We previously reported the presence of small LD (<1 μm) at the apical tip of the testis (Wat et al., 2020), a finding we reproduced in w¹¹¹⁸ males using neutral lipid stain BODIPY (Figure 1A). These LD were present in the region that contains stem cells, early-stage somatic cells, and germline cells (Figure 1A, A', arrows). LD were also present in the hub, an organizing center and stem cell niche in the Drosophila testis (Figure 1A'', A''', arrows) (de Cuevas and Matunis, 2011), but largely absent within the area occupied by spermatocytes (Figure 1A, A', arrowheads). This LD distribution was reproduced in two independent genetic backgrounds and at two additional ages (Figure 1B, C). While LD may contain multiple neutral lipid species (Walther et al., 2017), cholesterol-binding fluorescent polyene antibiotic filipin III did not detect cholesterol within testis LD (Figure 1—figure supplement 1A), suggesting triglyceride is the main neutral lipid in Drosophila testis LD.

Drosophila spermatogenesis requires the co-development and differentiation of the germline and the somatic lineages (Boyle and DiNardo, 1995). To identify LD in each lineage, we used the GAL4/UAS system to overexpress a GFP transgene fused to the LD-targeting motif of motor protein Klarsicht (UAS-GFP-LD) (Yu et al., 2011). Somatic overexpression of UAS-GFP-LD using Traffic jam (Tj)-GAL4 revealed that the majority of somatic LD in 0-day-old males were located <30 μm from the hub (Figure 1D, E). Because the somatic LD distribution coincided with a marker for somatic stem cells and their immediate daughter cells (Zinc finger homeodomain 1, Zfh-1) (Figure 1F; two-sample Kolmogorov–Smirnov test) (Leatherman and Dinardo, 2008), but not with a marker for late somatic cells (Eyes absent, Eya) (Amoyel et al., 2016; Fabrizio et al., 2003), our data suggest LD are present in early somatic cells. Germline overexpression of UAS-GFP-LD using nanos (nos)-GAL4 demonstrated the presence of LD within germline cells near the apical tip of the testis in 0-day-old males (Figure 1G, H). Specifically, the disappearance of germline LD coincided with peak expression of a GFP reporter that reflects the expression of Bag-of-marbles (Bam) protein in the testis (Bam-GFP) (Chen and McKearin, 2003; Figure 1I, J). Because peak Bam expression signals the last round of transient amplifying mitotic cell cycle prior to the germline’s transition into the meiotic cell cycle (Insco et al., 2009; Chen et al., 2014b; Gönczy et al., 1997), our data suggest that germline LD, like somatic LD, are present in cells at early stages of development.

Adipose triglyceride lipase (ATGL) is a critical regulator of neutral lipid metabolism and LD (Athenstaedt and Daum, 2003; Chitraju et al., 2013; Grönke et al., 2005; Haemmerle et al., 2006; Haemmerle et al., 2011; Huijsman et al., 2009; Korbelius et al., 2019; Kurat et al., 2006; Zimmermann et al., 2004; Attané et al., 2016). Loss of ATGL in many cell types triggers LD accumulation, and ATGL overexpression decreases LD number (Grönke et al., 2007; Grönke et al., 2005; Haemmerle et al., 2006; Korbelius et al., 2019; Zimmermann et al., 2004; Lee et al., 2014; Tuohetahuntila et al., 2016). Given that the Drosophila ATGL homolog brummer (bmm) regulates testis LD induced by mitochondrial stress (Sênos Demarco et al., 2019), we explored whether bmm regulates testis LD under normal physiological conditions. We first examined bmm expression in the testis by isolating this organ from flies carrying a bmm promoter-driven GFP transgene (bmm-GFP) that recapitulates many aspects of bmm mRNA regulation (Men et al., 2016). GFP expression was present in the germline of bmm-GFP testes, and we found germline GFP levels were higher in spermatocytes than at earlier stages of sperm development (Figure 2A, B; one-way analysis of variance [ANOVA] with Tukey multiple comparison test). In further support of this observation, we analyzed a publicly available single-cell RNA sequencing dataset from the male reproductive organ (Li et al., 2021). Using pseudotime analysis, we arranged the germline (Figure 2—figure supplement 1A) and somatic cells (Figure 2—figure supplement 1B) based on their annotated developmental trajectory. The expression pattern of bmm in the germline matched our observation with the bmm-GFP reporter (Figure 2—figure supplement 1C). While levels of the bmm-GFP reporter were lower in somatic cells, single-cell RNA sequencing data identified bmm expression in the somatic lineage that was higher in cells at later stages of development (Figure 2—figure supplement 1D). Additional neutral lipid- and lipid droplet-associated genes such as lipid storage droplet-2, Seipin, Lipin, and midway also showed differential regulation during differentiation in the testis (Figure 2—figure supplement 1C, D). Combined with our data on the location of testis LD, these gene expression data suggest that bmm upregulation in both somatic and germline cells during differentiation corresponds to the downregulation of testis LD. Supporting this, germline GFP levels were negatively correlated with testis LD in bmm-GFP flies (Figure 2A, C), suggesting regions with higher bmm expression had fewer LD.

To test whether bmm regulates testis LD, we compared LD in testes from 0-day-old males carrying a loss-of-function mutation in bmm (bmm¹) to control male testes (bmm^rev) (Grönke et al., 2005). bmm¹ males had significantly more LD across all LD sizes compared with control males at the apical tip of the testis (Figure 2D–G; Welch two-sample t-test with Bonferroni correction) and showed a significantly expanded LD distribution (Figure 2D–F, H; two-sample Kolmogorov–Smirnov test). This suggests bmm normally restricts LD to the apical tip of the testis, an observation we confirmed in both somatic and germline lineages using lineage-specific expression of GFP-LD (Figure 2—figure supplement 2A–D). Importantly, after inducing homozygous bmm^rev or bmm¹ clones in the testes using the FLP-FRT system (Figure 3A, B; Xu and Rubin, 1993), we found bmm¹ spermatocyte clones had significantly more LD at 3 days post-clone induction (Figure 3C; Welch two-sample t-test), a stage at which LD were absent from bmm^rev clones. Because we observed no significant effect of cell-autonomous bmm loss on LD at any other stage of germline development (Figure 3C), this suggests bmm function is not required to regulate LD at early stages of germ cell development. Instead, our data suggest bmm plays a role in regulating LD at the spermatogonia–spermatocyte transition. While we were unable to assess LD in bmm¹ somatic clones, our data reveal a previously unrecognized cell-autonomous role for bmm as a regulator of LD in germline cells.

To determine the physiological significance of bmm-mediated regulation of testis LD, we investigated testis and sperm development in males without bmm function. In 0-day-old bmm¹ males reared at 25°C, testis size was significantly smaller than in age-matched bmm^rev controls (Figure 4A, B; Welch two-sample t-test), and the number of spermatid bundles was significantly lower (Figure 4C; Kruskal–Wallis rank sum test). When the animals were reared at 29°C, a temperature that exacerbates spermatogenesis defects associated with changes in lipid metabolism (Ben-David et al., 2015), bmm¹ phenotypes were more pronounced (Figure 4—figure supplement 1A, B; Welch two-sample t-test, Kruskal–Wallis rank sum test). Defects in testis size were also observed at 14 days post-eclosion; suggesting testis size defects persist later into the life course (Figure 4—figure supplement 1C; Welch two-sample t-test). In contrast, the number of spermatid bundles per testis was not significantly different between bmm¹ and bmm^rev males at this age (Figure 4—figure supplement 1D; Welch two-sample t-test), potentially due to a large decrease in the number of spermatid bundles in 14-day-old bmm^rev males (Figure 4C, Figure 4—figure supplement 1D). Together, these data suggest loss of bmm affects testis development and spermatogenesis. Similar phenotypes are observed in male mice without ATGL (Masaki et al., 2017), and supplementing the diet of bmm¹ males with medium-chain triglycerides partially rescued the testis and spermatogenic defects we observed in flies (Figure 4—figure supplement 1E, F; one-way ANOVA with Tukey multiple comparison test), as it does in mice (Masaki et al., 2017; Kim et al., 2017). This identifies similarities between flies and mice in fertility-related phenotypes associated with whole-body loss of bmm/ATGL.

To explore spermatogenesis in bmm¹ animals, we used an antibody against the germline cell-specific marker Vasa to visualize the germline in the testes of bmm¹ and bmm^rev males (Figure 4D, E; Lasko and Ashburner, 1988). We observed a significant increase in the number of GSCs (Figure 4F; Kruskal–Wallis rank sum test) and higher variability in GSC number in bmm¹ males (p = 5.7 × 10⁻¹² by F-test). Given that GSC number is affected by hub size and GSC proliferation (Resende et al., 2013; Kiger et al., 2000), we monitored both parameters in bmm¹ and bmm^rev controls. While hub size in bmm¹ testes was significantly larger than in testes from bmm^rev controls (Figure 4—figure supplement 1G, H; Welch two-sample t-test), the number of phosphohistone H3-positive GSCs, which indicates proliferating GSCs, was unchanged in bmm¹ animals (Figure 4—figure supplement 1I; Kruskal–Wallis rank sum test). While this indicates a larger hub may partly explain bmm’s effect on GSC number, bmm also plays a cell-autonomous role in regulating GSCs, as we recovered a higher proportion of bmm¹ clones in the GSC pool compared with bmm^rev clones at 14 days after clone induction (Figure 4G; Welch two-sample t-test). Given that we detected no effect of cell-autonomous bmm loss on the number of GSC LD (Figure 3C), more work will be needed to understand how bmm regulates GSCs at a stage prior to its effects on LD number. Future studies will also need to confirm whether bmm¹ mutant GSCs show an increased ability to occupy space at the hub.

Beyond GSCs, we uncovered additional spermatogenesis defects in bmm¹ testes. Peak Bam-GFP expression in germline cells of the testes from 0-day-old bmm¹ and bmm^rev males showed GFP-positive cysts were significantly further away from the hub in bmm¹ testes (Figure 4H, Figure 4—figure supplement 1J; Welch two-sample t-test). Indeed, 15/18 bmm¹ testes contained Vasa-positive cysts with large nuclei in the distal half of the testis (Figure 4I, arrows), a phenotype not present in bmm^rev testes (0/8) (p = 0.0005 by Pearson’s Chi-square test). Because these phenotypes are also seen in testes with differentiation defects (Fairchild et al., 2017; Lin et al., 1996), we recorded the stage of sperm development reached by the germline in bmm¹ testes. Most bmm¹ testes contained post-meiotic cells in males raised at 25°C (Figure 4—figure supplement 1K); however, germline development did not progress past the spermatocyte stage in most bmm¹ testes from animals raised at 29°C (Figure 4—figure supplement 1K). Testes from bmm¹ males reared at 25°C also had a smaller Boule-positive area (Figure 4J, Figure 4—figure supplement 1L; Welch two-sample t-test) and fewer individualization complexes and waste bags (Figure 4—figure supplement 1M, N; Kruskal–Wallis rank sum test). Because Boule-positive area, individualization complexes, and waste bags are all markers for later stages in sperm development, these data indicate that loss of bmm caused a reduction in differentiated cell types. Because we observed significantly fewer bmm¹ spermatocyte and spermatid clones at 14 days after clone induction (Figure 4K, L; p = 0.0496, Kruskal–Wallis rank sum test), these effects on germline development may represent a cell-autonomous role for bmm in regulating spermatogenesis in this cell type. Given that the statistical significance of this finding was not as strong as for our other data, future studies should repeat this experiment with more samples. We also reveal a potential non-cell-autonomous role for somatic bmm. While there was no difference in the ratio of Zfh-1-positive cells between homozygous clones and heterozygous clones in animals carrying the bmm¹ or bmm^rev alleles at 14 days post-clone induction (Figure 4—figure supplement 1O; Kruskal–Wallis rank sum test), the distance from the hub to the Zfh-1-positive clones was significantly decreased in bmm¹ homozygous clones (Figure 4—figure supplement 1P; Kruskal–Wallis rank sum test). Together, these data indicate bmm may play a cell-autonomous role in germline cells, and potentially a non-cell-autonomous role in somatic cells, to regulate spermatogenesis.

ATGL catalyzes the first and rate-limiting step of triglyceride hydrolysis (Zimmermann et al., 2004; Eichmann et al., 2012; Schweiger et al., 2006). Loss of this enzyme or its homologs leads to excess triglyceride accumulation (Wat et al., 2020; Grönke et al., 2007; Grönke et al., 2005; Zimmermann et al., 2004; Lee et al., 2014) and shifts in multiple lipid classes (Chitraju et al., 2013; Missaglia et al., 2017; Williams et al., 1991; Yang et al., 2020). To determine how loss of bmm affects spermatogenesis, we carried out whole-body mass spectrometry (MS)-based untargeted lipidomic profiling of bmm¹ and bmm^rev males. Hierarchical clustering of lipid species suggests that bmm¹ and bmm^rev males show distinct lipidomic profiles (Figure 5A). Overall, we detected 2464 and 1144 lipid features with high quantitative confidence in positive and negative ion modes, respectively. By matching experimental m/z, isotopic ratio, and tandem MS spectra to lipid libraries, we confirmed 293 unique lipid species (Supplementary file 5). We found 107 lipids had a significant change in abundance between bmm¹ and bmm^rev males (p_adj < 0.05): 85 species were upregulated in bmm¹ males and 22 lipid species were downregulated. Among differentially regulated species from different lipid classes, triglyceride had the largest residual above expected proportion (p = 5.00 × 10⁻⁴ by Pearson’s Chi-squared test). This suggests triglyceride was the lipid class most affected by loss of bmm (Figure 5B, C).

In bmm¹ males, the majority of triglyceride species (55/97) were significantly higher in abundance compared with bmm^rev control males. Because we observed a positive correlation between the fold increase in triglyceride abundance with both the number of double bonds (p = 7.52 × 10⁻⁸ by Kendall’s rank correlation test; Figure 5—figure supplement 1A) and the number of carbons (p = 2.77 × 10⁻¹⁰ by Kendall’s rank correlation test; Figure 5—figure supplement 1B), our data align well with bmm/ATGL’s known role in regulating triglyceride levels (Grönke et al., 2005; Haemmerle et al., 2006; Zimmermann et al., 2004) and its substrate preference of long-chain polyunsaturated fatty acids (Eichmann et al., 2012). While we also detected changes in species such as fatty acids, acylcarnitine, and membrane lipids (Figure 5—figure supplement 1C–H), in line with recent Drosophila lipidomic data (Nazario-Yepiz et al., 2021; Giedt et al., 2021), the striking accumulation of triglyceride in bmm¹ males suggested that excess testis triglyceride in bmm¹ males may contribute to their spermatogenic defects. To test this, we examined spermatogenesis in bmm¹ males carrying loss-of-function mutations in midway (mdy). mdy is the Drosophila homolog of diacylglycerol O-acyltransferase 1 (DGAT1), and whole-body loss of mdy reduces whole-body triglyceride levels (Beller et al., 2010; Buszczak et al., 2002; Martínez et al., 2020). Importantly, testes isolated from males with global loss of both bmm and mdy (mdy^QX25/k03902;bmm¹) had fewer LD than testes dissected from bmm¹ males (Figure 5D, Figure 5—figure supplement 1I; one-way ANOVA with Tukey multiple comparison test).

We found that testes isolated from mdy^QX25/k03902;bmm¹ males were significantly larger and had more spermatid bundles than testes from bmm¹ males (Figure 5E–G; one-way ANOVA with Tukey multiple comparison test). The elevated number of GSCs in bmm¹ male testes was similarly rescued in mdy^QX25/k03902;bmm¹ males (Figure 5H; one-way ANOVA with Tukey multiple comparison test). These data suggest that defective spermatogenesis in bmm¹ males can be partly attributed to excess triglyceride accumulation. Notably, at least some of the effects of global mdy loss on bmm¹ males can be attributed to the germline: RNAi-mediated knockdown of mdy in the germline of bmm¹ males partially rescued the defects in testis size (Figure 5I; Kruskal–Wallis rank sum test with Dunn’s multiple comparison test) and GSC variance (Figure 5—figure supplement 1J; p = 4.5 × 10⁻⁵ and 8.2 × 10⁻³ by F-test from the GAL4- and UAS-only crosses, respectively). While future studies will need to test whether germline-specific loss of mdy also rescues spermatid number defects in bmm¹ males, our data suggest bmm-mediated regulation of testis triglyceride plays a previously unrecognized role in regulating sperm development.

Drosophila strains and their source are listed in the Key Resources table. Further information and requests for resources and reagents should be directed to, and will be fulfilled by, lead contact Dr. Elizabeth J. Rideout (elizabeth.rideout@ubc.ca).

Fly stocks were maintained at room temperature in 12:12 hr light:dark cycle. Unless otherwise indicated, all flies were raised at 25°C with a density of 50 larvae per 10 ml fly media. Because this project examines sperm development, we used male flies in all experiments. Fly media contained 20.5 g sucrose (SU10, Snow Cap), 70.9 g Dextrose (SUG8, Snow Cap), 48.5 g cornmeal (AO18006, Snow Cap), 30.3 g baker’s yeast (NB10, Snow Cap), 4.55 g agar (DR-820-25 F, SciMart), 0.5 g calcium chloride dihydrate (CCL302.1, BioShop Canada), 0.5 g magnesium sulfate heptahydrate (MAG511.1, BioShop Canada), 4.9 ml propionic acids (P1386, Sigma-Aldrich), and 488 μl phosphoric acid (P5811, Sigma-Aldrich) per 1 l of media. For diets with medium- or long-chain triglyceride, 4 g of coconut oil (medium-chain triglyceride) or olive oil (long-chain triglyceride) was added per 100 ml of media described above prior to cooling. Males were collected and dissected within 24 hr of eclosion unless otherwise indicated. Fixations were performed at room temperature with 4% paraformaldehyde (CA11021-168, VWR) in phosphate-buffered saline (PBS) for 20 min on a rotating platform followed by washing in PBS twice before staining. Fly strains used in our study are listed in a Key Resources table, and fly strains prepared in this study were made using standard Drosophila genetic crossing techniques.

Cells at an early stage of development (stem cells and early-stage somatic and germline cells) were located in the apical region of the testis, and were identified by their small and dense nuclei (White-Cooper, 2004). GSCs were defined as Vasa-positive cells in direct contact with the hub; proliferating GSCs were identified as Vasa-positive cells in direct contact with the hub that were also phospho-H3 positive. Cells in the testis region occupied by primary spermatocytes were identified by their large cell size and decondensed chromosome staining occupying three nuclear domains (White-Cooper, 2004). Spermatid bundles were identified by their condensed and needle-shaped nuclei, which roughly corresponds to nuclei with protamine-based chromatin (Fabian and Brill, 2012). The hub was identified as the FasIII-positive area of the testis. Hub size was estimated by measuring the FasIII-positive area in a Z-projected image of the hub in each testis. Z-projections were made using the ‘sum slices’ function in Fiji. Testis size was measured by quantifying the length of a line drawn down the middle of a testis image; starting from the apical tip of the testis and ending where the testis meets the seminal vesicle.

Adult males were collected at 3–5 days post-eclosion and heat shocked three times at 37°C for 30 min followed by a 10-min rest period at room temperature between heat shocks. After heat shock, the flies were incubated at room temperature until dissection.

Fixed samples were rinsed three times with blocking solution containing 0.2% bovine serum albumin (A4503, Sigma-Aldrich), 0.3% Triton-X in PBS, then blocked for 1 hr on a rotating platform at room temperature. During the incubation, the blocking solution was changed every 15 min. After blocking, the sample was resuspended in blocking solution with the appropriate concentration of primary antibody (see Key Resources table), and incubated overnight at 4°C. Samples were rinsed three times with blocking solution after removing primary antibody, and blocked for 1 hr on a rotating platform in blocking solution. Secondary antibody was applied in blocking solution and left on the rotating platform at room temperature for 40 min. The sample was rinsed with blocking solution three more times, and washed four times for 15 min per wash in blocking solution. Testis samples were resuspended in Vectashield mounting media with DAPI (H-1200-10, Vector Laboratory) or SlowFade Diamond mounting media (S36972, Thermo Fisher Scientific) prior to mounting.

Fixed testes were briefly permeabilized with 0.1% Triton-X in PBS for 5 min prior to applying phalloidin. For BODIPY (4,4-Difluoro-1,3,5,7,8-Pentamethyl-4-Bora-3a,4a-Diaza-s-Indacene) staining, samples were suspended in PBS containing 10 μg/ml DAPI (2879083-5 mg, PeproTech), 1:500 BODIPY 495/503 (Thermo Fisher Scientific D3922), and 1:1000 phalloidin iFluor647 (ab176759, Abcam) or 1:40 phalloidin TexasRed (T7471, Thermo Fisher Scientific). For staining with LipidTox Red, samples were suspended in PBS containing 10 μg/ml DAPI (2879083-5 mg, PeproTech), 1:200 LipidTox Red (H34476, Thermo Fisher Scientific), and 1:1000 phalloidin iFluor647 (ab176759, Abcam). For staining free sterols, samples were prepared as for BODIPY staining with 50 μg/ml filipin in place of BODIPY for 30 min. Samples were incubated on a rotating platform for 40 min at room temperature. After incubation, samples were washed twice with PBS, then resuspended in SlowFade Diamond mounting media (Thermo Fisher Scientific S36972) prior to mounting.

All images were acquired on a Leica SP5 confocal microscope system with ×20 or ×40 objectives and quantified with Fiji image analysis software (Schindelin et al., 2012).

Drosophila extracts were prepared following the previously reported protocol (Yu et al., 2020). Briefly, 10 Drosophila males (~10 mg) were weighed, 300 µl of ice-cold methanol/water mixture (9:1, vol:vol) was added to these males, and the samples were homogenized with glass beads using a bead beater (mini-beadbeater-16, BioSpec, Bartlesville, OK, USA). Sample weight was used for sample normalization. Fly lysate was kept at −20°C for 4 hr for protein precipitation. Then, 900 µl of methyl tert-butyl ether was added and the solution was shaken for 5 min to extract lipids. To induce phase separation 285 µl of water was added, followed by centrifugation. The upper layer was separated, dried, and reconstituted in isopropanol/acetonitrile (1:1, vol:vol) for liquid chromatography (LC)–MS analysis. The volume of reconstitution solution was proportional to sample weight for normalization. Quality control (QC) samples were prepared by pooling 20 μl aliquot from each sample. The method blank sample was prepared using an identical workflow but without adding Drosophila.

Drosophila extracts were analyzed on an UHR-QqTOF (Ultra-High Resolution Qq-Time-Of-Flight) mass spectrometry Impact II (Bruker Daltonics, Bremen, Germany) interfaced with an Agilent 1290 Infinity II LC Systems (Agilent Technologies, Santa Clara, CA, USA). LC separation was performed using a Waters reversed-phase (RP) UPLC Acquity BEH C18 Column (1.7 µm, 1.0 mm × 100 mm, 130 Å) (Milford, MA, USA) maintained at 30°C. For positive ion mode, the mobile phase A was 60% acetonitrile in water and the mobile phase B was 90% isopropanol in acetonitrile, both containing 5 mM ammonium formate (pH = 4.8, adjusted by formic acid). For negative ion mode, the mobile phase A was 60% acetonitrile in water and the mobile phase B was 90% isopropanol in acetonitrile, both containing 5 mM NH₄FA (pH = 9.8, adjusted by ammonium hydroxide). The LC gradient for positive and negative ion modes was set as follows: 0 min, 5% B; 8 min, 40% B; 14 min, 70% B; 20 min, 95% B; 23 min, 95% B; 24 min, 5% B; 33 min, 5% B. The flow rate was 0.1 ml/min. The injection volume was optimized to 2 µl in positive mode and 5 µl in negative mode using QC sample. The electrospray ionization (ESI) source conditions were set as follows: dry gas temperature, 220°C; dry gas flow, 7 l/min; nebulizer gas pressure, 1.6 bar; capillary voltage, 4500 V for positive mode, and 3000 V for negative mode. The MS1 analysis was conducted using following parameters: mass range, 70–1000 m/z; spectrum type: centroid, calculated using maximum intensity; absolute intensity threshold: 250. Data-dependent MS/MS analysis parameters: collision energy: 16–30 eV; cycle time, 3 s; spectra rate: 4 Hz when intensity <10⁴ and 12 Hz when intensity >10⁵, linearly increased from 10⁴ to 10⁵. External calibration was applied using sodium formate to ensure the m/z accuracy before sample analysis.

The raw LC–MS data were processed using MS-DIAL (ver. 4.38) (Tsugawa et al., 2015). The detailed MS-DIAL parameters are: MS1 tolerance, 0.01 Da; MS/MS tolerance, 0.05; mass slice width, 0.05 Da; smoothing method, linear weighted moving average; smoothing level, 3 scans; minimum peak width, 5 scans. Lipid features with high quantitative confidence were selected by the following criteria: retention time was within the gradient elution time (<23 min); average intensity in QC samples is larger than fivefold of the intensity in method blank sample. Lipid identification was performed by matching experimental precursor m/z, isotopic ratio and MS/MS spectrum against the LipidBlast libraries embedded in MS-DIAL. To improve the quantification accuracy, the measured MS signal intensities were corrected using serial diluted QC samples following the reported workflow (Yu and Huan, 2021).

All microscopy images were quantified using Fiji software (Schindelin et al., 2012). For lipid droplet counts, a single optical slice through the middle of the testis containing the hub was used with the exception of FLP-FRT experiment where all LD within a GFP-negative cyst were counted (Figure 2I). All statistical analyses were done using R (obtained from https://cran.r-project.org). With exception of data concerning spatial distribution, and lipidomic data, Shapiro–Wilk test (via shapiro.test in base R) was used to assess normality of distribution prior to testing for significance. Kruskal–Wallis rank sum test (from the R package coin) and Dunn’s test (from the R package dunn.test) were used in place of Welch two-sample t-test and Tukey’s multiple comparison test when the assumption of normality was not met. For testing differences in variance between two populations, F-test (via var.test in base R) was used. For testing differences in spatial distribution, two-sample Kolmogorov–Smirnov test (via ks.test in base R) was used. All p-values are indicated in figures; extremely small p-values are listed as p < 2.2 × 10⁻¹⁶.