An important role for triglyceride in regulating spermatogenesis

Abstract

Drosophila is a powerful model to study how lipids affect spermatogenesis. Yet, the contribution of neutral lipids, a major lipid group which resides in organelles called lipid droplets (LD), to sperm development is largely unknown. Emerging evidence suggests LD are present in the testis and that loss of neutral lipid- and LD-associated genes causes subfertility; however, key regulators of testis neutral lipids and LD remain unclear. Here, we show LD are present in early-stage somatic and germline cells within the Drosophila testis. We identified a role for triglyceride lipase brummer (bmm) in regulating testis LD, and found that whole-body loss of bmm leads to defects in sperm development. Importantly, these represent cell-autonomous roles for bmm in regulating testis LD and spermatogenesis. Because lipidomic analysis of bmm mutants revealed excess triglyceride accumulation, and spermatogenic defects in bmm mutants were rescued by genetically blocking triglyceride synthesis, our data suggest that bmm-mediated regulation of triglyceride influences sperm development. This identifies triglyceride as an important neutral lipid that contributes to Drosophila sperm development, and reveals a key role for bmm in regulating testis triglyceride levels during spermatogenesis.

Keywords: ATGL; D. melanogaster; adipose triglyceride lipase; brummer; developmental biology; lipid droplet; spermatogenesis; testis; triglyceride.

Figure 1.. Lipid droplets (LD) are present in early-stage somatic and germ cells.

(A) Testis LD in w¹¹¹⁸ animals visualized with neutral lipid dye BODIPY. (A, A') Scale bar = 50 μm; (A'', A''') scale bar = 15 μm. Asterisk indicates hub in all images. Arrows point to LD; arrowheads point to spermatocytes in A, B. Spermatocytes were identified as described in methods section. (B) Testis LD visualized with BODIPY in newly eclosed males from two wild-type genotypes. Scale bars: main image = 50 μm; inset image = 10 μm. (C) Testis LD from w¹¹¹⁸ animals at different times post-eclosion. Scale bars = 50 μm. (D) Testis LD visualized with LipidTox Red in animals with somatic cell overexpression of GFP-LD (Tj-GAL4>UAS-GFP-LD). Green fluorescent protein (GFP)- and LipidTox Red-positive punctae are somatic LD (D–D'' arrows); LipidTox punctae without GFP indicate germline LD (D–D'' arrowheads). Scale bars = 10 μm. (E) Histogram showing the spatial distribution of somatic cell LD; error bars represent standard error of the mean (SEM). (F) Cumulative frequency distributions of somatic LD (blue line, data reproduced from E), zfh-1-positive somatic cells (zfh-1⁺ cells, orange line), and Eya-positive somatic cells (Eya⁺ cells, gray line). (G) Testis LD visualized with LipidTox Red in males with germline overexpression of GFP-LD (nos-GAL4>UAS-GFP-LD). GFP- and LipidTox Red-positive punctae indicate germline LD (arrows); LipidTox punctae without GFP indicate non-germline LD (arrowheads). Scale bars = 10 μm. (H) Histogram representing the spatial distribution of LD within the germline; error bars represent SEM. (I) Histogram representing the spatial distribution of LD and GFP fluorescence (green line) (arbitrary units, a.u.) in a representative testis of a bam-GFP animal (panel J). (J) Testis LD in a bam-GFP animal; arrows point to LD and arrowheads point to spermatocytes. Scale bar = 50 μm. See also Figure 1—figure supplement 1.

Figure 1—figure supplement 1.. Cholesterol is absent from testis lipid droplets.

(A) Testes stained with BODIPY (A) to detect neutral lipids and Filipin III (A') to detect free cholesterol. Scale bars = 50 μm.

Figure 2.. brummer regulates lipid droplets (LD) in both germline and somatic cells of the testis.

(A–A'''') Testis LD indicated by LipidTox Red in bmm-GFP animals. Arrows point to LD in all images. Arrowheads point to spermatocytes. Scale bars = 50 μm. Asterisks indicate the hub in all images. (B) Quantification of nuclear GFP intensity in testes isolated from bmm-GFP animals (n = 3). Germline stem cell (GSC), spermatogonia (SG), spermatocyte (SC). (C) Spatial distribution of LD (gray histogram) and GFP expression (green line) in testes from bmm-GFP animals as a function of distance from the hub (n = 3). LD near the apical region of the testis in bmm^rev (D) or bmm¹ (E) animals. (F) LD further away from the apical tip in bmm¹animals. (D–F) Scale bars = 50 μm. (G) Histogram representing testis LD size distribution in bmm^rev (gray) and bmm¹ (orange). (B,C,G) Error bars represent standard error of the mean (SEM). (H) Apical tip of the testes is at the left of the graph; individual dots represent a single LD and its relative position to the hub marked by an asterisk. Cumulative frequency distribution of the distance between LD and the apical tip of the testes are drawn as solid lines. See also Figure 2—figure supplement 1 and Figure 2—figure supplement 2.

Figure 2—figure supplement 1.. Expression of brummer and selected lipid metabolic genes during spermatogenesis in germline and somatic lineages.

(A) Pseudotime trajectory of germline (black line) based on single-cell RNA sequencing data (Li et al., 2021). Individual cells are labeled according to the annotation within the dataset. (B) Pseudotime trajectory of the somatic cells (black line) based on publicly available single-cell RNA sequencing data (Li et al., 2021). Individual cells are labeled according to the annotation within the dataset. Only one trajectory (branch b) is marked and used for panel D. (C) Rolling average of normalized transcript counts in the germline along the trajectory shown in panel A are plotted as a black line on the upper panel. Composition of cell types mapped on to the trajectory at each time point is shown at the bottom of panel B. (D) Rolling average of normalized transcript counts in somatic cells plotted as a black line along the trajectory shown in C (upper panel). Composition of cell types mapped on to the trajectory at each time point is shown at the bottom of panel D.

Figure 2—figure supplement 2.. brummer regulates lipid droplets (LD) in both germline and somatic cells of the testis.

Representative images of bmm^rev (A and B) and bmm¹ (C and D) testes with somatic overexpression of GFP-LD (Tj-GAL4>UAS-GFP-LD). Panels B and D contain magnified images of the area indicated by the boxes in panels A and C, respectively. In bmm^rev testes, LD were restricted to a region near the apical tip (A) of the testis in both somatic (B–B''', arrows) and germline cells (B–B''', arrowheads). In bmm¹ testes, LD were present in both somatic (C, D, arrows) and germline cells (C, D, arrowheads), near the apical tip of the testis in a region corresponding to early-stage germ cells and in the region corresponding to spermatocytes. (A, C) Scale bars = 50 μm; (B, D) scale bars = 20 μm.

Figure 3.. bmm regulates germline lipid droplets (LD) in a cell-autonomous manner.

(A, B) Single confocal slices through a representative testis isolated from an individual carrying clones induced using the FLP-FRT system at 3 days post-clone induction. Clones are homozygous for an allele that encodes a functional bmm protein product (bmm^rev; A–A''') or a loss-of-function bmm allele (bmm¹, B–B'''). GFP-negative areas mark homozygous clones in panels A and B; the boxed areas in A, A'' and B, B'' are shown in A', A''' and B', B''', respectively. In homozygous bmm^rev spermatocyte clones we detected no LD using neutral lipid dye LipidTox (A'', A'''). In contrast, spermatocyte clones homozygous for bmm¹ have detectable LD (B'', B''', arrows). Scale bars = 50 μm in A, A'' and B, B''; scale bars = 10 μm in A', A''' and B', B'''. (C) Number of testis LD in bmm^rev (gray) or bmm¹ (orange) in FLP-FRT clones 3 days post-clone induction; dots represent measurements from a single clone. The number of cells in each cyst (CC) counted is indicated. There were significantly more LD in bmm¹ spermatocyte (SC) clones (p = 0.026; Welch two-sample t-test) but not at other stages of development. Error bars represent standard error of the mean (SEM).

Figure 4.. A cell-autonomous role for bmm in regulating spermatogenesis.

Testes isolated from bmm^rev (A) and bmm¹ (A') animals raised at 25°C stained with phalloidin. Scale bars = 100 μm. (B) Testis size in bmm¹ and bmm^rev animals raised at 25°C. (C) Spermatid bundle number in bmm¹ and bmm^rev testes from animals reared at 25°C. Representative images of bmm^rev (D) or bmm¹ (E) testes stained with 4′,6-diamidino-2-phenylindole (DAPI) and anti-Vasa antibody. Arrows indicate germline stem cells (GSCs). Scale bar = 50 μm. The hub is marked by an asterisk in all images. (F) GSC number in bmm¹ and bmm^rev testes. (G) Proportion of GSCs that were either bmm¹ or bmm^rev clones at 3 and 14 days post-clone induction. (H) Representative images of bmm^rev (H) and bmm¹ (H') testes carrying bam-GFP; data quantified in Figure 4—figure supplement 1J. Arrows indicate regions with high Bam-GFP. Scale bars = 50 μm. (I) Representative images of bmm^rev (I) or bmm¹ (I', I'') testes stained with anti-Vasa antibody. Arrows indicate Vasa-positive cysts in bmm¹ testis. Panel I'' is magnified from the boxed region in I'. (I, I') Scale bars = 100 μm; (I'') scale bar = 50 μm. Maximum projection of bmm^rev (J) or bmm¹ (J') testes stained with anti-Boule antibody (green) and DAPI (blue). Scale bars = 100 μm. Number of bmm¹ and bmm^rev spermatocyte clones (K) or post-meiotic clones (L) at 3 and 14 days post-clone induction. (B,C,F,G,K,L) Error bars indicate standard error of the mean (SEM). See also Figure 4—figure supplement 1.

Figure 4—figure supplement 1.. Additional characterization of testis development and spermatogenesis in animals lacking bmm.

(A) Testis size was smaller in bmm¹ mutant animals compared with bmm^rev controls at <24 hr post-eclosion when raised at 29°C (Welch two-sample t-test). (B) The number of spermatid bundles was significantly lower in bmm¹ mutant animals compared with bmm^rev controls at <24 hr post-eclosion when raised at 29°C (Kruskal–Wallis rank sum test). (C) Testis size was significantly smaller in bmm¹ mutant males compared with bmm^rev control males at 14 days post-eclosion (Welch two-sample t-test). (D) While the median number of spermatid bundles was not significantly different between bmm¹ mutant males and bmm^rev control males at 14 days post-eclosion (Welch two-sample t-test), 8/27 bmm¹ testis had no spermatid bundles, a phenotype absent in age-matched bmm^rev males (0/22) (p = 0.0163, Pearson’s Chi-squared test), suggesting a subtle defect is present. (E) Food supplemented with 4% medium-chain triglyceride (MCT), but not long-chain triglyceride (LCT), significantly increased testis length in bmm¹ animals but had no effect on this phenotype in bmm^rev control animals (one-way analysis of variance [ANOVA] with Tukey multiple comparison test). (F) Food supplemented with 4% medium-chain triglyceride significantly increased the number of spermatid bundles in bmm¹ testes but had no effect on this phenotype in bmm^rev control animals (one-way ANOVA with Tukey multiple comparison test). (G) Representative images of bmm^rev (G, G') or bmm¹ (G'', G''') testes stained for FasIII (G, G'') and Vas (G', G'''). Scale bars = 25 μm. (H) Quantification of hub area in bmm^rev or bmm¹ testes showed a significantly larger hub size in bmm¹ testes (Welch two-sample t-test). (I) The number of germline stem cells (GSCs) undergoing mitosis (phospho-histone H3⁺ GSC/total GSC) was not significantly different between bmm¹ and bmm^rev testes (Kruskal–Wallis rank sum test). (J) The distance between the hub and the first Bam-GFP-positive cyst (Figure 3H) was significantly higher in bmm¹ testes than in bmm^rev testes (Welch two-sample t-test). (K) All bmm^rev testes and most bmm¹ testes contained spermatids when raised at 25°C; however, the most advanced stage of spermatogenesis observed in the majority of bmm¹ testes isolated from animals reared at 29°C was the spermatocyte stage. (L) Testes isolated from bmm¹ animals showed a significantly smaller Boule-positive area than control testes (Welch two-sample t-test). (M) Testes isolated from bmm¹ animals contain fewer individualization complexes than bmm^rev control testes (Kruskal–Wallis rank sum test). (N) Fewer waste bags were present in testes isolated from bmm¹ animals compared with bmm^rev control testes (Kruskal–Wallis rank sum test). (O) Proportion of Zfh-1-positive clones that were homozygous for either bmm¹ or bmm^rev at 3 and 14 days post-clone induction (Kruskal–Wallis rank sum test). (P) The location of Zfh-1-positive clones homozygous for either bmm¹ or bmm^rev at 3 and 14 days post-clone induction measured as the distance from the center of the hub. Clones homozygous for bmm¹ locate significantly closer to the hub than the wild-type clones (Kruskal–Wallis rank sum test). Error bars indicate standard error of the mean (SEM).

Figure 5.. Loss of bmm disrupts triglyceride homeostasis and leads to spermatogenic defects.

(A) Hierarchical clustering of lipid species detected in bmm^rev and bmm¹ animals. (B) Histograms showing the proportion of significant species in each lipid class with different levels between bmm¹ and bmm^rev. Numbers on histograms indicate the number of species with differences in abundance. (C) Volcano plot showing fold change in abundance of triglyceride (green; 97 species) and non-triglyceride lipids (gray; 186 species) in our dataset. (D) Arrows indicate testis lipid droplets (LD) stained with LipidTox Red in bmm^rev (D), bmm¹ (D'), or mdy^QX25/k03902; bmm¹ (D'') animals. (E) Whole testes isolated from bmm^rev (E), bmm¹ (E'), or mdy^QX25/k03902;bmm¹ (E'') animals stained with anti-Vasa antibody (red) and DAPI (blue). Arrowheads indicate spermatid bundles. Scale bars = 100 μm. (F) Testis size in bmm^rev, bmm¹, and mdy^QX25/k03902;bmm¹ animals. Spermatid bundles (G) and number of germline stem cells (H) in bmm^rev, bmm¹, and mdy^QX25/k03902;bmm¹ animals. (I) Testis size in animals with germline-specific mdy knockdown (nos-GAL4>mdy RNAi; bmm¹) compared with controls (nos-GAL4>+; bmm¹ and +>mdy RNAi; bmm¹). Error bars indicate standard error of the mean (SEM). See also Figure 5—figure supplement 1.

Figure 5—figure supplement 1.. Lipidomic analysis of animals lacking bmm.

(A) Higher fold-changes of triglycerides in bmm¹ animals were associated with less saturation in the acyl-groups (Kendall’s rank correlation test). (B) Higher fold-changes of triglycerides in bmm¹ animals were associated with higher number of carbons in the acyl-groups (Kendall’s rank correlation test). Each dot represents a single triglyceride species for panels B and C. (C) Volcano plot of identified lipids; monoglycerides shown in blue and diglycerides shown in orange. Many monoglycerides and diglycerides show increase in fold-change in bmm¹ males. (D) The number of carbon and the degree of saturation of monoglycerides (MAG) and diglycerides (DAG) with significant changes in abundance between bmm¹ and bmm^rev males. (E) Volcano plot of identified lipids; fatty acids shown in magenta and acyl-carnitine shown in green. Many fatty acids show an increase in fold-change while many acyl-carnitines show a decrease in fold-change in bmm¹ males. (F) The number of carbon and the degree of saturation of fatty acids (FA) and acyl-carnitines (ACar) with significant changes in abundance between bmm¹ and bmm^rev males. (G) Volcano plot of identified lipids; membrane lipids shown in yellow. (H) The number of carbon and the degree of saturation of membrane lipids with significant changes in abundance between bmm¹ and bmm^rev males. For panels G and H, PC: phosphatidylcholine; PE: phosphatidylethanolamine; PI: phosphatidylinositol; LPC: lysophosphatidylcholine; LPE: lysophosphatidylethanolamine; SM: sphingomyelin; PG: phosphatidylglycerol. (I) Loss of mdy function rescued the elevated number of lipid droplets (LD) in bmm¹ testes to control levels (one-way analysis of variance [ANOVA] with Tukey multiple comparison test). (J) Germline-specific loss of mdy in bmm¹ animals did not reduce germline stem cell (GSC) numbers, but the variance in GSC numbers was significantly rescued (nos-GAL4>+; bmm¹ vs nos-GAL4>mdy RNAi; bmm¹: p = 4.5 × 10⁻⁵;+>mdy RNAi; bmm¹ vs nos-GAL4 >mdy RNAi; bmm¹: p = 0.0082 by F-test). Error bars indicate standard error of the mean (SEM).

Figure 6.. Model of bmm-mediated lipid droplet regulation in the Drosophila testis.

Schematic representation summarizing bmm-mediated lipid droplet regulation in the testis during development.