To address the molecular function(s) of Drosophila Hsl, we expressed a tagged version of the recombinant fly protein (His₆-Hsl) in heterologous cultured cells and measured lipid hydrolase activities in cell extracts. Murine Hsl (His₆-MmHsl) was expressed in parallel to serve as a positive control in these assays. Immunoblotting analysis confirmed the ectopic expression of recombinant His₆-Hsl, His₆-MmHsl, and of the His₆-β-Galactosidase (His₆-β-Gal) negative control (Figure 1A). Expression of His₆-MmHsl increased cellular hydrolysis rates toward monoacylglycerol (MG), DG, TG, and SE by 6-, 9-, 6-, and 66-fold, respectively, as compared to His₆-β-Gal expressing control cell extracts (Figure 1B). Expression of fly His₆-Hsl elevated cellular hydrolysis rates toward MG, DG, TG, and SE by 14-, 16-, 10-, and 17-fold, respectively, as compared to His₆-β-Gal expressing controls indicating a similar substrate spectrum of murine and fly Hsl (Figure 1B). To confirm these results in vivo, we engineered transgenic flies and expressed Hsl via the UAS/Gal4 system using the ubiquitous Act5c-GAL4 driver. Ectopic expression of Hsl increased DG, TG, and SE hydrolase activities of abdominal extracts by 7-, 2-, and 8-fold, respectively, when compared to control samples. In contrast, abdominal MG hydrolase activities were not significantly altered by Hsl expression (Figure 1C). Finally, we assessed the subcellular localization of a fluorescently-tagged version of Hsl (Hsl-EGFP) in fat body cells of adult flies and found that Hsl-EGFP concentrated at the surface of LDs (Figure 1D). The intensities of the Hsl-EGFP signal at individual droplets were highly variable suggesting enrichment of the protein at a subset of LDs, although this finding awaits confirmation via detection of endogenous Hsl. Taken together, our data show that Hsl is a LD-associated multifunctional lipid ester hydrolase, which implies a function of the enzyme in Drosophila storage lipid breakdown.

To investigate the role of Hsl in organismal lipid metabolism, we used imprecise P-element excision to generate the Hsl¹ deletion mutant, which lacks 2.74 kb genomic Hsl sequences, including most of the open-reading frame and the start codon (Figure 2A, Figure 2—figure supplement 1). This results in the absence of detectable Hsl mRNA whereas mRNA concentrations of the neighboring genes Ate1 and PCNA are unaffected (Figure 2—figure supplement 1). Homozygous Hsl¹ mutant animals are viable and were used to address the function of the enzyme in TG metabolism. Inspection of mature adult Hsl¹ mutant fat body tissue revealed no apparent difference in the size or abundance of LDs when compared to age-matched controls (Figure 2B). Consistently, whole-body TG and DG levels of these Hsl¹ mutants were indistinguishable from age-matched controls (Figure 2C). Species with 40–50 acyl carbon atoms and 0–3 double bonds accounted for >80% of the TG pool with comparable species distributions in both genotypes (Figure 2D). To exclude that Hsl functions in redundancy with other TG lipolytic pathways, we next assessed the expression of genes involved in TG storage regulation. The mRNA concentrations of bmm, Akh, perilipin1 (plin1), and perilipin2 (plin2) were similar in Hsl¹ mutant and control animals arguing against a compensatory transcriptional dysregulation of these genes in response to Hsl deficiency (Figure 2E). We then combined the Hsl¹ mutant allele with loss-of-function mutations of Akh or bmm and assessed TG levels in the ad libitum fed state and starvation. Unlike Hsl¹, the Akh^A and bmm¹ alleles were associated with 1.9-fold increased TG levels under ad libitum fed conditions (Figure 2F). As previously reported, starvation-induced consumption of TG was impaired in bmm¹ mutants and abolished in Akh^A bmm¹ double mutants, illustrating the strong genetic interaction between both lipolytic systems (Grönke et al., 2007). In contrast, Hsl¹ bmm¹ or Hsl¹ Akh^A double mutants were capable of starvation-induced TG mobilization and largely resembled bmm¹ or Akh^A single mutants in their TG storage phenotypes (Figure 2F). Next, we more broadly assessed energy metabolism in Hsl¹ mutant animals and found levels of free glycerol, free FAs, and glucose unchanged in ad libitum fed Hsl¹ mutants as compared to age-matched control animals whereas glycogen levels were 26% higher in Hsl¹ mutants as compared to controls (Figure 2G). Microcalorimetry revealed similar heat dissipation rates of Hsl¹ and control animals under ad libitum fed conditions. Upon fasting, control and Hsl¹ mutant animals reduced heat dissipation rates by 40% and 46%, respectively, indicating similar whole-body bioenergetics in both genotypes (Figure 2H). Also, the response of Hsl¹ mutants to starvation stress, a sensitive readout of perturbed energy metabolism, was indistinguishable from controls (Figure 2I). Taken together, these data argue against a function of Hsl in the regulation of TG and energy metabolism in adult Drosophila in parallel or downstream of other lipolytic systems.

The finding that Hsl possesses strong lipolytic activity toward SE in vitro prompted us to investigate sterol metabolism in Hsl¹ mutant animals. Compared to controls, SE levels of Hsl¹ mutants were similar at 1 day after eclosion, but 2.1- and 2.5-fold higher at 7 days and 14 days, respectively (Figure 3A). SE overstorage was also observed in 7-days-old females and males homozygous for the genetically independent Hsl^b24 allele (Figure 3B; Bi et al., 2012). Moreover, neutral SE hydrolase activity in Hsl¹ abdominal extracts was 24% lower compared to controls (Figure 3C). To address SE turnover more directly, we labeled sterol lipids during the larval period by feeding either ³H-cholesterol or ¹⁴C-FAs and followed the turnover of labeled lipids after eclosion. ³H-cholesterol is incorporated into free and esterified sterols and thus allows the simultaneous measurement of both sterol pools. Since the SE pool has a mixed sterol composition (see below), we used ¹⁴C-FA as a complementary tracer to label also ergosteryl and phytosteryl esters. Consistent with reduced SE hydrolysis Hsl¹ mutants exhibited slower turnover rates of SE prelabeled with ¹⁴C-FAs or ³H-cholesterol (Figure 3D,E). In contrast, turnover of free ³H-cholesterol was unaffected by impaired SE catabolism in Hsl¹ mutant animals (Figure 3F). In line with this observation, 7-days-old Hsl¹ mutant animals exhibited unaltered levels of total free sterols when compared to age-matched controls suggesting that defective SE catabolism does not limit free sterols due to dietary sterol supply under ad libitum fed conditions (Figure 3G). Drosophila is able to use a variety of sterols including cholesterol, phytosterols, and ergosterol. The molecular composition of its sterol pool depends on dietary availability and sterol turnover (Carvalho et al., 2012; Knittelfelder et al., 2020). We therefore asked whether defective SE catabolism in Hsl¹ mutants affects sterol composition. However, both Hsl¹ and control animals showed a similar sterol composition dominated by ergosterol and β-sitosterol with lower levels of campesterol and stigmasterol and traces of brassicasterol, desmo-/zymosterol, and lanosterol (Figure 3G). To test if SE accumulation in Hsl¹ mutants is accompanied by misexpression of other sterol metabolic genes, we assessed mRNA concentrations of the sterol-responsive transcription factor Hr96, the Niemann-Pick-related genes Npc1a, Npc1b, Npc2a, and Npc2b, the putative sterol-O-acyltransferase gene CG8112 and the intestinal SE hydrolase gene magro. We observed a 1.8-fold increase in magro mRNA levels in Hsl¹ mutants as compared to controls, whereas expression of all other sterol metabolic genes investigated remained largely unaltered in response to Hsl-deficiency (Figure 3H). Taken together, these data suggest that Hsl specifically regulates sterol homeostasis in adult Drosophila by promoting the catabolism of SE.

To characterize a potential differential role of Hsl in the SE catabolism of specific tissues, we next measured SE in dissected body segments and organs of Hsl¹ mutant animals. In comparison to controls, SE levels were increased 2.6-, 2.8-, 2.4-, and 6.8-fold in Hsl¹ mutant heads, thoraces, abdomen, and carcasses, respectively. Hsl¹ mutant animals accumulated SE at lower levels also in the intestine but not in ovaries (Figure 4A). Moreover, Hsl deficiency did not provoke accumulation of LDs in the larval ring gland, a steroidogenic tissue with highly active sterol metabolism (Figure 4B). These findings imply an accumulation of SE in the fat body of Hsl¹ mutant animals as this tissue is present in all body segments and enriched in carcass preparations. To more directly test this hypothesis, we restored Hsl expression in specific tissues of the Hsl¹ mutant animals by means of the UAS/Gal4 system. Ubiquitous or fat body-specific re-expression of Hsl by means of the Act5c-GAL4 or FB-GAL4 driver, respectively, rescued the SE overstorage phenotype of Hsl¹ mutant animals (Figure 4C). In contrast, neuronal, intestinal or oenocyte-specific re-expression of Hsl by means of elav-GAL4, Myo31DF-GAL4 or Desat1-GAL4 drivers, respectively, did not revert the SE overstorage phenotype of Hsl¹ mutants suggesting that Hsl acts autonomously in the fat body to control the organismal SE levels (Figure 4C). In line with this observation, ectopic expression of Hsl in the fat body of control flies decreased total body SE storage by 46% as compared to controls (Figure 4D). Since Hsl has been conserved during evolution, we next asked if an analogous pathway of sterol mobilization exists in mammalian adipocytes. To this end, we used mice with adipocyte-specific HSL-deficiency (AHKO) and analyzed storage lipid levels in the adipose tissue depots of these animals. As shown in Figure 4E, AHKO mice lack detectable HSL immunoreactivity in subcutaneous (SCAT), perigonadal (PGAT), and perirenal (PRAT) white adipose tissue depots but not in liver, kidney, or muscle. AHKO animals had similar amounts of TG as controls in their SCAT, PGAT, and PRAT but accumulated 10-, 12-, and 14-fold more SE, respectively, in these adipose tissue depots (Figure 4F). SE accumulation in AHKO adipose tissue depots was associated with a 63% decrease in neutral SE hydrolase activity as compared to controls (Figure 4G). To assess how disrupted SE hydrolysis affects sterol metabolic gene expression, we measured mRNA concentrations of genes involved in de novo synthesis, transport, and esterification of cholesterol. Notably, mRNA concentrations of the de novo cholesterol synthesis gene Hmgcs1 were decreased by 83% in PGAT of AHKO mice as compared to controls arguing for a dysregulation of adipocyte sterol metabolic gene expression upon disrupted SE hydrolysis. Conversely, mRNA concentrations of the sterol-O-acyltransferase gene Soat1 were increased by 4.8-fold, whereas Hmgcr, Srebp2, Star, Abca1 were similarly expressed in both genotypes (Figure 4H). We conclude that Hsl acts as evolutionarily conserved adipocyte-autonomous regulator of SE catabolism and thereby controls whole-body SE levels in Drosophila.

Hsl is expressed throughout development with particularly high levels during early embryogenesis (Bi et al., 2012). To test if Hsl function in SE hydrolysis is relevant for the early Drosophila development, we compared embryonic free and esterified sterol levels of Hsl¹ mutant and control embryos. To discriminate between maternal and zygotic effects in embryos, we also included reciprocal crosses between control and Hsl¹ mutant parents. In controls, embryonic SE levels dropped by 66% within 6–8 hr after egg laying (AEL) and were almost undetectable at the end of embryogenesis (18–20 hr; Figure 5A). Embryos completely devoid of Hsl exhibited strongly delayed SE catabolism resulting in 1.6-, 4.0-, and 7.4-fold higher SE levels than controls at 0–2 hr, 6–8 hr, and 18–20 hr AEL, respectively. Heterozygous embryos from Hsl¹ mutant mothers and control fathers had 1.5-, 4.1-, and 4.1-fold higher SE levels than control embryos at 0–2 hr, 6–8 hr, and 18–20 hr AEL, respectively. In contrast, heterozygous embryos from Hsl¹ mutant fathers and control mothers exhibited similar SE catabolism rates than control embryos (Figure 5A) suggesting that maternal rather than zygotic Hsl determines embryonic SE breakdown. SE catabolism in control embryos correlated with an increase in free sterol content from 12.4 to 20.7 pmol per animal between 0–2 hr and 18–20 hr AEL. Hsl¹ mutant embryos had lower free sterol content than controls throughout embryogenesis and exhibited a blunted increase from 7.8 to 11.6 pmol per animal between 0–2 hr and 18–20 hr AEL (Figure 5B). This suggests that loss of Hsl causes impaired conversion of SE to free sterols during embryogenesis. Since embryonic SE catabolism was essentially dependent on the maternal genotype, we next asked if germline Hsl autonomously controls embryonic SE catabolism. To this end, we re-expressed Hsl specifically in the germline of mutant animals using the nos-GAL4 driver. Enzyme assays revealed a 72% decrease in embryonic SE hydrolysis rates of Hsl¹ as compared to control animals, which was largely reverted by re-expressing Hsl in the germline of mutant mothers (Figure 5C). This indicates that our strategy was successful in restoring active Hsl enzyme in mutant embryos. Germline-specific expression of Hsl fully prevented the SE accumulation in Hsl¹ mutant embryos but not in mothers consistent with a germline-autonomous role of the enzyme in SE catabolism (Figure 5C). To investigate if defective sterol homeostasis affects success rate or speed of the embryonic development of Hsl¹ mutants, we first measured hatching rates and found that under our experimental conditions a similar fraction of control and Hsl¹ mutant embryos hatched into 1st instar larvae (Figure 5D). Between 4 hr and 24 hr AEL, thermal dissipation rates of control and Hsl¹ mutant embryos increased from 138 nW to 188 nW and from 132 nW to 182 nW, respectively, indicating similar metabolic rates in both genotypes (Figure 5E). Moreover, immunostainings of the segment-polarity protein Engrailed revealed that control and Hsl¹ mutant embryos reached comparable stages of embryogenesis at 4–6 hr, 7–9 hr, or 12–14 hr AEL (Figure 5F). Thus, although Hsl is rate-limiting for embryonic SE catabolism and required to adjust sterol levels during development, loss of Hsl is largely compatible with normal embryogenesis.

Sterols are essential for the biophysical properties of biomembranes. To understand the significance of sterols in the context of other membrane lipids, we monitored the embryo lipidome in the early (0–2 hr AEL), middle (6–8 hr AEL), and late (18–20 hr AEL) phases of embryogenesis. In control animals, embryogenesis was associated with a gradual decrease in TG (by 12% and 58% at 6–8 hr and 18–20 hr AEL, respectively, compared to 0–2 hr AEL) but not DG (Figure 6A). This was accompanied by moderate increases in the levels of major glycerophospholipids like phosphatidylcholine (PC, 13% and 23% at 6–8 hr and 18–20 hr AEL, respectively), phosphatidylethanolamine (PE, 19% and 26% at 6–8 hr and 18–20 hr AEL, respectively), phosphatidylinositol (PI, 21% and 34% at 6–8 hr and 18–20 hr AEL, respectively), and phosphatidylserine (PS, 44% and 54% at 6–8 hr and 18–20 hr AEL, respectively; Figure 6A). Within this time frame, we observed even more pronounced increases in sphingolipids and ether-linked glycerophospholipids including ceramide (Cer, 79% and 75% at 6–8 hr and 18–20 hr AEL, respectively), ceramide phosphoethanolamine (CerPE, 59% and 114% at 6–8 hr and 18–20 hr AEL, respectively), and ether-linked PE (PE O-, 17% and 350% at 6–8 hr and 18–20 hr AEL, respectively; Figure 6A). An exception to this general trend was phosphatidylglycerol (PG), which exhibited a subtle decrease during embryogenesis (3% and 7% at 6–8 hr and 18–20 hr AEL, respectively; Figure 6A). The net increase in membrane lipid content was associated with distinct shifts in the molecular composition of each lipid class. PC, PE, PI, and PS shifted toward increased acyl chain length at 18–20 hr AEL when compared to 0–2 hr AEL (Figure 6B). A similar shift was observed for Cer, CerPE, and PE O-, whereas PG shifted toward decreased acyl chain length (Figure 6—figure supplement 1A). Moreover, PC, PI, Cer, CerPE, and PE O- pools remodeled toward decreased acyl chain saturation whereas PE, PS, and PG became slightly more saturated (Figure 6—figure supplement 1A). Although these general shifts in total level, acyl chain length, and saturation degree of each lipid class were often moderate, distinct lipid species including many polyunsaturated glycerophospholipids exhibited more dynamic changes. Examples include PC 36:4, PE 36:4, and PI 36:4, which increased by 169%, 125%, and 194%, respectively, between early and late embryogenesis (Figure 6C, Supplementary file 1d). However, it should be noted that these species were low abundant and therefore contributed little to the overall saturation degree of each lipid class. When compared with control animals, Hsl¹ mutants exhibited similar decreases in TG and increases in membrane lipid levels during embryogenesis (Figure 6A,B; Figure 6—figure supplement 1; Supplementary file 1). Total PC increased slightly less in Hsl¹ compared to control animals between early and late embryogenesis (16% vs 23%), which was largely due to lower amounts of medium-chain PC species containing 30 or less carbon atoms (Figure 6C; Supplementary file 1). Instead, increases in PC species with higher chain length and more double bonds were comparable between Hsl¹ and control embryos (Figure 6C, Supplementary file 1). Likewise, the molecular composition of other phospholipids and ceramides changed with similar dynamics in control and Hsl¹ mutant animals (Figure 6B,C; Figure 6—figure supplement 1; Supplementary file 1). Taken together, these data suggest that (1) Hsl acts as specific regulator of sterols but not of other lipids during embryogenesis and (2) decreased sterols as a consequence of defective SE catabolism are largely compatible with normal remodeling of membrane lipids in Hsl¹ mutant embryos.

In order to more comprehensively understand the consequences of defective sterol handling for reproductive success we analyzed the quantity and quality of eggs laid by control and Hsl¹ mutant females. We used a lipid-depleted medium (LDM), which is naturally low in sterols and other lipids, to increase the reliance of the animals on endogenous sterol reserves and counted eggs and 1st instar larvae produced by individual females. Hsl¹ females had 24% and 20% less cumulative egg output than control females after 4 days and 7 days on LDM, respectively (Figure 7A). In both genotypes, the dietary switch to LDM was associated with a drastic time-dependent decline in egg hatchability reaching a plateau after 7 days. Hsl¹ females produced 37% and 25% less viable progeny than control females within 4 days and 7 days on LDM, respectively, indicating slightly higher developmental success of control embryos in the early period on LDM (Figure 7A). Indeed, hatching rates of Hsl¹ mutant animals were 14% lower after 4 days on LDM as compared to controls (Figure 7—figure supplement 1). After dietary supplementation with exogenous cholesterol, Hsl¹ females showed 16% and 14% lower egg output than control females at 4 days and 7 days, respectively, suggesting that dietary cholesterol partially rescued the fecundity defect of Hsl¹ mutant animals (Figure 7B). Cholesterol addition to LDM rescued the decline in embryo hatchability in both genotypes. As a consequence of similar hatching rates in both genotypes, Hsl¹ females produced 17% and 14% less viable offspring than control females at 4 days and 7 days, respectively (Figure 7B, Figure 7—figure supplement 1). To understand how defective SE catabolism in the maternal organism affects fecundity, we next assessed lipid transfer to control and Hsl¹ mutant oocytes. Vitellogenic follicles of control but also Hsl¹ females produced numerous LDs at stage 10 of oogenesis suggesting that oocyte lipid loading is generally functional in Hsl¹ mutant animals (Figure 7C). To more specifically monitor sterol transport, we followed the incorporation of labeled cholesterol into embryos produced by mothers of both genotypes and found that Hsl¹ mutants transferred 20% and 24% less labeled cholesterol into the embryos laid at 3 days and 4 days after female mating (Figure 7D). This suggests that Hsl indeed promotes the mobilization and transfer of maternal sterols to the progeny. Since Hsl contributes to SE catabolism in both, maternal fat body and germline, we finally used genetic approaches to address nonautonomous contributions of maternal fat body Hsl to fecundity. Fat body-specific knockdown of Hsl expression by means of a Cg-GAL4 driven Hsl shRNA transgene reduced cumulative egg and 1st instar numbers by 27–29% and 32–34%, respectively, after 4 days on LDM as compared to controls harboring either the Hsl shRNA or the Cg-GAL4 only (Figure 7E). Similar to complete Hsl-deficiency the fecundity defects induced by fatbody-specific Hsl knockdown were less pronounced after 7 days on LDM (25–27% less egg numbers, 14–24% less 1st instar larvae compared to controls, Figure 7E). In a complementary approach, we re-expressed Hsl in the fat body of the Hsl¹ mutant animals via the FB-GAL4 driver and assessed fecundity on LDM. After 4 days on LDM, Hsl¹ mutant animals harboring either the UAS-Hsl transgene or the FB-GAL4 element produced significantly less eggs (21–29%) and 1st instar larvae (24–29%) than controls with the FB-GAL4 element (Figure 7F). Fat body-specific Hsl re-expression was sufficient to ameliorate fecundity defects associated with complete Hsl-deficiency (14% less eggs and 8% less 1st instar larvae compared to controls at 4 days on LDM, Figure 7F). Taken together, these data indicate that optimal egg production and reproductive success requires Hsl-mediated SE catabolism in the fat body of the mothers, particularly when dietary sterol is limited.

An ECL Anti-Mouse IgG-Horseradish Peroxidase antibody from sheep s (Cat#: NA931; RRID: AB_772210) and a monoclonal Anti-His antibody from mouse (Cat#: 27-4710-01; RRID: AB_771435) were purchased from GE Healthcare (Chicago, IL). A monoclonal Anti-engrailed/invected antibody from mouse (Cat#: 4D9; AB_528224) was obtained from Developmental Studies Hybridoma Bank (Iowa City, IA). A polyclonal HSL antibody from rabbit (Cat#: 4107; RRID: AB_2296900) and a monoclonal Anti-GAPDH antibody from rabbit (Cat#: 2118, clone 14C10; RRID: AB_561053) were obtained from Cell Signaling Technology (Cambridge, UK). A polyclonal Anti-Rabbit IgG (H+L)-Horseradish Peroxidase antibody from goat (Cat#: PI-1000; RRID: AB_2336198) was purchased from Vector Laboratories (Peterborough, UK). TG 54:3 (triolein/trioleoylglycerol; Cat#: T7140), DG 36:2 (1,2-dioleoyl-rac-glycerol; Cat#: D8397), MG 18:1 (1-oleoyl-rac-glycerol; Cat#: M7765), egg yolk phosphatidylcholine (Cat#: P3556), soybean phosphatidylinositol (Cat#: 79401), CE 18:1 (cholesteryl oleate; Cat#: C9253), and cholesterol (Cat#: C3045) were obtained from Sigma Aldrich (St. Louis, MO). TG 54:0 (tristearin/tristearylglycerol; Cat#: 33–1800) and CE 18:0 (cholesteryl stearate; Cat#: 64–1800) were obtained from Larodan (Solna, SWE). Cholesterol [25,26–3H] (Cat#: ART-1987) was from American Radiolabeled Chemicals Inc (St. Louis, MO), [9,10–3H] Triolein was from PerkinElmer Life Sciences (Waltham, MA) and [1–14C] Palmitic acid (Cat#: MC121) was from Moravek Biochemicals (Brea, CA).

Synthetic lipid standards and ergosterol were purchased from Avanti Polar Lipids, Inc (Alabaster, AL), Sigma Aldrich (St. Louis, MO) and Toronto Research Chemicals (Toronto, CA). ¹³C uniformly labeled glucose was purchased from Euriso-top (Saint Aubin, FR) and yeast nitrogen base without amino acids from BD Difco (Le Pont de Claix, FR). All used solvents were of at least HPLC grade. Stocks of internal standards were stored in glass ampoules at −20°C until used for the preparation of internal standard mix in 10:3 methyl-tert-butyl ether (MTBE)/methanol. A total of 700 µl internal standard mix contained: 356 pmol cholesterol D₇, 224 pmol zymosterol D₅, 215 pmol campesterol D₆, 207 pmol sitosterol D₆, 201 pmol lanosterol D₆, 418 pmol stigmasterol D₆, 233 pmol desmosterol D₆, 196 pmol ¹³C ergosterol, 443 pmol CE 16:0 D₇, 417 pmol TG 50:0 D₅, 116 pmol DG 34:0 D₅, 220 pmol PC 25:0, 77 pmol LPC 13:0, 107 pmol PS 25:0, 354 pmol PE 25:0, 85 pmol LPE 13:0, 96 pmol PI 25:0, 109 pmol PG 25:0, 145 pmol Cer 30:1, 123 pmol PA 25:0, 91 pmol LPA 13:0, 178 pmol CerPE 29:1, 38 pmol LPI 13:0, 54 pmol CL 56:4, 59 pmol LPS 13:0, 75 pmol LPG 13:0. ¹³C uniformly labeled ergosterol was produced in the prototrophic yeast strain W303 Y3358 according to Knittelfelder et al., 2020.

The strains Akh^A, bmm¹ and FB-GAL4 have been described previously (Gáliková et al., 2015; Grönke et al., 2007; Grönke et al., 2005). Hsl¹, UAS-Hsl and UAS-Hsl-EGFP were generated in this study as described below. Hsl^b24 was a kind gift of Xun Huang (Bi et al., 2012). Act5c-GAL4 (Cat#: 4414) Myo31DF-GAL4 (Cat#: 67088), Desat1-GAL4 (Cat#: 65404), elav-GAL4 (Cat#: 8760), Cg-GAL4 (Cat#: 7011), and nos-GAL4 (Cat#: 64277) strains were obtained from Bloomington Drosophila Stock Center (BDSC, Bloomington, IL). A Hsl shRNA transgenic strain and a control strain were obtained from the TRiP collection of BDSC (Cat#: 65148, Cat#: 36304). A w¹¹¹⁸ strain (Cat#: 60000) from Vienna Drosophila Resource Center (VDRC, Vienna, AUT) was used as control and for backcrossing of transgenic flies. More details are described in Table 1.

The Hsl¹ allele was generated by imprecise excision of a P-element inserted in the 5′ UTR of the Hsl gene (GE15823) using a conventional P-element mobilization procedure. Imprecise excision events were identified by PCR using 5’-AAAGATCTGAGCCGCAATAGGTGGAC-3’ and 5’- AAGGTACCCTGATGAAGCGGCTAGACTTG-3’ as flanking primers. This resulted in the identification of the Hsl¹ allele, which encompasses a deletion of 2743 bp (2R:20,257,711.20,260,454; FlyBase FB2020-04) that removes most of the Hsl open-reading frame including the start codon. Sequencing of the PCR products revealed the presence of 51 bp residual P element sequence in the Hsl¹ allele. The same PCR reaction was routinely used for genotyping of the Hsl locus (Figure 2—figure supplement 1). To this end, individual flies were crushed in 50 µl 10 mM Tris/HCl buffer pH 8.2 containing 1 mM EDTA, 25 mM NaCl, and 200 µg/ml proteinase K. Homogenates were incubated for 30 min at 37°C followed by 5 min at 95°C. Debris was removed by centrifugation for 1 min at 10,000 x g and 1 µl of the supernatant was subjected to PCR using the Phire Tissue Direct PCR Master Mix (Cat#: F170S; Thermo Scientific, Waltham, USA) using the following cycling conditions: 5 min 98°C; 35 cycles with 30 s 98°C, 30 s 55°C, 90 s 72°C; 10 min 72°C; hold at 4°C. The Hsl¹ allele was backcrossed for eight generations into the w¹¹¹⁸ background prior to analysis. Transgenic flies for the expression of Hsl under UAS-control were generated by P-element-mediated germline transformation. To construct plasmids for the expression of Hsl under the control of the UAS element a pFLC-1 plasmid encoding the Hsl cDNA (clone#: RE52776, Drosophila Genomics Resource Center, Bloomington, IL) was digested with NotI and KpnI and the fragment encoding Hsl was ligated into pUAST. A Hsl-EGFP fusion gene was engineered by PCR amplification of the Hsl cDNA with the primer pairs 5’- AAAGATCTGAGCCGCAATAGGTGGAC-3’ and 5’- AAGGTACCCTGATGAAGCGGCTAGACTTG-3’ and subsequent ligation of the PCR product to the pEGFP-N2 vector by means of the KpnI and BglII restriction sites. The region encoding Hsl-EGFP was then cut with BglII and NotI and ligated to pUAST. Finally, pUAST plasmids were injected into embryos. Eye color was used to identify transgenic flies, which were backcrossed into the w¹¹¹⁸ background for six generations. Mice lacking the Lipe/HSL gene specifically in adipocytes (AHKO) were generated by crossbreeding HSL^flox/flox mice with Adiponectin-Cre transgenic mice (The Jackson Laboratory, Bar Harbor ME; JAX stock number: 010803; genetic background: C57BL/6J) (Eguchi et al., 2011; Haemmerle et al., 2002a). Mice were backcrossed onto C57BL/6J for 10 generations. Mice homozygous for the HSL^flox/flox allele were used as controls.

Flies were maintained at 25°C, 60% humidity, and a 12 hr:12 hr dark/light cycle. If not stated otherwise, flies were reared on standard food containing 69.6 g/l corn flour (Haindl Mühle, Kalsdorf bei Graz, AUT), 8.7 g/l soy flour (Soja Austria, Vienna, AUT), 19.1 g/l molasses/beet syrup (Cat#: 01936, Grafschafter, Meckenheim, DE), 69.6 g/l malt (Cat#: 728985, CSM Austria, Vienna, AUT), 15.7 g/l yeast (Cat#: 03462, Gewürzmühle Brecht, Eggenstein, DE), 5.4 g/l agar-agar (Cat#: 00162, Gewürzmühle Brecht, Eggenstein, DE), 5.4 ml/l propionic acid (Cat#: P5561 Merck, Darmstadt, DE) and 1.3 g/l methyl 4-hydroxybenzoate (Cat#: W271004, Merck, Darmstadt, DE). Flies were starved on tap water containing 6 g/l agar-agar (Cat#: 2266.2, Carl Roth, Karlsruhe, DE). Lipid-depleted medium (LDM), which is naturally low in sterols and other lipids, contained 100 g/l yeast extract (Cat#: 212720; Bacto Thermo Fisher Scientific, Waltham, MA), 100 g/l sucrose, 6 g/l agar-agar (Cat#: 00162, Gewürzmühle Brecht, Eggenstein, DE), 5.4 ml/l propionic acid (Cat#: P5561, Sigma Aldrich, St. Louis, MO), 1.3 g/L methyl 4-hydroxybenzoate (Cat#: W271004, Sigma Aldrich, St. Louis, MO), and (optional) 0.01% cholesterol (w/v). For density seeding and embryo collections flies were reared on apple juice agar plates containing 20 g/l sucrose, 20% commercial apple juice, 17 g/l agar-agar (Cat#: 2266.2, Carl Roth, Karlsruhe, DE), and 1.3 g/l methyl 4-hydroxybenzoate. Plates were supplemented with fresh yeast paste to promote egg laying. If not otherwise indicated mated female flies were used for all assays. Mice were maintained under specific pathogen free conditions with ad libitum access to a chow diet (R/M-H Extrudate, Cat#: V1126-037, Ssniff Spezialdiäten GmbH, Soest, DE). Regular housing temperatures were maintained between 22–23°C with a 14 hr:10 hr light/dark cycle. Females were used for all analyses unless indicated otherwise. All animal experiments were approved by the Austrian Federal Ministry for Science, Research, and Economy (protocol number BMWFW-66.007/0026/-WF/V/3b/2017) and the ethics committee of the University of Graz, and were conducted in compliance with the council of Europe Convention (ETS 123). Animals were allocated into experimental or control groups depending on the genotype. Individuals of the same genotype and age were randomely allocated to different experimetal conditions (e.g. food source).

For the expression of His₆-Hsl in tissue culture cells the open reading frame of the Hsl gene was amplified by PCR using the primer pair 5’-GCACTCGAGATTGACGCGGCTTCCG-3’ and 5’- GCATCTAGACTATGAAGCGGCTAGACTTG-3’ and ligated to the vector pcDNA4/Hismax C (Cat#: V86420, Thermo Fisher Scientific, Waltham, MA) using the restriction sites XhoI and XbaI. A plasmid for the expression of His₆-β-Galactosidase was provided by the supplier. A plasmid encoding for His₆-Hsl from Mus musculus (MmHsl) has been described previously (Lass et al., 2006). All plasmids were purified using the QIAGEN Plasmid Midi Kit (Cat#: 12143, QIAGEN, Hilden, DE). COS-7 cells (Cat#: CRL-1651; RRID: CVCL_0224; American Type Culture Collection, ATCC, Manassas, VA) were cultured in Dulbecco's modified Eagle's medium (Cat#: 41965039; Gibco Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 μg/ml streptomycin at 37°C, 95% humidity, and 5% CO₂. Mycoplasma tests were routinely performed to ensure that cells are devoid of contamination. Cells were transfected with DNA complexed to Metafectene (Cat#: T020-5.0; Biontex GmbH, DE) according to the manufacturer's instructions and used for experiments 24 hr thereafter.

Cell culture samples for lipid hydrolase assays were prepared by sonication of COS-7 cells in 0.25 M sucrose, 1 mM EDTA, 1 mM DTT containing 20 μg/ml leupeptin, 2 μg/ml antipain, and 1 μg/ml pepstatin (solution A) followed by centrifugation at 4°C and 1,000 × g. Fly abdominal samples were prepared by gross dissection of 10–15 abdomen per replicate in 250 µl solution A followed by sonication for 30 s with 15% output power using a Sonoplus sonicator (Bandelin electronic GmbH, Berlin, DE). These abdominal homogenates were first centrifuged at 4°C and 1000 x g for 10 min to remove debris. The resulting supernatants were then centrifuged at 4°C and 20,000 x g for 30 min and the soluble fraction was used for lipid hydrolase assays. Staged embryos were disrupted by sonication in solution A, centrifuged at 4°C and 1000 x g for 10 min and the resultant supernatants were used for lipid hydrolase assays. Murine tissue samples were homogenized in solution A using an UltraTurrax tissue homogenizer (IKA, Staufen, DE) and centrifuged at 20,000 x g and 4°C for 30 min to collect soluble infranatans. Protein concentrations were determined with the Bio-Rad Protein Assay kit according to the manufacturer's instructions (Cat#: 5000001; Bio-Rad Laboratories, Inc, Hercules, CA) using bovine serum albumin (BSA) as standard. Substrates for assaying TG and MG hydrolase activities were prepared as described using a Bis-tris propane buffer pH 7.5 (Heier et al., 2016; Heier et al., 2015). DG hydrolase substrate was prepared by emulsifying 0.3 mM 1,2-dioleoyl-rac-glycerol with 37 µM egg yolk PC and 11 µM soybean PI by sonication in BTP buffer pH 7.5 followed by the addition of essentially fatty acid free BSA to a final concentration of 2%. SE hydrolase substrate was prepared by emulsifying 0.45 mM cholesteryl oleate (containing 1 µCi/ml ¹⁴C-cholesteryl oleate as tracer) with 0.75 mM egg yolk phosphatidylcholine and 0.23 mM soybean phosphatidylinositol by sonication in BTP buffer pH 7.5 followed by the addition of essentially fatty acid free BSA to a final concentration of 5%. For TG and CE hydrolase assays, 100 µl samples were incubated with 100 µl substrate aliquots at 37°C for 60 min in a shaking water bath. Enzyme reactions were terminated by the addition of 3.3 ml methanol/chloroform/heptane (10:9:7, v/v/v) and 1 ml 0.1 M potassium carbonate/boric acid buffer pH 10.5. After centrifugation radioactivity in 1 ml of the upper phase was determined by liquid scintillation counting. MG and DG hydrolase activities were determined by incubating 10 µl protein samples with 50 µl substrate aliquots for 30–60 min at 37°C in a 96-well plate. The generation of free fatty acids was then determined with a colorimetric kit (NEFA-HR(2) reagent, Cat#: 999–34691, Cat#: 995–34791, Cat#: 991–34891, Cat#: 993–35191, and Cat#: 276–76491, Fujifilm Wako Diagnostics, Mountain View, CA). Mock substrates without MG or DG were used to determine background FA formation. All lipid hydrolase assays were repeated at least twice. All assays were performed with 3–8 biological replicates except for assays with recombinant proteins expressed in COS-7 cells, in which three technical replicates were used.

Protein extracts of transfected cells and murine tissue samples for immunoblotting analysis were generated as for the lipid hydrolase assays. Equal amounts of cell or tissue protein were separated by SDS-PAGE and electroblotted onto nitrocellulose membranes. Unspecific binding sites were blocked by incubating membranes with 10% milk powder in 150 mM NaCl, 50 mM Tris-HCl, 0.1% Tween-20, pH 7.6 (TBST) for 1 hr at room temperature under mild agitation. Proteins with N-terminal His₆-tags were detected by consecutive incubations with a monoclonal Anti-His antibody from mouse and an Anti-Mouse IgG-Horseradish Peroxidase antibody from sheep diluted 1:5,000 in TBST containing 5% milk powder. Murine HSL and GAPDH were detected using a polyclonal HSL antibody from rabbit (1:1,000) and a monoclonal Anti-GAPDH antibody from rabbit (1:10,000), respectively, as first antibodies, and a polyclonal Anti-Rabbit IgG (H+L)-Horseradish Peroxidase antibody from goat (1:5,000) as secondary antibody. Chemiluminescence was detected using the Pierce Supersignal West Pico PLUS chemiluminescent substrate (Cat#: 34577, Thermo Fisher Scientific, Waltham, MA) and the Bio-Rad ChemiDoc imaging system (Bio-Rad Laboratories, Inc, Hercules, CA).

For metabolite assays, microcalorimetry, and fecundity assays 1st instar larvae were collected on apple juice agar plates and grown at non-crowding densities in standard food (~50 animals per 7 ml food in a 28 ml vial or ~150 animals per 20 ml food in a 68 ml vial). After eclosion, flies were maintained in 68 ml vials at densities of 50 males plus 50 females for the indicated time with regular transfer to fresh food. Flies were frozen in liquid nitrogen and either processed immediately or stored at −80°C. The determination of glycogen and glucose was performed according to Tennessen et al. using the Glucose (GO) Assay kit (Cat#: GAGO20-1KT, Sigma Aldrich, St. Louis, MO) and Amyloglucosidase (Cat#: A1602-25MG, Sigma Aldrich, St. Louis, MO) (Tennessen et al., 2014). Glycerol was determined from the same homogenates using the Free Glycerol Reagent (Cat#: F6428, Sigma Aldrich, St. Louis, MO). TG equivalents were determined according to Hildebrandt et al. using the Triglycerides reagent (Cat#: 981786, Thermo Fisher Scientific, Waltham, MA) (Hildebrandt et al., 2011). Free FA were determined from the same homogenates as TG using the NEFA-HR(2) reagent (Cat#: 999–34691, Cat#: 995–34791, Cat#: 991–34891, Cat#: 993–35191, and Cat#: 276–76491, Fujifilm Wako Diagnostics, Mountain View, CA). For microcalorimetry measurements five adult flies or 100 embryos were transferred to 2 ml glass ampoules supplemented with 500 µl complex medium or agar, respectively, and heat dissipation was measured in a TAM IV microcalorimeter (TA instruments, New Castle, DE). Flies were pre-starved for 8 hr before measurements of the starved status. For determination of hatching rates on apple juice agar plates with yeast paste batches of ~100 embryos were collected 2 hr AEL and scored for hatching events 24–48 hr later. For fecundity assays, single female flies were collected within 1d after eclosion and maintained with two males in 28 ml vials containing 2 ml food. Flies were flipped daily. Eggs were counted directly after each transfer and hatching rates were determined 48 hr later. To monitor starvation sensitivity, female flies were aged on standard food for 7d and then transferred to 68 ml vials containing 10 ml tap water with 0.6 g/l agar-agar (Cat#: 2266.2, Carl Roth, Karlsruhe, DE). 4 replicates with 10 flies were set up and dead flies were scored every 2–12 hr. All metabolite and microcalorimetry assays were performed at least twice with 4–12 biological replicates per experiment. Starvation experiments were performed at least twice with a total of 39–40 flies in four replicates. Hatching rates were determined twice with ~500 individual embryos in five replicates. Fecundity assays were typically performed with 10–17 biological replicates per genotype and data of two independent experiment was pooled resulting in a total of 25–34 biological replicates. Individual flies or cohorts were excluded from analysis upon the following criteria: (1) complete sterility, (2) early death unrelated to the applied stress (e.g. squeezing or sticking to food), (3) early escape from experimental vial.

Abdominal fat body tissue associated with the spermatheca of ad libitum fed females was dissected in ice-cold phosphate-buffered saline (PBS) and incubated for 30 min in PBS containing 5 µg/ml Hoechst 33342 (Cat#: 14533, Sigma Aldrich, St. Louis, MO), 5 µg/ml BODIPY 493/503 (Cat#: B2103, Thermo Fisher Scientific Invitrogen, Carlsbad, CA) and 1 µg/ml CellMask Deep Red (Cat#: C10046, Thermo Fisher Scientific Life Technologies, Carlsbad, CA). Fat body tissue expressing Hsl-EGFP was incubated for 30 min in LipidTOX Deep Red (Cat#: H3477, Thermo Fisher Scientific Life Technologies, Carlsbad, CA) diluted 1:1000 in PBS. Samples were then mounted in PBS and imaged immediately by confocal fluorescence microscopy. Ovaries were dissected in PBS and fixed for 30 min in PBS containing 4% paraformaldehyde. Ovaries were then incubated for 30 min in PBS containing 5 µg/ml Hoechst 33342, 5 µg/ml BODIPY 493/503 and 1 µg/ml CellMask Deep Red. After transfering to glass slides, egg chambers were separated with fine forceps and mounted in PBS containing 30% glycerol. Ring glands of white prepupae were processed and stained as ovaries and were mounted in PBS. Samples were imaged on a Leica SP8 confocal laser scanning microscope using a HC PL APO 20x multi-immersion objective with a 0.75 NA. Embryos were collected on apple juice agar plates supplemented with yeast and aged at 25°C for the indicated time. Embryos were dechorionated with 50% bleach (DanKlorix Hygienereiniger mit Chlor), washed extensively with water and fixed for 20 min in a mixture of 500 µl heptane and 500 µl PBS with 4% paraformaldehyde. After removal of the aqueous phase, 1 ml methanol was added and samples were vortexed for 1 min. Embryos were washed twice with methanol and three times for 5 min with 0.5 ml 10 mM Tris pH 7, 55 mM NaCl, 7 mM MgCl₂, 5 mM CaCl₂, 20 mM glucose, 50 mM sucrose, 0.1% BSA, and 0.1% Tween-20 (BBT). Unspecific binding sites were blocked by incubating samples for 1 hr with 0.5 ml BBT containing 2% horse serum. Primary antibodies were diluted 1:100 in 100 µl BBT and incubated with the samples over night at 4°C. Primary antibody binding was detected using the Vectastain Elite ABC-HRP system according to the manufacturer’s instructions (Cat#: PK-6200, Vector Laboratories, Peterborough, UK) with 3,3’-diaminobenzidine as peroxidase substrate.

RNA was isolated from cohorts of 5 adult female flies using the RNeasy Mini Kit (Cat#: 74104, QIAGEN, Hilden, DE). Murine adipose tissue (~100 mg) was homogenized in 1 ml of TRIzol reagent (Cat#: 15596026, Thermo Fisher Invitrogen, Carlsbad, CA) using an Ultra-Turrax homogenizer (IKA, Staufen, DE) and incubated at room temperature for 5 min. Phase separation was induced by addition of 100 μl 1-bromo-3-chloropropane and centrifugation at 12,000 x g and 4°C for 15 min. Clear supernatant was transferred and total RNA was precipitated by addition of 500 μl isopropyl alcohol and centrifugation at 12,000 x g and 4°C for 15 min. cDNA was synthesized using the SuperScript III First-Strand Synthesis Supermix (Cat#: 18080051, Thermo Fischer Invitrogen, Carlsbad, CA) for fly samples and QuantiTect Reverse Transcription Kit (Cat#: 205313, QIAGEN, Hilden, DE) for murine samples. qPCR was performed on the StepOnePlus Real-Time PCR System (Applied Biosystems, Waltham, MA) using iTaq Universal SYBR Green Supermix (Cat#: 1725120, Bio-Rad, Hercules, CA). The thermal program for the qPCR included stage 1: 95°C, 10 min and stage 2: 95°C, 0.5 min and 60°C, 1 min for a total of 40 cycles. Non-template control (RNA) and non-reaction control (H₂O) were routinely performed. qPCR experiments were performed twice with 4–5 biological replicates. Relative mRNA expression levels were determined by means of the ΔΔCt method using rp49 (FlyBase name RpL32) and 36B4 as normalization control for fly and murine samples, respectively. A comprehensive list of primers used can be found in Table 2.

For the determination of total SE and TG content, lipid extraction followed essentially the protocol of Matyash et al. with minor modifications (Matyash et al., 2008). Batches of 5 flies, five adult segments, five adult tissues, 100 embryos, or pieces of 10–20 mg murine adipose tissue were collected in 2 ml safe-seal micro tubes (Cat#: 72.695.500, Sarstedt, Nürmbrecht, DE) and were disrupted in a vibration mill (Retsch MM400, Retsch, Haan, DE) with a metal bead (5 mm diameter; Cat#: 504942, Askubal Korntal-Münchingen, DE) in 1 ml MTBE/methanol (10:3, v/v) for 60 s followed by incubation at 600 rpm and 25°C for 30 min in a ThermoMixer (Eppendorf, Hamburg, DE). After the addition of 200 µl distilled H₂O samples were vortexed and phase separation was induced by centrifugation at 16,000 x g for 5 min. Upper phases were collected and lower phases were re-extracted with 300 µl artificial upper phase. The combined upper phases were evaporated in a stream of nitrogen, dissolved in chloroform/methanol (2:1, v/v) and aliquots of the extracts were applied onto silica gel 60-coated TLC plates (Cat#: 1.05554.0001, Merck, Darmstadt, DE), which were developed in petrolether/diethylether/acetic acid (25:25:1, v/v/v) until the solvent front reached ~1/3 of the plate. The plates were briefly dried and further developed in petrolether/diethylether (49:1, v/v) until the solvent front reached the top of the plate. The plates were dried thoroughly and immersed in a derivatization solution containing 50 g/l copper sulfate, 10% phosphoric acid, and 25% ethanol. Carbonization of lipids was achieved by incubating plates for 25 min at 120°C. SE content was determined by scanning densitometry at 400 nm using a Camag TLC scanner 3 (Camag, Muttenz, CH) and 3+1 mixtures of cholesteryl oleate and cholesteryl stearate or trioleoylglycerol and tristearylglycerol, respectively, as references. Data were normalized to animal number (adult flies, fly embryos) or mg tissue weight (murine adipose tissue). Lipid analysis by TLC was performed at least twice with 3–8 biological replicates per experiment.

For radiolabeling experiments, standard food was melted, mixed with ³H-cholesterol (0.05% final concentration and 50 µCi/ml final specific activity) or ¹⁴C-palmitic acid (200 µM final concentration and 10 µCi/ml final specific activity), and transferred à 500 µl into 24-well plates. 1st instar larvae were collected on apple juice agar plates supplemented with yeast paste and 20 larvae per genotype were transferred into each well containing radiolabeled food. Adult flies were sampled either within 1 day after eclosion or reared on unlabeled food for the indicated time. For lipid extraction, single flies were crushed with a pipet tip in 500 µl MTBE/methanol (3:1, v/v) and incubated for 30 min at 25°C and 600 rpm. After the addition of 100 µl H₂O, samples were vortexed and phase separation was induced by centrifugation at 16,000 x g for 5 min. Upper phases were collected and lower phases were re-extracted with 150 µl artificial upper phase. The combined upper phases were evaporated in a stream of nitrogen and the extracts were applied onto Silica G TLC plates, which were developed as described above. For the detection of sterols and SE, samples were spiked with cholesterol and cholesteryl oleate. Lipid bands were detected by exposure to iodine vapor and the associated radioactivity was determined by liquid scintillation counting. For the measurement of cholesterol transfer to embryos, individual females were mated with two males of the same genotype in vials containing 2 ml LDM. Flies were flipped daily to fresh food and their egg cohorts were harvested in 100 µl distilled water. Embryos were disrupted by sonication and lipids were extracted by incubating the embryo homogenates with 500 µl MTBE/methanol (3:1, v/v) at 25°C and 600 rpm for 30 min. After centrifugation at 16,000 x g for 5 min the upper phases were collected and the lower phases were re-extracted with 150 µl artificial upper phase. Radioactivity in the combined upper phases was determined by liquid scintillation counting. Radiolabling experiments were performed at least twice with 4–5 biological replicates of adult flies and 9–10 replicates of embryos per assay. Individual flies or cohorts were excluded from analysis upon the following criteria: (1) complete sterility, (2) early death unrelated to the applied stress (e.g. squeezing or sticking to food), (3) early escape from experimental vial.

Four batches of 5 flies each were mixed with 700 µl MTBE/methanol (10/3, v/v) in 2 ml safe-seal micro tubes (Cat#: 72.695.500, Sarstedt, Nürmbrecht, DE) and disrupted with a metal bead (5 mm diameter; Cat#: 504942, Askubal Korntal-Münchingen, DE) in a Retsch MM 400 mixer mill (3 min, 30 Hz, 4°C) and lipids were extracted by shaking for 20 min at 1400 rpm and 4°C in a ThermoMixer (Eppendorf, Hamburg, DE). After the addition of 200 µl H₂O, samples were again incubated 20 min at 1400 rpm and 4°C. Phase separation was induced by centrifugation for 10 min at 16,000 x g and 4°C. The upper organic phase was collected and dried under a stream of nitrogen. Lipids were washed by dissolving in 500 µl chloroform/methanol (2/1, v/v) and again dried under a stream of nitrogen and stored at −20°C or prepared immediately for LC-MS analysis. For MS analysis, lipid extracts were dissolved in 0.5 ml chloroform/methanol (2/1, v/v), 50 µl of the sample were diluted with 100 µl isopropanol and 10 µl were injected for analysis using UPLC-QTOF-MS as described (Knittelfelder et al., 2014). Data analysis was performed using the Lipid Data Analyzer (LDA) software (Hartler et al., 2011). The abundance of each TG species was normalized to the intensity of the internal standard TG 51:0 (Larodan, Solna, SWE) and to animal number.

A total of 100 embryos or five female flies were homogenized with 1 mm zirconia beads in a cooled tissuelyzer for 2 × 5 min at 30 Hz in 200 µl isopropanol. The embryo homogenate was evaporated in a vacuum desiccator to complete dryness and subjected to lipid extraction. In case of the fly homogenate, an aliquot corresponding to one fly was transferred to a new sample tube and vacuum dried. Lipid extraction was performed according to Sales et al., 2017. In brief, 700 µl internal standard mix in MTBE/methanol (10:3, v/v) was added to each sample and vortexed for 1 hr at 4°C. After the addition of 140 µl H₂O, samples were vortexed for another 15 min. Phase separation was induced by centrifugation at 16,000 x g for 15 min. The organic phase was transferred to a glass vial and evaporated. Samples were reconstituted in 300 µl chloroform/methanol (2:1, v/v). To a new vial, 100 µl were transferred and used for lipidome analysis. To quantify sterols, 150 µl of lipid extract were evaporated and acetylated with 300 µl chloroform/acetyl chloride (2:1, v/v) for 1 hr at room temperature (modified from Liebisch et al., 2006). After evaporation, sterol samples were reconstituted in 150 µl isopropanol/methanol/chloroform (4:2:1, v/v/v) and 7.5 mM ammonium formate (spray solution). For sterol measurements, samples were 1:5 diluted with spray solution. For lipidome measurements, samples were 1:10 diluted with spray solution. Mass spectrometric analysis was performed on a Q Exactive instrument (Thermo Fischer Scientific, Waltham, MA) equipped with a robotic nanoflow ion source TriVersa NanoMate (Advion BioSciences, Ithaca, NY) using nanoelectrospray chips with a diameter of 4.1 µm. The ion source was controlled by the Chipsoft 8.3.1 software (Advion BioSciences, Ithaca, NY). Ionization voltage was + 0.96 kV in positive and − 0.96 kV in negative mode; backpressure was set at 1.25 psi in both modes. Samples were analyzed by polarity switching (Schuhmann et al., 2012). The temperature of the ion transfer capillary was 200°C; S-lens RF level was set to 50%. All samples were analyzed for 10 min. FT MS spectra were acquired within the range of m/z 400–1000 from 0 min to 0.2 min in positive and within the range of m/z 350–1000 from 5.2 min to 5.4 min in negative mode at the mass resolution of R m/z 200 = 140,000; automated gain control (AGC) of 3 × 10⁶ and with the maximal injection time of 3000 ms. t-SIM in positive (0.2–5 min) and negative (5.4–10 min) mode was acquired with R m/z 200 = 140,000; automated gain control of 5 × 10⁴; maximum injection time of 650 ms; isolation window of 20 Th and scan range of m/z 400–1000 in positive and m/z 350–1000 in negative mode, respectively. The inclusion list of masses targeted in t-SIM analyses started at m/z 355 in negative and m/z 405 in positive ion mode and other masses were computed by adding 10 Th increment (i.e. m/z 355, 365, 375, …) up to m/z 1005. Acetylated sterols were quantified by parallel reaction monitoring (PRM) FT MS/MS in an additional measurement. FT MS spectra within the range of m/z 350–1000 were acquired from 0 min to 0.2 min and t-SIM ranging from m/z 350 to 500 were acquired from 0.2 min to 4 min with the same settings as described above. PRM spectra were acquired from 4 min to 10 min. For PRM micro scans were set to 1, isolation window to 0.8 Da, normalized collision energy to 12.5%, AGC to 5 × 10⁴ and maximum injection time to 3000 ms. All spectra were pre-processed using repetition rate filtering software PeakStrainer and stitched together by an in-house developed script (Schuhmann et al., 2017b; Schuhmann et al., 2017a). Lipids were identified by LipidXplorer software (Herzog et al., 2012). Molecular Fragmentation Query Language (MFQL) queries were compiled for acetylated sterols, PA, LPA, PC, PC O-, LPC, LPC O-, PE, PE O-, LPE, PI, LPI, PS, LPS, PG, LPG, CL, CerPE, Cer, TG, DG lipid classes. The identification relied on accurately determined intact lipid masses (mass accuracy better than 5 ppm). Lipids were quantified by comparing the isotopically corrected abundances of their molecular ions with the abundances of internal standards of the same lipid class. For acetylated sterols, the specific fragment (loss of acetyl group) was used for quantification. Shotgun lipidomics experiments were performed twice with 3–4 biological replicates and both datasets were pooled for analysis resulting in a total number of 7–8 biological replicates. Data were normalized to animal number.

Group differences were analyzed by Student’s unpaired t-tests and one-way ANOVAs using the Bonferroni correction for multiple comparisons. Differences were considered statistically significant at p<0.05. Lipidomic data were analyzed by multiple t-tests using the Benjamini-Hochberg method to control for a false discovery rate of 0.01. Survival was analyzed by log-rank tests. All statistical analyses were performed with the GraphPad Prism 8.0 software.