The fat body cortical actin network regulates Drosophila inter-organ nutrient trafficking, signaling, and adipose cell size

To determine how the actin network is organized within the Drosophila larval fat body (FB), we extracted FBs from w¹¹¹⁸ (laboratory control strain) late feeding L3 larvae, fixed and labeled these with the fluorescent F-actin stain phalloidin, and conducted confocal z-section microscopy. We found that most F-actin was highly polarized to FB cell surfaces exposed to the hemolymph, as well as intra-tissue cell-cell boundaries, with a dim intracellular pool also detectable (Figure 1A). Imaging revealed that the cortical F-actin localized to microridge protrusions at the tissue surface, giving the FB surface a ‘foamy’ appearance consistent with previous reports showing the plasma membrane of FB cells is densely ruffled (Diaconeasa et al., 2013; Ugrankar et al., 2019, Figure 1A).

Figure 1 with 1 supplement see all

Download asset Open asset

FB cells contain abundant triglyceride (TG) stores in cytoplasmic lipid droplets (LDs). A sub-population of small LDs are localized to the cell periphery next to the tissue surface (defined as peripheral LDs, pLDs), whereas larger LDs are packed in the cell interior close to the tissue mid-plane (defined as medial LDs, mLDs) (Ugrankar et al., 2019). We examined how this F-actin framework was spatially arranged in relation to pLDs and mLDs. Z-section confocal microscopy of FBs stained with the fluorescent F-actin dye Cell Mask Actin Tracking Stain, together with LD dye monodansylpentane (MDH), revealed a bright ruffled cortical F-actin cytoskeleton that appeared to intimately encircle many pLDs adjacent to the cell surface (Figure 1B, red arrows). Side-profile yz-axis views revealed the actin:pLD arrangement was particularly evident at hemolymph-exposed FB cell surfaces. In contrast, in the tissue interior the actin network did not encircle individual mLDs, but rather localized along cell edges at cell-cell boundaries.

Since the functional relationship between the actin cytoskeleton and LDs in adipose tissues is not well understood, we began to dissect their interplay using larval FBs. First, we utilized a candidate-based visual screen to identify the specific actin isoforms required for the FB cortical actin network. Drosophila express six actin isoforms (Act5C, Act42A, Act57B, Act79B, Act87E, and Act88F) in various tissues (Burn et al., 1989; Röper et al., 2005; Wagner et al., 2002). We RNAi depleted each actin isoform specifically in the FB using the FB tissue-specific driver Dcg-Gal4, and stained for F-actin. Strikingly, targeted RNAi depletion indicated that Act5C (Dcg-Gal4 >UAS-Act5C^RNAi), but not any other actin isoforms, dramatically altered F-Actin polarization to cell surfaces and cell-cell boundaries (Figure 1C, Figure 1—figure supplement 1A). Line-scans confirmed that the major F-actin surface enrichment was ablated in Act5C^RNAi FB tissue, indicating the Act5C isoform was necessary for the FB cortical actin network (Figure 1—figure supplement 1A). Additionally, the FB cell surface now appeared smooth and lacked its microridge patterning, suggesting Act5C, but not other actin isoforms, was required for this ‘foamy’ surface appearance (Figure 1C, red boxes are single FB cells).

To confirm that loss of the FB cortical F-actin network was specifically due to depletion of the Act5C isoform, we utilized quantitative PCR (QPCR) to confirm an ~70% knockdown of Act5C mRNA with our TRiP UAS-Act5C^RNAi line (Figure 1—figure supplement 1B). We also confirmed that Act5C expression was not significantly affected in the Act42A^RNAi line (the other major cytoplasmic actin isoform of the FB). We further confirmed specificity of the Dcg-Gal4 driver to the FB by examining other tissues, including the gut and brain, extracted from Dcg-Gal4 >UAS-GFP-expressing Drosophila larvae. We detected GFP signal only in the FB but not in other tissues, validating Dcg-Gal4 as a FB-specific driver, as previously reported (Suh et al., 2008; Ugrankar et al., 2015, Figure 1—figure supplement 1C). To rule out off-target effects from the Bloomington TRiP Act5C^RNAi line, we tested two additional independent UAS-RNAi lines for Act5C obtained from the Vienna Drosophila Research Center (Act5C^{RNAi VDRC#7139}, Act5C^{RNAi VDRC#101438}) using the Cg-Gal4 FB driver. Both recapitulated the initial observed Act5C^RNAi phenotype, similarly ablating the cortical actin network (Figure 1—figure supplement 1D).

To determine whether the TRiP Act5C^RNAi line was specifically depleting only Act5C in the FB, and not targeting the other actin isoforms, we co-expressed GFP-tagged versions of all six actins (UAS-ActXX-GFP) together with UAS-Act5C^RNAi in the FB. Indeed, while we observed that Act5C-GFP normally localized to the cell surface and cell-cell boundaries, the Act5C-GFP protein signal in Act5C^RNAi FB was diminished and did not localize to the cell periphery or cell-cell boundaries. Instead, a signal was observed only in the FB cell nuclei. In contrast, the other five GFP-tagged actins, ectopically expressed in Act5C^RNAi FB were capable of localizing to the cortical actin network. Thus, none of the other ActXX-GFP proteins appeared to be impacted by Act5C^RNAi, indicating that the TRiP Act5C^RNAi line specifically targeted only the Act5C isoform (Figure 1—figure supplement 1D and E). Indeed, co-staining Dcg-Gal4 >Act5C^RNAi/Act5C-GFP FB with phalloidin confirmed that the cortical actin was disrupted in these cells (Figure 1—figure supplement 1F). Collectively, this suggests that Act5C is the pertinent isoform necessary for forming the FB cortical actin network, although we cannot rule out that other actin isoforms may play a minor role. These experiments validate our genetic tools, a complete list of which is available in table Supplementary file 1.

To further dissect how Act5C forms the FB cortical actin network, we again imaged GFP-tagged Act5C in the FB by expressing UAS-Act5C-GFP using the Dcg-Gal4 driver. Analogous to the phalloidin and Cell Mask Actin staining of wildtype fat tissue, Act5C-GFP decorated FB cell surfaces and enriched at cell-cell interfaces (Figure 1D). Act5C-GFP also intimately wrapped around pLDs at the cell periphery (Figure 1D, red arrows). To understand how this cortical actin network relates to the cell surface, we co-expressed UAS-Act5C-RFP together with the plasma membrane (PM) marker UAS-mCD8-GFP. This revealed that the Act5C-RFP cortical actin network and PM are closely overlaid, and both exhibit the microridge pattern intimately encircling pLDs (Figure 1E, red arrows). Collectively, we conclude that Act5C is required for forming the FB cortical actin network, and that it is in close proximity to the cell PM, as well as pLDs adjacent to the cell surface.

Given that Act5C localizes in close proximity to LDs at the cell surface, we next examined how FB-specific Act5C depletion impacted LDs and fat storage. To quantitatively access this, we measured size changes of the peripheral LDs (pLDs) and mid-plane LDs (mLDs) in larval FBs in which each actin isoform was depleted. Notably loss of Act5C significantly altered both pLD and mLD morphologies and organization (Figure 2A and B, Figure 2—figure supplement 1A and B). Specifically, Act5C loss significantly increased the average size of pLDs, but drastically reduced the average size of mLDs (Figure 2A–D). In contrast, RNAi depletion of the other actin isoforms did not measurably alter the average sizes of either LD pool, although there was a mild increase in average mLD size in Act57B^RNAi tissues (Figure 2D, Figure 2—figure supplement 1A). We confirmed that this LD morphology defect was due to loss of the Act5C isoform, as we observed similar LD disruptions upon Act5C RNAi depletion with the two other VDRC UAS-RNAi lines (Figure 2—figure supplement 1C). We further confirmed that FB tissue targeted Act5C loss-of-function -resulted in defective LD storage and morphology by driving Act5C knockdown using two other validated FB drivers (Cg-Gal4, and Lpp-Gal4), both of which caused similar LD perturbations as Dcg-Gal4 (Figure 2—figure supplement 1D, red arrows).

Figure 2 with 1 supplement see all

Download asset Open asset

To determine how Act5C loss impacted larval triglyceride (TG) storage, we measured TG abundance in Act5C^RNAi and control whole larvae. Consistent with their altered LD morphologies, FB-specific Act5C depletion significantly reduced larval TG levels, which were ~10% of 6-day age-matched controls. A significant decrease in total TG was also observed in the two VDRC RNAi lines (Figure 2E, Figure 2—figure supplement 1E). Since TG levels are influenced by both lipogenesis and TG lipolysis, we employed QPCR to evaluate the expression levels of known lipogenesis and lipolysis factors in isolated FB tissues. Whereas major de novo lipogenesis genes like acetyl-CoA carboxylase (ACC) and fatty acid synthase (FASN) were expressed at normal levels in Act5C^RNAi FB compared to controls, genes for TG production including GPAT4, the diacylglycerol acyltransferase (DGAT) enzyme MDY, and the fatty acid desaturase DESAT1 were significantly reduced in expression, suggesting a defect in TG synthesis (Figure 2—figure supplement 1F). To determine whether this TG reduction was caused in part by increased lipid mobilization, we also examined mRNA levels of lipolysis-associated genes. The expression of the TG lipolysis enzyme Brummer (BMM), as well as mitochondrial fatty acid oxidation importer CPT1, were significantly reduced, suggesting the observed TG reduction was not a result of elevated lipolysis nor enhanced fatty acid oxidation. Collectively, this suggests that Act5C loss in the FB drastically alters LD morphologies and reduces TG pools, likely due to reduced TG synthesis.

To evaluate how loss of Act5C would influence LDs in a cell autonomous manner, we next utilized the FLP/FRT system to generate chimeric FBs where Act5C was RNAi depleted in only subsets of FB cells (Konsolaki et al., 1992). As expected, Act5C-mutant cells from successful mitotic recombination (identified by loss of the RFP marker) displayed a disrupted cortical actin network when stained with phalloidin or Cell Mask compared to adjacent control cells (Figure 2F and G, red arrow). Notably, Act5C depleted cells also exhibited perturbed pLD morphologies compared to adjacent control cells (Figure 2G, red arrow). LDs appeared sparser in individual Act5C-depleted cells (similar to general loss of Act5C in the entire FB), and a clear pLD layer was missing in the cell periphery. Some cells instead contained an amorphous fat mass in the periphery near cell-cell boundaries, potentially from the fusion or collapse of the pLD layer (Figure 2—figure supplement 1G). This is consistent with the role of the cortical actin cytoskeleton in maintaining LD morphology and organization within FB cells, and suggests that pLDs are encased by the actin meshwork that surrounds them in the cell periphery.

Drosophila Act5C global null mutants die as first instar larvae due to arrested growth (Wagner et al., 2002), but we noted that FB-specific Act5C depletion produced viable larvae that could proceed through larval development and sometimes pupate. However, we found that age-matched L3 Act5C-deficient larvae were undersized compared to late feeding/pre-wandering controls, suggesting a developmental arrest or significant delay (Figure 3A). Consistent with this, control larvae reached critical weight, and commenced wandering and pupating between days ~5–6 after egg laying (AEL), whereas there was a significant delay in the growth of Act5C-deficient larvae (Figure 3B, and see table Supplementary file 2 for developmental details). In fact, Act5C^RNAi larvae lingered in the food for an additional 7–14 days (~2–3 x slower relative to controls), gradually increasing their overall size (taking up to 3 weeks) to nearly match controls. However, these developmentally delayed larvae appeared transparent, and upon imaging were lipodystrophic with small FBs compared to controls (Figure 3E). TG levels of these 3 week old larvae were higher than at day 6, but still much less than L3 pre-wandering controls (Figure 2E). A few larger larvae did eventually wander and pupate (<10%), but became necrotic during metamorphosis and failed to eclose to adult flies (Figure 3B, C and D, red arrows denote necrotic pupae). Indeed, FBs extracted from wandering Act5C^RNAi larvae showed underdeveloped tissue lobes with significantly smaller cells, suggesting Act5C was required for normal FB development and cell size (Figure 3F, G and H). Collectively, this indicates that the Act5C-dependent cortical actin network is required for proper TG storage in the FB, as well as contributes to proper FB tissue and larval growth, and FB cell size.

Figure 3

Download asset Open asset

Next, we investigated when Act5C is specifically required in the larval FB, hypothesizing that Act5C may assist in FB cell size expansion during larval development. To enable temporal control over Act5C expression during development we employed the Gal4/Gal80 system (McGuire et al., 2004). Gal80 is a temperature sensitive allele that at lower temperatures (18 °C) prevents the Gal4 driver from binding to the UAS (Upstream Activating Sequences), thereby preventing activation of the Act5C RNAi transgene, and thus allowing endogenous Act5C expression (Act5C^ON). Upon switching to 29 °C, Gal80 is degraded, allowing Gal4 to bind the UAS and drive FB-specific (Dcg-Gal4 mediated) knockdown of Act5C, defined here as Act5C^OFF (Figure 4A).

Figure 4

Download asset Open asset

First, we allowed Dcg-Gal4; tub-Gal80ts X UAS-Act5C^RNAi parents to lay eggs at 29 °C, which were maintained at this non-permissive (Act5C^OFF) temperature through embryogenesis (i.e. 1 day). We then lowered the culturing temperature to the permissive 18 °C for the remainder of larval development to permit Act5C expression (Act5C^ON). Surprisingly, this 1 day Act5C depletion during embryogenesis still enabled larvae to successfully develop into flies (Figure 4A), suggesting that expression of Act5C in fat body progenitor cells was not essential for early adipogenesis or FB cell differentiation during late embryogenesis (Hoshizaki et al., 1994; Yongmei Xi, 2015). However, we cannot rule out that maternally inherited actin may contribute to very early stages of embryo development or tissue morphogenesis (Burn et al., 1989).

Next, we extended the Act5C^OFF 29 °C treatment for 2 days and 3 days, depleting Act5C during embryogenesis and through the L1 and/or L2 stages of larval development. Indeed, vials cultured at 29 °C for 2 days still successfully progressed into flies. However, Act5C^RNAi cultures maintained at 29 °C for the initial 3 days began displaying varying degrees of arrested larval development and pupal necrosis, although some still progressed into flies, indicating Act5C may be functionally required during the post-L1 stages of larval development (Figure 4A). Extending our 29 °C treatment for 4 days resulted in progressively fewer larvae able to develop into flies, and 5 days at 29 °C resulted in the developmental arrest and necrotic pupae observed with constitutive Act5C depletion. Collectively, this indicates that larval expression of Act5C is required for post-embryonic larval development.

To further dissect the developmental time-period during which Act5C is required, we monitored Act5C^RNAi larvae hatched and maintained at the permissive (Act5C^ON) 18 °C, then later transferred to the Act5C^OFF 29 °C at different ages. Because 18 °C incubation has been shown to predictably slow embryonic and larval developmental times by a factor of two (Ashburner et al., 2005; Shellenbarger and Mohler, 1978; Yadav and Sharma, 2013), an adjusted developmental timespan that accounts for this delay was employed. Thus, embryogenesis at 18 °C actually requires 2 ‘real’ days (denoted here as 1 ‘adjusted development day’, or 1 day*). Intriguingly, when cultures were initially placed at the permissive 18 °C for only 0.5 day* or 1 day*, then switched to 29 °C, larvae exhibited developmental arrest and failed to generate flies, suggesting that Act5C expression in the FB only during embryogenesis is not sufficient for complete Drosophila development (Figure 4B).

Next, we extended the permissive 18 °C treatment for 1.5 days*, and observed a small, but variable (≤30%) proportion of flies hatching across several culture vials analyzed. Further extension to 2 days* at 18 °C enabled ~70% of pupae to eclose to flies (Figure 4B). An almost complete rescue was noted with vials maintained for 2.5 days* at 18 °C. Together, both Gal80ts temperature-shift experiments suggest that larval Act5C is functionally important after embryogenesis and during the L1-L3 larval stages of development when the FB tissue is rapidly expanding to reach critical weight (CW). During this developmental period, larvae are actively feeding, growing, and storing gut-derived dietary nutrients in the FB, which requires inter-organ lipid trafficking and signaling to coordinate nutrient storage and FB tissue growth.

As an orthogonal experiment to dissect the developmental time-window when Act5C was required, we took advantage of previous studies showing that simply maintaining Drosophila stocks at lower temperatures can limit the degree of RNAi-mediated protein depletion in UAS-RNAi/Gal4 systems, even without the involvement of Gal80 (Duffy, 2002; Perkins et al., 2015). Indeed, we observed that while FBs of Dcg-Gal4 >UAS-Act5C^RNAi larvae maintained at 25 °C throughout development did not contain a cortical actin network (phalloidin-stained), those maintained at 18 °C exhibited a partial F-actin network, indicating some Act5C was still present at this cooler temperature (Figure 4C, red arrows). Furthermore, larvae constitutively maintained at 18 °C were generally larger and successfully developed into flies, whereas those maintained at 25 °C exhibited developmental arrest and pupal necrosis, indicating that at 18 °C Act5C was expressing above its functional threshold (Figure 4D and E). The Act5C^RNAi larvae cultured at 18 °C also displayed improved FB tissue morphology and normal FB cell sizes (Figure 4D and F).

Equipped with this additional temperature-sensitive experimental approach, we queried how culturing Dcg-Gal4 >UAS-Act5C^RNAi at 25 °C temperature during egg-laying and early development (Act5C^OFF), then switching to 18 °C at which RNAi efficiency is reduced (Act5C^functional), would impact development. Like the Gal80ts system, maintaining cultures for 1 day at 25 °C, then switching to 18 °C permitted full development into flies (Figure 4G). Extending the 25 °C timeframe to 2 days still allowed fly development, whereas 3 days at 25 °C led to developmental arrest and necrotic pupae similar to continually maintaining them at this temperature (Figure 4G). This further supports a model where Act5C is required later during larval development when the animal is feeding and the FB growing.

Finally, we monitored Dcg-Gal4 >Act5C^RNAi cultures initially reared at the Act5C^functional 18 °C, then transitioned to 25 °C. Like the Gal80ts system, cultures kept at 18 °C for only 1 day* (Act5C expressed during embryogenesis only) failed to develop into flies, whereas some larvae from cultures maintained at 18 °C for 1.5 days* could successfully complete metamorphosis into flies (Figure 4H). Cultures maintained for the initial 2.5 days* at 18 °C completed normal development into flies, underscoring that expressing Act5C only during embryogenesis is not sufficient for full development. Collectively, this indicates that larval FB-Act5C expression is required during the post-embryonic feeding phases of larval growth and development when the FB tissue is rapidly expanding and storing nutrients.

Given that larvae are actively feeding and growing during the L1-L3 developmental time window, and that FB-depletion of Act5C^RNAi results in developmental delay, shrunken FBs, and small FB cells (Figure 3), we hypothesized that Act5C may be required to expand the size of Drosophila adipose tissue during feeding and development. Larval FB cells are post-mitotic (Meschi and Delanoue, 2021; Musselman and Kühnlein, 2018), and therefore FB tissue expansion is accomplished by increasing individual cell size but not cell number. This growth is necessary to accommodate the vast amounts of lipids and biomass for achieving critical weight (CW) and successful progression through metamorphosis (Merkey et al., 2011; Tennessen and Thummel, 2011). Indeed, our hypothesis is supported by work in mice suggesting that actin reorganization is required to expand adipocyte cell sizes during high-fat diets (Hansson et al., 2019). Reinforcing this, our imaging revealed cortical actin networks at cell-cell boundaries within FB tissues, implying intra-tissue actin assemblies may be dedicated to cell ultra-structural support and mechanical cell growth.

To dissect how Act5C dynamics influences FB cell size, we quantified the sizes of FB cells from age-matched L3 control larvae and those with FB-specific depletion of Twinfilin (Twf), an actin effector protein that promotes F-actin disassembly (Hilton et al., 2018). Indeed, depletion of Twf in the FB (Dcg-Gal4 >UAS-Twf^RNAi) led to significantly decreased FB cell size resembling Act5C loss, indicating that actin dynamics are necessary in the FB for normal FB cell expansion (Figure 5A).

Figure 5

Download asset Open asset

Next, we again employed the Gal80ts system to dissect how temporally altering Act5C expression in the FB impacted cell size. We quantified FB cell sizes of age-matched L3 larvae cultured at the permissive 18 °C (Act5C^ON) for the initial 1.5 days*, then moved to 29 °C (non-permissive, Act5C^OFF) for the remainder of development. Of note, this 1.5 day* temperature scheme enables only a few larvae to successfully develop into flies (Figure 4B, Figure 5B). Indeed, these larvae exhibited significantly smaller FB cells, indicating that loss of Act5C expression after only 1.5 days* suppressed FB cell growth during larval development (Figure 5B). Next, we examined FBs from larvae cultured initially at 18 °C for 2 days* or 2.5 days* (followed by Act5C silencing at 29 °C), which in the previous experiments enabled majority of larvae to develop into flies. Indeed, the 2 days* larvae exhibited normal-sized FB cells, and surprisingly the 2.5 days* larvae exhibited slightly larger FB cells than controls (Figure 5B). These experiments together indicate that Act5C expression during larval growth is required for normal FB cell enlargement, which appears to correlate with successful development.

To query how these time-specific experiments related to FB TG storage, we also measured larval TG levels of these larvae. As expected, there was a direct correlation between smaller FB cells and reduced TG storage. Larvae from cultures kept at 18 °C for the initial 1 day* or 1.5 days* accumulated significantly less TG than controls. TG levels progressively increased with longer times at the permissive temperature, that is, 2 day* and 2.5 day* (Figure 5C). This suggests Act5C is required for FB cell size expansion as well as TG storage during the feeding phases of larval development.

Next, we examined FB cell sizes and TG levels of larvae from cultures initially maintained at the non-permissive 29 °C temperature (Act5C^OFF). Loss of Act5C expression during embryogenesis only (1 day at 29 °C) did not significantly impact FB cell sizes of L3 larvae, and their TG levels mirrored controls (Figure 5D and E). This is consistent with their ability to develop into flies. However, extension of this non-permissive 29 °C to 3 days and 5 days caused progressive reductions in both FB cells size and TG stores in these larvae (Figure 5D and E). Collectively, this indicates that Act5C is required for FB cell size expansion and TG storage, specifically in the post-embryonic phases of larval development as larvae feed and grow.

The FB is the major nutrient storage organ of Drosophila, and controls nutrient signaling pathways for the animal (Yongmei Xi, 2015). Since FB-specific Act5C depletion led to smaller FB tissue, reduced TG stores, delayed growth, and developmental arrest, we investigated whether Act5C-deficient larvae displayed altered feeding behavior and/or insulin signaling. L3 control and Act5C^RNAi larvae were allowed to feed on a 20% sucrose solution supplemented with Coomassie blue dye. After 1 hr and 4 hr, larvae were examined qualitatively (visually) and quantitatively (by dye extraction and colorimetric measurements) to determine the amount of dye in their gut. Following 1 hr of feeding,<1% of Act5C^RNAi larvae had visible dye in their guts compared to >90% of controls (Figure 6A). However, after 4 hr in the dyed food,~20% to 40% of Act5C^RNAi larvae displayed some gut dye, but even amongst those actively feeding, dye amounts were ~50% reduced relative to controls. This suggested a feeding defect in Act5C^RNAi animals (Figure 6A and B).

Figure 6

Download asset Open asset

As feeding stimulates insulin signaling, we next evaluated how Act5C loss from the FB impacted insulin signaling pathways. QPCR analysis revealed that Act5C^RNAi tissues displayed a ~50% reduction in Target-Of-Brain-Insulin (Tobi) transcripts as well as a ~threefold increase in the expression of ImpL2, a DILP2 antagonist, suggesting insulin signaling was blunted in Act5C-deficient FBs (Buch and Pankratz, 2009; Lee et al., 2021, Figure 6C). We also monitored insulin-stimulated phosphatidylinositol-3,4,5-P3 (PIP3) signaling in FB cells by expressing a GFP-fused PI3K activity reporter (tGPH reporter) (Britton et al., 2002). This tGPH reporter contains a PIP3-binding GFP-tagged PH domain ubiquitously expressed under the control of the Drosophila tubulin promoter. Re-distribution of tGPH from the cytoplasm to PIP3 at the plasma membrane (PM) is a bioindicator of active insulin-induced PI3K signaling. Indeed, Act5C-deficient FBs displayed reduced PM-localized tGPH relative to cytoplasmic and nuclear signals, indicating reduced insulin signaling in Act5C^RNAi larvae (Figure 6D). Finally, we directly monitored insulin signaling by extracting larval hemolymph and western blotting for the IPC-secreted insulin-like peptide DILP2 that mediates brain:FB inter-organ insulin signaling. Act5C^RNAi larvae displayed significantly reduced DILP2 in their hemolymph compared to controls, again indicating reduced insulin signaling (Figure 6E). Collectively, this indicates that larvae lacking Act5C in their FBs display muted feeding and insulin signaling.

Given that FB Act5C loss perturbed insulin signaling and reduced fat stores, we next focused on mechanistically dissecting how the cortical actin cytoskeleton of adipose cells contributes to nutrient storage. A recent model using mice adipocytes suggests that the F-actin network regulates fat storage by influencing the trafficking of carbohydrate transporters like Glut4 to the cell surface, and thus regulating carbohydrate uptake into adipocytes for eventual storage as TG (Brozinick et al., 2007; Kim et al., 2019). To determine whether the FB cortical actin network regulated nutrient uptake into FB cells, we developed ex vivo fluorescence-based nutrient uptake assays.

First, we examined whether Act5C loss impacted FB lipid absorption. We exposed extracted larval FB tissues to PBS media containing BODIPY-C12, a fluorescent-conjugated fatty acid that can be internalized by cells and incorporated into TG (Kasurinen, 1992). Intracellular BODIPY-C12 could be detected in both control and Act5C-depleted FB tissues (Figure 7A, B and C). Surprisingly, the BODIPY-C12 signal in FBs from Act5C^RNAi larvae was significantly higher compared to controls, suggesting Act5C loss did not inhibit the capability of FBs to absorb extracellular lipids, and in fact Act5C loss may promote extracellular lipid absorption. Next, we examined whether Act5C loss affected FB carbohydrate absorption. We extracted FB tissues from pre-starved to deplete glycogen stores (Yamada et al., 2018), control or Act5C-deficient larvae, and exposed them to PBS media containing 2-NBD-glucose (2-NBDG; Figure 7D). 2-NBDG was dimly detected in the cytoplasm of FB cells, and appeared at least as abundant in Act5C-depleted FB tissues as in controls, suggesting Act5C loss did not reduce the ability of FB tissue to absorb extracellular carbohydrates (Figure 7E and F).

Figure 7

Download asset Open asset

Carbohydrate uptake and its metabolic conversion to TG help Drosophila maintain circulating blood/hemolymph sugar levels and regulate FB TG storage (Havula et al., 2013; Sassu et al., 2012). Therefore, we also examined whether Act5C loss altered hemolymph carbohydrate levels in vivo. We monitored circulating hemolymph levels of both glucose and trehalose in control and FB-specific Act5C^RNAi larvae, and surprisingly found neither was altered (Figure 7G and H). Like the ex vivo experiments above, this suggested carbohydrate uptake into the FB was not significantly perturbed with Act5C loss, and rather may even be more efficient; this is intriguing since these larvae have reduced total fat tissue mass, which is the major carbohydrate sink (Musselman et al., 2013). The fact that Act5C-deficient larvae exhibited normal circulating carbohydrates was also surprising given that they also manifested reduced insulin signaling (Figure 6C and D), suggesting Act5C loss in the FB may have non-canonical impacts on nutrient signaling.

Next, we evaluated FB carbohydrate stores by monitoring tissue glycogen levels. Consistent with the tissue’s ability to absorb circulating carbohydrates ex vivo, FB glycogen levels were not significantly different between control and FB-specific Act5C^RNAi larvae (Figure 7I and J). This underscored that the FB was still operationally capable of absorbing and storing carbohydrates. To further confirm that FB tissues could deliver carbohydrate transporters to the cell surface for uptake from circulation, we imaged a GFP-tagged model glucose transporter, hGlut4-GFP (Crivat et al., 2013) expressed in larval FBs using the UAS-Gal4 system. As expected, hGlut4-GFP could be successfully trafficked to the surfaces of Act5C-deficient FB cells, suggesting the transport of glucose transporters to the cell surface was not significantly impacted by Act5C loss (Figure 7K).

Collectively this indicates that Act5C loss in the FB does not interfere with the capability of the FB to physically absorb extracellular carbohydrate and lipid nutrients, and thus cannot be solely responsible for the reduced TG levels observed in Act5C-deficient FBs.

Since Act5C-depleted FBs appeared capable of absorbing extracellular nutrients, we next examined whether there was an in vivo defect in delivering nutrients to the FB via inter-organ trafficking. The primary mechanism for inter-organ FB lipid delivery is the trafficking of lipophorin (Lpp) lipoprotein-bound DAG via the hemolymph from the animal gut to the FB and other tissues. Lpp particles composed of protein ApoLpp (a homolog of human ApoB) are solely manufactured within FB cells, then secreted into the hemolymph where they travel to the gut, dock, and receive DAG to deliver to other tissues (Heier and Kühnlein, 2018; Matsuo et al., 2019; Palm et al., 2012, Figure 8A). A block in Lpp-mediated lipid trafficking leads to lipid accumulation in the mid-gut, which can be visualized as LDs (Palm et al., 2012). Since Act5C-deficient FBs displayed reduced TG storage but were capable of nutrient uptake, we next interrogated whether there was a defect in this gut:FB inter-organ lipid transport.

Figure 8 with 1 supplement see all

Download asset Open asset

First, we isolated L3 larval guts and stained for LDs using Nile Red. Strikingly, whereas control mid-guts exhibited small sparse LDs, mid-guts from larvae with FB-specific Act5C-depletion displayed significant LD accumulation, and closely resembled larvae lacking ApoLpp (ApoLpp^RNAi), implying a defect in mid-gut lipid export (Figure 8B). To determine if other perturbations to F-actin dynamics in the FB could lead to similar defects in gut:FB inter-organ transport, we also RNAi-depleted the F-actin effector Twf in the FB (Dcg-Gal4 >Twf^RNAi). Like Act5C depletion, FB-specific Twf^RNAi loss also resulted in significant fat accumulation in larval mid-guts, indicating that disrupted F-actin turnover in the FB resulted in gut fat accumulation (Figure 8B). Indeed, Twf deficiency in the FB resulted in surface actin accumulation in Twf^RNAi FB cells, confirming it perturbed actin dynamics at the FB (Figure 8—figure supplement 1A). To quantify relative gut fat accumulation, we isolated guts from these larvae and measured TG levels. Indeed, both Act5C and Twf FB-depletion significantly elevated gut TG compared to controls (Figure 8C). As expected, loss of ApoLpp also increased gut TG levels (Figure 8—figure supplement 1B).

To further dissect how FB actin perturbation impacted circulating lipids and inter-organ lipid trafficking, we measured larval hemolymph DAG levels, the main lipid in Lpp inter-organ trafficking. Consistent with defective inter-organ trafficking, Act5C loss in the FB significantly reduced hemolymph DAG (Figure 8D). Hemolymph DAG pools also trended downward in Twf FB-deficient conditions, but were just below significance threshold due to sample variability. Since hemolymph isolation from ApoLpp-depleted larvae proved challenging, we instead measured whole-larval DAG pools, which were significantly decreased from controls, consistent with the established role of Lpp in trafficking DAG between larval tissues (Figure 8—figure supplement 1C). Collectively, this suggests that perturbing F-actin in the FB leads to perturbed gut:FB inter-organ lipid trafficking, reduced circulating DAG, and TG accumulation in the gut.

In addition to the cortical actin network, FB cells contain a network of spectrin filaments at their surface that provide shaping and structural support for the cell architecture. Since previous work from our lab and others has shown that beta-spectrin localizes to this spectrin network and helps shape the highly-ruffled FB cell surface (Diaconeasa et al., 2013; Ugrankar et al., 2019), we used this opportunity to also examine how loss of beta-spectrin and its partner alpha-spectrin impacted FB LDs and inter-organ lipid trafficking. Consistent with their role in organizing cell surface architecture that intimately associates with pLDs, loss of both beta-spectrin and alpha-spectrin altered pLD sizes at the cell periphery (Figure 8—figure supplement 1D–F). However, mLDs were not significantly affected, distinct from Act5C and Twf loss (Figure 8—figure supplement 1F). In line with this, beta-spectrin loss in the FB did not lead to fat accumulation in the larval gut (in fact, gut TG levels were actually reduced in larvae that lacked FB beta-spectrin) (Figure 8B and C). Moreover, hemolymph of beta-spectrin depleted larvae displayed normal circulating DAG levels, indicating that loss of the FB spectrin network did not impact inter-organ lipid trafficking like Act5C loss (Figure 8D). Furthermore, we find that F-actin still localized to FB cell-cell boundaries in beta-spectrin RNAi-depleted FBs, although the FB cell surface appeared significantly less ruffled (Figure 8—figure supplement 1G). This collectively suggests that spectrin and actin have distinct roles at the FB cell surface; while the spectrin network also helps to shape the highly ruffled surface architecture of the FB, it appears surprisingly dispensable for inter-organ lipid transport.

How does Act5C in the FB contribute to FB:gut inter-organ trafficking? Since hemolymph DAG is bound to Lpp lipoprotein particles that are synthesized and secreted by the FB (Heier and Kühnlein, 2018; Palm et al., 2012), we hypothesized that Act5C loss impacted Lpp production and/or secretion from FB cells. To interrogate this, we immuno-stained for Lpp particles in larval FB tissue. Notably, in control tissues Lpp particles were concentrated in the FB cell periphery just below the cell surface (Figure 9A). In fact, high resolution imaging indicated that Lpp particles formed distinct clusters in close proximity to pLDs adjacent to the cell surface (Figure 9A, red arrows). This is in agreement with the known spatial distribution of ApoB lipoprotein particles in human hepatocytes. ApoB particles are assembled and lipidated with TG in ER sub-domains immediately adjacent to LDs (Taghibiglou et al., 2000; Ye et al., 2009). Relatedly, our previous work indicated that depletion of Drosophila Lpp particles in larvae impacted pLD homeostasis, suggesting a functional connection between pLDs and Lpp particles at the FB surface (Ugrankar et al., 2019).

Figure 9 with 1 supplement see all

Download asset Open asset

To further probe the similarities between Drosophila Lpp particles and human ApoB, we cultured human Huh7 hepatocytes and immuno-stained for ApoB in cells also co-stained for the ER network and LDs. This confirmed that ApoB forms puncta and crescent-like structures in ER sub-domains adjacent to LDs that appeared very similar to the Lpp puncta we observed in larval FB cells (Figure 9—figure supplement 1A). Next, we tested whether Act5C was necessary to maintain this LD:Lpp particle spatial orientation by IF staining Act5C^RNAi FBs for Lpp. In striking contrast to controls, Act5C^RNAi FBs displayed a bright Lpp immuno-stain throughout the tissue interior, and the Lpp signal was no longer polarized to the cell periphery near pLDs (Figure 9A). The Lpp immuno-signal appeared to localize along a reticulated network that resembled the ER, implying Act5C loss caused some defect in Lpp intra-cellular organization and/or secretion.

To gain higher resolution insights into how Act5C loss impacted Lpp organization and secretion, we conducted thin-section transmission electron microscopy (TEM) on larval FBs. In control tissues, we observed grape-like particle clusters in close proximity to some LDs that resembled Lpp fluorescence immuno-staining (Figure 9B, red arrows). The clustering of these particles into patches adjacent to LDs also appeared very similar to accumulations of ApoB punctae during their lipidation. By TEM, individual particles were ~60 nm in diameter, in close agreement to the expected sizes of Lpp particles (Wojczynski et al., 2011). In further support of this, previous work showing TEM of locust FBs revealed like-sized Lpp particles (Dantuma et al., 1998). Moreover, we observed some particles in the microridge folds at the FB cell surface, indicating they may be secreted by FB cells (Figure 9C, blue stars). To investigate whether these Lpp-like particles were in fact Lpp particles, we did TEM of ApoLpp-depleted FB TEM sections (Dcg-Gal4 >ApoLpp^RNAi), and found these particles were no longer observed, supporting that they were Lpp particles (Figure 9B and C). A few studies have suggested that similar electron-dense particles may be glycogen particles (Érdi et al., 2012), but when we labeled glycogen in the FB using a UAS-glycogenin-YFP construct, we found glycogenin had a very different cellular distribution, indicating these LD-associated particles were not glycogen (Figure 9—figure supplement 1B).

Strikingly, TEM of Act5C^RNAi FBs revealed that the Lpp particles were no longer sequestered in grape-like clusters adjacent to LDs. Instead, these particles now appeared to accumulate inside FB cells and distributed throughout the cell interior, similar to the anti-Lpp IF signal (Figure 9B and C). This collectively supports the model that these particles are intracellular Lpp particles prior to their secretion, and that loss of Act5C perturbs their proper spatial organization and eventual secretion out of the FB, leading to defective gut:FB and other (whole-body) inter-organ lipid trafficking.

Since our observations support a model where Act5C is required for FB Lpp secretion, perturbation of this process may explain (at least in part) the reduced LD and TG storage defects observed in Act5C-depleted FB tissues. This may also account for the reduced insulin signaling in Act5C-deficient FBs, as Lpp trafficking to the brain and insulin secretion from brain IPCs have been shown to directly correlate in other studies (Brankatschk et al., 2014; Kelly et al., 2022).

To further test this model, we compared Act5C depletion to loss of known secretory pathway factors in the FB. We RNAi depleted the known secretory protein Sar1, an Arf family GTPase that promotes ER-to-Golgi trafficking, as well as Tango1, an ER-to-Golgi trafficking factor (Yang et al., 2021). Indeed, Tango1 has previously been associated with human ApoB/VLDL secretion (Santos et al., 2016). Strikingly, FB-specific loss of either Sar1 or Tango1 and Act5C produced common defects. Both included LD accumulations and TG increases in the larval gut (Figure 8B and C). Sar1 and Tango1 loss also mirrored Act5C FB depletion in that they exhibited decreased whole-larval TG levels (Figure 9D), as well as lower hemolymph DAG pools (Figure 8D). Tango1-depleted larvae also developed L3 developmental arrest, consistent with a block in Lpp secretion and inter-organ lipid transport (table Supplementary file 2). In line with this, staining for Lpp particles in Tango1-depleted FBs revealed an identical build-up of intracellular Lpp in FB cells as Act5C loss, indicating that loss of Act5C or Tango1 led to similar defects in Lpp secretion (Figure 9E, red arrows). As an additional interrogation of how Act5C loss impacted the FB cell secretory pathway, we imaged UAS-dGRASP65-GFP, a known Golgi marker protein. As expected, dGRASP65-GFP marked numerous Golgi foci in control FB tissues, particularly in the cell periphery just below the surface (Figure 9—figure supplement 1C). However, dGRASP65-GFP foci were reduced in abundance in Act5C-depleted FBs, particularly in the mid-plane, indicating Act5C loss impacted Golgi homeostasis in the FB.

To quantitatively access how Act5C loss impacted Lpp and general protein secretion, we isolated larval hemolymph and western blotted for circulating Lpp protein ApoLII. Indeed, RNAi depletion of either Act5C or Tango1 led to reductions in secreted ApoLII in the hemolymph, consistent with a block in Lpp release from the FB (Figure 9F). Similarly, general protein staining (Ponceau stain) also revealed reduced secreted proteins in Act5C^RNAi and Tango1^RNAi hemolymph, indicating loss of Act5C possibly impacted secretion of other proteins and/or adipokines from the FB. In contrast, beta-spectrin-depleted FBs showed no major defects in ApoLII and protein secretion.

Collectively, this suggests that Act5C and the cortical actin network are required for normal Lpp particle secretion into the larval hemolymph, subsequent lipid release from the gut and Lpp-mediated lipid transport required for normal TG storage in the FB. In line with this, both ApoLpp and Act5C FB depletion produce analogous phenotypes such as reductions in larval FB cell size, and developmental arrest (Figure 9—figure supplement 1D, table Supplementary file 2). Previous work has also indicated that defects in Lpp trafficking are closely correlated with reduced insulin signaling in Drosophila, indicating functional crosstalk between lipoprotein and insulin signaling (Brankatschk et al., 2014). Thus, Act5C appears to play an essential role in Lpp-dependent inter-organ lipid trafficking and FB development in Drosophila that ultimately influences nutrient signaling and storage (Figure 10).

Figure 10

Download asset Open asset

Altogether, our work suggests that Act5C is necessary for larval FB cell and tissue expansion during larval development, as well as normal Lpp-mediated lipid trafficking and insulin signaling (Figure 10). Loss of Act5C or defects in F-actin turnover in the FB (Twf^RNAi) lead to disrupted Lpp secretion into the hemolymph, and consequently reduced inter-organ lipid transport, gut fat accumulation, and lower FB fat storage. This Lpp secretion defect mimics loss of canonical secretory pathway factors like Tango1 or Sar1. Additionally, we find that FB Act5C depletion leads to reduced larval feeding and suppressed insulin signaling. The mechanism underlying these signaling changes is not fully understood, but it is possible that these are primarily due to disrupted nutrient signaling. Past work has shown that Lpp particles accumulate at neurons that harbor insulin-like peptides and influence their release (Brankatschk and Eaton, 2010). More recent studies reveal that high sugar Drosophila feeding promotes Lpp-mediated crosstalk with the brain and ultimately influences insulin signaling and feeding behavior (Kelly et al., 2022). A key emerging question is how Lpp lipid trafficking influences inter-organ insulin signaling and feeding, and how this impacts organismal homeostasis and development. Future studies will dissect how Act5C, FB adipose tissue growth, and Lpp signaling regulate nutrient signaling and feeding.

All materials related to this study are accessible either through this source or via the BioRxiv preprint. Data also available via Dryad.

Please contact Mike.Henne@utsouthwestern.edu regarding any questions.

Drosophila L3 late feeding larvae were used for larval assays. All Drosophila animal research was performed in accordance with NIH recommended policies. Flies were reared and maintained on standard fly food containing cornmeal, yeast, molasses (~5% sugars), and agar. The Dcg-Gal4 fat body (FB)-specific driver, previously described (Suh et al., 2008), and Dcg-Gal4;tub-Gal80ts strains were provided by Jonathan M. Graff (UTSW). The Cg-Gal4;UAS-Dcr2 FB driver was received from Steve Jean (Université de Sherbrooke), and UAS-Glycogenin-YFP from Michael Welte (University of Rochester), respectively. UAS-Act5C^RNAi and other TRiP RNAi stocks, Gal4 drivers, hs-FLP, and protein/organelle marker strains used were obtained from the Bloomington Stock Center (Bloomington, IN). We obtained additional Act5C^RNAi strains from Vienna Drosophila Research Center, and FRT line for mosaic analysis from Kyoto Drosophila Stock Center. The UAS-hGlut4-GFP line was made by BestGene by injecting a UAS plasmid that was a kind gift of Dr. Leslie Pick (University of Maryland). All Gal4/UAS crosses were cultured at 25 °C to maintain uniform transgene expression. Crosses involving Cg-Gal4;UAS-Dcr2 with RNAi lines were cultured at 29 °C to maximize knockdown efficiency. A list of genetic tools is available in Supplementary file 1.

To track development of Act5C^RNAi and other FB-specific RNAi strains (Supplementary file 2), we crossed 20–25 FB-driver (Dcg> or Cg>) females with 7–8 UAS-RNAi males for ≤24 hr (to ensure similar egg densities across the genotypes examined). Life cycles of these RNAi progeny were closely monitored, and the developmental timeline/ status was recorded.

Late feeding L3 larvae were gently removed from inside the food media using a paintbrush and rinsed in water to remove food particles. Larval fat bodies (FBs) or guts were dissected in PBS using Dumont #5 forceps (Electron Microscopy Sciences). All tissues were fixed in 4% paraformaldehyde for 20 min at room temperature and rinsed briefly in PBS prior to staining with organelle labeling dyes. Lipid droplets in FBs were stained by incubating FBs in 100 µM monodansylpentane (MDH) LD stain (Abgent) for 30 min at room temperature, in the dark. F-actin structures were stained with Phalloidin (1:400) or Cell Mask Green Actin Tracking Stain (1:1000) according to manufacturer’s instructions (Invitrogen). LDs in guts were stained with 1 µM Nile Red for 30 min at room temperature, in the dark. Next, FBs and guts were rinsed in PBS and mounted on slides in SlowFade Gold antifade reagent with DAPI (Invitrogen) for subsequent imaging.

Mated flies (Dcg-Gal4;tub-Gal80ts X UAS-Act5C^RNAi) were allowed to lay eggs for 8 hr at 18 °C (permissive temperature) or 4 hr at 29 °C (nonpermissive temperature) and allowed to develop for 1/2/3/4/5 days before moving culture vial from 18° to 29 °C or vice versa for the remainder of fly development. Developmental progress was closely monitored on a daily basis, and larvae from both temperature-shift set-ups were harvested on different days for FB dissection/confocal microscopy (MDH, phalloidin), and TG measurements by TLCs.

FBs dissected from larvae were fixed in 4% paraformaldehyde at room temperature for 30 min on a nutator, then washed four times with 1 X PBST (PBS in 0.1% Tween), 10 min each. Tissues were blocked in 1 X PBST +2% normal goat serum for 2 hr at room temperature, then incubated with Rb-ApoII primary antibody (1:500 dilution, provided by Dr. Akhila Rajan) added to the blocking buffer, overnight at 4 °C. Next, FBs were washed 4 times with 1 X PBST 15 min each, followed by incubation with anti-rabbit Alexa546 secondary antibody (1:1000 dilution) in blocking buffer at room temperature in the dark for 2 hr on a nutator. FBs were rinsed again four times in 1 X PBST 15 min each, and subsequently mounted in SlowFade Gold antifade reagent with DAPI. Fluorescent structures were observed and imaged by confocal microscopy using standard RFP channel filter at 1 µm step-sizes.

Prepared slides were imaged using Zeiss LSM880 inverted laser scanning confocal microscope. Tissues stained with organelle dyes and those expressing transgenic fluorescent-tagged proteins (e.g. UAS-mcd8-GFP, UAS-Act5C-RFP, UAS-hGlut4-GFP) were imaged using the appropriate channel filter DAPI/RFP/GFP. Most FBs were imaged using a ×40 oil immersion objective, as 1 µm z-stack sections, or at lower magnification using a ×10 objective. Larval guts were imaged using a ×20 objective.

Whole larvae whose FBs were illuminated by FB-expressing fluorescent-GFP markers were imaged using EVOS FL Cell Imaging System (ThermoFisher) at low ×2.5 magnification. Pupal imaging was performed on Leica S8AP0 stereoscope.

Larvae were weighed, then homogenized in 2:1:0.8 of methanol:chloroform:water. Samples were incubated in a 37 °C water bath for 1 hr. Chloroform and 1 M KCl (1:1) were added to the sample, centrifuged at 3000 rpm for 2 min, and the bottom layer containing lipids was aspirated using a syringe. Lipids were dried using argon gas and resuspended in chloroform (100 µl of chloroform/7 mg of fly weight). Extracted lipids alongside serially diluted standard neutral lipids of known concentrations were separated on TLC plates using hexane:diethyl ether:acetic acid solvent (80:20:1, v/v/v). TLC plate was air dried for 10 min, spray stained with 3% copper (II) acetate in 8% phosphoric acid and incubated at 145 °C in the oven for 30 min to 1 hr to allow bands to develop for scanning and imaging. Neutral lipid (TAG and DAG) band intensity was quantified using Fiji/ImageJ software, and lipid concentrations were calculated from the standard curve generated with standard mixture. In the case of dissected larval guts, dried lipids for all samples were re-suspended in 120 µl of chloroform and final lipid concentrations were calculated by normalizing to protein in sample, measured by standard Bradford assay. Hemolymph (4 µl/sample) for DAG measurements collected from larvae was diluted in 0.1% N-Phenylthiourea (Sigma-Aldrich) in 50 µl PBS. The same lipid extraction protocol was followed, dried lipids were re-suspended in 75 µl of chloroform, and equal volumes were loaded on TLC plates.

We generated homozygous Act5C mutant clones in the fat body using the FLP/FRT-mediated mitotic recombination system (Theodosiou and Xu, 1998). Briefly, P{lacW}Act5CG0025 P{neoFRT}19 A/FM7c females carrying the Act5C mutation and Flippase Recognition Target (FRT) sites were crossed with P{Ubi-mRFP.nls}1, w*, P{hsFLP}12 P{neoFRT}19 A males that contain heat-shock inducible flippase (FLP) recombinase, a similarly positioned FRT site, and an RFP cell marker. After egg laying, 0–8 hr embryos were heat shocked at 37 °C for 1 hr to induce mitotic recombination between the FRT sites on homologous chromosome arms. Following heat shock, embryos were allowed to develop normally at 25 °C. Feeding third instar larvae were harvested for fat body dissection and stained with monodansylpentane (MDH) LD stain and Cell Mask Green Actin Tracking stain, or phalloidin (no MDH). Since the cell marker segregates with the wild-type gene, KO cells with 2 copies of mutant Act5C are recognized by the lack of nuclear RFP. Mosaic fat bodies were imaged by confocal microscopy using DAPI (MDH-LDs), GFP (Cell Mask Actin), and RFP (phalloidin and cell marker) filters.

Whole RNA was extracted from FB tissues using Trizol reagent (Ambion Life Technologies) according to manufacturer’s instructions. RNA was quantified by Denovix spectrophotometer, and 1 mg of RNA was reverse transcribed to cDNA using ReadyScript cDNA Synthesis Mix (Sigma). QPCR was performed with cDNA and SYBR Green JumpStart TAQ ReadyMix on the BioRad CFX96 Real-Time System; mRNA expression data were normalized to that of the fly housekeeping gene rp49. Primer sequences for all transcripts amplified are available in the supplement data file for primers (table Supplementary file 3).

FB tissue was removed from Drosophila larvae, fixed in 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer paraformaldehyde, and processed in the UT Southwestern Electron Microscopy Core Facility. They were post-fixed in 1% osmium tetroxide and 0.8% K3[Fe(CN6)] in 0.1 M sodium cacodylate buffer for 1 hour at room temperature. Cells were rinsed with water and en bloc stained with 2% aqueous uranyl acetate overnight. Next, they were rinsed in buffer and dehydrated with increasing concentration of ethanol, infiltrated with Embed-812 resin and polymerized in a 60 °C oven overnight. Blocks were sectioned with a diamond knife (Diatome) on a Leica Ultracut UCT (7) ultramicrotome (Leica Microsystems) and collected onto copper grids, post stained with 2% aqueous uranyl acetate and lead citrate. Images were acquired on a JOEL 1400 Plus transmission electron microscope using a voltage of 120 kV.

Duplicate samples of 3 L feeding larvae (~100 each) were transferred to 9-well plates containing 20% sucrose +Coomassie blue dye (200 µg/ml). After 1 hour of feeding on the dyed sucrose, larvae from the first well were removed and rinsed in water. Larval guts were visualized using a dissecting scope and scored for the presence or absence of Coomassie dye. Larvae in the second well were allowed to feed for an additional 3 hr (total 4 hr), before removing from sucrose, washing, and scoring guts for food consumption. To quantify relative feeding,~size-matched larvae with visually detectable gut dye were washed in water, then rinsed briefly in methanol, blotted dry, and transferred to 1.5 ml chilled Eppendorf tubes (30 larvae/tube, in triplicate). Larvae were homogenized in 200 µl of 100% methanol. The homogenate was centrifuged at maximum speed for 10 min at 4 °C. Supernatant was transferred to fresh tubes containing equal volume of dH2O, and centrifuged again. Absorbance of supernatant was measured at 595 nm, and dye ingested was determined by comparing to a standard curve of Coomassie dye concentrations.

FB tissues dissected from L3 feeding larvae starved in PBS for 4 hr (or FB tissues extracted from larvae and pre-incubated in PBS for 4 hr) to deplete glycogen stores, were incubated in 300 µM of a fluorescently labeled glucose analog 2-NBDG (2-Deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-D-glucose) (Sigma-Aldrich, #72987) for 45 min at room temperature, in the dark. FBs were then washed with PBS, fixed in 4% paraformaldehyde for 20 min, washed again in PBS, and mounted in SlowFade on slides. 2-NBDG was imaged using the GFP channel filter by confocal microscopy. To measure BODIPY-C12 uptake, FB tissues from larvae were incubated in BODIPY FL C12 (4,4-Difluoro-5,7-Dimethyl-4-Bora-3a,4a-Diaza-s-Indacene-3-Dodecanoic Acid) (Invitrogen, #D3822) 1:1000 in PBS for 20 min at room temperature, in the dark. Subsequent to fixation with 4% paraformaldehyde, FBs were also stained with MDH LD dye to visualize if BODIPY-C12 was incorporated into LDs. Prepared slides were imaged using DAPI(MDH), and RFP(BODIPY-C12) channel filters on the confocal microscope. Both 2-NBDG and BODIPY-C12 FB cell fluorescence was quantified using the Fiji/Image J software.

Feeding larvae washed out from culture vials were collected in 630 µm mesh-fitted baskets (Genesee) and rinsed to get rid of adherent food particles. Larvae were dried and divided onto into piles (10–12 larvae each) on a strip of parafilm. Larvae were bled by tearing the cuticle with Dumont 5 forceps (Electron Microscopy Sciences). Two µl of colorless hemolymph was aspirated from each pile and separately transferred to 96-well plates (Thermo-Scientific) containing 0.1% N-Phenylthiourea (Sigma-Aldrich) in 50 µl PBS. 150 µl of Autokit Glucose reagent (Wako) was added to each well and incubated at room temperature for 20 min before measuring absorbance at 505 nm. Glucose concentration was calculated from a standard curve generated with manufacturer’s glucose standards. For trehalose assays, 8 µl of dilute hemolymph was treated with 5 µl of (diluted 8 X) porcine kidney trehalase (Sigma) overnight at 37 °C. Ten µl of treated sample was assayed for trehalose as described for glucose.

Dissected FBs were homogenized in 300 µl ice-cold PBS using sonication followed by a syringe (29G1/2 needle). After reserving 20 µl of the homogenate for protein Bradford quantification, the rest of the homogenate was heat inactivated at 70 °C for 10 min. Homogenate was then centrifuged at maximum speed at 4 °C for 10 min and supernatant was collected in a new tube. In a 96-well plate, 30 µl of each sample was loaded in duplicate rows. Then, 100 µl of Autokit Glucose reagent +amyloglucosidase (1 µl amyloglucosidase {Sigma A1602; 25mg} per 1 ml of Glucose reagent) was added to one row of samples, and 100 µl of Glucose reagent alone (without amyloglucosidase) was added to the duplicate row of samples, and to the glucose standards. The plate was incubated at 37 °C for 30 min, after which color intensity was measured using a microplate reader at 505 nm. Free glucose concentration in treated and untreated samples was calculated based on the glucose standard curve. Glycogen concentration (as free glucose) was determined by subtracting glucose in the untreated samples from those treated with amyloglucosidase. Finally, total glycogen was normalized to protein measured by the Bradford assay.

Glycogen in FB tissues was stained using the Periodic Acid-Schiff (PAS) kit (Sigma). Larval FBs were dissected in 1% BSA in PBS, fixed with 4% paraformaldehyde for 20 min, and washed twice with 1% BSA in PBS. FBs were then incubated in 0.5% period acid solution for 5 min, washed twice with 1% BSA in PBS, then stained with Schiff’s reagent for 15 min, washed again, and mounted on slides. PAS-stained FB tissues were imaged using the Leica DM6B light microscope.

Feeding larvae were bled on parafilm, and 20 μl of pooled hemolymph was collected per sample and transferred to a chilled 1.5 ml Eppendorf tube containing 40 μl of Ringer’s+protease inhibitor cocktail. Samples were centrifuged at 300 X g for 10 min at 4 °C to pellet hemocytes. The supernatant was transferred to fresh tubes on ice. Sample for loading was prepared by mixing 20 μl of dilute hemolymph (equal volumes for each sample) with 10 μl of 4 X LDS buffer containing 0.2 M DTT and 10 μl Ringer’s solution, and boiling at 95 °C for 5 min. Sample was run on a 10% pre-cast polyacrylamide gel (BioRad) at 120 V for 1 hr. Gel proteins were transferred to nitrocellulose membrane by wet transfer. After briefly rinsing membrane in Milli-Q water, Ponceau stain was added to blot for 15 min, then rinsed again with MQ-water, and proteins on membrane imaged using BioRad Imager. Membrane was blocked with 5% Casein in 1 X TBS-T (0.1% Tween) for 1 hr, then incubated overnight at 4 °C with rabbit primary antibody against ApoII (provided by Dr. Akhila Rajan) at a 1:1000 dilution. After washing with TBS-T (4 X for 5 min), membrane was incubated with HRP-conjugated goat anti-rabbit secondary antibody (Abcam ab6721) (1:5000 dilution) for 1 hr at room temperature. Membrane was washed in TBS-T (4 X for 5 min), developed with Clarity Western ECL blotting substrate (Bio-Rad, ab1705061) and imaged with BioRad Imager.

For the DILP2 immunoblot, a 4 to 20% polyacrylamide gel was employed due to the low molecular weight of DILP2 (12 kDa). Further, LI-COR secondary antibodies (1:5000 dilution) were used against chicken DILP2 (1:1000 dilution, provided by Dr. Akhila Rajan). Here we used Immobilon PVDF-FL PVDF membrane (Millipore Sigma, IPFL0010) activated in methanol for 5 min, and 1% fish gelatin in 1 X TBS as blocking buffer. Secondary antibody used was donkey anti-chicken IRDye 800CW (LI-COR, 926–32218).

Huh7 cells were cultured in DMEM (Corning; 10–017-CV) supplemented with 10% Fetal calf serum (Corning; 35–010-CV) and 1% penicillin streptomycin solution (Corning; 30–002- Cl). To induce lipid droplet biogenesis, cells were incubated with 600 μM of Oleic acid (Sigma Aldrich; O1008) conjugated with 100 μM of fatty acid-free BSA (Sigma-Aldrich; A3803) for 16 hr.

After OA treatment, the cells were fixed with 4% paraformaldehyde solution in PBS for 15 min at room temperature. For IF staining, fixed cells were washed with 1 x PBS and permeabilized with 0.1% Triton X-100 in 1 X PBS at room temperature for 5 min. The permeabilized cells were blocked in IF buffer (PBS containing 3% BSA and 0.1% Triton X-100) for 30 min. The cells were then incubated with primary antibody in IF buffer for 1 hr and washed 3 X with PBS. This was followed by incubation with secondary antibody in IF buffer for 30 min and 3 X washes with PBS. The primary antibodies used are goat anti-Apolipoprotein B (1:200; Rockland Immunochemicals, Inc; 600-101-111), and mouse anti-Hsp90B1 (1:100; Sigma Aldrich; AMAb91019). The secondary antibodies are donkey anti-mouse AF594 (ThermoFisher Scientific; R37115), and donkey anti-goat AF488 (ThermoFisher Scientific; A11055) used at a dilution of 1:400. Visualization of LDs were achieved by staining the cells with MDH (1:1,000; Abgent; SM1000a) for 15 min at room temperature. The images of the cells were taken using a ×63 oil-immersion objective in a laser scanning confocal Zeiss LSM880+Airyscan microscope. Fiji/ImageJ was used for the representation and quantification of images.

Fiji/ImageJ was used for quantifications involving size of FB lipid droplets and area of larval fat cells imaged by confocal microscopy. First, the area was measured for the entire field of view or a region-of-interest (ROI) of defined size. For lipid droplets (LDs) size measurements, prior to converting image to binary, a ‘smooth’ function was applied to the image to remove inherent graininess of the lipid stain and allow for more accurate quantification of LD cross-sectional area. The method for binary conversion was selected based on which one best represented LDs in the region of interest. ‘watershed’ was performed on binary images, and ‘analyze particles’ was used to quantify LD number and size. For fat cell area, the ‘polygon selections’ tool was used to draw boundaries around individual cells highlighted by fluorescent protein-tagged markers (e.g. phalloidin, Cell Mask, mcd8-GFP, hGlut4-GFP), and ‘analyze’, ‘measure area’ functions were applied. All volumetric analysis and multi-color linescans were conducted using Fiji/ImageJ software.

All experiments were conducted at least three times to provide experimental rigor. For the figures, all comparative statistics were either student t-tests or ANOVA tests for multiple data point experiments. Statistical significance was applied if p<0.05. Error bars indicate SD. In general a single * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, **** indicates p<0.0001.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

The authors thank members of the Henne and Friedman labs for insights and discussion. We also thank Dr. Akhila Rajan for the kind use of her anti-ApoLII Lpp and anti-DILP2 antibodies, Dr. Michael Welte for the UAS-Glycogenin-YFP line, and Dr. Steve Jean (Université de Sherbrooke) for the Cg-Gal4;UAS-Dcr2 driver. We would also like to thank Dr. Anza Darehshouri, Phoebe Doss, and Rebecca Jackson for assistance with TEM. We are also grateful to Dr. Saikat Mukhopadhyay for use of his light and stereo microscopes and Real-Time QPCR machine. WMH is supported by funds from the Welch Foundation (I-1873), the NIH NIGMS (GM119768), NIDDK (DK126887), and the UT Southwestern Endowed Scholars Program.

© 2023, Ugrankar-Banerjee et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.