Abstract
Sleep is a universal and tightly regulated process that is controlled by both circadian and homeostatic mechanisms[1]. Work in Drosophila melanogaster has shown that sleep homeostasis is largely governed by the dorsal fan-shaped body (dFB). Within this region, some dFB neurons monitor the need to sleep through changes in intrinsic excitability. As sleep pressure builds, their input-output function becomes biased toward spike generation, whereas excitability returns toward baseline after rebound sleep[2] in a process linked to mitochondrial reactive oxygen species (ROS)[3]. Prolonged periods of wakefulness elevate ROS-derived carbonyls, which are reduced by Hyperkinetic, an aldoketoreductase enzyme binding the cofactor NADP(H)[3][,][4]. The resulting change in cofactor redox state decelerates potassium channel inactivation, increases excitability and promotes sleep[3][,][4]. In line with this mechanism, dampening ROS levels and disrupting the excitability shift, or the Hyperkinetic-dependent redox-sensing mechanism, results in insomnia[2][,][3]. Conversely, production of non-radical ROS at the plasma membrane, i.e., where the functional Shaker-Hyperkinetic ion channel complex is localized, increases the excitability of dFB neurons and promotes sleep[3][,][4]. Together, these observations suggest that changes in mitochondrial oxidation in dFB neurons convey sleep need by coupling metabolic state to neuronal excitability[3][,][4][,][5]. However, the original R23E10-GAL4 driver line used to identify this mechanism has been recently shown to also label sleep-promoting ventral nerve cord (VNC-SP) cells in addition to dFB neurons[6]. Although prior electrophysiological and imaging experiments only targeted dFB neurons[3][,][4][,][5], the interpretation of the sleep phenotypes may be confounded by contributions from both dFB and VNC-SP neurons. Indeed, a recent study suggested that dFB neurons may in fact not have a sleep-promoting role and that the redox-dependent mechanism may act in the ventral nerve cord instead[7]. To resolve this uncertainty, we directly compared the functional roles of dFB and VNC-SP neurons in redox-dependent sleep control using behavioral sleep assays, redox-state manipulations, and split-GAL4 (Sp-GAL4) lines that segregate these neuronal populations. As a result, we demonstrate that the redox-sensing mechanism operates specifically in dFB neurons to promote sleep.
