Polycomb regulates sleep homeostasis in a specific group of cells in Drosophila brain

Sleep is regulated by the circadian rhythm and homeostasis. Previous studies have shown that epigenetic factors play a crucial role in controlling sleep homeostasis. However, a global view of the epigenetic target genes involved in sleep homeostasis is still lacking. In this study, we have discovered that Polycomb (Pc) regulates sleep in EB1-Gal4 and R69F08-Gal4 co-expressing cells, which are important for controlling sleep homeostasis. Using Targeted DAM-ID (TaDa) and RNAseq, we have identified Pc targets in EB1-Gal4 and R69F08-Gal4 expressing cells and analyzed Pc targets that are crucial for the control of sleep homeostasis. Furthermore, identifying common targets in EB1-Gal4 and R69F08-Gal4 marked cells provides a global view of the cluster-specific target genes of Pc, which potentially play a crucial role in controlling sleep homeostasis. The clustering of these target gene profiles reveals their enrichment distributions. Further experiments demonstrated that Pc stabilized the expression levels of their targets under conditions of sleep deprivation. This study sheds light on the role of Pc in sleep homeostasis control and identifies its target genes in a cell-specific manner. It also lays the foundation for further studies on the mechanisms of epigenetic regulation of sleep homeostasis

Methods
Fly strains and rearing

All fly stocks were reared on standard corn meal/agar fly food at 25°C, humidity of 60% and light dark cycle of 12:12 LD (12 hours of light followed by 12 hours of darkness). The following fly lines were used in this study: w1118 (BDSC: 5905), EB1-Gal4 (BDSC: 44409), R69F08-Gal4 (BDSC: 39499), UAS-RedStinger (BDSC: 8547) are from the Bloomington stock center (BDSC). UAS-PcRNAi-1 (THU1990), UAS-PcRNAi-2 (THU1306), UAS-GABA-B-R1RNAi-1 (THU2990), UAS-GABA-B-R1RNAi-2 (THO1472.N), UAS-GrdRNAi (THU5922), UAS-SecClRNAi (THU2480), UAS-CG12344RNAi (THU2249), UAS-nAChRalpha3RNAi (THU2756), UAS-HdcRNAi (THU2140), UAS-HisCl1RNAi (THU2851), UAS-Daao1RNAi (THO4274.N), UAS-Daao2RNAi (THO4297.N) are from the TsingHua Fly Center (THFC). UAS-stop>kir2.1; LexAop-Flp, UAS-stop>TNT; 8×LexAop-Flp, LexAop-Flp; UAS-stop>shits, 8×LexAop-Flp; UAS-stop>TrpA1, LexAop-kir2.1, UAS-StingerGFP, lexAop-tomato::nls, 13×LexAop-FLP; UAS-FRT>stop>FRT-myrGFP, tub-Gal80ts, UAS-inactive-TNT, UAS-Strong-TNT, UAS-kir2.1 and UAS-dTrpA1 are gifts from Dr. Yi Rao’s lab (Peking University, Beijing, China). R69F08-lexA was generated based on R69F08-Gal4 in this study by using the Site-specific recombination [28]. All behavior tests were performed in isogenous backgrounds.

Activity measurement, sleep deprivation and sleep analysis

Sleep measurement was conducted using the Drosophila Activity Monitor (DAM) system (Trikinetics, MA, US). The methodology for sleep measurement was consistent with previous studies conducted in our laboratory [19, 45, 46]. Briefly, flies were subjected to a 12-hour light/dark (LD) cycle at a temperature of 25°C, unless otherwise stated. Male flies aged between 4-7 days were placed in 65 mm × 5 mm tubes containing fly food and allowed to acclimate for 24-36 hours prior to sleep recording. Subsequently, sleep behavior was recorded continuously for 3 days under LD conditions without any disruptions. For the experiments with* tub-Gal80ts* (Fig. 1A-D; Fig. S1 and Fig. 5), to prevent Pc downregulation during development, flies were raised at 18°C in LD cycle. For sleep behavior with UAS-PcRNAi-1 and UAS-PcRNAi-2, the newly eclosed male flies were transferred to 29°C to allow Gal4 activity. Sleep behavior was recorded 3 days at 29°C in LD condition. The DAM system records each time a fly breaks an infrared beam that bisects the center of each glass tube, and activity counts are collected in a 1 min bin. Sleep was defined as resting for more than 5 minutes [47, 48]. Sleep curve and total sleep were analyzed using Pysolo software 

For sleep deprivation experiments, sleep deprivation was carried out using the DAMs monitors as described previously [45]. For sleep deprivation experiments involving *tub-Gal80ts *(Fig. 1E-P; Fig. 5C-I), conditions for the induction of Gal4 activity were performed as described above, sleep measurement was performed at 29°C until the end of the experiment. For sleep deprivation experiments at 25°C (Fig. 3D-I; Fig. 8; Fig. 14A-B), sleep measurement was performed at 25°C until the end of the experiment. Baseline sleep was measured for 2 days prior to sleep deprivation. Sleep deprivation was achieved by loading flies in DAMs monitors into the shaker (Oscillation mixer, #QB-600, Kylin-Bell lab instruments) fitted with a custom base. Flies were vortexed from ZT12 (Zeitgeber time 12: beginning of the dark phase) to ZT24 (Zeitgeber time 24: beginning of the light phase) at the lower intensity setting for 3sec/min. Zeitgeber time is used to indicate the phase of the circadian rhythm in relation to a specific Zeitgeber. ZT12 is the time when the dark phase begins (or dusk). ZT0 (ZT24) corresponds to the time when the light/dark cycle begins (often referred to as the “lights-on” time or dawn). The 24-hour cycle is divided into 24 Zeitgeber times, where ZT0 represents the start of the light phase, and ZT12 corresponds to the start of the dark phase. The day time increase of total sleep amount after sleep deprivation was calculated by subtracting the baseline daytime sleep amount of each repeat from the corresponding daytime sleep amount after the sleep deprivation. The night time decrease of total sleep amount after sleep deprivation was calculated by subtracting the baseline night time sleep amount of each repeat by the corresponding night time sleep amount after the sleep deprivation. The percentage of sleep recovery was calculated by (day time increase of total sleep amount) / (night time decrease of total sleep amount). Sleep recovery (min) was calculated by (daytime sleep amount after sleep deprivation - baseline daytime sleep amount before sleep deprivation).

For experiments involving LexAop-Flp; UAS-stop>shits, 8×LexAop-Flp; UAS-stop>TrpA1 and UAS-dTrpA1, flies were raised at 21°C in LD 12:12 condition until 4-7 days old (Fig. 6C-L; Fig. 7). Male flies were monitored at 21°C for at least one day followed by shifting to 29°C for 12 hrs at night (ZT12-ZT24) to activate the cells. After the heat pulse, the incubator temperature dropped to 21°C and the recovery sleep was measured. For experiments involving UAS-stop>kir2.1; LexAop-Flp, UAS-stop>TNT; 8×LexAop-Flp, LexAop-kir2.1, UAS-kir2.1, UAS-inactive-TNT and UAS-Strong-TNT, flies were raised at 25°C in LD 12:12 condition until 4-7 days old, sleep measurement was performed at 25°C until the end of the experiment. Sleep propensity was measured by examining the percentage of flies keep asleep in five minutes after light on. Waking activity was measured by examining activity counts in 24 hrs after being fully aroused.

All statistical tests were conducted using GraphPad prism. The statistical tests for each experiment are shown in the figures. The sleep parameters were tested by unpaired Student’s t-test and One-way ANOVA followed by Tukey post hoc test. n.s. indicates no significant difference, * indicates P < 0.05, ** indicates P < 0.01, and *** indicates P < 0.001.

The sample size for all behavioral tests was determined previously [47, 48]. Flies were randomly selected for behavioral tests upon eclosion. Steps for blinding of the investigators were taken to minimize the subjective bias when analyzing the data. All sleep tests were done in males.

Immunostaining and imaging

Imaging was performed as previously described [27] with minor modifications. Briefly, male fly brains were dissected in cold 0.03% PBST and fixed in 2% PFA for 55 min at RT. After brief washes in washing buffer (1×PBS with 1% Triton X-100), the brain samples were mounted on glass slides with Mounting medium containing anti-fading and DAPI (Solarbio, #S2110). For Fig. 4 and Fig. 3A-C, 10-day-old fly brains were incubated with antibodies. The concentration used were 1:1000 for rabbit anti-GFP (Thermo Fisher Scientific, #A11122), 1:200 for Mouse anti-Repo (Developmental Studies Hybridoma Bank, #8D12) and 1:200 for Rat anti-Elav (Developmental Studies Hybridoma Bank, #7E8A10). After brief washes in washing buffer, brains were incubated with secondary antibodies. The concentration used were 1:200 for Alexa 488 anti-rabbit (ABclonal, #AS011) and 1:200 for Alexa 568 anti-mouse (Thermo Fisher Scientific, #A1104) and Alexa FluorTM 647 anti-mouse (Thermo Fisher Scientific, #A21247). The imaging was performed on Leica SP8 confocal microscope.

Quantitative real-time PCR

The process for total RNA extraction from fly heads involved the use of 40 fly heads from 3-5-day-old flies for each sample. Total RNA was extracted using Trizol Reagent, and the samples were then subjected to reverse transcription and quantitative real-time PCR. Specifically, we employed the HiScript III All-in-one RT SuperMix Perfect (Perfect for qPCR) (Vazyme, #R333) for reverse transcription and the ChamQ SYBR qPCR Master Mix (High ROX Premixed) (Vazyme, #Q341) for real-time PCR.

For qPCR quantification, Rp49 was used as normalization control, the delta-delta CT method was used for quantification. The sequence of primers is shown in Table S4. The significance of differences between genotypes was tested by Student’s t-test (GraphPad Prism). All experiments were done with at least 3 biological repeats, and three technical repeats were done for each biological repeat. All quantitative RT-PCR were performed using Step One Real-Time PCR system (Applied Biosystems).

TaDa sample preparation and data analysis

Tada plasmids were obtained from Andrea H Brand’s laboratory [20]. UAS-PcDam was generated in our previous study [19]. The genotypes used for the final experiments were tub-Gal80ts/+; UAS-Dam/EB1-Gal4, tub-Gal80ts/+; UAS-PcDam/EB1-Gal4, tub-Gal80ts/+; UAS-Dam/R69F08-Gal4, and tub-Gal80ts/+; UAS-PcDam/R69F08-Gal4. The experimental protocol details were described by Marshall et al., 2016 [20] with some minor adjustments. Briefly, flies were raised at 21°C until eclosion. Subsequently, the newly eclosed flies underwent a 72-hour heat shock at 29°C. Fly head tissue was collected at ZT12, and genomic DNA was extracted from approximately 100 fly heads. The TaDa experiments were then conducted as per the description provided by Marshall et al., 2016 [20]. The TaDa samples were utilized for TaDa-seq, which was carried out by Novogene in Beijing, China.

Tada data analysis: The damidseq_pipeline was used to analyze Tada data [20, 49]. Transcriptome sequencing data was compared with Drosophila genome annotation release 6.37 using Bwa software, genome sequence was divided into 395891 intervals of different lengths according to methylated GATC sites by the damidseq_pipeline. Additionally, Log2 ratio files were generated by comparing Dam-Pc over Dam-only samples and subjected to median normalization for further analysis. In this study, we used the coverage ratio of genes as the threshold (coverage ratio≥0.182 for DamPc/EB1-Gal4 and coverage ratio ≥ 1.028 for DamPc/R69F08-Gal4) to screen out meaningful regions. To ensure the accuracy of our results, macs2 software was also used to predict the enrichment of transcripts on the genome besides using Damidseq_pipeline software. After macs2 software screening (macs2 threshold is 0.2), there are 32817 peaks in the R69F08-Gal4/PcDam samples and 77481 peaks in the EB1-Gal4/PcDam samples. Further, analysis revealed that 7781 peaks and 13128 peaks fell within the gene region, respectively. In our study, we intersect Damidseq_pipeline and macs2 results, we found that 6089 Pc binding target genes exist in the R69F08-Gal4 cells and 9252 target genes exist in the EB1-Gal4 cells.

Pc binding genes and down-regulated Pc genes in the R69F08-Gal4 cells and EB1-Gal4 cells were used for volcano plots, dot plots, KEGG enrichment, and Gene Ontology (GO) enrichment analysis. Tada data peaks were annotated to genomic features and the nearest genes of TSS using R package Chipseeker (v1.31.3.900) [50]. Volcano plots were plotted by R-package ggplot2 (v3.2.1) [51].  KEGG enrichment, GO enrichment and dot plots were performed by using the R package clusterProfiler (v4.3.3) [52] and org.Dm.eg.db (v3.7.0) dictionary. The simplify output from compareCluster by removing redundancy (cutoff=0.7, by=”p. adjust”) of enriched GO terms. Venn diagrams were generated by R package VennDiagram (v1.7.1) [53].

RNA-seq sample preparation and data analysis

For the RNA-seq analysis in this study, Drosophila heads of specific genotypes at ZT12 were utilized. The RNA-seq experiments were conducted by Biomics Company in Beijing, China, following standard protocols. To perform differential expression analysis, the EBSeq method [54] was employed. This approach allows for the identification of genes that exhibit significant differences in expression between control and treatment conditions. Specifically, genes with a mean fold change greater than 1.2 were considered as significantly changed and selected for further analysis. Each sample used in the analysis consisted of 50 fly heads. 

Genotypes for RNA-seq:

tub-Gal80ts/+; EB1-Gal4/+

tub-Gal80ts/+; UAS-PcRNAi/EB1-Gal4

tub-Gal80ts/+; R69F08-Gal4/+

tub-Gal80ts/+; UAS-PcRNAi/R69F08-Gal4