Cytochrome c₁-like is required for mitochondrial morphogenesis and individualization during spermatogenesis in Drosophila melanogaster

The fruit fly Drosophila melanogaster is a popular model organism because of the availability of powerful genetic tools to dissect the mechanisms involved in various biological processes. The D. melanogaster genome sequence shows more than 60% of human genes (including disease genes) have functional orthologues in flies (Brandt and Vilcinskas, 2013; St Johnston, 2002). Flies and humans share many fundamental cellular and developmental processes, including part of spermatogenesis, despite their obvious differences. Therefore, D. melanogaster has been widely used for understanding the molecular mechanisms of different human diseases, including tumorigenesis, infertility and neurodegenerative disease.

The Drosophila testis is an excellent tool to study the complex fundamental processes of testis development and spermatogenesis, as the sperm development stages have been clarified, and various molecular genetic techniques, such as the UAS/Gal4 system, allow the tissue- and time-specific control of transgene expression, such as overexpression or knockdown in germ cells or cyst cells (Fabian and Brill, 2012). The Drosophila testis is a long blunt-ended tube with a niche containing hub cells at the apical tip to support germline stem cells (GSCs) and somatic cyst stem cells (Lowe and Montell, 2022). The GSC divides asymmetrically to produce a gonialblast (GB), which starts the cellular differentiation programme. Each GB undergoes four mitotic divisions and two meiotic divisions to produce 64 interconnected haploid spermatids in a cyst (Gleason et al., 2018). Then, accompanying nuclear compaction and shaping, mitochondria aggregate and condense to form the nebenkern (Zhong and Belote, 2007; Tokuyasu, 1975). During spermatid elongation, both axonemal microtubules and mitochondrial nebenkern extend to form the flagellum (Fabian and Brill, 2012). Finally, the syncytial spermatids are individualized with the help of an individualization complex (IC) composed of 64 F-actin cones (Fabian and Brill, 2012). When the actin cones move from the sperm head toward the tail, the unnecessary cytoplasmic components are removed, forming the ‘cystic bulge’ and then ‘waste bag’ at the tip of the sperm tail (Noguchi and Miller, 2003; Noguchi et al., 2008). The individualized sperm, with condensed needle-shaped nuclei and long tails, then become coiled and are released into the seminal vesicles (Zhong and Belote, 2007).

Mitochondria, well known as for producing a large fraction of cellular energy, are actually also involved in various cellular functions, such as metabolism, immunity and cell death (Eslamieh et al., 2022). Normally, unlike oocytes, which have unusually small and simple mitochondria with the suppression of electron transport and free radical production for maintaining the fidelity of mitochondrial DNA inheritance, sperm contain mitochondria that are metabolically active, transcribe genes for respiratory electrons and also produce free radicals (de Paula et al., 2013). Therefore, mitochondria in sperm are unusually kept active and undergo special morphogenesis during spermatogenesis for the function of sperm in movement and fertilization, though they are not passed to the next generation. Many mitochondria-related genes have evolved in D. melanogaster with testis-specific expression, and loss of function of these genes dramatically reduces male fertility, mainly related to defects in spermatogenesis (Vedelek et al., 2016; Eslamieh et al., 2022).

The structure of the actin cones and their synchronous movement is essential for spermatid individualization. Evidence shows that mutations in components of the dynein–dynactin complex, including CDIC, DDLC1, DLC90F and Dynamitin, disturb the synchronous movement of the actin cones in Drosophila (Ghosh-Roy et al., 2004; Li et al., 2004; Ghosh-Roy et al., 2005; Wu et al., 2016). In addition to dynein genes, many other genes involved in plasma membrane reorganization, lipid metabolism and apoptotic removal of cytoplasmic contents also play a critical role during the individualization process (Noguchi and Miller, 2003; Yuan et al., 2019). The individualization process has also been found to be a caspase-dependent apoptosis-like event, where some caspases and caspase regulators participate actively (Cagan, 2003; Noguchi et al., 2011). Dronc, Ice and Hid proteins, as essential members of apoptotic pathways, have been identified in elongated spermatids (Arama et al., 2003; Huh et al., 2004).

ocnus (ocn) encodes a 14 kDa phosphohistidine phosphatase, specifically expressed in the testis during the postmeiotic stages of Drosophila (Parsch et al., 2001; Dorus et al., 2006). We have previously found that knockdown of ocn in the testis of D. melanogaster caused much smaller testes and severe defects in male fertility. RNA sequencing showed that many genes were significantly downregulated as a result of ocn knockdown in fly testes, including the Cytochrome c1-like gene cyt-c1L (Zheng et al., 2018). Recently, with iTRAQ (isobaric tag for relative and absolute quantification) proteome sequencing, the cyt-c1L protein was again found to be significantly downregulated by ocn knockdown. However, the precise role of cyt-c1L in male fly fertility is unknown. Unlike its parent gene, cyt-c1 (CG4769), which is expressed in all tissues of D. melanogaster, cyt-c1L (CG14508) is highly expressed in the testis (https://flybase.org/reports/FBgn0039651.html). Furthermore, cyt-c1L has been identified to be a fly sperm protein (Wasbrough et al., 2010). These findings imply that cyt-c1L might be involved in spermatogenesis in D. melanogaster. In addition, the orthologue of cyt-c1L was also identified to be one of the sperm proteins in mice and humans (Baker et al., 2007, 2008); thus, cyt-c1L in Drosophila spermatogenesis could become a useful model for revealing the function of this gene in the human testis.

This study used both the UAS/Gal4 system for gene knockdown and the CRISPR/Cas9 technique for gene knockout to investigate the role of cyt-c1L in spermatogenesis. Either knockdown or knockout of cyt-c1L resulted in male sterility associated with no individualization complex and no mature sperm in male D. melanogaster. Lack of cyt-c1L disrupted the normal mitochondrial morphogenesis during flagellum formation, and led to the downregulation of expression of several cytoskeleton-related genes but upregulation of apoptosis-related genes. These results suggest that cyt-c1L plays a critical role in spermatogenesis in D. melanogaster.

The transgenic cyt-c1L RNAi (cyt-c1L-hp) D. melanogaster line (THU1499), which interferes with mRNA sequence positions from 1255 to 1275, was obtained from the Tsinghua Fly Center (Beijing, China). The hairpin sequence of cyt-c1L-hp is TTGGTGAACTGTTTAAATAAA. Transgenic knockout line cyt-c1L^KO (82026), carrying a CRISPR construct with a single guide RNA (sgRNA), was purchased from Bloomington Drosophila Stock Center (BDSC, Bloomington, IN, USA). The sgRNA sequence of the knockout line cyt-c1L^KO is TGTCCGCTGGATGCCGCTCATGG.

A UAS-cyt-c1L fly line was generated as previously described (Liu et al., 2014). Briefly, the coding sequence of cyt-c1L was amplified and cloned into the pUAST vector (Addgene; https://www.addgene.org/). The recombinant plasmid was injected together with the Δ2-3 plasmid (transposase) into w¹¹¹⁸ (with white eyes) fly embryos. Progeny with red eyes were selected for balancing and mapping insertions. We generated the rescue flies (UAS-cyt-c1L; cyt-c1L-hp) with both cyt-c1L RNAi and overexpression elements by a sequence of crossing; the detailed protocols are shown in Fig. S1.

We knocked down cyt-c1L in the testes by use of Drosophila strains expressing inducible hairpin RNAi constructs under the control of nosGal4. This method was developed based on the upstream activating site (UAS)/Gal4 system to control the expression of a gene fragment that is dimerized to produce a dsRNA hairpin (hp) structure, which then triggers a sequence-specific post-transcriptional silencing and RNAi response. As nanos (nos) is specifically expressed in early germ cells, we used the nosGal4 driver to direct gonad (early germ line)-specific RNAi. The nosGal4 driver was kindly provided by Professor Zhaohui Wang (Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, CAS). Virgin females of nosGal4 were crossed with the cyt-c1L-hp flies to obtain cyt-c1L knockdown flies in germlines (gonads) (nosGal4>cyt-c1L-hp). The male flies (UAS-cyt-c1L; cyt-c1L-hp) were crossed with nosGal4 virgin females to try to obtain cyt-c1L rescue flies (nosGal4>UAS-cyt-c1L; cyt-c1L-hp). Flies from the cross of nosGal4 females and wild-type males (nosGal4>+) were used as the control.

Vasa-Cas9 fly line, a gift from Professor Shan Jin (Hubei University, Wuhan, China), was crossed with cyt-c1L^KO flies to obtain gene knockout (with 23 bp deletion in the anterior region of coding sequence) flies (Cas9>cyt-c1L^KO).

All fly lines were reared on a standard cornmeal/yeast medium at 25°C and under non-crowded conditions (200±10 eggs per 50 ml vial of media in a 150 ml conical flask) (Yamada et al., 2007).

For each biological replicate, 1 day old males (1d; n=15) were left to mate with 3–4 day old w¹¹¹⁸ virgin females (n=10) overnight (∼12 h). The following males were used in this experiment: (1) nosGal4>cyt-c1L-hp (knockdown), (2) nosGal4>UAS-cyt-c1L; cyt-c1L-hp (rescue), and (3) nosGal4>+ (control). The males were then removed and the mated females were maintained to lay eggs for 4 days. For Cas9>cyt-c1L^KO (knockout) flies, 1d males were individually paired with three w¹¹¹⁸ virgin females to distinguish knockout efficiency. Eggs were collected and counted, and then incubated at 25°C and 45–70% humidity for around 30 h. Hatch rates were determined by counting the number of hatched eggs out of the total eggs (Wu et al., 2016; Biwot et al., 2020; Horard et al., 2022). Under the microscope, unhatched eggs (looking like plump rice grains) are easily distinguished from hatched eggs (shrivelled eggshells). At least three independent trials per cross-type were conducted.

After mating, the Cas9>cyt-c1L^KO male flies were used to extract DNA individually to test knockout efficiency. Flies were homogenized in solution A (100 mmol l⁻¹ Tris-HCl, 100 mmol l⁻¹ EDTA, 100 mmol l⁻¹ NaCl, 0.5% SDS, pH 7.5), followed by centrifugation at 2000 g for 10 min at 4°C. Then, 20 μl protease K (20 μg ml⁻¹) was added to the supernatant and incubated at 56°C for 3 h. DNA was extracted by gently mixing with an equal volume of a mixture of phenol/chloroform/isoamyl alcohol (25:24:1) followed by centrifugation at 10,000 g for 10 min at 4°C. The aqueous layer was aspirated to a fresh tube, and isopropanol was added to precipitate DNA. After centrifugation, the supernatant was discarded and the DNA pellet was washed with 70% ethanol. After drying for around 5 min, the DNA pellet was dissolved in 50 μl of TE buffer (10 mmol l⁻¹ Tris-HCl, 1 mmol l⁻¹ EDTA, pH 8). PCR was performed to test whether the target region was deleted, in a 20 μl reaction volume containing 10 μl of 2×Es Taq MasterMix (cat. no. CW0690M, CoWin Biosciences, Inc., Cambridge, MA, USA: which contained Es Taq DNA Polymerase, 3 mmol l⁻¹ MgCl₂, 400 µmol l⁻¹ each dNTP and PCR buffer), 0.4 μl (0.4 µmol l⁻¹) each of forward and reverse Primer (Tsingke Biotechnology Co., Ltd, Beijing, China), 1 μl (400 ng) template DNA and ddH₂O. The sequence of one of the primers was the same as the sgRNA (Table S1). The PCR cycling program was as follows: 95°C for 3 min, followed by 35 cycles of 95°C for 15 s, 60°C for 30 s, 72°C for 20 s. PCR products were evaluated by 1% agarose gel electrophoresis at 150 V for 20 min.

Total RNA was extracted using Trizol (cat. no. 15596026, Invitrogen, Waltham, MA, USA) from flies at various developmental stages (Fig. 1A) and 1d adult gonads. First-strand cDNA was synthesized from 2 μg of total RNA using EasyScript first-strand cDNA synthesis SuperMix Kit (cat. no. AT321-01, TransGen Biotech, Beijing, China). qPCR was performed using a Miniopticon system (Bio-Rad, Hercules, CA, USA) with a Platinum SYBR Green qPCR SuperMix (cat. no. Q711-02, Vazyme, Shanghai, China) as described previously (Zheng et al., 2018). Specific primers for tested genes were designed based on the sequences from FlyBase and shown in Table S1. The qPCR cycling programme was as follows: 95°C for 3 min, followed by 40 cycles of 95°C for 10 s, 55–60°C (depending on different primers) for 30 s and 65°C for 5 s. Then, a melting curve was constructed from 55°C to 98°C. The relative expression of the gene was calibrated against the reference gene rp49 using the 2^−ΔΔCT calculation method: ΔΔCt=(Ct_Target−Ct_rp49)cyt-c1L-kd−(Ct_Target−Ct_rp49)control (Livak and Schmittgen, 2001). For the various developmental stages (DS), the 1st instar larval stage (L1), which had the lowest level of transcript of cyt-c1L, was used as a calibrating sample; hence, ΔΔCt=(Ct_Target−Ct_rp49)DS−(Ct_Target−Ct_rp49)L1 was used in this calculation method.

Immunofluorescence staining was conducted as described previously (Mao et al., 2022). For cyt-c1L knockout flies, only Cas9>cyt-c1L^KO male flies with a successful knockout were used. Testes were dissected in phosphate-buffered saline (PBS, pH 7.4) and fixed in 4% paraformaldehyde (in PBS) for 30 min at room temperature (RT). Then, the samples were washed 3 times (15 min each) in PBST (PBS with 0.1% Triton X-100). The samples were blocked for 30 min in 5% normal goat serum at RT, incubated overnight at 4°C with primary antibody, and washed 4 times (each for 15 min) in PBST at RT. The secondary antibody was added and incubated at RT in the dark for 2 h, and the samples were washed 4 times (each for 10 min) in PBST at RT (Perrimon, 2014).

Primary antibodies were used at the following dilutions: rabbit anti-Vasa (1:100, cat. no. AB760351, Developmental Studies Hybridoma Bank, Iowa, IA, USA), mouse anti-spectrin (1:100, cat. no. AB528473, Developmental Studies Hybridoma Bank), rabbit anti-β-tublin (1:200, cat. no. ABP0128, Abbkine, Wuhan, China) (Vedelek et al., 2016), anti-ATP5A and anti-caspase3 (1:100, cat. no. AB14748 and AB32351, Abcam, Boston, MA, USA). Secondary antibodies were used at the following dilutions: phalloidin (1:200, cat. no. BMD00084, Abbkine), rabbit 594 and mouse 488 (1:200, cat. no. A23420 and A23210, Abbkine). All samples were mounted on a glass slide with 4′-6-diamidino-2-phenylindole (DAPI) (2 μg ml⁻¹, cat. no. S2110, Solarbio, Beijing, China) solution (Yang et al., 2017). Fluorescence images were collected using a Leica SP8 laser confocal microscope (Germany).

Western blots were performed following standard methods. Each protein sample from 80 pairs of nosGal4>+ or nosGal4>cyt-c1L-hp fly testes, respectively, was used for gel electrophoresis and then transferred to a nitrocellulose membrane. For blocking, 5% non-fat dried milk dissolved in Tris-buffered saline with 0.05% Tween 20 as the detergent (TBST) was used. The membrane was incubated overnight with the primary anti-rabbit polyclonal antibody caspase-3 (1:1000, cat. no. AB32351, Abcam) in 5% non-fat dried milk and TBST at 4°C (Gong et al., 2021), and then incubated with the secondary horseradish peroxidase-conjugated antibody (dilution 1:5000 in TBST; cat. no. BA1054, Boster, Wuhan, China). Finally, the PVDF membrane was scanned using an Odyssey near-infrared fluorescence scanning imaging instrument, and the grey value was analysed using BandScan (Glyko).

The testes were dissected from 1d nosGal4>+ or nosGal4>cyt-c1L-hp males in PBS and fixed in 2.5% glutaraldehyde (0.1 mol l⁻¹ phosphate buffer, pH 7.4) at 4°C overnight. Then, these testes were washed 3 times (15 min each) in phosphate buffer and post-fixed in 1% OsO₄ for 2–3 h. After double fixation, the samples were dehydrated through an ascending series of ethanol (30%, 50%, 70%, 80%, 85%, 90%, 95%, 100%, 15–20 min for each concentration, doubled for 100%), and then embedded in Araldite (EMbed 812, Electron Microscopy Sciences, Hatfield, PA, USA). Ultrathin sections (80 nm) were collected on copper grids and stained with uranyl acetate and lead citrate. Sections were observed using a Tecnai G2 20 TWIN transmission electron microscope (FEI, Hillsboro, OR, USA) operating at 200 kV.

For terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labelling (TUNEL), 30 pairs of 1d Drosophila testes were dissected and fixed in 4% paraformaldehyde at RT for 40 min, then washed 4 times (15 min each time) in PBST. The samples were then incubated with TUNEL reaction mixture (5 μl TdT enzyme solution and 45 μl fluorescein-dUTP tag solution; Roche, Mannheim, Germany) in a dark environment at 37°C for 3 h. After rinsing 3 times with PBST in the dark, we used an anti-fading medium (Solarbio) containing 2 μg ml⁻¹ DAPI to fix the testes on the glass slide. A Leica SP8 laser confocal microscope was used to observe and take photos, and the fluorescence density (intDen/pixel) of TUNEL was analysed using ImageJ.

Results are presented as means±s.e.m. (n≥3). Differences among various developmental stages were evaluated with a one-way analysis of variance followed by Student's t-test. Differences between means were analysed by Student's t-test. Differences were regarded as statistically significant when P<0.05.

To explore the role of cyt-c1L in Drosophila development, cyt-c1L expression levels at different developmental stages and in the gonads were detected by RT-qPCR. As shown in Fig. 1A, cyt-c1L expression levels were relatively low in embryonic stages though they were significantly higher (P<0.01) than expression levels in the 1st instar larval stage (L1). Levels significantly increased from the 2nd instar larval stage (L2) to the adult stage. In 1d adults, the expression level of cyt-c1L was significantly higher in males than in females. Consistently, there was considerably more cyt-c1L transcript in the testis than in the ovary (Fig. 1B). This result implies that cyt-c1L may play a role in male reproduction.

To examine the role of cyt-c1L in male fertility, we used a cyt-c1L RNAi (cyt-c1L-hp) line. We tested the efficiency of knocking down the gene by RT-qPCR. As shown in Fig. 2A, the cyt-c1L expression level was significantly decreased in the testis of nosGal4>cyt-c1L-hp flies compared with the control (nosGal4>+) (P<0.001). To investigate whether knockdown of cyt-c1L in fly testes affects male fertility, 1d cyt-c1L knockdown adult males (nosGal4>cyt-c1L-hp) were crossed with w¹¹¹⁸ virgin females, and the egg hatching rates from these crosses were calculated. Surprisingly, in three independent trials, the hatching rates of the eggs in groups crossed with cyt-c1L knockdown males were zero, while 91.67±1.26% of eggs from the crosses with control males hatched (Fig. 2B).

Then, we overexpressed cyt-c1L to see whether this could rescue cyt-c1L-hp knockdown male fertility. The hatching rate of eggs from the groups crossed with nosGal4>UAS-cyt-c1L; cyt-c1L-hp males was 67.66±7.73%, significantly higher than that in cyt-c1L knockdown (nosGal4>cyt-c1L-hp) groups (Fig. 2B), and the cyt-c1L expression level was also significantly increased in the testis of nosGal4>UAS-cyt-c1L; cyt-c1L-hp males when compared with that in the cyt-c1L knockdown groups (Fig. 2A), although the gene expression level and the egg hatching rate were still lower than that in control group.

To confirm the results from the knockdown experiments, and to study the function of cyt-c1L in male fertility in more detail, we used a cyt-c1L^KO fly line. Taking advantage of CRISPR-Cas9 technology, a small fragment of cyt-c1L was removed from the genome. We found that when the target fragment was successfully deleted (Fig. 2C), the male flies (Cas9>cyt-c1L^KO) were completely sterile. When the males were crossed with w¹¹¹⁸ virgin females, the egg hatching rates were zero (Fig. 2D). However, in some of the flies, the target DNA region was not completely excised, and these males had a certain level fertility (Fig. 2C,D). Therefore, we used the sterile males in which the fragment was successfully excised as knockout flies (Cas9>cyt-c1L^KO) for the testis staining experiment.

To avoid the possibility that the phenotypes caused by cyt-c1L knockdown or knockout arise from off-target effects, we examined the expression of its parent gene cyt-c1 in nosGal4>cyt-c1L-hp and successful knockout Cas9>cyt-c1L^KO male flies. RT-qPCR analysis showed that the expression of cyt-c1 was not significantly influenced in cyt-c1L knockdown or knockout flies (Fig. S2). These results indicate that cyt-c1L is essential for male fertility in D. melanogaster.

The fact that cyt-c1L knockdown in fly testes or knockout led to male sterility is reminiscent of defects in spermatogenesis. We thus dissected and stained fly testes with anti-Vasa antibody and DAPI to investigate the process of sperm production in D. melanogaster. There were no obvious differences in overall appearance of cyt-c1L knockdown or knockout testes from the control. The germ cells at various developmental stages were normally distributed in testes (Fig. 3A,B,E,F,I,J). However, in the basal region of the control testes, we observed many tightly bundled spermatid nuclei with a typical umbrella-like shape (arrows in Fig. 3C). The seminal vesicles were full of mature sperm with needle-like heads (Fig. 3D). In contrast, in cyt-c1L knockdown or successful knockout testes, there were less tightly grouped/more scattered spermatid nucleus bundles (arrows in Fig. 3G,K). In particular, there were no mature sperm at all in the seminal vesicles of cyt-c1L-hp knockdown or successful knockout testes (Fig. 3H,L). Thus, these results indicate that cyt-c1L plays an important role in spermatogenesis. We used nosGal4>cyt-c1L-hp male flies alone for the following experiments.

The male germline cells at different developmental stages contain subcellular structures called fusomes, which can be detected by antibodies against α-spectrin. GSCs and gonialblasts each contain a single round fusome. With germ cell division and growth, the fusomes appear progressively wider and more branched. During elongation, the spectrin scaffold of fusomes changes to become the elongation cones (ECs) at the growing end of the cyst (Hime et al., 1996). To investigate why mature sperm cannot be produced in cyt-c1L knockdown testes, we stained testes with anti-spectrin antibody. There were round fusomes at the tip of both control and cyt-c1L knockdown testes (arrowheads in Fig. 4B,E), and wider and branched fusomes in spermatocytes (arrows in Fig. 4B,E). This is consistent with Vasa staining in Fig. 3, reflecting the regular mitosis and growth in early stages. We also clearly observed spectrin signals marking the growing ends of the elongating spermatids, and some of the ECs had moved to the middle and head regions of both control and cyt-c1L knockdown testes (white arrows in Fig. 4A,D), indicating that spermatid elongation seems not to be impaired by cyt-c1L knockdown.

As the early stages of spermatogenesis were not apparently disrupted, and the spermatid nuclear bundles (despite being deformed) were visible in cyt-c1L knockdown testes, we asked whether the late stages of spermatogenesis were impaired. After completing spermatid elongation, the cyst containing the elongated spermatids undergoes individualization, which requires the formation of ICs composed of 64 actin cones. To visualize spermatid individualization, we used phalloidin to label the actin cones, and thus the ICs. In control testes, the ICs were correctly assembled around the nuclei at the base of the testes, and some had moved to middle and head regions in the testes (yellow arrows in Fig. 4A,C). However, ICs did not appear at all in nosGal4>cyt-c1L-hp testes (Fig. 4D,F). The final stage of spermatid differentiation involves the removal of bulk cytoplasm. During individualization, ICs progressively move from the sperm head toward the tip of the tail, striping off the unneeded cytoplasm and organelles into the cystic bulges (CBs) that ultimately form waste bags (WBs) at the end of individualization in the anterior end of the testis. Caspase activation in the CBs and the WBs has been shown to be required for sperm individualization (Fabian and Brill, 2012; Nakajima and Kuranaga, 2017). To further investigate whether cyt-c1L knockdown affected sperm individualization, we detected CBs and WBs with an antibody against caspase-3 protein. This showed that CBs and WBs could be visualized clearly in the control testis (Fig. 4G,I). In cyt-c1L knockdown testes, however, no CB or WB structures were apparent (Fig. 4J,L). This indicates that the knockdown of cyt-c1L impeded IC formation and thus prevented sperm individualization.

During spermatid elongation, the central axoneme flanked by major and minor mitochondrial derivatives extends to form the flagellum. To test whether the sperm flagellum was correctly formed, we stained testes with β-tubulin antibody and found that the well-parallelized axenomal microtubules were regularly bundled and were wrapped by actin cones near the sperm head region in control testes (Fig. 5A–C). By contrast, in nosGal4>cyt-c1L-hp testes, the knotted microtubules were not well parallelized and showed uneven thickness (Fig. 5E), and even in the basal region there were no ICs around the microtubule bundles (Fig. 5F). This result implies that knockdown of cyt-c1L causes defects in the formation of the flagellum.

To further investigate the defects during flagellum formation in cyt-c1L knockdown flies, we obtained ultrathin slices of the testes and characterized the abnormal morphogenesis during spermatogenesis by TEM. In control testes, the normal axonemes, containing an ordered ‘9+2’ microtubule structure (arrowheads in Fig. 5G–I), associated with two mitochondrial derivatives, were clearly observed in the spermatid. The major mitochondrial derivative was gradually deposited with dense paracrystalline material (arrows in Fig. 5G–I). In the cyt-c1L knockdown testes, in the early stages of elongation, the axoneme, associated with mitochondrial derivatives, could be found although the dense material deposition was scattered (arrows in Fig. 5J). As elongation progressed, the spermatid exhibited multiple types of disorganization. First, the two mitochondrial derivatives did not show one major and one minor form, but instead looked similar in size (arrows in Fig. 5K). Second, the axoneme and nebenkern were not paired one to one. The nebenkern appeared in triangular, quadrilateral or horseshoe shapes (arrows in Fig. 5L), rather than the black half circle shape seen in control testes (arrow in Fig. 5I). Third, several axonemes (arrowheads in Fig. 5L) were distributed within one cell membrane with or without paired mitochondrial derivatives. These data suggest that the IC movement was defective and that mitochondrial morphogenesis was disrupted in cyt-c1L knockdown fly testes.

Given that cyt-c1L localizes to the mitochondria and mitochondria-derived nebenkern assembly was disturbed by cyt-c1L knockdown, we asked whether cyt-c1L knockdown affected the oxidative phosphorylation pathway associated with mitochondria. To address this, we detected the expression of ATP5A (also named bellwether, blw), the alpha subunit of the mitochondrial F₁F₀ ATP synthase complex (complex V), which is the final enzyme of the oxidative phosphorylation pathway. In control testes, ATP5A was highly expressed in spermatocytes and round spermatids (Fig. 5M). In contrast, in cyt-c1L knockdown testes, weak ATP5A signal was detectable (Fig. 5P). Furthermore, unlike the continuous and uniform ATP5A synthase signal in elongated spermatids in the control testes (Fig. 5N), that of the cyt-c1L knockdown testis was intermittent (arrows in Fig. 5Q). Under phase contrast microscopy, elongated tightly bundled spermatid tails (arrow in Fig. 5O) could be clearly seen in control testes, but only loose spermatid tails with numerous vacuolated mitochondria (arrows in Fig. 5R) were observed in nosGal4>cyt-c1L-hp testes. These results indicate that the lack of cyt-c1L results in swelling and fracture of mitochondria during spermatid elongation in D. melanogaster.

Although there were spermatid nuclei bundles (albeit abnormal) in the base of cyt-c1L knockdown fly testes, no sperm were observed in the seminal vesicle. Considering the defects in mitochondrial dynamics observed during spermatid elongation, we speculate that cell death could be induced. We thus examined testes using the TUNEL technique. The results showed that the fluorescence signals of TUNEL in cyt-c1L knockdown testes were significantly stronger (Fig. 6D,F) than those in the control testes (Fig. 6A,C). In addition, the TUNEL signal was mainly concentrated on the spermatid nuclei bundles in the base of the cyt-c1L knockdown testes (Fig. 6D–F). The fluorescence intensity was significantly higher in cyt-c1L knockdown testes than in the control (Fig. 6G, P<0.01).

Multiple caspases and caspase regulators also participate the process of apoptosis. For example, caspase 3 plays a crucial role in the apoptosis process. Western blot analysis of testes also showed an apparent increase in cleaved caspase 3 proteins (17 kDa, activated caspase 3) in cyt-c1L knockdown fly testes relative to control testes (Fig. 6H). The fluorescence intensity of the cleaved caspase 3 (17 kDa) signal was significantly higher in cyt-c1L knockdown testes than in control testes (Fig. 6I, P<0.05). In Drosophila, the combined action of Dark, coding for death-associated APAF1-related killer, and Dronc results in the formation of apoptotic bodies, which initiates the proteolytic cascade to activate the effector caspases Dcp-1 and Drice, and eventually lead to apoptosis (Nakajima and Kuranaga, 2017; Shinoda et al., 2019). Fadd is an adaptor that mediates the recruitment of cysteine proteases to ligand-bound death receptors, thereby promoting caspase activation (Huh et al., 2004). Bcl-2 (Debcl) encodes a pro-apoptotic member of the Bcl-2 family involved in programmed cell death. Therefore, we tested the expression of these six apoptosis-related genes, including Dark, Dcp-1, p53, Fadd, Dredd and Bcl-2. RT-qPCR analysis showed that their expression levels were significantly increased in cyt-c1L knockdown testes relative to control testes (Fig. 6J). Therefore, these data suggest that knockdown of cyt-c1L results in excessive apoptosis in the testis of D. melanogaster.

In this study, we observed the loss of F-actin cones in ICs in the cyt-c1L knockdown testes. It has been reported that mutants for proteins associated with cytoskeleton assembly possibly affect spermatid individualization (Arama et al., 2006; D'Brot et al., 2013; Wu et al., 2016). We detected some genes encoding cytoskeleton-related proteins, including Klc (kinesin light chain), Dhc98D (dynein heavy chain at 89D), Dhc62B (dynein heavy chain at 62B), Dhc93AB (dynein heavy chain at 93AB), Ddlc1 (dynein light chain 1), DCTN1-p150 (the p150 subunit of Dynactin 1) and Btv (the minor cytoplasmic dynein heavy chain). Knockdown of cyt-c1L caused a significant reduction in expression of all these genes compared with controls (Fig. 7). Additionally, some studies showed that D. melanogaster with a parkin mutation also exhibits male sterility with similar phenotypes (Darios et al., 2003; Greene et al., 2003). Thus, we tested the expression levels of some mitochondrion-related genes, such as parkin and ATPsysGL in cyt-c1L knockdown fly testes and found that the expression levels of these genes were also significantly decreased as a result of the depletion of cyt-c1L in testes (Fig. 7). These results indicate that the knockdown of cyt-c1L disrupts expression of these cytoskeleton-related genes, consequently affecting sperm individualization and mature sperm production.

Spermatogenesis in Drosophila is a complex and continuous process that requires the cooperation of multiple gene functions to maintain the integrity of the entire process. Drosophila cyt-c1L, as a subunit of the mitochondrial respiratory chain complex III, participates in mitochondrial ATP synthesis and exchange proton transport. In this study, we used the UAS/Gal4 system to explore the role of cyt-c1L in male fertility and found that it is essential for spermatogenesis in D. melanogaster.

cyt-c1L, a duplicate of cyt-c1, is a nuclear-encoded mitochondrial gene in D. melanogaster that is transcribed in the nucleus and translated in the cytoplasm; the resulting proteins then enter the mitochondria to function. Cyt-c1L is a subunit of mitochondrial respiratory chain complex III, responsible for transferring electrons from CoQH2 to cytochrome c (Cyt-c), participating in the synthesis of mitochondrial ATP. Analysis of the D. melanogaster genome has revealed that all the nuclear-encoded mitochondrial gene duplicates with tissue-biased expression have testis-biased expression (Gallach et al., 2010; Eslamieh et al., 2017). Many of these genes code for proteins that are Drosophila sperm proteins (Wasbrough et al., 2010). In this study, we first found that cyt-c1L expression level was generally lower in early developmental stages, including embryos and the 1st and 2nd instar larval stages, but then gradually increased till the adult stage. It is much higher in fly testes than in ovaries. This expression pattern is different from that of its parent gene, cyt-c1 (CG4769), which is expressed in all developmental stages and almost all tissues of D. melanogaster, but is consistent with the RNA-Seq data of cyt-c1L presented in FlyBase (https://flybase.org/reports/FBgn0039651.html). The trait of testis-biased expression indicates its role in male reproduction. By knockdown, rescue and knockout experiments, we indeed showed that cyt-c1L plays a critical role in spermatogenesis, and its expression level in fly testes is positively related to male fertility. Although the phenotypes induced by depletion of cyt-c1L are not exactly the same as those in ocn knockdown fly testes, where the defects are more serious, both conditions affect sperm production. This suggests that cyt-c1L functions in fly spermatogenesis but not under the direct control of ocn.

Some nuclear-encoded mitochondrial gene duplicates have been demonstrated to be critical in spermatogenesis in D. melanogaster. For example, an amino acid substitution (Ala278→Thr) in cytochrome B of complex III leads to male sterility as a result of the failure of IC formation (Clancy et al., 2011). A single amino acid mutation in Cytochrome Oxidase I (R301L) causes a defect in sperm storage in females and thus male sterility, and R301S males are also sterile because of a defect in spermatogenesis (Xu et al., 2008). Recently, COX4L, a duplicate of the Cytochrome c oxidase 4 (COX4) gene, was demonstrated to be required for spermatogenesis. A defect in sperm individualization was observed in both COX4L knockdown and knockout fly lines (Eslamieh et al., 2022). In addition, mitochondrial iron metabolism plays an essential role in Drosophila spermatogenesis, and mitoferrin gene deficiency also leads to male sterility (Metzendorf and Lind, 2010; Bettedi et al., 2011). Our previous work revealed that knockdown of ATPsyn-b, coding for the b subunit of ATP synthase, resulted in disrupted nuclear bundles during spermatogenesis and oval nuclei arrested in canoe stage (Chen et al., 2015). In the present study, we found that knockdown of cyt-c1L caused male sterility with no mature sperm in the seminal vesicle. Cytological analysis revealed that no ICs were observed during spermatogenesis. These phenotypes were in accord with those caused by substitution (Ala278→Thr) in the Cyt B of complex III (Clancy et al., 2011) and loss of function of COX4L (Eslamieh et al., 2022). In addition, we also showed that during spermatid elongation, mitochondrial morphogenesis was disrupted by cyt-c1L knockdown in fly testes, though the axenome was seemingly normal and long axenomal tubules could be observed (Fig. 5E,F). Axoneme and nebenkern elongation during spermatogenesis can proceed independently (Fabian and Brill, 2012), which may explain the appearance of elongation caps moving to the anterior region of the testis.

We found two downregulated genes in cyt-c1L knockdown testes: mitochondrial protein coding gene bb8, which encodes glutamate dehydrogenase involved in sperm mitochondrial development (Vedelek et al., 2016), and parkin, which prevents mitochondrial swelling (Darios et al., 2003). Similar defects in mitochondrial morphogenesis during spermatogenesis observed in cyt-c1L knockdown testes were also seen in bb8 or parkin mutants (Vedelek et al., 2016; Greene et al., 2003; Riparbelli and Callaini, 2007). Mitochondria are the most efficient source of ATP, which is required for the growth of germ cells. We found that ATP synthases, including ATP5A and ATPsynGL expression, were significantly depleted when cyt-c1L was knocked down, which could cause an energy deficit. Other studies have revealed that the depletion of ATP synthase subunits or energy deficiency results in severe defects in spermatogenesis (Chen et al., 2015; Vedelek et al., 2016; Sawyer et al., 2017; Yu et al., 2019). Given the fact that this duplicate, cyt-c1L, has energy-related functions, these findings suggest that an energy deficit, caused by loss of this nuclear-encoded mitochondrial gene duplicate, may inhibit proper mitochondrial elongation and the initiation of spermatid individualization in D. melanogaster (Fig. 8). Our findings support the hypothesis that the testis-biased expression pattern of the energy-related duplicated genes in the fly genome might have evolved for resolving intralocus sexually antagonistic conflict at the parental gene (Gallach et al., 2010). Because of the way spermatogenesis and fertilization occur in Drosophila, the testis has evolved a special set of mitochondrial genes, such as cyt-c1L, for roles in higher/specialized energy production that would make the sperm more competitive. But these genes might be detrimental to the soma and/or female functions; therefore, they show a testis-biased expression pattern.

Both TUNEL staining and caspase 3 detection revealed excessive apoptosis occurring in cyt-c1L knockdown fly testes in comparison to the control testes. RT-qPCR results also demonstrated that some genes related to the apoptosis process, including Dark, Dcp-1, p53, Fadd, Dredd, Bcl-2, were all significantly upregulated when cyt-c1L was knocked down. Greene et al. (2003) found a dramatic increase in TUNEL-positive nuclei in the major flight muscles of 1d adult parkin mutants relative to age-matched control flies; thus, they suggest that the muscle mitochondrial defects cause cell apoptosis, resulting in muscle degeneration. They also found a male sterile phenotype as a result of a defect in spermatid individualization (Greene et al., 2003), and Riparbelli and Callaini (2007) found irregular mitochondrion dynamics during spermatid elongation, but did not detect apoptosis in fly testes. Here, we observed similar defective phenotypes of mitochondrial morphogenesis during spermatogenesis, raising the possibility that the abnormal mitochondrial dynamics at the late stage of spermatogenesis might lead to the excessive apoptosis in the spermatid bundles in the base of cyt-c1L knockdown fly testes.

Some cytoskeletal regulatory proteins regulate not only spermatid elongation but also individualization. For example, Ddlc1 is required for tight bundling of the elongated nuclei and proper actin cone organization. Furthermore, Dhc proteins also function in maintaining nuclear positions at the beginning of sperm individualization (Ghosh-Roy et al., 2005). Btv encodes the minor cytoplasmic dynein heavy chain, used as the motor of retrograde transport in flagella (Pascale et al., 2011). The dynein–dynactin complex plays an essential role in positioning the F-actin cones around nuclei (Ghosh-Roy et al., 2005; Wu et al., 2016). Normally, the bundling of nuclei indicates the beginning of IC formation. However, this study found that the spermatids were not tightly bundled, and the ICs composed of actin cones were utterly lost. Furthermore, the transcriptional level of all these protein-coding genes, including Ddlc1, Dhc62B, Dhc98D, Dhc93AB, btv and Dhc1, was significantly downregulated in cyt-c1L knockdown fly testes. These findings imply that cyt-c1L plays an essential role in forming ICs by influencing the expression of these cytoskeletal regulators during the late steps of spermatogenesis. Whether and how cyt-c1L affects expression of these genes needs to be further investigated. Taken together, our data suggest that cyt-c1L might play an essential role in mitochondrial morphogenesis and sperm individualization during spermatogenesis in D. melanogaster.

In this study, we knocked down cyt-c1L in fly testes and demonstrated that it has a critical function in spermatogenesis. Depletion of cyt-c1L resulted in male sterility, with no mature sperm appearing in the seminal vesicle. Knockdown of cyt-c1L caused severe defects in mitochondrial morphogenesis and the formation of ICs, probably through inhibition of the production of ATP by oxidative phosphorylation, thus inducing excess apoptosis of abnormal spermatids. These results suggest that cyt-c1L plays an important role in spermatid morphogenesis and individualization, and consequently male fertility. Our findings support the hypothesis that the mitochondria-related gene duplicates have been retained for a special role in spermatogenesis and/or fertilization in males, which could be involved in the special requirement for energy production.

We thank Professor Zhaohui Wang (Institute of Genetics and Developmental Biology, CAS) for providing nosGal4 flies, Professor Shan Jin (Hubei University) for providing Vasa-Cas9 flies. We are grateful to Bi-Chao Xu of the Core Facility and Technical Support, Wuhan Institute of Virology, for her technical support in ultrathin sections and micrography for TEM. We sincerely acknowledge the microinjection and screening process performed by the Core Facility of Drosophila Resource and Technology, CEMCS, CAS.