Peroxisomal DBP deficiency causes male infertility through disruption of lipid homeostasis in Drosophila

Abstract

Peroxisomes are essential organelles in almost all eukaryotic cells with a variety of functions in cellular metabolism, including β-oxidation of very-long-chain fatty acids (VLCFAs). Impairment in peroxisome biogenesis or single enzyme function often causes human metabolic diseases. Adult male patients with mutations in HSD17B4, which encodes for D-bifunctional protein (DBP), the central enzyme of peroxisomal β-oxidation pathway, present with infertility defects in addition to neurological abnormalities. However, the pathogenic mechanism for DBP deficiency associated male infertility is not known. Here, we employ a Drosophila DBP deficiency model combined with scRNA-seq analysis to examine the pathogenic mechanism of male infertility in DBP deficient animals. We show that loss of Mfe2, the Drosophila ortholog of human HSD17B4, results in male infertility. Further analysis reveals that Drosophila Mfe2 mutant males display several defects in the progression of spermatogenesis, including a delay of cell cycle progression in spermatogonia and spermatocytes as well as defective meiotic cytokinesis. Moreover, aberrant central spindle and contractile ring behaviors are observed in Mfe2 mutant spermatocytes. Lipidomic analysis shows that many lipid derivatives, including phospholipids and sterols, decrease in Mfe2 mutant testes. The levels of phosphatidylinositol 4,5-biphosphate (PI(4,5)P2) and phosphatidylinositol 3,4,5-triphosphate (PI(3,4,5)P3) are both reduced at the cleavage furrow as well as the membrane of Mfe2 mutant spermatocytes. Importantly, diets supplemented with Asolectin, a mixture of polyunsaturated phospholipids, partially rescue the spermatocyte cytokinesis defects in Mfe2 mutants. Together, these results demonstrate the critical roles of Mfe2 during Drosophila spermatogenesis and shed light on our understanding of the pathogenic mechanism of male infertility caused by peroxisomal DBP deficiency.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-026-06232-y.

Introduction

D-bifunctional protein (DBP) deficiency is an autosomal recessive genetic disorder caused by HSD17B4 gene mutation, with an incidence rate of approximately one in 100,000 [1]. HSD17B4, also known as DBP, plays an important role in β-oxidation of very-long-chain fatty acids (VLCFAs) in peroxisomes. The 2-enoylcoenzyme A (CoA) hydratase domain and 3-hydroxyacylCoA dehydrogenase domain of DBP catalyze the second (enoylCoA hydratase) and third steps (3-hydroxyacyl-CoA dehydrogenase) of β-oxidation [2]. DBP deficiency is classified into three categories based on the affected enzymatic activity: Type I lacks both hydratase and dehydrogenase activities, while Types II or III display isolated deficiencies in hydratase or dehydrogenase, respectively [3]. The vast majority of infants carrying type I deficiency die within one year after birth and those with type II or III deficiency often survive beyond age of 10 [4]. The prototypical clinical features of DBP deficiency include muscular hypotonia, craniofacial anomalies, hepatomegaly, and seizures [5]. Adult-onset patients experience infertility symptoms, such as amenorrhea in females and decreased testosterone level and azoospermia in male patients [4, 6]. Compound heterozygous mutations in the HSD17B4/DBP gene have been documented in adult males presenting with infertility [4, 7]. These cases typically involve one mutation that disrupts two protein domains, and a second, missense mutation specifically within the dehydrogenase domain, consistent with a Type III deficiency classification [4, 7]. Animal experiments in mice have shown that impairment of HSD17B4/DBP function leads to spermatogenesis defects and male infertility [8]. Nevertheless, the pathogenic mechanism of male infertility caused by DBP deficiency remains unclear.

Drosophila melanogaster is a valuable model organism in the study of human rare genetic diseases due to its high degree of genetic conservation with humans, sharing pathways and cellular processes that enable the translation of findings to human contexts [9]. In Drosophila, the sperm developmental stages are readily observable during spermatogenesis and diverse molecular genetic tools available enable detailed analysis of the underlying cellular processes [10]. Male germ cell differentiation in Drosophila is strikingly comparable to mammalian spermatogenesis at a broad level. The process in both flies and mammals begins with a relatively small population of self-renewing germline stem cells. These cells divide persistently to self-renew and give rise to spermatogonia cells. Following four-cycle mitotic amplification divisions, spermatogonia cells enter meiosis and differentiate into primary spermatocytes. Ultimately, post-meiotic spermatid differentiation marked by profound morphological changes remodels virtually all cellular structures [11].

In this study, we took advantage of a Drosophila DBP deficiency model, in which Mfe2, the Drosophila ortholog of DBP/HSD17B4, was largely deleted, and combined single-cell RNA-sequencing technology to explore the pathogenic mechanism of male infertility caused by DBP deficiency. Our study showed that Mfe2 mutant male displayed reduced fertility. The phenotype could be rescued by the wild type but not the peroxisome targeting signal peptide sequence deleted form of Mfe2. Analysis using scRNA-seq data combined with immunofluorescence staining results uncovered that the cell cycle progression of germ cells in Mfe2 mutants was impaired, and spermatocytes in Mfe2 mutants displayed meiotic cytokinesis defect. Lipidomic analysis revealed a marked increase in the levels of VLCFAs with carbon chain lengths exceeding 26 in the Mfe2 mutant testes. Conversely, many other lipid derivatives were significantly reduced in Mfe2 mutant testes. In addition, the levels of phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) and phosphatidylinositol 3,4,5-trisphosphate (PI(3,4,5)P3), two molecules involved in phosphoinositol signaling, were reduced on the membrane of spermatocytes in Mfe2 mutant. Adding Asolectin in the diets was able to partially suppress the meiotic cytokinesis defects. Collectively, our data reveal that Mfe2 disruption causes meiotic cytokinesis defects in spermatocytes by disrupting lipid homeostasis and provide novel insights into the pathogenic mechanisms underlying male infertility in patients with DBP deficiency.

Results

Loss of Mfe2 causes male infertility

To examine the role of Mfe2 in spermatogenesis and male fertility, we analyzed the Mfe2 loss-of-function mutant line, Mfe2, which was generated previously through the CRISPR/Cas9 approach. In the Mfe2 mutant, a 7 base-pair deletion at position c.39–45 of Mfe2 locus results in a frameshift mutation, leading to premature termination of translation and the formation of a truncated protein of 14 amino acids. The male fertility assay showed a dramatic reduction of fertility in Mfe2 mutant males compared to the control. Majority of control male flies (95%) were fertile while only 10% of Mfe2 mutant males were fertile (Fig. 1A). Introducing a genomic DNA fragment containing full Mfe2 gene was able to rescue the fertility defect in Mfe2 mutant males, indicating that reduction of fertility is specifically caused by Mfe2 mutation (Fig. 1A). Western blot assay using anti-Mfe2 antibody confirmed that Mfe2 protein was not present in Mfe2 mutant testes, and was restored in the rescued fly testes (Fig. 1B). We next dissected and examined the testes under bright-field microscope. Testes from control males showed a regular arrangement of multiple elongating spermatid bundles, while testes from Mfe2 mutant males contained few and sparse spermatid bundles (Fig. 1C). This defect was fully rescued by the transgene expressing Mfe2 (Fig. 1C). Statistical analysis showed that 80% of Mfe2 mutant testes displayed this abnormal morphology (Fig. 1D). Moreover, seminal vesicles were full of mature sperm in control and rescued male fly testes, no morphologically normal mature sperm was observed in Mfe2 mutant testes (Fig. 1E). Further immunofluorescence staining showed that the nuclei were scattered in the elongating spermatid cyst of Mfe2 mutant testes, while the nuclei of spermatids from the control and rescued flies at this stage were aligned in one direction (Fig. 1F), indicating that the spermatid polarity was defective in Mfe2 mutant testes. Anti-cleaved Caspase 3 antibody staining revealed that there were fewer cystic bulges and no waste bags in elongated spermatids undergoing individualization in Mfe2 mutant testes (Fig. 1F). The absence of these structures in Mfe2 mutants suggests that the spermatid individualization process fails to proceed normally.

Fig. 1.

Fertility, morphology and cytological analyses in Mfe2 mutants. (A) Fertility assay of w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male flies. The results were obtained from three independent experiments (n = 50 flies per group). The 1-way ANOVA multiple comparison was used to calculate P-values, where **** P ≤ 0.0001, and ns P > 0.05. (B) Protein level of Mfe2 in w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes. (C) The whole morphology of w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue fly male testes. White arrowhead indicates the distal end of the flagellum. (D) Statistics on the proportion of abnormal testes. Numbers at the bottom indicate the number of samples examined. (E) The seminal vesicle morphology of w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes. (F) Left: Early elongating spermatids of w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes were visualized with DAPI (DNA) and E7 antibody (anti-β-tub, mark early elongating spermatids). Right: Whole mount testis stained with anti-cleaved Caspase 3 (CC3) antibody. Cystic bulges and waste bags were indicated with arrows

Peroxisomal localization of Mfe2 is required for male fertility

We next examined the expression patterns of Mfe2 in the testes. Immunofluorescence staining revealed that Mfe2 was widely expressed in the testes (Fig. 2A). As Mfe2 is a peroxisomal β-oxidation enzyme and functions in the peroxisomes, we then wanted to determine whether it was co-localized with the peroxisomes. A Catalase reporter line (Cat-GFP) was used to mark the peroxisomes in the testes and as expected, Mfe2 and Cat-GFP were co-localized in the testes (Fig. 2B). Like most other enzymes that exert their functions within the peroxisomes, Mfe2 contains a peroxisomal targeting signal peptide (PTS1) at the C-terminus, which is required for the transportation into the peroxisome under the assistance of a cargo binding protein Pex5 [12] (Fig. 2C). The core of this PTS1 sequence for Mfe2 is a tripeptide (Alanine–Lysine–Leucine, AKL) [13]. To test whether peroxisomal localization of Mfe2 is required for its function in maintaining male fertility, we constructed two transgenic fly lines carrying UAS-Mfe2 CDS with (UAS-Mfe2) or without (UAS-Mfe2-ΔAKL) the core of PTS1 sequence. Tubulin-Gal4 was used to drive the expression of these transgenes. Effective expression of Mfe2 was confirmed by Western blot assay for both transgenes (Fig. 2D).

Fig. 2.

Peroxisomal localization of Mfe2 is required for male fertility. (A) Localization of Mfe2 in testes. Whole mount wild-type testis stained with anti-Mfe2 antibody. (B) Colocalization of Mfe2 and peroxisomes in testes. Whole mount wild-type testis stained with anti-GFP and anti-Mfe2 antibodies. Cat-GFP was used to label peroxisomes. (C) The schematic diagram showing the importation of Mfe2 into the peroxisome. The C-terminal of Mfe2 contains a signal peptide of PTS1 (Peroxisomal targeting signal type 1), which helps Mfe2 to enter the peroxisome under the synergistic effect of Pex5. (D) Protein levels of Mfe2 in testes dissected from w¹¹¹⁸, Mfe2 mutant, Tub-Gal4 > UAS-Mfe2 rescue and Tub-Gal4 > UAS-Mfe2-ΔAKL rescue male flies. (E) Fertility assay of w¹¹¹⁸, Mfe2 mutant, Tub-Gal4 > UAS-Mfe2-AKL rescue and Tub-Gal4 > UAS-Mfe2-ΔAKL rescue male flies. The results were obtained from three independent experiments (n = 50 flies per group). The 1-way ANOVA multiple comparison was used to calculate P-values, where **** P ≤ 0.0001, and ns P > 0.05. (F, G) The whole testis morphology (F) and seminal vesicle morphology (G) of testes dissected from w¹¹¹⁸, Mfe2 mutant, Tub-Gal4 > UAS-Mfe2 rescue and Tub-Gal4 > UAS-Mfe2-ΔAKL rescue male flies. (H) Localization of Mfe2 in testes dissected from Tub-Gal4 > UAS-Mfe2 rescue (upper) and Tub-Gal4 > UAS-Mfe2-ΔAKL rescue (lower) male flies. Whole mount testis stained with anti-Mfe2 and anti-GFP antibodies. Cat-GFP was used to label peroxisomes

Expression of UAS-Mfe2 with PTS1 was able to rescue the fertility of the majority of male Mfe2 mutants (approximately 60% of flies were fertile), while expression of UAS-Mfe2 without PTS1 failed to rescue the fertility defects (Fig. 2E). Microscopic analysis of dissected testes showed that the Mfe2 mutant teste defects including sparse spermatid bundles and empty seminal vesicle were rescued by UAS-Mfe2 with PTS1, but not UAS-Mfe2 without PTS1 (Fig. 2F and G). We also examined the cellular localization of these two different forms of Mfe2 and found that Mfe2 with PTS1 was co-localized with Cat-GFP in the peroxisomes and Mfe2 without PTS1 failed to do so, indicating that it was the peroxisomal localized Mfe2 that rescued phenotypes (Fig. 2H).

Given the peroxisomal import of Mfe2 is crucial in maintaining male fertility, we next asked whether impairment of peroxisomal biogenesis and function causes the similar fertility defects as in Mfe2 mutants. Pex12 acts with Pex2 and Pex10 to form a complex at the peroxisomal membrane and functions to facilitate peroxisomal protein import. Pex16 is likely required for peroxisome biogenesis and proliferation [12]. It has been shown that knockdown in either Pex12 or Pex16 gene leads to abnormal peroxisome assembly phenotypes in Drosophila S2 cells [14]. Drosophila loss-of-function mutant lines for Pex12 or Pex16 were generated with CRISPR/Cas9 method (Supplementary Fig. S1A-D). In the Pex12 mutant, a 12-base-pair fragment containing a stop codon TGA was inserted at the position 390 bases away from the ATG in the first exon of Pex12, resulting in premature termination of protein translation. In the Pex16 mutant, two base pairs were deleted at the position 61 bases away from the ATG in the first exon of Pex16, leading to a frameshift mutation and premature translation termination. Male fertility assay results showed that both Pex12 and Pex16 mutants had a much more severe phenotype than Mfe2 mutants with complete male sterility (Supplementary Fig. S1E). Moreover, testes from these two mutants were smaller than those from control and Mfe2 mutants as evidenced by the reduced testis length (Supplementary Fig. S1F and G). In addition, spermatid bundle was not present in both Pex12 and Pex16 mutants (Supplementary Fig. S1F). Similar to Mfe2 mutants, the seminal vesicles from Pex12 and Pex16 mutants were empty and lacked mature sperm (Supplementary Fig. S1H). Our data on the effects of Pex12 and Pex16 mutations on Drosophila male fertility are consistent with previous studies [15, 16]. Overall, these results suggest that the peroxisomal localization of Mfe2 and overall peroxisome function are required for male fertility.

Single-cell transcriptomes from wild-type and Mfe2 mutant testes cluster into distinct cell types

To identify the molecular changes across different cell types during the progression of spermatogenesis and investigate the underlying mechanism of male infertility in Mfe2 mutants, we performed single-cell RNA sequencing (scRNA-seq) on wild-type control and Mfe2 mutant testes. Testes were isolated from 40 wild-type control (w¹¹¹⁸) or Mfe2 mutant adult males, and single-cell RNA profiling was carried out using the 10x Genomics Chromium platform (Fig. 3A). After strict filtration, 24,627 high-quality cells from the two genotypes were used for subsequent bioinformatic analysis.

Fig. 3.

Identification and annotation of testis cell types by single-cell RNA-seq. (A) Schematic of the experimental workflow. (B) Cell type-annotated uniform manifold approximation and projection (UMAP) plot of the integrated w¹¹¹⁸ and Mfe2 mutant adult testis scRNA-seq data sets. (C) Dot plot displaying scaled average expression of somatic (upper panel) and germline (lower panel) cell type marker genes. (D) UMAP visualization of representative marker genes for somatic cells. (E) UMAP visualization of representative marker genes for germline cells

Using uniform manifold approximation and projection (UMAP) in Seurat, we reduced the dimension of the gene/cell expression matrix to two main axes and classified the cells based on their similarities in thousands of unique gene expression patterns. We used recently reported and classical marker genes to discern cell types [17–20]. For instance, Fas3 and aop are highly expressed in the hub cells [21, 22]. tj is highly expressed in cyst stem cells, while Wnt4 is highly expressed in early cyst cells [19]. nanos is highly expressed in early germ cells and has been used as a GSC marker [23]. bam is specially expressed in transit-amplifying (TA) spermatogonia and plays an important role in the switch from TA proliferation to meiotic differentiation [24]. Expression of Cyclin B (CycB) varies along mature spermatocytes. fzo and twe denote early and late spermatocytes, respectively [16, 25]. As a result, 14 cell populations were identified, including 8 somatic cell types (hub cells, cyst cells at each stage, epithelial cells, pigment cells and muscle cells) and 6 germline cell types (germline stem cells (GSCs), spermatogonia, spermatocytes, spermatids and their transitional period) (Fig. 3B and C). To accurately annotate the cell clusters, the expression patterns of multiple marker genes were exhibited (Fig. 3D and E).

Mfe2 mutant testes display abnormal germline cell proportion

We performed clustering and analyzed the proportions of distinct germ cell types for each of the two genotypes (wild-type control and Mfe2 mutant). Notably, late spermatocytes made up the predominant part in the Mfe2 mutant testes. (Fig. 4A and B). To verify this, we further performed immunofluorescence staining analysis using an antibody against Vasa, a germline cell specific DEAD-box helicase. The result revealed an aberrant accumulation of late spermatocytes, characterized by the extremely large cell size, in the testes from Mfe2 mutants compared to the control and rescued fly testes, corroborating with the cell ratio analysis (Fig. 4C and D). From the immunofluorescence results, we also noticed that there was an increased number of cells in the region of spermatogonia and GSCs, characterized by the increased area of the dense nuclear region in Mfe2 mutant testes (Fig. 4E and F). However, it is worth noting that the scRNA-seq data showed a decrease in the proportion of GSCs/spermatogonia in the testes from Mfe2 mutants (Fig. 4B). This might be due to the fact that the population of spermatocytes increased much more than that of GSCs/spermatogonia. To confirm the increase of spermatogonia, we stained the testes using an antibody against Bam, a protein required for the transition from proliferating spermatogonia to spermatocytes [26]. Bam is expressed in 4-to-16-cell spermatogonia and can act as a marker for these cells. We found that Bam-positive area in Mfe2 mutant testes was significantly larger than that in control fly testes, supporting the notion that the number of spermatogonia cell was increased in the Mfe2 mutants (Supplementary Fig. S2A-C). Consistent with this, bulk RNA-seq analysis showed that the expression of genes related to GSCs/spermatogonia characterization increased significantly in the Mfe2 mutant testes (Supplementary Fig. S2D). An obvious trend of enrichment in GSEA (Gene Set Enrichment Analysis) was observed (Supplementary Fig. S2E). In conclusion, these data indicate that the proportions of both late spermatocytes and GSCs/spermatogonia are increased in Mfe2 mutant testes.

Fig. 4.

Alteration in the proportion of different types of germline cells in the testes of Mfe2 mutants. (A) Separate UMAP plots from w¹¹¹⁸ and Mfe2 mutant adult testis scRNA-seq data sets. The germline cell types corresponding to different colors are marked below. Somatic cell clusters and unknown clusters are marked in light grey and dark grey respectively. (B) Cell ratio of germline cell types from (A). Colors denote different cell populations, as in (A). (C) Whole mount testes from two-day-old w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male flies immunostained for anti-Vasa antibody(germline), anti-Arm antibody (membranes) and DAPI (DNA). The white lines mark the length of the late spermatocyte region. (D) Statistics on the regional length of late spermatocytes for w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes. This measurement was first done as an absolute length, and the data showing in the graph was normalized to the mean value of the control group. Each dot represents an individual sample. Data are presented as mean ± SD (n = 3 biologically independent samples per group). The 1-way ANOVA multiple comparison was used to calculate P-values, where **** P ≤ 0.0001, and ns P > 0.05. (E) The apical area for w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes. The area highlighted by the white dotted line is the nuclei dense region containing germline stem cells and spermatogonia. (F) Statistics on the size of nuclei dense region for w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes, and the final data were normalized to the mean value of the control group. Each dot represents an individual sample. Data are presented as mean ± SD (n = 10 biologically independent samples per group). The 1-way ANOVA multiple comparison was used to calculate P-values, where **** P ≤ 0.0001, and ns P > 0.05

Loss of Mfe2 impairs cell cycle progression in spermatocytes and spermatogonia

During spermatogenesis, spermatocytes undergo a long G2 phase of the cell cycle, which increases the cell volume by 25-fold. Late spermatocytes are in the late G2 phase of the cell cycle and characterized by their large size [27]. Increased proportion of late spermatocytes in Mfe2 mutant testes indicates impairment of cell cycle progression in these cells. However, the basis for the increased number of GSCs/spermatogonia in Mfe2 mutant testes was unknown. We focused on this and tried to figure out what caused the increase in the number of GSCs/spermatogonia in Mfe2 mutant testes.

In the Drosophila testis, germline stem cells undergo asymmetric division to generate a daughter GSC and another differentiating daughter cell called gonialblast, which divides to form a cyst of spermatogonia [28]. As asymmetric division is critical for GSCs to maintain the balance between self-renewal and differentiation, impairment in GSC asymmetric division can cause an alteration in the number of GSCs/spermatogonia [29]. We first asked whether there was any defect in asymmetric division of germline stem cells. The proper orientation of centrosomes and spindles is a critical step during asymmetric division of GSCs [30]. We used γ-tubulin to mark centrosomes and analyzed the distribution of centrosomes during asymmetrical division in wild-type control and Mfe2 mutant germline stem cells. While misoriented centrosomes were present in a low percentage of GSCs in both the control and the Mfe2 mutant testes, there was no statistical difference between these two genotypes (Supplementary Fig. S3A and B). This indicates that there is no major defect in asymmetric division of germline stem cells in Mfe2 mutant testes. Consistent with this, there was no difference in the number of GSCs between the two genotypes, although it seemed to be more variable for Mfe2 mutants (Supplementary Fig. S3C and D). Based on this result, we conclude that the increase in the number of GSCs/spermatogonia in Mfe2 mutant testes is not due to overproliferation of germline stem cells.

In order to further address this issue, we extracted GSCs/spermatogonia information from scRNA-seq data of the two genotypes for further analysis. We re-clustered the GSCs/spermatogonia dataset to identify distinct states of spermatogonia. After re-clustering, these cells could be divided into four new clusters. Combined with the results of the pseudotime analysis and expression profile of nanos and bam (Supplementary Fig. S3E-G), we named these four groups as state0, state1, state2 and state3/4 according to the chronological order of their developmental trajectory (Fig. 5A and B). Based on bam expression, differences can be observed within state3/4, but there are no specific markers to distinguish them. We then conducted cell ratio analysis and found that the Mfe2 mutant spermatogonia had a significantly lower proportion of state 3/4 cells compared to the wild-type controls (Fig. 5C). This result suggested that spermatogonia in Mfe2 mutants might exhibit cell cycle progression defects. We further conducted GSVA (Gene Set Variation Analysis) on these states and evaluated the scores for S phase related gene sets such as base excision repair, nucleotide excision repair as well as regulation of DNA replication. We found that state3/4 cells had lower scores for these gene sets compared to other states (Fig. 5D). Moreover, GO analysis of the whole spermatogonia showed that the enriched pathways among upregulated genes in Mfe2 mutants included cell proliferation, cell cycle and DNA replication (Fig. 5E). Conversely, the enriched pathways among downregulated genes in Mfe2 mutants contained spermatogenesis, flagellum elongation and sperm motility (Fig. 5F).

Fig. 5.

Impairment of cell cycle progression in Mfe2 mutant GSCs/spermatogonia. (A)UMAP plot showing four distinct states of spermatogonia (merged germline stem cells and spermatogonia), each represented by a different color. (B) Violin plots visualizing selected gene marker expression in different spermatogonia states. (C) State ratio of state types from (A). Colors denote different cell states, as in (A). (D) The GSVA scores of classic S-phase-related pathways in the w¹¹¹⁸ and Mfe2 mutant spermatogonia in different states. (E) GO term enrichment of GSCs/spermatogonia in the Mfe2 mutant group in up-regulated genes compared with w¹¹¹⁸. (F) GO term enrichment of GSCs/spermatogonia in the Mfe2 mutant group in down-regulated genes compared with w¹¹¹⁸. (G, H) The apical area of testes labeled for EdU (cells in the proliferative phase) and DAPI (DNA) in w¹¹¹⁸ and Mfe2 mutant testes. The white dotted lines mark the nuclei dense region, and the red dotted lines mark the cells in S phase. (I) Statistics on the ratio of the EdU positive region area to the nuclei dense area of w¹¹¹⁸ and Mfe2 mutant testes. Each dot represents an individual sample. Data are presented as mean ± SD (n = 8 biologically independent samples per group). The P-value was calculated by Student’s t-test, where *** P ≤ 0.001

To further analyze cell cycle progression, we performed EdU incorporation to visualize DNA synthesis and label S phase cells. The results showed that the proportion of spermatogonia in S phase increased significantly in Mfe2 mutant testes compared to the controls (Fig. 5G-I). Increased S phase cells might reflect spermatogonia overproliferation. In wild-type testis, the gonialblast and its progeny undergo four rounds of mitotic divisions to generate a cyst of 16 interconnected spermatogonia before entering meiosis [10]. Spermatogonial overproliferation can lead to the generation of cell cysts with more than 16 germ cells, which has been detected in bam and bgcn mutant or Cbc knocking-down males [11, 31]. To test this, we used Armadillo (Arm) as a cell membrane marker to visualize each individual cyst. Cysts containing more than 16 germ cells were not observed in the nuclei dense area in Mfe2 mutant testes, indicating that loss of Mfe2 does not cause spermatogonia overproliferation (Supplementary Fig. S3H). These data confirm that more spermatogonia cells are in S phase of the cell cycle without overproliferation in Mfe2 mutants. Collectively, these findings indicate that Mfe2 depletion leads to germ cell accumulation and impairs cell cycle progression in both spermatocytes and spermatogonia.

Meiotic cytokinesis is defective in spermatocytes to Mfe2 mutants

Analysis of germ cell trajectories via conventional inference revealed dynamic expression of 341 differentially expressed genes (DEGs) in Mfe2 mutant testes, with the majority exhibiting strong downregulation during development from late spermatocytes onward, indicating a transcriptomic change at this stage in Mfe2 mutants (Fig. 6A).Through the GSVA of the late spermatocytes, we found that the scores of the pathways related to inflammation were increased, while the scores of the pathways related to cytokinesis and membrane fusion were significantly reduced, indicating a potential defect in Mfe2 mutant male meiotic cytokinesis (Fig. 6B).

Fig. 6.

Spermatocytes of Mfe2 mutants display meiotic cytokinesis defect. (A) Heatmap displays the smoothed expression patterns of DEGs (differentially expressed genes) (Wilcox test with Bonferroni correction, adjusted P-value < 0.01, absolute value of Log2 fold change > 1) between w¹¹¹⁸ and Mfe2 mutant testes, along the developmental trajectory from GSCs to late spermatids (bottom panel). Paired with this is the smoothed proportion of annotated cell types along the same trajectory (top panel). (B) Bar plot showing the exchanged GO terms of interest enriched in late spermatocytes in the Mfe2 mutant group compared with w¹¹¹⁸, as assessed by GSVA. (C) Schematic of morphological changes of mitochondria during the spermatocyte and round spermatid phases. Green represents mitochondria and blue represents the cell nucleus. Mitochondria are represented schematically and not to actual size. (D) Phase contrast microscopy of live testes to visualize round spermatids. Dark circles are mitochondrial derivatives, and light circles are nuclei. (E) Quantification of cytokinesis defects in w¹¹¹⁸, Mfe2 mutant and Mfe2 genomic rescue male fly testes. (F) Immunofluorescence staining to visualize round spermatids. ATP5A was used to mark mitochondria, DAPI was used to mark DNA (nuclei). (G) Maximum intensity projection of Z-stack images of late spermatocytes undergoing meiosis I cytokinesis. Feo-mCherry marked the central spindle, and Sqh-GFP-RLC marked the contractile ring. (H) Pattern diagrams of the meiosis period in w¹¹¹⁸ and Mfe2 mutant spermatocytes. (I) GSEA of spindle-related gene sets in late spermatocytes of Mfe2 mutant compared with w¹¹¹⁸. Each plot displays the running enrichment score (purple line) as it moves down the rank-ordered list of all genes (ranked by differential expression between genotypes). Vertical bars indicate the positions of genes belonging to the specified gene set. The names of three gene sets analyzed are shown on top: “spindle”, “spindle assembly”, and “spindle elongation”

To further examine the process of meiotic cytokinesis in the spermatocytes, we made live testis squash preparations and analyzed the phenotype under phase contrast microscopy. During spermatogenesis in the Drosophila testis, 16 late spermatocytes undergo two meiotic divisions to produce 64 interconnected spermatids [10]. During meiotic divisions, both chromosomes and mitochondria partition equally between two daughter cells and the mitochondria in each spermatid fuse to form a complex structure called the nebenkern at the end of meiosis. Each wild-type spermatid contains a single phase-light round nucleus and a phase-dark nebenkern of similar size and shape [32] (Fig. 6C). Defects in meiotic cytokinesis are revealed by assessing the pattern of nuclei and nebenkern in the round spermatids under phase contrast microscopy [33]. For instance, failure in cytokinesis but not in chromosome segregation during meiosis disrupts proper mitochondria partitioning, producing spermatids with a large nebenkern paired with two or four normal-sized nuclei. Analysis by phase contrast microscopy revealed that Mfe2 mutant testes contained spermatids with a large nebenkern and two or four normal-sized nuclei, while both wild-type control and rescued fly male spermatids display a nebenkern and a single nucleus of similar size (Fig. 6D). Quantification of this defect showed that the frequency of Mfe2 mutant spermatids showing abnormal nebenkern to nucleus ratio of 1:2 and 1:4 increased to 34.5% and 35.3% respectively, indicating cytokinesis defects in the meiotic divisions (Fig. 6E). Further immunofluorescence staining analysis with an antibody against ATP5A, which labels mitochondria, also showed that an abnormally large nebenkern surrounding by multiple nuclei appeared in the Mfe2 mutant round spermatids, but not in the control and rescued fly male spermatids, confirming failure of meiotic cytokinesis (Fig. 6F).

Having established that Mfe2 is required for meiotic cytokinesis in the spermatocytes, we next wanted to elucidate the underlying mechanisms. Similar to many other animal cells, the integrity of the central spindle and the contractile ring are essential for cytokinesis in the Drosophila spermatocyte meiotic cell divisions [34, 35]. To explore this, mCherry-tagged microtubule cross-linker Feo (Feo-mCherry) was used to label the central spindle, and GFP-tagged myosin regulatory light chain Sqh (Sqh-GFP-RLC) was used as a marker for the contractile ring. In wild-type control spermatocytes undergoing meiotic anaphase, accumulation of Feo-mCherry appeared at the central spindle as a bright band and Sqh-GFP-RLC signals were present as a ring structure at the cell cortex as well as a tight constricted ring after furrow ingression. In contrast, central spindle was impaired and the contractile ring was fragmented in some Mfe2 mutant spermatocytes (Fig. 6G), as also illustrated in the model diagram (Fig. 6H).

To better investigate the dynamics of central spindle and contractile ring behavior, we monitored cytokinesis by live-cell imaging. Analysis of living spermatocytes revealed that the central spindle and contractile ring were initially formed in late anaphase of Mfe2 mutants but they were destabilized and collapsed along with the failure of furrow ingression at the later stage in some spermatocytes (Supplementary Fig. S4, Supplementary Video S1 and S2). To investigate the transcriptional basis of the spindle defects observed in Mfe2 mutants, we performed GSEA on late spermatocytes. Strikingly, we observed divergent regulation among spindle-related gene programs. The “spindle” (a broad category of spindle-related genes) and “spindle assembly” gene sets were significantly downregulated in Mfe2 mutants compared to controls (Fig. 6I, left and middle panels), suggesting a transcriptional deficit in initiating or organizing the spindle apparatus. In contrast, the “spindle elongation” gene set was significantly upregulated (Fig. 6I, right panel). This uncoupling between assembly and elongation programs reflects a specific dysregulation of spindle dynamics in late spermatocytes of Mfe2 mutants. Taken together, we conclude that Mfe2 depletion affects central spindle behavior and leads to cytokinesis defects in the meiotic division of spermatocytes.

Mfe2 mutation leads to defective lipid homeostasis

As Mfe2 functions as an essential enzyme in β-oxidation of VLCFAs, we conducted a targeted lipid approach to analyze the levels of VLCFAs in testes from 2-day-old control and Mfe2 mutant flies. The levels of most VLCFAs with carbon chain lengths greater than 26 increased obviously in Mfe2 mutants, indicating impairment of β-oxidation of VLCFAs (Fig. 7A). We also performed untargeted lipidomics on these testes to assay levels of a broader array of lipid species. The result showed that the levels of a majority of lipid derivatives, including phospholipids and sterols, were reduced in Mfe2 mutants (Fig. 7B). As illustrated in the model diagram (Fig. 7C), the successful completion of cytokinesis is critically dependent on the unique lipid composition of the cleavage furrow, where specific lipids create the high membrane curvature [36]. For example, phosphatidylethanolamine (PE) localizes to the outer leaflet of the cleavage furrow promoting concave bending, while phosphatidylinositol phosphates (PIPs) localize to the inner leaflet of the cleavage furrow driving convex bending [36, 37]. The trafficking of specific lipids, such as phosphatidylinositol (PI) and phosphatidylinositol 4-phosphate (PI(4)P), via the endocytic pathway is essential for cytokinesis [36]. In Drosophila spermatocytes, multivesicular endosomes (MVEs) release intraluminal vesicles near the furrow rather than fusing with the cleavage furrow directly [38].

Fig. 7.

Disruption of lipid homeostasis leads to cytokinesis defect. (A) The level changes of VLCFAs in Mfe2 mutant male fly testes compared with w¹¹¹⁸ male fly testes. (B) The level changes of lipid derivatives in Mfe2 mutant male fly testes compared with w¹¹¹⁸ male fly testes. (C) Schematic diagram of lipid composition and membrane trafficking via multivesicular endosomes at the cleavage furrow during cytokinesis. (D) Maximum intensity projection of Z-stack images of late spermatocytes expressing Feo-mCherry and PLCδ-PH-EGFP from w¹¹¹⁸ and Mfe2 mutant male fly testes. Feo-mCherry marked the central spindle, and PLCδ-PH-EGFP marked phosphatidylinositol-4,5-bisphosphate (PI(4,5)P2). (E) Fluorescence intensity measurements of PI(4,5)P2. The results showed the PI(4,5)P2 membrane/cytoplasm fluorescence ratio. Each dot represents an individual measurement. Data are presented as mean ± SEM (n = 21 for w¹¹¹⁸ group, n = 24 for Mfe2 mutant group). The P-value was calculated by Student’s t-test, where *** P ≤ 0.001. (F) Maximum intensity projection of Z-stack images of late spermatocytes expressing Feo-mCherry and tGPH from w¹¹¹⁸ and Mfe2 mutant male fly testes. Feo-mCherry marked the central spindle, and tGPH marked phosphatidylinositol-3,4,5-triphosphate (PI(3,4,5)P3). (G) Fluorescence intensity measurements of PI(3,4,5)P3. The results showed the PI(3,4,5)P3 membrane/cytoplasm fluorescence ratio. Each dot represents an individual measurement. Data are presented as mean ± SEM (n = 22 for w¹¹¹⁸ group, n = 20 for Mfe2 mutant group). The P-value was calculated by Student’s t-test, where **** P ≤ 0.0001. (H) Phase contrast microscopy of live testis to visualize round spermatids of w¹¹¹⁸ and Mfe2 mutant flies raised on normal food or food supplemented with 1.5% Asolectin. Dark circles are mitochondrial derivatives, and light circles are nuclei. (I) Quantification of cytokinesis defects in w¹¹¹⁸ and Mfe2 mutant flies raised on normal food or food supplemented with 1.5% Asolectin. (J) Immunofluorescence staining to visualize round spermatids of w¹¹¹⁸ and Mfe2 mutant flies raised on normal food or food supplemented with 1.5% Asolectin. ATP5A was used to mark mitochondria, DAPI was used to mark DNA (nuclei)

We next tested whether the distribution of phosphatidylinositol phosphates at the cleavage furrow was altered in Mfe2 mutant spermatocytes. We used phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) reporter UAS-PLCδ-PH-EGFP (EGFP fused to the PH domain of phospholipase Cδ (PLCδ) driven by Tub-Gal4) and phosphatidylinositol 3,4,5-trisphosphate (PI(3,4,5)P3) reporter tGPH (EGFP fused to the PH domain of Step) to determine their distribution during cytokinesis. The fluorescence intensity for both reporters was significantly reduced at the cleavage furrow as well as the plasma membrane in Mfe2 mutant spermatocytes compared to the controls, indicating defective PI(4,5)P2 and PI(3,4,5)P3 distribution (Fig. 7D-G). These results suggest impairment of PI(4,5)P2 and PI(3,4,5)P3 is associated with meiotic cytokinesis defects in Mfe2 mutants.

Our results revealed a decrease in the levels of various lipid derivatives, including phospholipids, and a decline in PIP signaling, which is known to be associated with cytokinesis failure. Asolectin is composed of approximately equal proportions of lecithin, cephalin, and phosphatidylinositol, along with minor amounts of other phospholipids and polar lipids. On one hand, an Asolectin-based diet can replenish the generally reduced lipid derivatives, particularly phospholipids. On the other hand, since phosphatidylinositol serves as a precursor for PIP synthesis, an Asolectin-based diet containing phosphatidylinositol can promote PIP production to some extent. Thus, we reason that the use of Asolectin supplementation is likely to rescue the cytokinesis defects in Mfe2 mutant spermatocytes. To further verify this, we fed the flies with Asolectin from the larval stage, and examined the adult male testis phenotype. Microscopic analysis revealed that diets supplemented with 1.5% Asolectin were able to partially mitigate the cytokinesis defect in Mfe2 mutants, with reduced proportion of round spermatids containing abnormal nebenkern to nucleus ratio of 1:2 and 1:4 and increased proportion of round spermatids containing 1:1 ratio of nebenkern to nucleus (Fig. 7H and I). Wild-type control flies, fed with the same diet, displayed no defects in cytokinesis (Fig. 7H and I). Further immunofluorescence staining with an anti-ATP5A antibody also confirmed this rescue effect of Asolectin (Fig. 7J). Together, these results demonstrate that lipid homeostasis is disrupted in Mfe2 mutants and reduced lipid derivatives, including PI(4,5)P2 and PI(3,4,5)P3, are associated with the failure of cytokinesis in spermatocytes lacking of Mfe2.

Discussion

Male infertility, a clinical feature of DBP deficiency, has not been well studied. it is possible that male infertility is less noticeable compared to female infertility, and thus the number of cases of male adult patients with infertility recorded is not as common as those of female patients. In this study, we have characterized a Drosophila DBP deficiency model and investigated the pathogenic mechanism of male infertility. We employed an array of phenotypic analyses, immunofluorescence imaging and single-cell profiling methods and discovered major defects in Mfe2 mutant testes, including impaired cell cycle of germ cells and abnormal meiotic cytokinesis of spermatocytes (Fig. 8). Our lipidomic analysis revealed that the lipid homeostasis of Mfe2 mutant testes was disrupted, and Asolectin feeding partially rescued the spermatocyte cytokinesis defects. Furthermore, we provide evidence that the function of Mfe2 is dependent on its peroxisomal localization, indicating that its role in the peroxisome is critical for male fertility. Consistent with this finding, loss of function in other peroxisome-related genes also leads to male sterility, as demonstrated by other research on peroxisomal disorders [15, 16, 39]. The Drosophila loss-of-function mutant strains of Pex12 and Pex16 constructed in our work also exhibit the phenotype of male infertility, underscoring the essential role of peroxisomes in maintaining male fertility.

Fig. 8.

A working model explaining the role of Mfe2 in maintaining lipid homeostasis, cell cycle progression and meiotic cytokinesis in the Drosophila testis. Loss of Mfe2 function leads to the block of β-oxidation of very-long-chain fatty acids (VLCFAs) in peroxisomes and a decrease of lipid derivatives. Loss of Mfe2 function leads to abnormal germ cell cycle progression, resulting in the accumulation of germ cells. The reduction of some special lipids results in meiotic cytokinesis defects. Failed cytokinesis results in multinucleated spermatids and disrupts subsequent spermatogenesis stages, such as flagellum elongation and individualization, thereby resulting in failed spermatogenesis in the testis terminus and a complete absence of mature sperm

scRNA-seq analysis has been widely used in characterizing cell types within the Drosophila reproductive system and enables the construction of high-resolution transcriptional atlases of tissues such as the testis and ovary, revealing novel insights into gametogenesis and fertility [19, 20, 40]. Here, we performed single-cell transcriptome profiling on wild-type control and Mfe2 mutant testes. We found that the proportion of late spermatocytes in Mfe2 mutants was significantly increased, which indicated an impairment in the cell cycle progression of spermatocytes.

By re-clustering the GSCs/spermatogonia data and analyzing the molecular characteristics of the new pseudotime stages, we found that the number of Mfe2 mutant spermatogonia in S phase was increased. The results of immunofluorescence staining revealed that no cyst containing more than 16 cells was detected in nuclei dense area, confirming that it was not caused by overproliferation but an impairment of cell cycle progression. However, our scRNA-seq data suggested that the proportion of late spermatocytes increased while the proportion of GSCs/spermatogonia decreased, which appeared to be contradictory to immunofluorescence staining. The possible explanation for this phenomenon is likely due to the faster population expansion of late spermatocytes than that of GSCs/spermatogonia. It is known the initiation and completion of the premeiotic S phase occur rapidly in spermatocytes, and the premeiotic G2 phase of spermatocytes lasts 80 h, accompanied by an approximately 25-fold increase in the volume of spermatocytes in wild type [27]. We speculate that in the absence of Mfe2 functions, spermatocytes at G2 phase are most vulnerable and cell cycle is delayed. As a result, an abnormal increase in the proportion of late spermatocytes was observed. Further exploration is required to determine the causes of the impairment of the cell cycle process.

Cytokinesis in animal cells is a highly coordinated process that relies on the precise regulation of different cytoskeletal components and the plasma membrane [33]. The central spindle (CS), composed of antiparallel microtubule bundles, sends stimulatory signals to the cell equator and promotes the assembly of the contractile ring (a structure of F-actin and myosin II) at the cell equator during anaphase in animal cells [41]. Contractile ring (CR), on the inner face of the plasma membrane around the cell equator generates the constricting force that produces a cleavage furrow and separates the cell into two [42]. During cytokinesis, lipids play an important role. To accommodate the two rapid and successive divisions in meiosis, a cell must expand its surface area by approximately 60%. This considerable challenge requires the coordinated action of multiple membrane trafficking and lipid remodeling pathways [36]. It has been suggested that endosomal transport and fusion with the plasma membrane during cytokinesis delivers both new membrane and specific lipids to the cleavage furrow and intercellular bridge [43]. Among the lipids that affect cytokinesis, phosphatidylinositol phosphates have been extensively investigated [44]. Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) exerts direct control over the actin cytoskeleton in vivo by modulating the activity and targeting of actin-regulatory proteins [45]. It is unknown whether phosphatidylinositol 3,4,5- triphosphate (PI(3,4,5)P3) affects the spermatocyte meiotic cytokinesis. It has been previously shown that PI(3,4,5)P3 exhibits a polar localization in Dictyostelium cells, suggesting that a PIP gradient may guide division [46, 47]. Interestingly, we observed that PI(3,4,5)P3 localized to the plasma membrane and was particularly enriched at the cleavage furrows of spermatocytes in Drosophila (Fig. 7C), which is consistent with the experimental results of Govind Kunduri et al [38]. However, when Drosophila spermatocytes were cultured individually ex vivo, PI(3,4,5)P3 was not concentrated on cellular membranes during cleavage [48]. Therefore, it is likely that this enrichment of PI(3,4,5)P3 is depended on the integrity of the cyst. This suggests that PI(3,4,5)P3 might play an unknown but important role at the cleavage furrow in Drosophila spermatocytes. Mutations related to central spindle components or F-actin ring regulation, usually fail to form structural assemblies of the central spindle and F-actin ring, leading to early cytokinesis arrest [49, 50]. However, defects in membrane trafficking genes(which finally affected plasma membrane composition) allow initial ring assembly but prevent full constriction due to impaired vesicle-mediated membrane addition [33]. Similarly, PI(4,5)P2 and PI(3,4,5)P3 are not required for initiation of cytokinesis, as initial furrow formation proceeds normally in cells with altered PI(4,5)P2 or PI(3,4,5)P3 levels, even though cleavage furrow regression is observed later [45, 46]. In our results, during the meiotic division of Mfe2 mutant spermatocytes, complete actin rings and central spindle structures were observed in the early stage. Combined with our lipidomic analysis and the signaling profiles of PI(4,5)P2 and PI(3,4,5)P3, these results confirm that the decrease in the levels of various lipid derivatives (which affected membrane trafficking) in the testis, along with the reduction in PI(4,5)P2 and PI(3,4,5)P3 signaling, is associated with cytokinesis failure in Mfe2 mutant spermatocytes.

We have shown that Mfe2 mutant had an increased level of 26-carbon VLCFAs, which corresponded to the increase in VLCFAs in the blood of infant-onset patients with DBP deficiency [6, 51]. In patients with DBP deficiency, neurodegeneration is believed to be associated with the accumulation of VLCFAs [4], while the connection between VLCFAs accumulation and male infertility has not been well studied. In Drosophila, the bond gene encodes a VLCFA elongase protein which is essential for VLCFA biosynthesis. However, bond mutants, which presumably have low VLCFA levels, also exhibit strong cytokinesis defects of spermatocytes [52]. This indicates that insufficient amount of VLCFA can lead to the cytokinesis defect. Although the accumulation of VLCFAs in Mfe2 mutants suggest that excessive VLCFA levels could lead to spermatocyte cytokinesis defects due to the lipid toxicity, this explanation is somewhat limited. Given the decrease in various lipid derivatives observed in the Mfe2 mutant strains and the phenotype of cytokinesis defect can be alleviated by supplementing Mfe2 mutants with Asolectin, it is likely that lipid derivatives contribute to the normal process of cytokinesis. These findings provide significant novel insights into the molecular pathogenic mechanism of male infertility caused by DBP deficiency and hopefully expand therapeutic option for patients with DBP deficiency.

Materials and methods

Fly husbandry and stocks

Drosophila stocks were raised and maintained on standard fly food under controlled conditions (25 °C, 40–60% RH). Sources of Drosophila stocks in this study are listed below: Tubulin-Gal4 (BDSC#5138), tGPH (BDSC#8163), Sqh-GFP-RLC (BDSC#57145), UAS-PLCδ-PH-EGFP (BDSC#39693), p(Ubi-p63E-Feo-mCherry)3 (BDSC#59277), y[1] M{RFP[3xP3.PB] GFP[E.3xP3] = vas-Cas9}ZH-2 A w[1118]/FM7c (BDSC#51323), y[1] sc[*] v[1] sev[21]; P{y[+ t7.7] v[+ t1.8] = TKO.GS00684}attP40 (BDSC#76954), Cat-GFP (BDSC51546, a gift from Prof. Zongzhao Zhai (Hunan Normal University)); If/CyO, TM3/TM6B, w¹¹¹⁸, Mfe2^−/−, Mfe2^gDNA, Pex16^−/−, Pex12^−/−, UAS-Mfe2, UAS-Mfe2-ΔAKL. Stocks listed after w¹¹¹⁸ were made in our lab.

Generation of Mfe2 mutant and Mfe2 genomic rescue transgene

The Mfe2 mutant line and the Mfe2^gDNA line were generated by our group, and the details for Mfe2 mutant and Mfe2^gDNA generation are described in another manuscript which is currently under review. The Mfe2 mutant contains a 7 base-pair deletion at position C.39–45 of Mfe2 locus. The Mfe2^gDNA transgene contains a genomic fragment covering Mfe2 gene along with the 2-kb upstream and 2-kb downstream sequences.

Construction of Mfe2 overexpression Drosophila strains

The pUASt-Mfe2-attB and pUASt-Mfe2-ΔAKL-attB plasmid constructs were produced by introduction of the corresponding Mfe2 cDNA fragments into pUASt-attB backbone. The primers used for amplifying the Mfe2 fragment (with PTS1 sequence) were CGGAATTCATGTCCTCATCCGATGGAA (Forward primer) and GCTCTAGATTACTGTGAGCTCTTCAGGTC (Reverse primer). The same forward primer and GCTCTAGATTACAGCTTGGCCTGTGAGCT (Reverse primer) were used for amplifying the Mfe2-ΔAKL fragment (without PTS1 core sequence). Total RNA was extracted from precisely dissected wild-type third-instar larvae and reverse-transcribed into cDNA using the Vazyme HiScript III RT SuperMix kit (Cat# R211-01/02). The Mfe2 target fragment was amplified via PCR using TaKaRa Ex Taq DNA Polymerase (Cat# R045A). The PCR product was purified using the TIANGEN Universal DNA Purification Kit (Cat# DP214-02). Both the empty vector and the purified insert fragment were digested with EcoRI and XbaI restriction enzymes (NEB, Cat# R3101V and Cat# R0145V). After separation by agarose gel electrophoresis, the desired gel bands were excised and the DNA was recovered using the AXYGEN DNA Gel Extraction Kit (AP-GX-250). The target fragment and the digested vector were then ligated using the TaKaRa DNA Ligation Kit (Cat# 6022). The ligation product was transformed into Trans5α E. coli (transgen, Cat# CD201) for amplification. Cells were plated and incubated overnight at 37 °C, and single colonies were subsequently picked for sequence verification. Plasmids confirmed to have the correct insert sequence without frameshift mutations were selected. The correct plasmid was expanded, and isolated using the QIAGEN Plasmid Midi Kit (Cat# 12143). The purified plasmid constructs were sent to the Shanghai Drosophila Resource & Technology Platform for microinjection. Embryos from the background strain vas-phi-Zh2A-VK5 (75B1) were injected. Red-eyed progeny emerging from these injections were selected and used to establish UAS-Mfe2 and UAS-Mfe2-ΔAKL transgenic lines (inserted into 75B1 attP docking site).

Generation of Pex12 and Pex16 mutants

The construction of Pex12 mutant was carried out as described previously [53–55]. The chosen guide RNA target sequence for Pex12 was GGCACGAGGACACCAGTCCC TGG. The DNA template for single-guide RNA (sgRNA) transcription was prepared by PCR amplification following the manufacturer’s protocol, using the forward and reverse primers (TAATACGACTCACTATAGGCACGAGGACACCAGTCCCTGGGTTTTAGAGCTAGAAATAGC and AAAAAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC), respectively. Subsequently, the sgRNA was transcribed in vitro from the purified DNA template using the T7 Transcription Kit (Vazyme, R101-01) and further purified with the RNeasy^® Mini Kit (QIAGEN, 74104). The purified sgRNA was then microinjected into embryos of the vas-Cas9 flies (BDSC#51323). The resulting adult flies were individually crossed with the balancer flies. Genomic DNA was extracted from the progeny of these crosses, and mutations were identified by PCR analysis. For the construction of Pex16 mutant, sgRNA line of Pex16 was obtained from Bloomington Drosophila Stock Center (BDSC#76954) and crossed with the vas-Cas9 flies (BDSC#51323). The progenies were individually crossed to the balancer flies. The subsequent steps are the same as the previous ones.

Male fertility assay

For male fertility assays, individual crosses were set up between 1 male and 2 wild-type females. Each experiment comprised 3 independent replicates of 50 crosses each (150 males tested in total). One week later, if there are offspring in the tube, that male fly is marked as fertile.

Testis morphology assay

To assess testis morphology, testes were observed under a stereomicroscope. Control testes displayed a distinctive silvery/refractile appearance, indicative of abundant, bundled sperm flagellum. Most of Mfe2 mutant testes were characterized by sparse sperm flagellum lacking this obvious refractile property. For statistical analysis, testes exhibiting no clear refractility were classified as abnormal.

Phase-contrast microscopy observation

Testes from the 2-day-old flies were dissected in Schneider’s Drosophila Medium. Testes were placed on the cover glass and forceps were used to make a small opening. Then the microscope slide was inverted over the testes to allow the cells in the testes to be fully released. Squashed testes on slides were observed under a phase-contrast microscope.

Immunostaining and confocal imaging

Dissected testes were fixed with 4% formaldehyde in 0.1% PBST for 30 min at RT, washed (3 × 10 min each), blocked with 2.5% goat serum in 0.1% PBST (1 h), and incubated with primary antibody (The specific dilution ratio is described in Table S1.) of 2.5% goat serum in 0.1% PBST) overnight (8 to 12 h). Testes were then washed with 0.1% PBST (3 × 10 min each) and incubated with secondary antibody (1:500 dilution in 2.5% goat serum in 0.1% PBST) (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) for 2 h at RT. Testes were washed with 0.1% PBST (3 × 10 min each), and mounted with mounting medium (Beyotime P0131 containing DAPI). The slides were imaged on Olympus confocal laser scanning microscope FV1000. All primary antibodies and dye used in this study were listed in supplementary Table S1. The anti-Mfe2 antibody is a rabbit polyclonal antibody generated in our laboratory, and the details for the antibody generation are described in another manuscript which is currently under review. For the staining with the anti-Bam antibody, the 0.1% PBST was changed to 1% PBST buffer for all the steps.

Live cell imaging

Live cell imaging was performed as described previously [56]. A 40x oil immersion microscope (Olympus CSU-W1) was used for observation.

Elongated spermatid staining and observation

The staining and observation of elongated spermatid was carried out as described previously [38, 57]. Briefly, the dissected testes were placed onto the coverslip and forceps were used to tear a small opening in each testis. The microscope slide was inverted and put onto the coverslip carefully, sandwiching the sample. This slide-coverslip sandwich was quickly transferred into liquid nitrogen. A razor blade was then used to gently pry the slide and coverslip apart, and the testes tissue was adhered to the microscope slide. Pre-chilled 95% ethanol was added and the sample was incubated at −20 °C for 10 min. Next, the slide was transferred into 4% formaldehyde (diluted in 0.1% PBST) for 15 min at room temperature. After two washes in PBS and each for 5 min, the sample was added 2.5% goat serum and incubated for 45 min to block non-specific binding. A PaP pen was used to circle the target area on the slide. Then, the primary antibody was added and the sample was incubated at room temperature for 2 h. After three washes with PBST and each for 10 min, the secondary antibody was added and the sample was incubated at room temperature for 1.5 h. After two additional washes with PBST and each for 5 min, mounting medium containing DAPI was added to the sample before observation.

EDU staining

Dissection of testes was performed in Schneider’s medium (ThermoFisher Scientific, #21720024). The dissected testes were subsequently incubated at 25℃ in the dark for 1 hour in Schneider’s medium supplemented with 10 µm EdU (5-ethynyl-2’-deoxyuridine). Fixation and permeabilization were then carried out just like described in “Immunostaining and confocal imaging” part.

Asolectin feeding

Asolectin (Merck, 11145) was dissolved in PBS, and then subjected to ultrasonic treatment (100 W, 5s/5s, 10 min) to form a uniform suspension with a concentration of 3%. The standard food was evaporated by microwave oven to half of its original moisture content, and then mixed with an equal amount of the suspension to prepare food with an Asolectin final concentration of approximately 1.5%. The control group was added with an equal amount of PBS. Wild-type and mutant fruit flies were placed in the two types of food respectively, and the male offspring’s testes were dissected for further analysis.

Western blot analysis

80 pairs of testes from 2-day-old adult flies were collected and homogenized in 4% sodium dodecyl sulfate (SDS) solution containing cOmplete™ protease inhibitor cocktail (Roche, Cat# 4693132001) at 4 °C for 30 min. The resulting lysates were then separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. For immunoblotting, the membranes were incubated with the following primary antibodies: mouse anti-α-Tubulin (1:1000, Proteintech, Cat# 66031) and rabbit anti-Mfe2 (1:1000). Horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibody (1:5000, Abcam, Cat# ab6721) was subsequently applied.

Lipidomics analysis

Lipids were extracted from approximately 400 pairs of testes from 2-day-old adult flies for each group using a modified Bligh and Dyer’s method [58, 59]. Briefly, tissues were homogenized in 500 µL of chloroform: methanol: MilliQ H₂O (3:6:1, v/v/v). The homogenate was incubated at 1500 rpm for 1 h at 4 °C. Following incubation, phase separation was induced by adding 150 µL of deionized water and 350 µL of chloroform. After centrifugation, the lower organic phase containing lipids was collected. A second extraction was performed on the remaining aqueous phase by adding 450 µL of chloroform. The combined lipid extracts were pooled and dried in a SpeedVac under OH mode. The dried organic phase was then derivatized with 3-Nitrophenylhydrazine [60].

For fatty acid, metabolite analysis was performed using a Jasper HPLC coupled to a Sciex 4500 MD system. Separation was achieved on a Phenomenex Kinetex C18 column (100 × 2.1 mm, 2.6 μm) with mobile phase A (0.05% formic acid in acetonitrile: water, 1:9) and mobile phase B (0.05% formic acid and 2 mM ammonium acetate in acetonitrile: methanol: isopropanol, 1:2:2). Quantitation utilized the following internal standards: SCFAs and MCFAs were quantitated using Octanoic acid-1-¹³C₁ (Sigma-Aldrich) and Butyric-2,2-d₂ (CDN Isotopes); free fatty acids and VLCFAs were quantitated using d₃₁−16:0 (Sigma-Aldrich) and d₈−20:4 (Cayman Chemicals).

For other lipid classes, lipidomic analyses were conducted at LipidALL Technologies using a Shimadzu Nexera 20-AD HPLC coupled with Sciex QTRAP 6500 PLUS as reported previously [61]. Separation of individual lipid classes of polar lipids by normal phase (NP)-HPLC was carried out using a TUP-HB silica column (i.d. 150 × 2.1 mm, 3 μm) with the following conditions: mobile phase A (chloroform: methanol: ammonium hydroxide, 89.5:10:0.5) and mobile phase B (chloroform: methanol: ammonium hydroxide: water, 55:39:0.5:5.5). MRM transitions were set up for comparative analysis of various polar lipids. Individual lipid species were quantified by referencing to spiked internal standards. d9-PC32:0(16:0/16:0), d7-PE33:1(15:0/18:1), d31-PS(d31-16:0/18:1), d7-PA33:1(15:0/18:1), d7-PG33:1(15:0/18:1), d7-PI33:1(15:0/18:1), d5-CL72:8(18:2)4, Cer d18:1/15:0-d7, C12-PECer, C8-GluCer, C8-GalCer, d3-LacCer d18:1/16:0, d7-LPC18:1, d7-LPE18:1, C17-LPI, C17-LPA, C17-LPS, C17-LPG, d17:1 Sph, d17:1 S1P were obtained from Avanti Polar Lipids. GM3-d18:1/18:0-d3 was purchased from Matreya LLC. Glycerol lipids including diacylglycerols (DAG) and triacylglycerols (TAG) were quantified using a modified version of reverse phase HPLC/MRM [62]. Separation of neutral lipids were achieved on a Phenomenex Kinetex-C18 column (i.d. 4.6 × 100 mm, 2.6 μm) using an isocratic mobile phase containing chloroform: methanol:0.1 M ammonium acetate 100:100:4 (v/v/v) at a flow rate of 300 µL for 10 min. Levels of short-, medium-, and long-chain TAGs were calculated by referencing to spiked internal standards of TAG(14:0)3-d5,TAG(16:0)3-d5 and TAG(18:0)3-d5 obtained from CDN isotopes, respectively. DAGs were quantified using d5-DAG17:0/17:0 and d5-DAG18:1/18:1 as internal standards (Avanti Polar Lipids). Free cholesterols and cholesteryl esters were analyzed under atmospheric pressure chemical ionization (APCI) mode on a Jasper HPLC coupled to Sciex 4500 MD as described previously, using d6-cholesterol and d6-C18:0 cholesteryl ester (CE) (CDN isotopes) as internal standards [63]. The levels of all detected substances were normalized by protein content. Lipidomic analysis was performed by LipidALL Technologies Co. Ltd. (Changzhou, China).

Bulk RNA-seq analysis

RNA sequencing experiments (executed by Novogene Experimental Department) were performed on RNA prepared from 2-day-old testes, and 40 pairs of testes for each sample. RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). RNA was extracted in Trizol (TRNzol Universal Reagent). The sequence libraries were generated using the Fast RNA-seq Lib Prep Kit V2 (Cat.No.RK20306), following the manufacturer’s instructions. After assessing library quality on the Agilent Bioanalyzer 2100 system, the library preparations were sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated. After quality control of the raw reads, alignment against the Drosophila reference genome was performed on HISAT2 v2.0.5. Differential expression analysis between two comparison groups was performed using the DESeq2 software (version 1.20.0). Pathway enrichment was analyzed by GSEA. The gene sets used in the text refer to a previous publication [38].

Tissue preparation for single cell RNA sequencing

40 pairs of testes were dissected from w¹¹¹⁸ or Mfe2 mutant flies in Schneider’s medium, and then washed in PBS for 3 times. Tissues were then enzymatically digested using 1 mg/ml dispase (D4818, Merck), 1 mg/ml trypsin (Y0002311, Merck) and 1 mg/ml DNase I (D8071, Solarbio) at 37 °C for 10 min. Afterward, 40 μm cell sieve was used to filter the sample, and the cell suspension was centrifuged at 300 rpm at 4 °C for 5 min. The supernatant was then removed and 1 mL of re-suspension buffer (Precooled 1X PBS) was added to fully re-suspend the cells. The cell counter (Countstar Rigel S2) was used to measure cell concentration. A cell concentration of 700–1200 cells/µL was obtained, ensuring viability > 90% and aggregation < 15%.

Single cell RNA-seq library preparation and sequencing

Single-cell RNA-seq libraries were prepared using Chromium Next GEM Single Cell 3ʹ Reagent Kits v3.1 on the Chromium Controller (10× Genomics). Cell suspensions were loaded onto Chromium Next GEM Chip G and processed to generate single-cell GEMs per manufacturer specifications. Captured cells were lysed for RNA barcoding via reverse transcription within individual GEMs. Barcoded full-length cDNA was synthesized, and libraries were constructed following the manufacturer’s protocol. Library quality was assessed using Qubit 4.0 and Agilent 2100. Sequencing was performed on Illumina NovaSeq 6000 (Biomarker Technologies Corporation, Beijing, China) with ≥ 50,000 reads per cell and 150 bp paired-end reads (PE150).

Processing of single cell RNA-seq data

Alignment was performed against the amended reference (dmel_r6.60, ftp.flybase.net/releases/current/dmel_r6.60/) using 10x Genomics Cell Ranger 6.0.0 with the STAR aligner [64]. Following quality control, 24,627 cells were retained using these thresholds: >500 and < 4500 genes per cell and mitochondrial ratio < 0.15. Genes detected (≥ 200 UMI) in ≥ 3 cells were included for analysis. Expression was quantified via unique molecular identifiers (UMIs) per cell barcode-gene combination. Cell-containing barcodes were identified using Cell Ranger’s filtering (v3.0.2) and used for downstream analyses: clustering, cell typing, and differential expression with Seurat v5.1.0. We integrated w¹¹¹⁸ and Mfe2 mutant samples and performed harmony package (v1.2.3) and got an integrated dataset. UMAP and clustering analysis were performed on the merged dataset using the top 5000 highly variable genes and 1–30 PCs. Selecting weighted Shared Nearest Neighbor (SNN) graph-based clustering method was used to find clusters. We identified the cell types by referring to the marker genes mentioned in previous publications^{17 19}.

Functional enrichment analysis

GO analysis was performed using R package clusterProfiler (v4.14.4). To score the special enriched functions or pathways. Gene set analysis between Mfe2 mutant and w¹¹¹⁸ group was performed using GSVA (R package GSVA, v2.0.5) and GSEA (R package clusterProfiler, v4.14.4). GSVA ranked all the genes in the scRNA-seq data using all genes as a basis, while GSEA ranked all the genes using DEGs from control and Mfe2 mutant group. GSVA assesses the relative enrichment of gene sets within individual samples, rather than absolute expression levels. It calculates activity scores by comparing the ranking of target pathway genes among all genes in a given sample, reflecting the relative importance of the pathway within the cell’s current transcriptional state. To visualize the result of differential analysis, enrichplot (R package, v1.26.6) and ggplot2 (R package, v3.5.2) were used. The gene sets involved in the GSVA and GSEA are listed in Supplementary Table S2.

Cell trajectory analysis

We applied Monocle (version 2.34.0) to order cells along pseudotime trajectories. This ordering enabled the visualization of their developmental trajectory within the reduced-dimensional space.

Quantification of tGPH and PLCδ-PH-EGFP signals

For quantification of tGPH and PLCδ-PH-EGFP signals, images were collected using the same imaging settings for different genotypes. The ratio of the membrane mean fluorescence to the cytoplasmic mean fluorescence measured within the same sized areas of membrane and cytoplasm was used to determine the signal intensity.

Statistical method

The length, area and fluorescence statistics of the immunofluorescence images were all carried out in the Image J software. Statistical comparisons were performed using GraphPad Prism software (Version 9.5; GraphPad Inc., La Jolla, CA, USA). Differences between datasets were assessed using Student’s t-test or one-way ANOVA, while proportional data were evaluated using the Chi-squared test. Significance levels are denoted as follows: ns (not significant) P > 0.05, * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, **** P ≤ 0.0001.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.W. and W.G. designed the research, J.W. and Y.M. performed the experiments and analyzed the data with the help of Y.W., J.Z., J.J., P.Z. and H.Z.. W.G. supervised the project, and J.W. and W.G. wrote the manuscript.

Data availability

All data and analyses reported in this paper will be provided upon request.

Declarations

Conflict of interest

The authors declare no competing interests.