1 µm interval) covering the entire brain tissue, the bottom panels are projections of 60 slices centering around IPCs. Similar results were obtained from 5 independent samples of the same genotype. (F) The effect of neuronal silencing confined to the HugPC subpopulation. ** p < 0.01, * p < 0.05 (Mann–Whitney U-test with Bonferroni correction).

Activation of Hug neurons triggers Ca2+ responses in larval IPCs.

(A) Schematic workflow for assessing intracellular Ca2+ levels in IPCs while thermogenetically activating Hug neurons. Larvae were exposed to low (18°C) or high (30°C) temperature for 1 h, followed by imaging of CRTC::GFP in IPCs and calculation of the CRTC::GFP nuclear localization index (NLI; see Methods). (B–C) Nuclear-localized CRTC::GFP signal (CRTC::GFP NLI) in larval IPCs under low-temperature (B) or high-temperature (C) conditions, with or without thermogenetic activation of Hug neurons. Each dot represents an individual cell; the thick line indicates the median and the thin error bars indicate the interquartile range (Q1–Q3). “N” indicates the number of cells analyzed per group. **** p < 0.0001, NS p ≥ 0.05 (Mann–Whitney U-test). (D) Working model of a neuronal network in which Hug neurons activate IPCs to regulate larval sleep

Hug peptides induce Ca2+ responses in larval IPCs via PK2-R1

(A) Schematic flow of peptide application followed by Ca2+ imaging. (B–C) Ca2+ responses in laval IPCs during ex vivo bath application of either glucose (B) or Hug peptides (C). In this and the following panels, ‘N’ indicates the number of cells used for each group, and the thick line and thin error bar represent the median and interquartile range (Q1–Q3), respectively. **** p < 0.0001, *** p < 0.001 (Mann–Whitney U-test with Bonferroni correction). (D) Representative images of larval IPCs upon peptide application. Scale bars, 2 μm. (E) Anti-Dilp3 signal intensity measured within the cytosolic areas of larval IPCs. *** p < 0.001, * p < 0.05 (Mann–Whitney U-test with Bonferroni correction). (F–G) Ca2+ responses in larval IPCs derived from PK2-R1 knockout mutants, measured after either glucose (F) or Hug peptide application (G). ** p < 0.01, NS p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction). (H) A model where Hug peptides, but not glucose, activate IPCs via the PK2-R1 receptor in the larval brain.

Distinct impacts of Hugin/PK2-R1 axis on wake/sleep control in larvae and adult

(A) Total sleep amounts in Hug knockout mutant adults. NS p ≥ 0.05 (Mann–Whitney U-test). In these and the following panels, ‘N’ indicates the number of biologically independent animals used for each group, and the thick line and thin error bar represent the median and interquartile range (Q1–Q3), respectively. NS p ≥ 0.05 (Mann–Whitney U-test). (B) Sleep amounts of adults in which HugPC neurons were silenced. *** p < 0.001, NS p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction). (C) Sleep amounts of Dilp3 or Dilp5 null mutant adults. *** p < 0.001 (Mann–Whitney U-test with Bonferroni correction). (D) Schematic flow of peptide application experiments using adult brains. (E–F) Ca2+ sponses (E) or anti-Dilp3 signal intensity (F) in adult IPCs after Hug peptide application. ** p < 0.01, NS p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction).

A schematic model of the Hugin/PK2-R1/Dilps axis in larval and adult sleep

Schematic summary comparing larvae and adults. In larvae, bath application of hugin peptides activates IPCs, and both hugin signaling (HugPC neurons/hugin peptides) and IPC output (including Dilp3) act to suppress larval sleep. In adults, hugin peptides do not activate IPCs (dashed arrow), whereas IPC/Dilp3 signaling promotes adult sleep.

Detailed experimental setup for analyzing larval sleep

(A) Schematic flow of automated larval sleep quantification. A second instar larva was placed in each of 24 wells, and all 24 wells were videotaped from above. For each well and frame, the image of larva was blurred and binarized to automatically locate the larva. Consecutive frames were then compared so that changes in pixel values were summed to estimate larval motion. This motion estimate was then used to classify every consecutive pair of frames as representing “active” or “inactive.” We found that inactivity across 12 consecutive frames meets the sleep criteria, as shown in Figure 1—figure supplement 2. (B) Schematic representation of the 24-well chamber used for larval sleep assays. (C) Flow of automated processes of larval motion detection and active/inactive state determination. For each well and frame, the image of larva was blurred and binarized. Consecutive frames were then compared so that changes in pixel values were summed to estimate larval motion. This motion estimate was then used to classify every consecutive pair of frames as representing “active” or “inactive.”

Larval “sleep state” in this study is consistent with behavioral criteria for sleep

(A) Relationship between “inactive” state duration and spontaneous arousal probability. Sustained quiescence for ≥ 12 frames (open circles) significantly reduced arousal probability compared to 1-frame inactivity. * p < 0.05, NS: p ≥ 0.05 (chi-squared test). (B) The percentage of larvae that showed a response within the next 10 frames after high-intensity blue light. The number of events analyzed was 36 and 60 with and without blue light irradiation, respectively. * p < 0.05 (Fisher’s exact test). (C) Sleep amounts before, during, and after blue light irradiation. The number of larvae tested was 21 and 21 for the control and experimental groups, respectively. ** p < 0.01, NS: p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction). (D) Percentage of larvae that showed a response in the next 1 or 12 frames after blue light stimulation. Number of events analyzed: 1960 or 111 for 1- or 12-frame continuous inactivity prior to blue light illumination, respectively. **** p < 0.0001 (chi-squared test).

CRISPR-knock-out screen for genes that regulate larval sleep.

Total sleep amounts in CRISPR-knock-out mutants were measured for 18 h from the beginning of the second instar. The data for Control, PK2-R1, and Oamb mutants are identical to those shown in Figure 1. Box plots are generated so that the center line indicates the median, the box limits indicate the upper and lower quartiles, and the whiskers indicate the minimum-to-maximum range. *** p < 0.005 (Kruskal–Wallis one-way ANOVA and post hoc Mann–Whitney U-test with Bonferroni correction). Number of larvae tested: 475 for the control and 10–52 for each mutant.

Effects of neuronal manipulation on larval sleep

Sleep amounts in larvae in which distinct neuronal populations were silenced. Kir2.1 expression was driven by each enhancer-GAL4 for “hit” genes identified in the CRISPR-knock-out screen. Each dot represents an individual animal in these panels, “N” indicates the number of biologically independent animals per group, and the thick line and thin error bars indicate the median and interquartile range (Q1–Q3), respectively. *** p < 0.001, ** p < 0.01, * p < 0.05, NS: ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction)

Larval sleep phenotypes in PK2-R1 mutants quantified over 18 h and the first 6 h.

(A–C) Sleep metrics quantified over an 18-hour period in second-instar larvae, whereas (D–F) show the same metrics quantified over the first 6 hours in the same animals. Total sleep is shown in (A) and (D), bout number in (B) and (E), and mean bout length in (C) and (F). Each dot represents an individual larva; the thick line indicates the median and the thin error bars indicate the interquartile range (Q1–Q3). Numbers in parentheses indicate N, the number of biologically independent animals per group. ** p < 0.01, * p < 0.05, NS: ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction).

Locomotion speeds are not consistently affected by genetical manipulations of PK2-R1, IPCs, or Dilps

(A) Larval locomotion speed in controls and PK2-R1 mutants. (B-C) Effects of silencing PK2-R1 neurons (B) and IPCs (C) on larval locomotion speed, shown together with the corresponding genetic controls. (D) Larval locomotion speed in Dilp3 or Dilp5 null mutants. Average locomotion speed during wake periods was calculated from the datasets shown in Figures 1A, 1B, 2C, and 2D. N indicates the number of biologically independent larvae per group. Each dot represents an individual larva; thick lines indicate the median and thin error bars indicate the interquartile range (Q1-Q3). *** p < 0.0001, * p < 0.05, NS: p ≥ 0.05 (Mann-Whitney U-test with Bonferroni correction).

Effects of Hug pathway manipulations on larval locomotion speed.

(A) Locomotion speed of Hug knockout mutants during wake periods. (B) Effects of silencing Hug neurons on locomotion speed. (C-D) Effects of activating Hug neurons on locomotion speed, shown as the light ON/light OFF ratio (C) and the absolute locomotion speed (D). (E) Effects of silencing HugPC neurons on locomotion speed. Locomotion speed was quantified from the same videos used for the larval sleep assays in Figure 3. Each dot represents an individual larva; thick lines indicate the median and thin error bars indicate the interquartile range (Q1-Q3). N indicates the number of biologically independent animals per group. NS p ≥ 0.05 (Mann-Whitney U-test) in (A). *** p < 0.0001, * p < 0.05, NS p ≥ 0.05 (Mann-Whitney U-test with Bonferroni correction) in (B-E).

Generation of HugPC-LexA transgenic lines

(A-D) Labeling patterns of four HugPC-LexA strains in larval CNS visualized by rCD2::RFP. Four transgenic lines were generated by injecting the same plasmid into the following landing sites individually: su(Hw)attP5, VK00005, attP2, and attP40. Magenta arrowheads represent cell bodies apparently located outside the previously classified HugPC subpopulation (Bader et al., 2007).

Silencing HugPC neurons reduces larval food intake measured by a dye-based assay.

After a 30 min starvation, larvae were allowed to feed on red yeast paste placed on an apple juice agar plate, and food intake was quantified as the ratio of the red dye-stained body surface area to the total body surface area (ImageJ). Each dot represents an individual larva; the thick line indicates the median and the thin error bars indicate the interquartile range (Q1–Q3). ** p < 0.01 (Mann–Whitney U-test).

Axonal projections of HuginPC neurons visualized with a presynaptic marker.

Representative whole-mount image of the larval central nervous system showing Syt-eGFP signals expressed in HuginPC neurons. Scale bar, 100 µm.

D-glucose but not Hug peptide cause Dilp2 reduction in larval IPCs

(A–B) Anti-Dilp2 signal intensity measured in larval IPCs after glucose (A) or Hug-γ application (B). In this and the following panels, ‘N’ indicates the number of cells used for each group, and the thick line and thin error bar represent the median and interquartile range (Q1–Q3), respectively. **** p < 0.0001, * p < 0.05 (Mann–Whitney U-test). (C) Anti-Dilp3 signal intensity measured within the cytosolic areas of larval IPCs. NS ≥ 0.05 (Mann–Whitney U-test).

Representative images of Dilp3 immunoreactivity in larval IPCs after bath application of Hug-γ peptide.

Larval IPCs were labeled with Dilp3>mCherry (top row), and endogenous Dilp3 was detected by anti-Dilp3 immunostaining (bottom row). Representative brains from buffer-only controls (Brain #1 and #2) and Hug-γ peptide application (Brain #3 and #4) are shown. Scale bars: 25 μm (Brains #1–3) and 10 μm (Brain #4).

Sleep patterns of adult flies with genetical manipulations in the Hugin and insulin pathways

(A) Experimental setup for monitoring adult sleep using the Drosophila Activity Monitor (DAM). (B) Total daily sleep duration in adult flies carrying CRISPR-knockout mutations in genes whose larval sleep was significantly altered in Figure 1B. Each dot represents an individual fly; the thick line indicates the median and the thin error bars indicate the interquartile range (Q1–Q3). *** p < 0.01, ** p < 0.05 (Kruskal–Wallis one-way ANOVA followed by post hoc Mann–Whitney U-test with Bonferroni correction). (C) Daytime (left) and nighttime (right) sleep amounts in flies in which HugPC neurons were silenced. (D) Daytime (left) and nighttime (right) sleep amounts in Dilp3 or Dilp5 null mutants. In (C) and (D), numbers in parentheses indicate N, the number of biologically independent animals per group. *** p < 0.001, ** p < 0.01, * p < 0.05, NS p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction).

Morphologies of PK2-R1 neurons, HugPC neurons, and IPCs in the adult brain.

(A) Dual labeling of PK2-R1 neurons and IPCs expressing nuclear-localized RFP (magenta) and GFP (green), respectively. The top panels show signals in the entire adult brain. The bottom panels show magnified images of the white-squared area in each corresponding top panel, focusing on the dorsomedial brain region where the cell bodies of IPCs are located. Similar results were obtained from 5 independent samples of the same genotype. (B) Visualization of HugPC neurons and IPCs labeled by mCD8::GFP (magenta) and the anti-Dilp3 antibody (green), respectively. Similar results were obtained from 4 independent samples of the same genotype.

Adult sleep architecture and waking activity following genetic manipulations of the hugin and insulin pathways.

(A) Bout number per day (left), mean bout length (middle), and mean beam crossings during wakefulness (right) in Hug<attP mutants. (B) The same metrics in flies with HugPC neuron silencing. (C) The same metrics in Dilp3[1] and Dilp5<1 null mutants. (D) Total daily sleep in flies with Dilp3-positive neuron silencing. Each dot represents an individual fly; the thick line indicates the median and the thin error bars indicate the interquartile range (Q1–Q3). Numbers in parentheses indicate N, the number of biologically independent animals per group. * p < 0.05, ** p < 0.01, *** p < 0.001, NS p ≥ 0.05 (Mann–Whitney U-test with Bonferroni correction).