The expression of UAS-Dp RNAi transgene driven by the muscle (Mef2-GAL4) or fat body (cg-GAL4) GAL4 drivers specifically depletes Dp protein in the corresponding tissue, which results in two readily observed phenotypes. First, the inactivation of Dp in muscle is accompanied by severely reduced muscle growth in both larval and thoracic muscles (Zappia and Frolov, 2016). Second, fat body-specific Dp depletion triggers a reduction in fat storage and an increase in DNA damage response that promotes uncontrolled DNA replication and, ultimately, leads to the occurrence of binucleated cells (Guarner et al., 2017). Given that the inactivation of Dp in either tissue had a similar impact on viability and caused pupal lethality (Guarner et al., 2017; Zappia and Frolov, 2016; Figure 1—figure supplement 1), we asked whether Dp depletion in muscle may also affect fat body development and, conversely, if Dp deficiency in fat body may cause muscle abnormalities.

We began by comparing muscle structure between Mef2>Dp-RNAi and cg>Dp-RNAi animals. The ventral longitudinal 3 (VL3) muscles in wandering third instar larva were visualized by staining the body walls with phalloidin and 4,6-diamidino-2-phenylindole (DAPI). As previously reported, the area of the VL3 muscles in Mef2>Dp-RNAi animals were significantly smaller than control (Zappia and Frolov, 2016). In contrast, no differences were detected in the VL3 muscle area between cg>RFP and cg>Dp-RNAi larval muscles (Figure 1A–B). To confirm efficiency of Dp depletion in Mef2>Dp-RNAi, the expression of Dp was monitored with the transgene Dp^GFP, which expresses the fusion protein Dp::GFP (Figure 1A; Zappia and Frolov, 2016). Thus, we concluded that the loss of E2F/DP in fat body does not cause the phenotype observed in Dp-depleted muscles.

One of the hallmarks of the loss of Dp in fat body is an increase in the DNA damage response that leads to the appearance of binucleated cells (Guarner et al., 2017). We confirmed that the depletion of Dp driven by the fat body-specific GAL4 driver cg-GAL4 resulted in the formation of binucleated cells (~4.2% in cg>Dp-RNAi, Figure 2A–B) while none were found in fat bodies from wild-type animals. This phenotype was also observed in E2f2; E2f1 double mutant animals (Figure 2—figure supplement 1A) and in Dp null mutants (Guarner et al., 2017). Given that the cg-GAL4 driver expresses GAL4 in both fat body and hemocytes (Pastor-Pareja and Xu, 2011), an additional independent fat body driver, Lpp-GAL4, was used to drive Dp depletion in fat body. The occurrence of binucleated cells in fat body were also observed in Lpp>DpRNAi larva (Figure 2—figure supplement 1B), which died at pupa stage (Figure 1—figure supplement 1), thus confirming the cell-autonomous effect of Dp loss in fat body, in concordance with previous work (Guarner et al., 2017).

Next, we examined the fat body in Mef2>Dp-RNAi larva following muscle-specific Dp depletion by staining the tissue with phalloidin and DAPI. Surprisingly, ablation of Dp in muscles led to the appearance of binucleated cells (7.7% in Mef2>Dp-RNAi, Figure 2A–B). We confirmed that Dp was depleted in a tissue-specific manner, as only cg>Dp-RNAi fat bodies showed reduced levels of Dp protein as revealed by immunofluorescence using anti-Dp antibodies (Figure 2A, bottom panel).

One of the functions of E2F in the fat body is to limit the response to DNA damage. In Dp-depleted fat body, there is an increased recruitment of the DNA damage proteins of the MRN sensor complex, such as Rad50 and Mre11 (Guarner et al., 2017). To determine whether the defects found in the fat bodies of Mef2>Dp-RNAi animals were related to the activation of the DNA damage response, tissues were immunostained with anti-Rad50 (Figure 2C) and anti-Mre11 (Figure 2—figure supplement 1C) antibodies. Notably, the MRN proteins were not recruited in the fat body of Mef2>Dp-RNAi animals, as opposed to cg>Dp-RNAi fat body cells.

We conclude that the loss of Dp in muscles elicits defects in fat body that might be similar to the phenotype seen in Dp-deficient fat body albeit not accompanied by the upregulation of MRN proteins. Thus, Dp-deficient muscle seems to exert a systemic effect on normal tissues, such as fat body.

Leaky expression of GAL4 drivers in other tissues during earlier developmental stages might provide a trivial explanation for the results described above. To exclude this possibility, we used three approaches to confirm the tissue specificity of cg-GAL4 and Mef2-GAL4 drivers used to knock down Dp. First, we examined the real-time and lineage tracing expression of the drivers. We used the system G-TRACE, which combines the FLP recombinase-FRT and the expression of GFP to trace earlier GAL4 expression, and the presence of RFP to identify real-time expression of GAL4 (Evans et al., 2009). G-TRACE showed that the cg-GAL4 driver is expressed in the fat body in agreement with previous report (Pastor-Pareja and Xu, 2011; Figure 3A, left panel). No GFP or RFP signal was detected in larval muscles (Figure 3A, right panel) suggesting that at no point during development cg-GAL4 was expressed in muscles. Similarly, Mef2-GAL4 expression was detected in larval skeletal and smooth muscles (Figure 3B, left panels), and in the adult muscle precursors (wing disc myoblasts), but not in the fat body (Figure 3B, right panels).

Second, the levels of Dp expression in fat body and muscles of Mef2>Dp-RNAi and cg>Dp-RNAi larva were examined by immunofluorescence using Dp antibodies. Third, Dp expression was monitored using a Dp^GFP transgene that expresses a Dp::GFP fusion protein from the endogenous Dp locus (Zappia and Frolov, 2016). The Dp::GFP was efficiently depleted in fat body of cg>Dp-RNAi compared to control (Figure 3C, quantified in Figure 3D). Importantly, the levels of the Dp::GFP remained unaffected in the fat body of Mef2>Dp-RNAi (Figure 3C, quantified in Figure 3D). This result was further confirmed by staining with anti-Dp antibody that showed no changes in the endogenous expression of Dp in fat body of Mef2>Dp-RNAi (Figure 3E). Similarly, the expression of Dp::GFP in muscles was not altered in cg>Dp-RNAi, whereas, as expected, Dp::GFP was significantly reduced in muscles of Mef2>Dp-RNAi compared to control (Figure 3F, quantified in Figure 3G). Using anti-Dp antibody, we further confirmed that the endogenous levels of Dp protein in muscles did not change upon Dp depletion in the fat body (Figure 3F, quantified in Figure 3H).

Thus, the occurrence of binucleated cells in Mef2>Dp-RNAi fat body is not due to altered expression of Dp in fat body of these animals and, therefore, reflects a systemic effect induced by muscle-specific Dp depletion.

The expression of UAS-Dp transgene with either the fat body- or muscle-specific drivers, cg-GAL4 or Mef2-GAL4, can significantly extend viability of Dp mutants (Guarner et al., 2017; Zappia and Frolov, 2016). Given the systemic effect of Dp described above, we asked whether the muscle-specific Dp expression suppresses the defects in fat body of Dp mutants. The UAS-Dp transgene was expressed in the trans-heterozygous Dp^a3/Df(2R)Exel7124 (Dp-/-) mutant animals under the control of cg-GAL4 or Mef2-GALl4. Larval fat bodies were stained with phalloidin and DAPI to visualize the occurrence of binucleated cells. In agreement with previously published data (Guarner et al., 2017), 7.1% of cells in fat body of Dp mutants were binucleated and this phenotype was fully rescued in the Dp-/-; cg>Dp animals (Figure 4A, quantified in Figure 4B). Strikingly, the binucleated phenotype was also largely rescued by re-expression of Dp in muscles of Dp mutants, in Dp-/-; Mef2>Dp animals (Figure 4A–B). We note, however, that fat bodies of Dp-/-; Mef2>Dp animals still contained fragmented and decondensed/large nuclei indicating that the rescue was incomplete. Staining with anti-Dp antibody confirmed the lack of Dp expression in Dp-/-; Mef2>Dp fat bodies (Figure 4A, bottom panel), thus excluding the possibility of a leaky expression of Mef2-GAL4 driver in the fat body.

Next, we asked whether re-expression of Dp in the fat body suppresses muscle defects of Dp mutants. The body walls of third instar larvae were dissected and VL3 muscles of the A4 segment were visualized by staining the tissue with DAPI and phalloidin. As previously reported, VL3 muscle area was smaller in Dp-/- mutant larvae and fully rescued in Dp-/-; Mef2>Dp (Figure 4C, quantified in Figure 4D; Zappia and Frolov, 2016). In contrast, the expression of Dp in the fat body was insufficient to suppress the small size of VL3 muscles in Dp-/- mutant larvae (Figure 4C, quantified in Figure 4D).

Thus, muscle-specific expression of Dp can rescue the binucleated phenotype of Dp mutant fat body, which is consistent with the idea that E2F/Dp in muscle exerts a systemic effect in larva that impacts the fat body. In contrast, re-expressing Dp in the fat body is insufficient to suppress the muscle defects in Dp mutants. This suggests that E2F/Dp modifies the inter-tissue communication between muscle and fat body.

A significant limitation of transcriptional profiles is that it is difficult to know whether a change in mRNA levels leads to a measurable difference in protein level or causes a change in pathway activity. To avoid these issues, we generated proteomic profiles of fat body and skeletal muscles and used these to compare how the loss of E2F affects protein levels in each tissue. Third instar larval fat bodies were collected from both wild type (w¹¹¹⁸) and Dp-/- mutant (Figure 5A, left panel) and subjected to multiplexed quantitative mass spectrometry-based proteomics using tandem-mass tag (TMT) technology (Edwards and Haas, 2016; McAlister et al., 2014). Collecting larval muscles in sufficient quantities for such proteomic profiling was not feasible due to technical challenge of separating larval muscle from adjacent tissue. Therefore, we turned to dissecting thoracic muscles from pharate pupa. We used Mef2>Dp-RNAi pharates since Dp-/- mutants die as early pupa (Figure 5A, right panel). Western blot analysis confirmed that the levels of Dp protein were low in lysates from Dp-/- fat bodies and from Dp-depleted muscles compared to controls (Figure 5—figure supplement 1A). Note that two distinct developmental stages are being compared in our study, and it could potentially introduce additional variations when comparing data between fat body and muscles.

The changes in the proteomic profiles of fat bodies between the control, w¹¹¹⁸, and Dp mutant larva, and in muscles between the control, Mef2>mCherry-RNAi, and Mef2>Dp-RNAi pharates were examined (Supplementary file 1). The pairwise correlation between replicate samples was found to be significant (Figure 5—figure supplement 1B). The mean of correlation values for fat body replicates was r=0.55 for w¹¹¹⁸ and r=0.42 for Dp-/-, while muscle replicates showed r=0.8 for Mef2>mCherry-RNAi and r=0.63 for Mef2>Dp-RNAi.

In the proteomics datasets, we identified and quantified 6578 proteins in fat body and 5370 proteins in muscle samples. We confirmed that the intensity level of Dp protein was downregulated in both Dp-/- fat bodies and Dp-depleted muscles compared to the control tissues, respectively. Concordantly, the expression levels of the well-known E2F-direct targets that are repressed by E2F, such as Arp53D, SpnE, and dATM, were upregulated in both Dp-/- fat bodies and Dp-depleted muscles (Figure 5B).

A set of 5024 proteins that were shared between both tissues was selected for integrative analysis (Supplementary file 2). A log₂fold change (FC) ratio was calculated to identify proteins that were changed upon the loss of Dp. These two datasets showed a Pearson correlation of r = 0.172 (Figure 5—figure supplement 1C) indicating that a relevant subset of the changes detected in fat body were also present in muscles. Statistically significant changes were revealed using a false discovery rate (FDR) < 5%, and either a log₂FC > 0.5 or a log₂FC < −0.5 as cutoff values for upregulated and downregulated proteins, respectively (Figure 5C). We found that 556 proteins increased and 456 proteins decreased in Dp-/- fat bodies, whereas 844 increased and 1208 proteins decreased in Dp-depleted muscles (these are visualized in heatmaps in Figure 5—figure supplement 1D–E).

KEGG pathways enrichment analysis was performed using the functional annotation tool DAVID (Figure 5—figure supplement 1F, Supplementary file 3) to obtain an overall picture of the changes resulting from the loss of E2F/Dp in the two tissues. Categories related to DNA repair, glutathione metabolism, and amino acid metabolism were significantly enriched for upregulated proteins in both tissues, while nucleotide metabolism was only significantly enriched in muscle (FDR<5%, Figure 5—figure supplement 1F, top panel). Similar changes have been linked to the loss of E2F/Dp in previous studies (Guarner et al., 2017; Nicolay et al., 2015; Nicolay et al., 2013). Additionally, proteins related to cytochrome p450 enzymes that catalyze detoxification and biosynthetic reactions in fat body (Chung et al., 2009) were upregulated in Dp mutants (Figure 5—figure supplement 1F, top panel). Among the downregulated proteins, the fat body proteome was enriched for pentose phosphate pathway and fatty acid metabolism, whereas citrate cycle, oxidative phosphorylation, and ribosome categories were significantly overrepresented among the muscle proteome upon the loss of Dp (Figure 5—figure supplement 1F, bottom panel).

Next, we focused on proteomic changes that were shared between these two tissues since these may reflect an E2F function that is common to both tissues (Figure 5D, Supplementary file 4). As expected, the upregulated proteins in Dp-deficient fat body and muscle showed a significant enrichment for DNA damage, DNA recombination, and homologous recombination (Figure 5D, top panel, FDR < 5%). The top annotation cluster for downregulated proteins displayed a significant enrichment for glycolysis, gluconeogenesis, and carbohydrate metabolic process (Figure 5D, bottom panel, FDR < 5%), thus indicating that the loss of Dp alters carbohydrate metabolism in both fat body and muscle.

To explore the metabolic defects triggered by the loss of Dp, we used targeted liquid chromatography tandem mass spectrometry (LC-MS/MS) to profile the metabolic changes upon Dp loss. Third instar Dp^-/- larva were collected and compared to two wild-type strains, w¹¹¹⁸ and Canton S, to account for differences in the genetic background (Figure 5E, Supplementary file 5). Fifty-five compounds showed significant changes in the Dp mutant compared to both controls (FDR < 5%, Figure 5—figure supplement 1G, Supplementary file 6). The increased and decreased metabolites were selected, and KEGG pathways enrichment analysis was performed. Interestingly, the major metabolic pathways that showed significant enrichment for downregulated compounds were glycolysis, gluconeogenesis, and ammonia recycling (FDR < 5%, Figure 5F, Supplementary file 7), which is largely consistent with the proteome analysis described above. Notably, four metabolites of the core module of the glycolytic pathway, (Figure 5G), 2,3-diphosphoglyceric acid, 3-phosphoglycerate, D-glyceraldehyde-3-phosphate, and phosphoenolpyruvate, were significantly reduced in Dp mutant compared to both controls (Figure 5H).

We conclude that the tissue-specific depletion of Dp results in extensive metabolic changes in both fat body and muscle. These changes were evident in proteomic profiles and confirmed by metabolomic profiling. The alterations indicate that E2F-depleted tissues undergo significant changes in carbohydrate metabolism affecting, in particular, glycolytic metabolites.

The changes in the proteomic and metabolomic profiles are very interesting but they raised the question of whether the metabolic changes observed contribute to the lethality of Mef2>Dp-RNAi or cg>Dp-RNAi animals. Since diet is known to impact metabolic phenotypes, we asked whether varying the levels of carbohydrates, protein, and fat in fly food could alter the lethal stage of these animals.

To properly control the food composition and the effect of nutrients, we switched to a semi-defined food, made of sucrose, lecithin, and yeast, as major sources of carbohydrates, fat, and protein, respectively (Lee and Micchelli, 2013). Control diet contained 7.9% carbohydrates, 0.08% fat, and 1.9% protein (Supplementary file 8). As expected, Mef2>Dp-RNAi or cg>Dp-RNAi did not survive on control diet and died at pupal and pharate stage (Figure 6A–B). Nutrient composition was then altered by varying the amount of a single component of the control diet and the viability of cg>Dp-RNAi and Mef2>Dp-RNAi animals relative to the control genotype was scored. Interestingly, while the survival of Mef2>Dp-RNAi animals was unaffected by different nutrient composition (Figure 6B, and Figure 6—figure supplement 1B), cg>Dp-RNAi were highly sensitive to dietary changes. The increase in protein content had a negative impact on the survival of cg>Dp-RNAi and resulted in significant developmental delay (Figure 6—figure supplement 1A, left panel). About 83% of animals showed melanotic masses at larval stages when reared on high protein diet, and only 35% of third instar larva progressed onto pupa stages and eventually died (Figure 6—figure supplement 1C, quantified in Figure 6—figure supplement 1D). The melanotic masses, which were also observed in the E2f1 mutant larvae (Royzman et al., 1997), are related to the immune system response (Watson et al., 1991). However, reducing the content of protein suppressed the occurrence of melanotic masses (Figure 6—figure supplement 1C–D) and consequently increased their survival rate (Figure 6—figure supplement 1A). In contrast, increasing lipid content up to 1% in diet was beneficial for survival as about 22% of cg>Dp-RNAi pupa reached adulthood (Figure 6—figure supplement 1A, right panel). However, no further rescue in viability was detected beyond 1% lipid. Strikingly, while no cg>Dp-RNAi animals survived on control diet (7.9% carbohydrates), almost half of them eclosed when reared in the presence of 12% carbohydrates and the lethality was fully rescued when food contained even higher amount of carbohydrates (16% or 24%) (Figure 6A, left panel). This result was confirmed using a second fat body-specific driver r4-GAL4 (Figure 6A, right panel).

One of the functions of fat body is to maintain homeostasis of trehalose, a main circulating sugar in hemolymph (Becker et al., 1996), allowing animals to adapt to a high sugar diet. The synthesis of trehalose occurs in fat body from glucose-6P and is regulated by trehalose-6-phosphate synthase (Tps1) (Figure 6C; Elbein et al., 2003). We noted that the levels of Tps1 proteins were significantly reduced in Dp mutant fat body, along with other enzymes that generate glucose-6P including phosphoglucomutase, glucose-6-phosphate isomerase, and hexokinase_A in the proteomic datasets (Figure 6C and Supplementary file 2). Additionally, trehalase, an enzyme that converts trehalose back to glucose, was significantly increased in Dp-deficient fat body. This suggested that animals lacking Dp in the fat body may be defective in the regulation of trehalose synthesis.

To test this idea directly, trehalose was measured in hemolymph of cg>Dp-RNAi and control, cg>mCherry-RNAi, third instar larvae fed either a control diet or a high sugar diet containing 16% sugar. Notably, trehalose was significantly reduced in hemolymph of cg>Dp-RNAi larva compared to matching control (Figure 6D) when animals were reared on control diet. Switching to high sugar diet significantly improved the levels of trehalose in cg>Dp-RNAi to levels comparable to cg>mCherry-RNAi larva that were fed a control diet (Figure 6D). In contrast, the levels of circulating trehalose were largely unchanged in Mef2>Dp-RNAi compared to control (Figure 6E).

The fat body is the principal site of stored fat in Drosophila. Dp mutants were previously shown to have lipid droplets of reduced and irregular size (Guarner et al., 2017). Therefore, we examined the impact of high sugar diet on fat body in cg>Dp-RNAi. Fat bodies of cg>Dp-RNAi and control larva were dissected and stained with BODIPY to visualize lipid droplets. In control diet, Dp-depleted fat bodies displayed smaller lipid droplet size compared to control (Figure 6F, quantification in bottom panel), which is indicative of defects in fatty acid synthesis. Concordantly, the expression of acetyl-CoA carboxylase, the rate-limiting enzyme for fatty acid synthesis, was significantly reduced in fat body of cg>Dp-RNAi compared to control (Supplementary file 2). As expected (Musselman et al., 2011; Pasco and Léopold, 2012), feeding control larva a high sugar diet resulted in significantly larger lipid storage droplets (Figure 6F). A similar trend was observed in cg>Dp-RNAi although the increase in droplet size was not as robust as in the wild type (Figure 6F, quantification in bottom panel). Nevertheless, feeding cg>Dp-RNAi animals on high sugar diet dramatically improved the lipid droplet size in comparison to control diet. This was further corroborated by measuring total triglycerides content at third instar larval development (Figure 6G). The levels of triglycerides normalized to total protein content were significantly reduced in cg>Dp-RNAi animals reared on control diet compared to cg>mCherry-RNAi. Moreover, the normalized levels of triglycerides were fully restored in cg>Dp-RNAi animals fed on high sugar diet. Thus, feeding animals with high content of sugar increased their lipid storage, as evidenced by examining both the lipid droplet size in fat body and the total content of triglycerides, and could, subsequently, contribute to rescue animal viability upon Dp knockdown in fat body.

In contrast, muscle-specific Dp depletion did not affect either the size of lipid droplets in the fat body or the total content of triglycerides in Mef2>Dp-RNAi animals (Figure 6—figure supplement 1G and Figure 6H, respectively). The Mef2>Dp-RNAi animals reared on high sugar showed subtle changes although not statistically different from control animals (Figure 6H and Figure 6—figure supplement 1E, quantification in bottom panel). This suggests that the loss of Dp in muscle does not severely impair the ability of fat body to regulate fat storage.

Dp-depleted fat body failed to maintain the proper level of circulating sugar in hemolymph. Since glucose is converted into glycogen in both fat body and muscle for storage, we measured the level of glycogen in cg>Dp-RNAi and in Mef2>Dp-RNAi. As shown in Figure 6—figure supplement 1F–G, there were no major defects in glycogen storage normalized to protein content in cg>Dp-RNAi and Mef2>Dp-RNAi compared to control animals fed either a control or a high sugar diet.

As mentioned above, Dp depletion in fat body leads to the appearance of binucleated cells (Figure 2A–B; Guarner et al., 2017). Therefore, we asked whether this phenotype can be suppressed by food composition. Binucleated cells were readily found in both cg>Dp-RNAi and Mef2>Dp-RNAi larvae fed a control diet, but the phenotype was not modified by rearing larvae on high sugar diet (Figure 6—figure supplement 1H–I). Since dietary changes suppress the defects and fully rescue animal lethality, we infer that the essential function of Dp in fat body is to maintain homeostasis of the circulating sugar trehalose and to regulate fat storage. The fact that binucleated cells persisted in the fat body of rescued animals suggests that the binucleated phenotype is most likely distinct from the metabolic defects.

Flies were raised in vials containing standard cornmeal agar medium at 25° C. The w¹¹¹⁸ flies were used as wild-type control flies. Either the Dp^a3 and Dp^a4 alleles or Dp^a3 and deficiency Df(2R)Exel7124, which deletes the entire Dp gene, were used in this work to obtain the trans-heterozygous Dp mutant larvae for proteome and metabolome experiments, respectively (Frolov et al., 2005; Royzman et al., 1997). The trans-heterozygous E2f2 ^{76Q1 /c03344}; E2f1 ^{91 / m729} mutant animals were used as double E2f1 and E2f2 mutants. The GAL4 drivers, P{Cg-GAL4.A}2, P{w[+mC]=Lpp-GAL4.B}3 and P{GAL4-Mef2.R}3, and the following control UAS-RNAi lines from the TRIP collection: Luciferase (P{y[+t7.7] v[+t1.8]=TRiP.JF01355}attP2) and mCherry (P{VALIUM20-mCherry}attP2) obtained from Bloomington Drosophila Stock Center (Bloomington, IN). The line UAS-Dp-RNAi was obtained from the library RNAi-GD (ID 12722) at the Vienna Drosophila Resource Center (Vienna, Austria). The stock UAS-G-TRACE (Evans et al., 2009) was used to trace the expression of the drivers. The P{PTT-GA}Dp^CA06954 line from the Carnegie collection (Buszczak et al., 2007), here annotated as DpGFP, contains a GFP-expressing protein trap insertion (Zappia and Frolov, 2016). The P{UAS-Dp.D} was used to overexpress Dp in the Dp mutant background (Du et al., 1996; Neufeld et al., 1998; Zappia and Frolov, 2016).

The total number of pupae, pharate pupae, and adult flies able to eclose out of the pupal case were scored. The pupal developmental stages were assessed by following markers of metamorphosis (Ashburner et al., 2005). At least 58 flies per group were scored in a minimum of three independent experiments.

All flies were raised on Bloomington standard cornmeal food. After eclosion, adults were transferred to different fly food composition, which was made based on the semi-defined control diet (Lee and Micchelli, 2013) with adjustments. Control diet was made of 1% agar (Lab Scientific, Fly 8020), 4.35% brewers yeast (MP Biomedicals, 2903312), 0.04% lecithin (soybean, MP Biomedicals, 102147), propionic acid (0.5% v/v), and 7.9% sucrose (FCC Food grade, MP Biomedicals, 904713). The adjustments for each food type are detailed in Supplementary file 8.

Hemolymph was collected from ~10 or 15 third instar larvae per sample. Protocol was adapted from Tennessen et al., 2014. Each animal was rinsed with ddH₂O, wipped to remove excess of water, carefully punctured in the mouth hook using a tungsten needle and placed in a 0.5 ml tube with a hole at the bottom of the tube. This tube was then placed in a 1.5 ml tube and centrifuged to maximum speed for 10 s. Approximately 1 µl of hemolymph was collected for each sample. Hemolymph was diluted 1:50 in trehalase buffer (TB) (5 mM Tris pH 7.6, 137 mM NaCl, 2.7 mM KCl). Samples were heat-treated for 5 min at 70°C and centrifuge for 3 min at maximum speed at 4°C. Supernatant was quickly snap-frozen and stored at −80°C until all samples were harvested.

Hemolymph samples were further diluted to final dilution 1:150 with buffer TB. Circulating trehalose was measured in hemolymph as previously described (Tennessen et al., 2014). Briefly, an aliquot of each sample was treated with porcine trehalase (Sigma, T8778-1UN) overnight at 37°C in a G1000 Thermal cycler (Bio-Rad) to digest trehalose and produce free glucose. In parallel, another aliquot was incubated with buffer TB to determine the levels of glucose. The total amount of glucose was determined using the glucose (HK) assay reagent (Sigma, G3293) following a 15 min incubation at room temperature. Trehalose (Sigma, 90208) and glucose standard solutions (Sigma, G3285) were used as standards. Plate reader BioTek Epoch was used to read absorbance at 340 mm. The trehalose concentrations for each sample were determined by subtracting the values of free glucose in the untreated samples. Each sample was measured twice, a total of six independent biological samples were collected by group, and three independent experiments were done.

Third instar larvae were harvested to measure glycogen as previously described (Tennessen et al., 2014). Briefly, seven animals were collected per sample, rinsed with ddH₂O, and homogenized in 100 µl PBS 1×. Samples were heat-treated at 70°C for 10 min and centrifuged at maximum speed for 3 min at 4°C. Supernatant was stored at −80°C until all samples were collected. Samples were diluted 1:6 in PBS 1× for the assay and transferred to two wells. One well was treated with amyloglucosidase (Sigma A1602) and the second well with PBS 1×. The plate was incubated at 37°C for 1 hr. Then, the total amount of glucose was determined using 100 µl of glucose (HK) assay reagent (Sigma, G3293) following a 15 min incubation at room temperature. Glycogen and glucose standard solutions were used as standards. Plate reader BioTek Epoch was used to read absorbance at 340 mm. The glycogen concentrations for each sample were determined by subtracting the values of free glucose in the untreated samples. Each sample was measured twice, a total of six independent biological samples were collected by group, and three independent experiments were done.

Total glycogen was normalized to soluble protein amount. Aliquots of larval homogenate were removed prior heat treatment to measure soluble protein using a Bradford assay (Bio-Rad 500–0006) with BSA standard curves.

A coupled colorimetric assay was used to quantify triglycerides by measuring free glycerol as previously described (Tennessen et al., 2014). Briefly, seven animals were collected per sample, rinsed with ddH₂O and homogenized in cold 100 µl PBS-T (PBS 0.05% Tween-20). Samples were heat-treated at 70°C for 10 min and stored at −80°C until all samples were collected. Samples were diluted 1:6 in PBS-T for the assay and transferred to two wells. One well was treated with triglyceride reagent (Sigma, T2449) and the second well with PBS-T. The plate was incubated at 37°C for 30 min in a G1000 Thermal cycler (Bio-Rad). Then, the total amount of free glycerol was determined using 100 µl free glycerol reagent (Sigma, F6428) following a 5 min incubation at 37°C. Glycerol (triolein, Sigma, G7793) standard solution was used as standard. Plate reader BioTek Epoch was used to read absorbance at 540 mm. The triglyceride concentrations for each sample were determined by subtracting the values of free glycerol in the untreated samples. Each sample was measured twice, a total of six independent biological samples were collected by group, and three independent experiments were done. Total triglyceride was normalized to soluble protein amount as described above.

Tissues were dissected and fixed in 4% formaldehyde in PBS for 30 min. Then, tissues were permeabilized during 10 or 15 min in 0.1% Triton X-100 in PBS or in 0.3% Triton X-100 in PBS for muscles tissues. Tissues were washed and blocked in 1% or 2% BSA PBS. Primary antibodies were incubated overnight at 4°C in 2% BSA and 0.1% Triton X-100 in PBS. After washing three or four times for 10 min each in 0.1% Triton X-100 (in PBS), secondary antibodies (Alexa Fluor, Cy3- or Cy5-conjugated anti-mouse and anti-rabbit secondary antibodies, Life Technologies and Jackson Immunoresearch Laboratories) were incubated for 60 or 90 min in 10% normal goat serum 0.1% Triton X-100 in PBS. After washing three times with 0.1% Triton X-100 (in PBS), tissues were mounted on glass slides in glycerol with antifade or in Vectashield with DAPI (Vector Laboratories). All steps were performed at room temperature, unless otherwise stated.

In the case of fat bodies, the fixation was done for 60 min and PBS 1× was used throughout the protocol instead of 0.1% Triton X-100 in PBS 1× (Guarner et al., 2017). For larval body wall musculature staining, larva was dorsally opened, pinned in a Sylgard dish, and fixed for 20 min. A minimum of five to eight animals per genotype was dissected per experiment, and the staining was carried out two or three times.

The primary antibodies were mouse monoclonal anti-Dp antibody (Yun6, dil 1:10, Du et al., 1996) used in fat bodies and rabbit polyclonal anti-Dp antibodies (#212, dil 1:300 Dimova et al., 2003) used in muscles, anti-GFP (FITC, 1:500, Abcam ab6662), Guinea pig anti-Rad50 (1:100, Gao et al., 2009), Guinea pig anti-Mre11 (1:100, Gao et al., 2009). Rhodamine–phalloidin or fluorescein isothiocyanate–phalloidin were used to counterstain, and DAPI for nucleus staining.

To visualize lipid droplets, dissected third instar larval fat bodies were fixed in 4% formaldehyde in PBS for 1 hr at room temperature and washed three times in PBS 1×. Fat bodies were incubated in solution containing both 0.5 µg/ml BODIPY 493/503 (Invitrogen, D3922) and DAPI diluted in PBS 1×, for 10 min at room temperature, then washed three times in PBS 1×.

Fluorescent images were acquired with the laser scanning confocal microscope (Zeiss LSM 700) using × 20/0.8, and × 40/1.2 objectives at University of Illinois at Chicago and 710 Zeiss Confocal microscope at MGH Cancer Center. Images were processed using ImageJ (1.52k5, National Institutes of Health, Bethesda, MD) and Photoshop CC 2019 (Adobe Systems). All images are confocal single plane images, except otherwise stated. Only representative images are shown.