Regulation of sleep by cholinergic neurons located outside the central brain in Drosophila

Abstract

Sleep is a complex and plastic behavior regulated by multiple brain regions and influenced by numerous internal and external stimuli. Thus, to fully uncover the function(s) of sleep, cellular resolution of sleep-regulating neurons needs to be achieved. Doing so will help to unequivocally assign a role or function to a given neuron or group of neurons in sleep behavior. In the Drosophila brain, neurons projecting to the dorsal fan-shaped body (dFB) have emerged as a key sleep-regulating area. To dissect the contribution of individual dFB neurons to sleep, we undertook an intersectional Split-GAL4 genetic screen focusing on cells contained within the 23E10-GAL4 driver, the most widely used tool to manipulate dFB neurons. In this study, we demonstrate that 23E10-GAL4 expresses in neurons outside the dFB and in the fly equivalent of the spinal cord, the ventral nerve cord (VNC). Furthermore, we show that 2 VNC cholinergic neurons strongly contribute to the sleep-promoting capacity of the 23E10-GAL4 driver under baseline conditions. However, in contrast to other 23E10-GAL4 neurons, silencing these VNC cells does not block sleep homeostasis. Thus, our data demonstrate that the 23E10-GAL4 driver contains at least 2 different types of sleep-regulating neurons controlling distinct aspects of sleep behavior.

Fig 1. The 23E10-GAL4 driver contains non-dFB sleep-promoting neurons.

(A) Cartoon depicting known sleep-regulating centers in the fly brain. (B) Diagram of the experimental assay. Sleep was measured at 22°C for 2 days to establish baseline sleep profile. Flies were then shifted to 31°C for 24 h at the start of day 3 to increase activity of the targeted cells by activating the TrpA1 channel and then returned to 22°C. White bars (L) represent the 12 h of light and black bars (D) represent the 12 h of dark that oscillate daily. (C, D) Sleep profile in minutes of sleep per hour for day 2 (22°C, blue line) and day 3 (31°C, red line) for parental control female flies: 23E10-GAL4/+ (C) and UAS-TrpA1/+ (D). (E) Sleep profile in minutes of sleep per hour for day 2 (22°C, blue line) and day 3 (31°C, red line) for 23E10-GAL4>UAS-TrpA1 female flies. (F) Box plots of total sleep change in % ((total sleep on day 3-total sleep on day 2/total sleep on day 2) × 100) for data presented in (C–E). Flies expressing UAS-TrpA1 in 23E10-GAL4 significantly increase sleep when switched to 31°C compared with parental controls, Kruskal–Wallis ANOVA followed by Dunn’s multiple comparisons. ****P < 0.0001, n = 25–31 flies per genotype. (G) Box plots of locomotor activity counts per minute awake for flies presented in (C–E). Two-way repeated measures ANOVA followed by Sidak’s multiple comparisons test found no differences in locomotor activity between 22°C and 31°C, n.s. = not significant, n = 25–31 flies per genotype. (H, I) Representative confocal stacks of a female 23E10-GAL4>UAS-mCD8GFP brain (H) and VNC, (I). Green, anti-GFP; magenta, anti-nc82 (neuropile marker). (J) Cartoon depiction of the original goal of our Split-Gal4 screen, which was to identify which dFB neurons modulate sleep. (K) Sleep profile in minutes of sleep per hour for day 2 (22°C, blue line) and day 3 (31°C, red line) for empty control (Empty-AD; 23E10-DBD>TrpA1) flies. (L) Sleep profile in minutes of sleep per hour for day 2 (22°C, blue line) and day 3 (31°C, red line) for 23E12-AD; 23E10-DBD>TrpA1 flies. (M) Box plots of total sleep change in % for female control (Empty-AD; 23E10-DBD) and 23E12-AD; 23E10-DBD flies expressing UAS-TrpA1. A two-tailed unpaired t test revealed that 23E12-AD; 23E10-DBD>TrpA1 flies increase sleep significantly more than control flies when transferred to 31°C. ****P < 0.0001, n = 44–51 flies per genotype. (N) Box plots of locomotor activity counts per minute awake for flies presented in (M). Two-way repeated measures ANOVA followed by Sidak’s multiple comparisons test found that locomotor activity per awake time is increased in 23E12-AD; 23E10-DBD>TrpA1 flies transferred to 31°C. ****P < 0.0001, n = 44–51 flies per genotype. (O) Representative confocal stacks of an Empty-AD; 23E10-DBD>UAS-mCD8GFP female brain (left panel) and VNC (right panel). Green, anti-GFP; magenta, anti-nc82 (neuropile marker). (P) Representative confocal stacks of an 23E12-AD; 23E10-DBD>UAS-mCD8GFP female brain (left panel), VNC (middle panel left), a side view of the VNC (middle panel right) as well as a magnified view of VNC “bowtie” processes in the brain as indicated by the gray rectangle. Yellow arrows indicate TPN1-like processes in the VNC. Yellow asterisks indicate TPN1-like cell bodies. Gray arrows indicate “bowtie” neurons processes in the VNC. Gray asterisks indicate “bowtie” neurons cell bodies. Green, anti-GFP; magenta, anti-nc82 (neuropile marker). (Q) Diagram of the experimental assay. Sleep was measured in retinal-fed and vehicle-fed flies for 2 days without 627 nm LED stimulation to establish baseline sleep profile. LEDs were then turned on for 24 h at the start of day 3 to increase activity of the targeted cells by activating the CsChrimson channel and then turned off on day 4. White bars (L) represent the 12 h of light and black bars (D) represent the 12 h of dark that are oscillating daily. (R) Box plots of total sleep change in % ((total sleep on day 3-total sleep on day 2/total sleep on day 2) × 100) for control (Empty-AD; 23E10-DBD) and 23E12-AD; 23E10-DBD female flies expressing CsChrimson upon 627 nm LED stimulation. Two-way ANOVA followed by Sidak’s multiple comparisons revealed that total sleep is significantly increased in 23E12-AD; 23E10-DBD>UAS-CsChrimson female flies stimulated with 627 nm LEDs when compared with vehicle-fed flies. ****P < 0.0001, n.s. = not significant, n = 24–32 flies per genotype and condition. The raw data underlying parts F, G, M, N, and R can be found in S1 Data. DANs, dopaminergic neurons; dFB, dorsal fan-shaped body; DN1, dorsal neurons 1; DPM, dorsal paired medial neurons; EB, ellipsoid body; FSB, fan-shaped body; l-LNv, large lateral neurons ventral; LPN, lateral posterior neurons; MB, mushroom body; MBONs, mushroom body output neurons; PI, pars intercerebralis; s-LNv, small lateral neurons ventral; vFB, ventral fan-shaped body; VNC, ventral nerve cord.

Fig 2. The 23E10-GAL4 driver contains sleep-promoting neurons that are located in the VNC.

(A-D) Representative confocal stacks of a female VT020742-AD; 23E10-DBD>UAS-GFP brain (A), brain sagittal view (B), VNC (C), and VNC sagittal view (D). The cell bodies panel in (C) is a magnified view of the metathoracic area of the VNC. Yellow arrows indicate TPN1-like processes and gray arrows “bowtie” processes. Green, anti-GFP; magenta, anti-nc82 (neuropile marker). A = anterior, P = posterior, V = ventral, and D = dorsal. (E, F) Representative confocal stacks of a female 30A08-AD; 23E10-DBD>UAS-GFP brain (E) and VNC (F). The cell bodies panel in (F) is a magnified view of the metathoracic area of the VNC. Yellow arrows indicate TPN1 processes in the brain and VNC. Green, anti-GFP. (G–J) Representative confocal stacks of a female VT013602-AD; 23E10-DBD>UAS-GFP brain (G), brain sagittal view (H), VNC (I), and VNC sagittal view (J). The cell bodies panel in (I) is a magnified view of the metathoracic area of the VNC. Gray arrows indicate “bowtie” processes and cell bodies in (I) and (J). Green, anti-GFP; magenta, anti-nc82 (neuropile marker). A = anterior, P = posterior, V = ventral, and D = dorsal. (K) Box plots of total sleep change in % for control (Empty-AD; 23E10-DBD>UAS-CsChrimson), VT020742-AD; 23E10-DBD>UAS-CsChrimson, 30A08-AD; 23E10-DBD>UAS-CsChrimson and VT013602-AD; 23E10-DBD>UAS-CsChrimson vehicle-fed and retinal-fed female flies upon 627-nm LED stimulation. Two-way ANOVA followed by Sidak’s multiple comparisons revealed that retinal-fed VT020742-AD; 23E10-DBD>UAS-CsChrimson and VT013602-AD; 23E10-DBD>UAS-CsChrimson flies increase sleep significantly when stimulated with 627-nm LEDs when compared with vehicle-fed flies. **P < 0.01, ****P < 0.0001, n.s. = not significant, n = 13–34 flies per genotype and condition. (L) Box plots of locomotor activity counts per minute awake for retinal-fed flies presented in (K). Two-way repeated measures ANOVA followed by Sidak’s multiple comparisons test show that locomotor activity per awake time is not affected when the flies are stimulated with 627-nm LEDs, n.s. = not significant, n = 15–34 flies per genotype. (M) Box plots of total sleep change in % for control (Empty-AD; 23E10-DBD>UAS-CsChrimson) and VT013602-AD; 23E10-DBD>UAS-CsChrimson vehicle-fed and retinal-fed female flies upon 627-nm LED stimulation measured with the multibeam MB5 system. Two-way ANOVA followed by Sidak’s multiple comparisons revealed that retinal-fed VT013602-AD; 23E10-DBD>UAS-CsChrimson flies increase sleep significantly when stimulated with 627-nm LEDs when compared with vehicle-fed flies. ***P < 0.001, n.s. = not significant, n = 19–24 flies per genotype and condition. (N) Arousal threshold in vehicle-fed and retinal-fed VT013602-AD; 23E10-DBD>UAS-CsChrimson female flies. Percentage of flies awakened by a stimulus of increasing strength (1, 2, and 4 downward movements in the SNAP apparatus) with and without 627-nm LEDs stimulation. Two-way ANOVA followed by Sidak’s multiple comparisons indicates that in retinal-fed flies activation of VT013602-AD; 23E10-DBD neurons reduce the responsiveness to the 1 and 2 stimulus strength when compared with non-activated flies. No difference in responsiveness is seen at the strongest stimulus (4). Two-way ANOVA followed by Sidak’s multiple comparisons indicates that in vehicle-fed flies, no difference in responsiveness is seen between LED stimulated and non-stimulated flies. ****P < 0.0001, n.s. = not significant, n = 16 flies per genotype and condition. The raw data underlying parts (K–N) can be found in S1 Data. SNAP, sleep-nullifying apparatus; TPN1, taste projection neurons 1; VNC, ventral nerve cord.

Fig 3. Silencing VNC-SP neurons reduces sleep.

(A) Sleep profile in minutes of sleep per hour for control (Empty-AD; 23E10-DBD>UAS-Kir2.1, blue line) and VNC-SP>Kir2.1 (VT013602-AD; 23E10-DBD>UAS-Kir2.1, red line) female flies. (B) Box plots of total sleep time (in minutes) for flies presented in (A). A two-tailed unpaired t test revealed that total sleep is significantly reduced in VNC-SP>Kir2.1 female flies compared to controls. ****P < 0.0001, n = 58–60 flies per genotype. (C) Box plots of locomotor activity counts per minute awake for flies presented in (A). Two-tailed Mann–Whitney U tests revealed that activity per minute awake is significantly reduced in VNC-SP>Kir2.1 female flies compared to controls. **P < 0.001, n = 58–60 flies per genotype. (D) Box plots of total sleep time (in minutes) for control (Empty-AD; 23E10-DBD>UAS-Kir2.1) and VNC-SP>UAS-Kir2.1 female flies measured with the multibeam MB5 system. A two-tailed unpaired t test revealed that total sleep is significantly reduced in VNC-SP>Kir2.1 female flies compared to controls, *P < 0.05, n = 28–31 flies per genotype. (E) Sleep profile in minutes of sleep per hour for control (Empty-AD; 23E10-DBD>UAS-Shi^ts1) female flies at 22°C (blue line) and 32°C (red line). (F) Sleep profile in minutes of sleep per hour for VNC-SP>UAS-Shi^ts1 female flies at 22°C (blue line) and 32°C (red line). (G) Box plots of nighttime sleep change in % for female flies presented in (E) and (F). Two-tailed unpaired t tests revealed that VNC-SP>Shi^ts1 flies lose significantly more sleep when transferred to 32°C compared with controls. *P < 0.05, n = 27–31 flies per genotype. The raw data underlying parts (B), (C), (D), and (G) can be found in S1 Data. VNC-SP, VNC sleep-promoting.

Fig 4. VNC-SP neurons are cholinergic.

(A–C) Representative confocal stacks focusing on the metathoracic ganglion of the VNC of female VT013602-AD; 23E10-DBD>UAS-GFP flies stained with antibodies to ChAT (A), VGlut (B), and GABA (C). Gray arrows in (A) indicate colocalization of GFP and ChAT staining in VNC-SP neurons. Yellow arrows in (B) and (C) indicate the localization of VNC-SP neurons. Green, anti-GFP; magenta, anti-ChAT (A), anti-VGlut (B), anti-GABA (C). (D) Sleep profile in minutes of sleep per hour for control (Empty-AD; 23E10-DBD>UAS-ChAT^{RNAi 60028}, blue line) and VNC-SP>ChAT^{RNAi 60028} (VT013602-AD; 23E10-DBD>UAS-ChAT^{RNAi 60028}, red line) female flies. (E) Box plots of total sleep time (in minutes) for flies presented in (D). A two-tailed unpaired t test revealed that VNC-SP> ChAT^{RNAi 60028} flies sleep significantly less than controls. ****P < 0.0001, n = 22–30 flies per genotype. (F) Box plots of daytime sleep bout duration (in minutes) for flies presented in (D). A two-tailed Mann–Whitney U test revealed that VNC-SP> ChAT^RNAi flies daytime sleep bout duration is significantly reduced compared with controls. ****P < 0.0001, n = 22–30 flies per genotype. (G) Box plots of nighttime sleep bout duration (in minutes) for flies presented in (D). A two-tailed Mann–Whitney U test revealed that VNC-SP> ChAT^RNAi flies nighttime sleep bout duration is significantly reduced compared with controls. ****P < 0.0001, n = 22–30 flies per genotype. (H) Box plots of locomotor activity counts per minute awake for flies presented in (D). A two-tailed Mann–Whitney U test revealed that there is no difference in waking activity between VNC-SP> ChAT^RNAi flies and controls, n.s. = not significant, n = 22–30 flies per genotype. (I) Box plots of total sleep change in % for vehicle-fed and retinal-fed VT013602-AD; 23E10-DBD>UAS-CsChrimson; UAS-GFP.VALIUM10 (control) and VT013602-AD; 23E10-DBD>UAS-CsChrimson; UAS-ChAT^{RNAi 60028} flies stimulated by 627-nm LEDs. A two-way ANOVA followed by Sidak’s multiple comparisons shows that sleep is significantly increased in control flies and that expressing ChAT RNAi in VNC-SP neurons completely abolishes the sleep-promoting effect of VNC-SP neurons activated by 627-nm LEDs. ****P < 0.0001, n.s. = not significant, n = 31–36 flies per genotype and condition. The raw data underlying parts (E–I) can be found in S1 Data. VNC, ventral nerve cord; VNC-SP, VNC sleep-promoting.

Fig 5. Multiple sleep-regulating regions in 23E10-GAL4 with functional specificity?

(A) Box plots of total sleep recovered in % during the first 24 h following 12 h of sleep deprivation at night for 23E10-GAL4/+ and UAS-Kir2.1/+ parental controls and 23E10-GAL4>UAS-Kir2.1 female flies. A one-way ANOVA followed by Tukey’s multiple comparisons demonstrate that both parental controls have a bigger homeostatic sleep rebound than 23E10-GAL4>UAS-Kir2.1 female flies. *P < 0.05, n.s. = not significant, n = 47–62 flies per genotype. (B) Box plots of total sleep recovered in % during the first 24 h following 12 h of sleep deprivation at night for Empty-AD; 23E10-DBD>UAS-Kir2.1 and VT013602-AD; 23E10-DBD>UAS-Kir2.1 female flies. A two-tailed unpaired t test revealed that there is no difference in homeostatic sleep rebound, n.s. = not significant, n = 24–29 flies per genotype. (C) Representative confocal stacks of a female 23E10-GAL4>UAS-GFP fly highlighting different 23E10-GAL4 expressing neurons and their different roles in sleep modulation. The raw data underlying parts (A) and (B) can be found in S1 Data. VNC-SP, VNC sleep-promoting.