(A) Third instar larval skeletal musculature stained with Phalloidin (F-actin). In this and subsequent figures, we assayed the ventral longitudinal muscles (highlighted in green). (B–D) GFP–Atg8, overexpressed using the Dmef2–Gal4 driver, labels autophagosomes. Dmef2–Gal4, UAS–GFP–Atg8 animals were fed on high-nutrient food (B), starved on low-nutrient food for 6 h (C), or starved on low-nutrient food +2.5 mg/ml CQ for 6 h (D). GFP–Atg8-labeled vesicles appeared only in the starved animals (C–D), localizing around the nucleus and between myofibers. (D) CQ treatment caused accumulation of bloated GFP–Atg8-labeled vesicles. (E–G) Dmef2–Gal4, UAS–GFP–Atg8/UAS–HRP–Lamp1 animals were assayed for Lamp1 and Atg8 localization (anti-HRP, red; GFP, green; DAPI, blue). . (E) High-nutrient food suppressed formation of both GFP–Atg8 and HRP–Lamp1-labeled vesicles. (F) Colocalization of GFP–Atg8 and HRP–Lamp in animals starved on low-nutrient food. The yellow arrowhead points to a vesicle positive for both Atg8 and Lamp. (G) Addition of CQ to the starvation diet resulted in accumulation of both GFP–Atg8 and HRP–Lamp-labeled vesicles, but they failed to colocalize. (H) Quantification of the number of GFP–Atg8, HRP–Lamp1, or GFP–Atg8+HRP–Lamp1 vesicles in starved or starved +CQ muscles. (I–M) The core Atg genes are required for starvation-induced autophagy in both wild-type and CQ-treated skeletal muscles. Dmef2–Gal4, UAS–GFP–Atg8/UAS–Atg1 larvae were starved on low-nutrient food for 6 h (I) or starved on low-nutrient food +2.5 mg/ml CQ for 6 h (J). Note that Atg1 knockdown completely abolished the formation of GFP–Atg8-labeled autophagosomes (compare I–J to C–D). (K–L) Dmef2–Gal4, UAS–GFP–Atg8, Atg1^Δ3d larvae failed to form GFP–Atg8 vesicles when starved or starved and treated with CQ. (M) Quantification of autophagy changes due to Atg gene knockdown in Dmef2–Gal4, UAS–GFP–Atg8 larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h. Each of the 10 UAS–Atg RNAi transgenes tested caused a highly significant decrease (p<.01) in the total area of GFP–Atg8 vesicles. SEM is indicated, with n = 5 ventral longitudinal muscles from individual animals. (N–O) EM of muscles from Dmef2–Gal4, UAS–whitei larvae. Animals starved on low-nutrient food +2.5 mg/ml CQ (O) accumulated vesicles in the intermyofibril spaces (red asterisk), disrupting the integrity of the sarcomere compared to non-CQ-treated control muscles (N). (P) CQ treatment increased the larval crawling time of Dmef2–Gal4, UAS–whitei larvae in starved animals, and weakly in fed animals. (Q) CQ treatment increased the larval righting time of Dmef2–Gal4, UAS–whitei larvae in starved but not fed animals. For both locomotor assays, SEM is indicated for n = 10 larvae (*p<.05, **p<.01).

(A) Sectioned third instar OreR larva stained with Periodic acid-Schiff (PAS). The muscles, but not the fat body, are stained purple, indicating high levels of glycogen (m, muscle; bw, body wall; fb, fat body). (B) Glycogen was also detected in muscle from Dmef2–Gal4, UAS–GFP–Atg8 larvae, immunostained with an antiglycogen monoclonal antibody. (C) GFP–Atg8 vesicles colocalized with glycogen in Dmef2–Gal4, UAS–GFP–Atg8 larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h (GFP, green; antiglycogen, red). (D) HRP–Lamp1 vesicles show less colocalization with glycogen in UAS–HRP–Lamp1;Dmef2–Gal4 larvae starved and treated with CQ. (E) Quantification of GFP–Atg8 or HRP–Lamp1 vesicles with glycogen. (F–G) EM from Dmef2–Gal4, UAS–whitei larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h. (F) Double- and single-membrane vesicles containing glycogen granules accumulated between myofibers (s, sarcomere; m, mitochondrion; AVs, autophagic vesicles). (G) Higher magnification view of region outlined in (E). (H) CQ treatment is not required for glycogen autophagy as seen in an EM from a Dmef2–Gal4, UAS–whitei larva starved on low-nutrient food for 6 h. Arrow points to double membrane.

(A–D) Time course of autophagy induction in Dmef2–Gal4, UAS–GFP–Atg8 muscles, accompanied by quantification of GFP–Atg8 and glycogen colocalization. Animals were fed for 18 h in high-nutrient food +2.5 mg/ml CQ, then starved on low-nutrient food +2.5 mg/ml CQ for 0–8 h (antiglycogen, red; GFP, green; DAPI, blue). (A) At time point 0, following 18 h in high-nutrient food +CQ, the muscles contained large amounts of glycogen with no apparent autophagy. (B) At 3 h of starvation, glycogen stores were still high, and GFP–Atg8-labeled vesicles began to appear. (C–D) At 6 and 8 h of starvation, the majority of GFP–Atg8-labeled vesicles colocalized with glycogen. (E) Time course of glycogen levels in Dmef2–Gal4 carcasses (muscle+body wall). Animals were fed for 24 h in high-nutrient food, then starved on low-nutrient food +/− 2.5 mg/ml CQ for 0–24 h. Starvation caused reduction of glycogen levels in both untreated and CQ-treated larvae over time. However, after 6 h of starvation, CQ treatment significantly increased glycogen levels compared to controls. SEM is indicated for n = 5–8 samples (*p<.05, **p<.01). (F–G) activation of the Tor pathway blocked autophagy in the muscles from larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h. (F) Autophagy levels were high in control Dmef2–Gal4/UAS–whitei larvae. Muscles from (G) UAS-Rheb/+; Dmef2-Gal4/+, (H) Dmef2–Gal4/UAS–Tsc1i, and (I) Dmef2–Gal4/UAS–gigi all failed to induce autophagy.

(A–B) Glycogen phosphorylase is not required for glycogen autophagy (antiglycogen, red; GFP, green; DAPI, blue). (A) UAS–GlyPi/+; Dmef2–Gal4, UAS–GFP–Atg8/+ larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h exhibited high levels of colocalization between GFP–Atg8 and glycogen. (B) Higher magnification of region outlined in (A). (C–F) Dmef2–Gal4, UAS–GFP–Atg8 larvae with GlyP and/or Atg1 knockdown were fed on high-nutrient food for 18 h before being starved on low-nutrient food (antiglycogen, red; GFP, green; DAPI, blue). (C) UAS–GlyPi/+; Dmef2–Gal4, UAS–GFP–Atg8/+ larval muscle contained high levels of glycogen prior to starvation, indicating no defect in glycogen synthesis. (D) Following 24 h starvation UAS–GlyPi/+; Dmef2–Gal4, UAS–GFP–Atg8/+ muscles contained no glycogen detectable by antibody staining. (E) Following 24 h of starvation Dmef2–Gal4, UAS–GFP–Atg8/UAS–Atg1i muscles contained no glycogen. (F) Double-mutant larvae UAS–GlyPi/+; Dmef2–Gal4, UAS–GFP–Atg8/UAS–Atg1i larval muscles contained high levels of glycogen after 24 h of starvation, indicating an inability to break down glycogen. (G) Time course of glycogen levels in Dmef2–Gal4 carcasses (muscle+body wall) with expression of UAS–RNAi transgenes targeting white, GlyP, Atg1, or GlyP+Atg1. Simultaneous knockdown of GlyP and Atg1, but not either gene alone, significantly reduced glycogen degradation compared to the white control after 24 h of starvation, consistent with immunostaining (C–F). Between 6 and 12 h of starvation, individual knockdown of GlyP or Atg1 caused a significant increase in glycogen levels, indicating a reduced rate of glycogen degradation. SEM is indicated for n = 5–8 samples. The p values were calculated relative to white RNAi control at each time point (*p<.05, **p<.01).

(A–D) Glycogen synthase (GlyS) is required for glycogen synthesis in D. melanogaster muscles. PAS staining for glycogen was absent in Dmef2–Gal4/UAS–GlySi muscles (B) compared to control Dmef2–Gal4/UAS–whitei muscles (A). Antiglycogen immunostaining for glycogen was absent in Dmef2–Gal4/UAS–GlySi muscles (D) compared to Dmef2–Gal4/UAS–whitei control muscles (C). (E–K) GlyS is required for the formation of large CQ-induced autophagosomes. Vesicles are much smaller in Dmef2–Gal4, UAS–GFP–Atg8/UAS–GlySi larval muscle starved 6 h in low-nutrient food +2.5 mg/ml CQ (G) than in control Dmef2–Gal4, UAS–GFP–Atg8/UAS–whitei larval muscle (E). (F, H) The difference in autophagosome size is clearly evident at high magnification. (I–K) Quantification of autophagy changes due to GlyS gene knockdown in Dmef2–Gal4, UAS–GFP–Atg8 larvae starved on low-nutrient food +2.5 mg/ml CQ for 6 h. SEM is indicated, with n = 5 (I) or n = 10 (J–K) ventral longitudinal muscles from individual animals (*p<.05, **p<.01). (I) Each of the four UAS-GlyS RNAi transgenes tested caused a significant decrease in the total area of GFP–Atg8 vesicles in the muscle compared to the UAS–whitei control. (J) Vesicle number was unchanged by GlyS knockdown. (K) UAS–GlyS RNAi caused a highly significant decrease in the mean vesicle size (area) compared to the control. (L–M) GlyS gene expression, monitored using a MiMIC transposon insertion (MI01490), showed expression in the larval muscle (L) but not the fat body (M) (green, GFP; blue, DAPI). (N–O) Larvae were starved on low-nutrient food for 6 h prior to dissection of the fat bodies. Autophagy in Cg-Gal4/+; UAS–GFP–Atg8/UAS–GlySi (O) was not substantially different from autophagy in Cg-Gal4/+; UAS–GFP–Atg8/UAS–whitei control fat bodies (GFP, green; DAPI, blue). (P) EM of muscle from Dmef2–Gal4/UAS–GlySi animal starved on low-nutrient food +CQ. Note that the intermyofibril spaces (red asterisk) and sarcomere structure are not distorted. (Q) GlyS or Atg1 knockdown significantly improved the crawling time of larvae treated with CQ and starved for 6 h. SEM is indicated for n = 10 larvae. The p values were calculated relative to white RNAi control larvae (*p<.05, **p<.01).

(A) Protein sequence alignment of the C-terminal region of D. melanogaster, human, and yeast Glycogen synthases. Identical residues are blue; all other residues are red. Conserved in all three species, R593, W609, and S651 are underlined. (B) Western blot/Co-immunoprecipitation (co-IP) showing that Flag–Atg8 binds to a Venus–GlyS or Venus–GlyS (S651A) protein complex in response to starvation. Flag–Atg8 is unable to co-IP with either Venus–GlyS (R593A) or Venus–GlyS (W609A). Venus–GlyS and Venus–GlyS mutants were co-IP'd from muscle lysate from Dmef2–Gal4/UAS–Flag–Atg8 or UAS–Venus–GlyS(WT or mutant)/+;Dmef2–Gal4/UAS–Flag–Atg8 third instar larvae. These were fed on high-nutrient food for 18 h, and then transferred to fresh high-nutrient food or low-nutrient food for 6 h. (C–F) UAS–Venus–GlyS (WT or mutant)/+;Dmef2–Gal4/UAS–Flag–Atg8 larvae were treated with starved 6 h in low-nutrient food +2.5 mg/ml CQ (Venus, green; α-Flag, red). Purple arrows mark examples of the presence or absence of colocalization. (C) Venus–GlyS was localized predominantly to the Flag-labeled autophagosomes, with weak staining in the rest of the cytoplasm. (D) Venus–GlyS (R593A) was found throughout the cytoplasm and did not colocalize with autophagosomes. (E) Venus–GlyS (S651A) was localized to the autophagosomes. (F) Venus–GlyS (W609A) did not colocalize with the Flag-labeled autophagosomes.