Evolutionarily Conserved Regulation of Sleep by the Protein Translational Regulator PERK

Abstract

Sleep is a cross-species phenomenon whose evolutionary and biological function remain poorly understood. Clinical and animal studies suggest that sleep disturbance is significantly associated with disruptions in protein homeostasis-or proteostasis-in the brain, but the mechanism of this link has not been explored. In the cell, the protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK) pathway modulates proteostasis by transiently inhibiting protein synthesis in response to proteostatic stress. In this study, we examined the role of the PERK pathway in sleep regulation and provide the first evidence that PERK signaling is required to regulate normal sleep in both vertebrates and invertebrates. We show that pharmacological inhibition of PERK reduces sleep in both Drosophila and zebrafish, indicating an evolutionarily conserved requirement for PERK in sleep. Genetic knockdown of PERK activity also reduces sleep in Drosophila, whereas PERK overexpression induces sleep. Finally, we demonstrate that changes in PERK signaling directly impact wake-promoting neuropeptide expression, revealing a mechanism through which proteostatic pathways can affect sleep and wake behavior. Taken together, these results demonstrate that protein synthesis pathways like PERK could represent a general mechanism of sleep and wake regulation and provide greater insight into the relationship between sleep and proteostasis.

Keywords: Drosophila; ER stress; PERK; protein synthesis; protein translation; proteostasis; sleep; unfolded protein response; zebrafish.

Figure 1.. Activation of the PERK pathway is associated with wakefulness in Drosophila

(A) Simplified schematic of PERK activation and inhibition of protein synthesis. Activated PERK phosphorylates the alpha subunit of eIF2 to inhibit protein synthesis. The phosphatase GADD34 returns eIF2α to its unphosphorylated state. (B) Quantification from western blot analysis of peIF2α expression in flies collected at the beginning of the lights-on period (ZT0) compared to the end of the day (ZT10). peIF2α protein level is 50% higher near the end of the day compared to the start of the day (N=13 animals, *P<0.05, student’s t-test). Representative peIF2α and eIF2α signal shown on the right. Bar graph shows mean ± SEM. See also Figure S1

Figure 2.. Pharmacological inhibition of PERK decreases sleep in Drosophila and zebrafish.

(A) GSK0606414 alters the ribosomal profile in the Drosophila brain. The proportion of actively translating polysomes is higher following GSK2606414 administration (Fraction 7–21) (a.u. = arbitrary units). (B) Sleep profile of flies treated with GSK2606414 or vehicle (0.5% DMSO). The PERK inhibitor GSK2606414 suppresses nighttime sleep in Drosophila(N=42 animals). (C) Quantification of daytime and nighttime sleep. GSK2606414 significantly reduces sleep at night (****P<0.0001, student’s t-test). (D) Sleep profile of zebrafish during two days and nights. GSK2606414 decreases sleep at night (Vehicle: N = 102; GSK2606414: N= 114). (F) Quantification of daytime and nighttime sleep after vehicle and GSK2606414 administration. Total sleep is reduced during the day and night in zebrafish following GSK2606414 administration (* P<0.05, ****P<0.0001, student’s t-test). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively. See also Figure S2

Figure 3.. Pharmacological activation of eIF2B decreases sleep in Drosophila and zebrafish

(A) ISRIB administration significantly reduces the expression of peIF2α in the brain. (N=9 animals, **P<0.01, student’s t-test). (B) Sleep profile of flies treated with ISRIB or vehicle (0.5% DMSO) (N=28 animals, shaded area represents SEM). (C) Quantification of daytime and nighttime sleep. ISRIB reduces the total amount of sleep at night (N=28, **P<0.01, student’s t-test).(D) Sleep profile of larval zebrafish over the course of two days and nights. ISRIB decreases sleep at night (Vehicle: N =232 animals; ISRIB: N = 240 animals; shaded area represents SEM). (E) Quantification of daytime and nighttime sleep. Total sleep is reduced at night in zebrafish larvae following ISRIB administration (**P<0.01, student’s t-test). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively. See also Figure S2

Figure 4.. Genetic inhibition of peIF2α reduces sleep

(A) Sleep profile of flies expressing PERK RNAi in nSyb-expressing neurons compared to uncrossed parental controls. Expression of PERK RNAi in neurons reduces sleep levels (N=27 animals) (B) Quantification of daily sleep amounts. Total sleep is significantly reduced following PERK RNAi expression in neurons (N=27 animals, P<.01, one-way ANOVA; *P<0.05, ***P<0.001, ****P<0.0001). (C) Sleep profile of flies expressing mGADD34 in nSyb-expressing neurons compared to uncrossed parental controls. Expression of mGADD34 in neurons reduces sleep levels (N=18 animals). (D) Quantification of daytime and nighttime sleep amounts. Daytime sleep amount is not significantly affected by genotype (P= 0.0953, one-way ANOVA). Total nighttime sleep is significantly reduced by mGADD34 expression in neurons (N=18 animals, P<.001, one-way ANOVA; **P<0.01,****P<0.0001, student’s t-test). Total daytime sleep is not significantly changed by mGADD34 expression in neurons (P=0.09, one-way ANOVA). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively. See also Figures S3 and S4

Fig. 5.. Neuronal knockdown of PERK inhibits sleep in Drosophila

(A) RU486 administration reduces the expression of PERK mRNA in the brain of flies expressing PERK RNAi in neurons. Expression is normalized to Act5c and GAPDH housekeeping controls (P<0.01, student’s t-test). (B) Sleep profile of neuron-specific GeneSwitch>PERK RNAi flies administered either RU486 or vehicle (N=28 animals). (C) Quantification of sleep during the day and night. RU486 administration reduces the total amount of sleep at night (N=28 animals, *P<0.05, student’s t-test). (D) RU486 administration does not significantly affect the number of sleep bouts during the day and night in flies expressing PERK RNAi in neurons (N=28 animals). (E) RU486 administration does not significantly alter the average length of sleep bouts during the day but reduces the average sleep bout length at night in flies expressing PERK RNAi in neurons (N=28 animals, **P<0.01, student’s t-test). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively.

Figure 6.. Neuronal overexpression of PERK induces sleep in Drosophila

(A) RU486 administration leads to a ~14-fold enhancement in PERK transcriptional expression in the brain in nSyb GeneSwitch>dPERK flies. Expression is normalized to Act5c and GAPDH housekeeping controls (****P<0.0001, student’s t-test). (B) Sleep profile of nSybGeneSwitch>dPERK flies on vehicle compared to RU486. The gray trace shows sleep amount in progeny of nSyb GeneSwitch and w1118 wildtype controls fed RU486. Successive days on RU486 food incrementally increases sleep time in flies in nSyb GeneSwitch>dPERK transgenic animals (N=28 animals). (C) Quantification of daily sleep amount over the course of the three-day experiment on vehicle/RU486 administration. RU486 significantly increases daily sleep amount in nSyb GeneSwitch>dPERK flies but not in the w1118/nSyb GeneSwitch progeny controls (N=28 animals, ****P<0.0001, student’s t-test) (D) RU486 administration increases the number of daytime sleep bouts in nSyb GeneSwitch>dPERK flies (N=28 animals, ****P<0.0001, student’s t-test) and (E) RU486 increases the average sleep bout length during both the day and night in nSyb GeneSwitch>dPERK flies (N=28, **P<0.01, ****P<0.0001, student’s t-test). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively. See also Figure S5

Figure 7.. Genetic knockdown and overexpression of PERK in PDF neurons reduces and increases sleep and alters PDF expression at projection terminals

(A) Sleep profile of flies expressing PERK RNAi in PDF neurons (PDF>PERK RNAi) compared to parental controls (N=18 animals). (B) Quantification of daytime and nighttime sleep. During the day, flies expressing PERK RNAi in PDF neurons (PDF>PERK RNAi) sleep less than PDF GAL4 parental controls but not UAS PERK RNAi controls, which also slept less than the PDF GAL4 parental line (P<0.05, one-way ANOVA; *P<0.05, **P<0.01). At night, flies expressing PERK RNAi in PDF neurons sleep less than both parental controls (P<0.01, one-way ANOVA; **P<0.01). (C) Sleep profile of flies overexpressing PERK in PDF neurons compared to parental controls (N=18 animals) (D) Quantification of daytime and nighttime sleep. Daytime sleep amount is not significantly affected by PERK overexpression in PDF neurons (P= 0.6642, one-way ANOVA). At night, flies overexpressing PERK in PDF neurons (PDF>dPERK) sleep more than parental controls (P<0.05, one-way ANOVA; *P<0.05). (E) Representative confocal images of sLNvs projections in transgenic and parental control flies and quantification of PDF expression. At ZT2 (upper panels), PERK overexpression in PDF neurons significantly suppresses PDF expression at projection terminals compared to parental controls(UAS dPERK, N=7; PDF GAL4, N=7; PDF>dPERK, N=9;, P<0.0001, one-way ANOVA; ****P<0.0001). At ZT14 (lower panels), PERK knockdown in PDF neurons significantly increases PDF expression at projection terminals compared to parental controls (UAS PERK RNAi, N=10; PDF GAL4, N=11; PDF> PERK RNAi, N=12; P<0.0001, one-way ANOVA; ****P<0.0001). (α-PDF, 1:1000; AlexaFluor 594 1:500). Line and bar graphs show mean ± SEM. White and black boxes under line graphs indicate day and night, respectively. See also Figures S6 and S7