Mob4 is essential for spermatogenesis in Drosophila melanogaster

Abstract

Gamete formation is essential for sexual reproduction in metazoans. Meiosis in males gives rise to spermatids that must differentiate and individualize into mature sperm. In Drosophila melanogaster, individualization of interconnected spermatids requires the formation of individualization complexes that synchronously move along the sperm bundles. Here, we show that Mob4, a member of the Mps-one binder family, is essential for male fertility but has no detectable role in female fertility. We show that Mob4 is required for proper axonemal structure and its loss leads to male sterility associated with defective spermatid individualization and absence of mature sperm in the seminal vesicles. Transmission electron micrographs of developing spermatids following mob4^RNAi revealed expansion of the outer axonemal microtubules such that the 9 doublets no longer remained linked to each other and defective mitochondrial organization. Mob4 is a STRIPAK component, and male fertility is similarly impaired upon depletion of the STRIPAK components, Strip and Cka. Expression of the human Mob4 gene rescues all phenotypes of Drosophila mob4 downregulation, indicating that the gene is evolutionarily and functionally conserved. Together, this suggests that Mob4 contributes to the regulation of the microtubule- and actin-cytoskeleton during spermatogenesis through the conserved STRIPAK complex. Our study advances the understanding of male infertility by uncovering the requirement for Mob4 in sperm individualization.

Introduction

The many causes of male infertility include defects in spermatogenesis. However, there is still limited knowledge of the molecular and cellular processes that regulate sperm production and that lead to its failure in humans (Tüttelmann et al. 2018). Drosophila melanogaster offers a powerful model in which to uncover genes that execute and regulate spermatogenic processes, and many of the genes involved in spermatogenesis in Drosophila are conserved in humans (White-Cooper 2010; Siddall and Hime 2017).

In Drosophila testes, cysts of 16 primary spermatocytes undertake meiosis to generate cysts of 64 syncytial haploid spermatids (reviewed in Fuller 1998). Cells within each cyst are interconnected by cytoplasmic bridges due to incomplete cytokinesis (Lin et al. 1994). The synchronous differentiation of spermatids within each cyst requires formation of flagellar axonemes and acrosomes, remodeling of mitochondria and nuclei, accompanied by elongation of the developing sperm and their plasma membranes (Tates 1971; Lindsley and Tokuyasu 1980). The fully elongated syncytium of 64 spermatids then undergoes a membrane remodeling process known as individualization, which requires that actin cones form around each of the 64 nuclei that assemble into an individualization complex (IC) (Fuller 1998). The IC moves processively along the spermatid bundle, stripping away unnecessary organelles and cytoplasm, forming the “cystic bulge” (Tokuyasu et al. 1972) and so remodeling the cyst membrane and resolving intercellular bridges in order that each sperm cell becomes encased in its own plasma membrane (Fabrizio et al. 1998). By the end of the individualization process, the cystic bulge is disposed of in the “waste bag” in a caspase-dependent apoptosis-like event, leaving the now fully individualized mature sperm.

The striatin-interacting phosphatase and kinase (STRIPAK) complex is an evolutionarily conserved supramolecular complex with functions in cell proliferation, migration, vesicular transport, cardiac development, and immune regulation (Goudreault et al. 2009; Hwang and Pallas 2014; Neisch et al. 2017). Although many isoforms and paralogs of STRIPAK subunits can assemble into various STRIPAK variants, the core STRIPAK complex of Drosophila includes striatin (Cka), protein phosphatase 2A (Pp2A-29B, Mts), protein kinase (Gck-III), and regulatory and structural subunits (Strip, Mob4). These core subunits are structurally conserved in all eukaryotes. In Drosophila and mammals, STRIPAK complexes act as platforms to integrate upstream inputs to the Hippo pathway (Chen et al. 2019). Several STRIPAK components have been reported to suppress Hippo signaling, thereby regulating cell proliferation (Couzens et al. 2013; Bae et al. 2017; Zheng et al. 2017).

Mob4 can inactivate the Hippo pathway as a core component of STRIPAK complex, resulting in enhanced expression of growth-promoting genes (Chen et al. 2018). Mob4 also regulates neurogenesis and synapse formation (Baillat et al. 2001; Florindo et al. 2023) and is required for both centrosome separation and focusing of kinetochore fibers (Trammell et al. 2008). Within the STRIPAK complex, Mob4 acts as a molecular glue to tether STRN1 (the mammalian homolog of Cka) to the main body of STRIPAK. Mob4 point mutations targeting phospho-threonine binding sites to Strip1 (the mammalian homolog of Strip), decrease the formation of STRIPAK, leading to deregulated Hippo signaling (Jeong et al. 2021).

Here, we report an essential role for Mob4 in spermatogenesis in D. melanogaster and show that the human Mob4 homolog can rescue the Drosophila phenotypes, indicating conserved molecular function.

Materials and methods

Fly husbandry and genetics

All Drosophila stocks used in this study were maintained on standard cornmeal–yeast–sucrose media and are listed in Supplementary Table 1. Unless otherwise indicated, experiments were performed at 25°C. The w¹¹¹⁸ flies were used as wild-type strain. The mob4^P mutant allele was generated using the P-element remobilization technique, in the line y; KG4509 P[y^mDint2w⁺ BR.E.BR = SUPor-P] (Gene Disruption Project, Bellen et al. 2004). The mob4^SVC mutant allele was generated by CRISPR/Cas9-catalyzed homology-directed repair by the sgRNA pair targeted by 5′-GGATCATCGCGTCTTCGAGC-3′ and 5′-GCGAGTCTGACTAATCTGGG-3′ sequences (Gratz et al. 2014).

Tissue-specific genetic manipulation

For early germ cell-specific gene depletion, flies carrying the nanos (nos)-GAL4:VP16 or bam-GAL4:VP16 were crossed to flies carrying dsRNA under UAS control. Control flies were the progeny of the cross between the driver and UAS-mCherry^RNAi.

Viability assay

Ubiquitous depletion was accomplished by crossing flies carrying the daughterless (da)-GAL4 driver with flies carrying dsRNA against the genes of interest. Each cross was established with 5 virgin females (aged 1–3 days) and 5 age-matched males. Mating was allowed for 3 days, and then flies were discarded. The number of pupae and the number of eclosed adults were scored in 6 independent crosses, and viability is expressed as the percentage of flies that eclosed from the pupae.

Fertility assay

To test female fertility, virgin female flies (aged 1–3 days) depleted for the genes of interest (see tissue-specific genetic manipulation) were collected and individually mated with 2 wild-type age-matched males. The crosses were kept at 25°C for 6 days, after which the adults were removed. The number of eclosed progeny in each vial was scored. Male fertility was similarly tested as described; individual males depleted for the genes of interest were mated with 2 wild-type virgin females. Vials in which any of the adults died were discarded. Ten crosses per genotype were evaluated in 2 independent experiments.

Statistical analysis

The sample size (n), mean, SEM, and number of times each experiment was repeated (N) are indicated in corresponding figure legends. Data collected from viability and fertility assays were analyzed with a 2-tailed, unpaired t-test by GraphPad Prism version 7.03 for Windows. P-values for each analysis are indicated in corresponding figure legends; a 99% confidence interval was applied in all statistical tests.

Immunostaining of whole-mount testes

Testes from unmated young adults (2-day-old males unless otherwise indicated) were fixed in 4% formaldehyde/ phosphate-buffered saline (PBS) for 20–30 min. Samples were washed 3 times in PBST for 20 min, and blocked in PBST/10% SFB/10% BSA for 1 h. The samples were incubated overnight in a humid chamber at 4°C with the following primary antibodies: mouse anti-pan polyglycylated tubulin (1:1000, Axo-49; Sigma) and rabbit anti-cleaved caspase 3 (1:500, Cell Signaling #9661). DNA and actin were stained with 1 µg/mL DAPI and phalloidin (1:1000, Flash Phalloidin™ Red 594; BioLegend).

Confocal microscopy

Images were acquired on a LSM710 confocal microscope (Zeiss). All processing and analysis of microscope images was performed with ImageJ 1.53q (Schneider et al. 2012). Figures were created using the QuickFigures2 plugin for FIJI/ImageJ (Mazo 2021).

Transmission electron microscopy

Testes from control and mob4^RNAi young males were dissected in PBS pH 7.2, and fixed in 2.5% glutaraldehyde in PBS overnight at 4°C. Samples were then fixed in 1% osmium tetroxide in PBS for 1–2 h at 4°C and dehydrated through a graded series of ethanol, infiltrated with a mixture of Epon-Araldite resin and polymerized at 60°C for 48 h. Silver-gray sections (70-nm thick) were cut and stained with 2% aqueous uranyl acetate for 20 min in the dark and then with lead citrate for 2 min. TEM preparations from control (n = 11) and mob4^RNAi (n = 14) males were observed with a Tecnai G2 Spirit EM equipped with an Osis Morada CCD camera.

GFP-Trap affinity purification

The 0–2 h embryos expressing GFP-only (negative control) or GFP-mob4 were dechorionated and snap-frozen in liquid nitrogen. The protocol for affinity purification of protein complexes from Drosophila embryos was previously published (Lipinszki et al. 2014). Isolated proteins were digested with trypsin, and the resulting peptide samples were analyzed using an Orbitrap-LTQ mass spectrometer (Thermo Fisher Scientific). Acquired data were searched using the Mascot program (Matrix Science) against the D. melanogaster database.

Results

Lack of mob4 causes lethality and male sterility in D. melanogaster

To study the potential functions of the mob4 gene, we generated 2 different mob4 mutant alleles (Fig. 1a). The first, mob4^P, generated by P-element mutagenesis, lacks most of the first exon, including the start codon. Homozygous mob4^P animals died during the third larval instar stage, but when heterozygous against a deficiency (mob4^P/Df), adult male flies were formed that were sterile whereas the adult females were fertile. Thus, the mutagenized chromosome carries a second site lethal mutation and loss of the N-terminal part of Mob4 results in male sterility. The second mob4 mutant allele (mob4^SVC) lacks the entire coding region and showed third instar lethality when homozygous or heterozygous to a deficiency lacking mob4. The lethality of mob4^SVC flies can be rescued by ubiquitous expression of wild-type GFP-mob4, indicating that the lethality is a consequence of the mutation in mob4.

Fig. 1.

Mob4 is essential for viability and male fertility in Drosophila melanogaster. a) Schematic of Mob4 gene and mutant alleles (red). The mob4^P deletion allele was generated by imprecise excision of the P-element in the line KG4509, and the mob4^SVC deletion allele was generated by CRISPR/Cas9-targeted mutagenesis. The endpoints of deletion in mob4^P and mob4^SVC were determined by sequencing and are indicated in basepairs (bp). b) Male fertility test of control (mCherry^RNAi), mob4^RNAi and rescued flies (mob4^RNAi and hmob4, both under UAS control). Males of each genotype were individually mated with wild-type females. Data points represent number of progeny from individual males. Means ± SEM are shown for at least n = 8 males per genotype. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level); ****P < 0.0001, ***P < 0.001, ns, not significant. c) Fertility test of control and mob4^RNAi females. Females of each genotype were individually mated with wild-type males. Data points represent numbers of progeny from individual females. Means ± SEM are shown for n = 10 females per genotype. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level). d) Viability of control, mob4^RNAi and rescued flies (as in b). Data points represent percentage of viable flies in 6 independent experiments. The da-GAL4 driver was crossed either to UAS-mCherry^RNAi (n = 439), UAS-mob4^RNAi (n = 338), UAS-mob4^RNAi; UAS-mCherry^RNAi (n = 201), or UAS-mob4^RNAi; UAS-hmob4 (n = 286). Means ± SEM are shown. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level).

Intrigued by the sterility of mob4^P/Df males, we sought to analyze the role of mob4 during spermatogenesis. To this end, we used 2 transgenic UAS-mob4^RNAi lines in which mob4 was downregulated in the male germline by 2 different drivers: nos-GAL4, which induces target expression in germinal stem cells and spermatogonia, and bam-GAL4, which acts in late spermatogonia and early spermatocytes. Identical results were obtained for both UAS-mob4^RNAi lines with both drivers, and therefore we present only the results obtained with the nos-GAL4 driver. Downregulation of mob4 in the male germline (hereafter mob4^RNAi) also resulted in male sterility (Fig. 1b), while depletion in the female germline had no effect on fertility (Fig. 1c), thus recapitulating the mob4^P/Df phenotype. Ubiquitous depletion of mob4 resulted in complete lethality at the late larval/pupal stage, phenocopying the mob4 null mutant phenotype (Fig. 1d). Together, these results indicate that mob4 is required for male fertility, but not for female fertility.

We found that all phenotypes of mob4 depletion could be rescued by provision of human MOB4 (hmob4, HGNC:17261) (Fig. 1), which is resistant to dsRNA directed against the Drosophila gene, indicating that the lethality and male sterility phenotypes observed upon mob4^RNAi are not off-target responses.

Mob4 is required for spermatid individualization

To explore the mechanisms underlying the sterility phenotype, we visualized spermatid axonemes by immunostaining in testes from 2-day-old mob4^RNAi males. In control testes, axonemes could be observed throughout the testes length (elongated cysts), in the terminal epithelium (cyst coiling) and inside the seminal vesicles (where mature sperm is stored until mating) (Fig. 2a). By contrast, cysts of mob4^RNAi testes were fully elongated, but their seminal vesicles were empty (Fig. 2d). In addition, most of the mob4^RNAi sperm bundles accumulated in the terminal epithelium and showed increased accumulation of coiled cysts. On the other hand, nuclear morphology appears to be normal at the round, canoe, and needle stage in mob4^RNAi spermatids, suggesting that both nuclear condensation and remodeling progress normally (Supplementary Fig. 1).

Fig. 2.

Individualization defects in mob4 depleted testes. Control (mCherry^RNAi) (a) and mob4^RNAi (d) testes from 2-day-old males were immunostained with antipolyglycylated tubulin (green) to visualize axonemes in individualizing/coiling cysts or mature sperm. The seminal vesicle (SV) and terminal epithelium (TE) are highlighted in white and pink, respectively. Control (b) and mob4^RNAi (e) testes were stained with phalloidin-594 and immunostained to reveal cleaved caspase 3 (cyan). Cystic bulges (CB, white arrows) and waste bags (WB, white arrowheads) resulting from spermatid individualization can be seen in control testes (b­–c) but are absent from mob4-depleted testes (e–f). Actin cones and spermatid bundles from control (c) and mob4^RNAi (f) testes visualized with phalloidin-594 and immunostained to reveal polyglycylated tubulin (green).

Spermatid individualization occurs in the final stage of spermiogenesis and is mediated by ICs (Lindsley and Tokuyasu 1980; Fuller 1993). The actin cones appeared to associate normally with spermatid nuclear bundles in mob4^RNAi animals (Supplementary Fig. 1). However, whereas in control testes, the distribution of actin cones suggested uniform movement along the sperm bundles forming the cystic bulge (Fig. 2c), in mob4^RNAi testes, actin cones were scattered along the sperm bundles suggesting loss of migration synchrony (Fig. 2f) and a failure of individualization.

As multiple caspases and caspase regulators act in nonapoptotic roles to mediate spermatid individualization (Arama et al. 2003), we looked whether effector caspases become activated in mob4^RNAi testes by immunostaining with anticleaved caspase-3 (CC3). In control testes, CC3 signal was mostly restricted to the cystic bulge and the waste bags (Fig. 2b). By contrast, there were neither cystic bulges nor waste bags in mob4^RNAi testes, although there was caspase activation throughout the whole length of the spermatid tail (Fig. 2e).

Together, these results suggest that in the absence of Mob4, cyst elongation occurs, but disruption of the ICs prevents spermatid individualization and migration of sperm into the seminal vesicle.

Axoneme disruption following mob4^RNAi

To determine defects in spermiogenesis that might underlie the failure of spermatid individualization in mob4^RNAi males, we examined the ultrastructure of the developing axoneme in 2-day-old males by transmission electron microscopy (TEM). During the elongation phase, the developing axoneme and major and minor mitochondrial derivatives of wild-type spermatids are encompassed by a plasma membrane, known as the ciliary sheath (Fig. 3a) (Fabrizio et al. 1998). The major mitochondrial derivative is filled with an electron dense material of the paracrystalline body. The elongation of the 2 mitochondrial derivatives is required for the elongation and function of the flagellum (Tates 1971; Noguchi et al. 2011).

Fig. 3.

Electron microscopy reveals aberrant axonemal structure and mitochondrial defects in mob4 depleted spermatid cysts. a) Transverse section of control (mCherry^RNAi) elongating spermatids showing the major mitochondrial derivative (MM) containing paracrystalline material and the minor mitochondrial derivative (mM) adjacent to the axoneme. Accessory microtubules are in the vicinity of the mitochondrial derivatives. b) mob4^RNAi spermatids with incomplete axonemes (red arrow points to missing microtubule doublets). c) mob4^RNAi spermatids with “expanded” axonemes showing loss of linkage between microtubule doublets. d) mob4^RNAi spermatids sharing the same ciliary sheath and spermatids where both mitochondrial derivatives accumulate paracrystalline material. (a'–d') shows the schematic representation of (a–d). Scale bar represents 500 nm.

We found that the elongating cysts of mob4^RNAi males displayed defects in axonemal structure before spermatid individualization was initiated. These defects included the loss of microtubule doublets (Fig. 3b) or preservation of stereotypical 9 + 2 microtubule-doublets of the axoneme but with radial expansion (Fig. 3c). Such abnormal axonemes were typically associated with mitochondrial derivatives of abnormal shape and/or size. Moreover, we also observed elongating spermatid cysts containing 2 paracrystalline body-filled mitochondrial derivatives paired with 1 axoneme, as well as elongating cysts with multiple axonemes and an irregular number of mitochondrial derivatives (Fig. 3d) suggesting an incomplete second meiotic division. Taken together, these findings suggest a role for Mob4 in the structural maintenance of axonemes and in forming or maintaining the integrity of associated cellular structures during spermatogenesis.

GFP-mob4 has a dynamic localization throughout spermatogenesis

Next, we examined the subcellular localization of GFP-mob4 in testes of 2-day-old males. The rescue of all mutant phenotypes by ubiquitous expression of a GFP-mob4 transgene indicates that the GFP-mob4 fusion protein is fully functional. Using GFP-mob4 transgenic flies, we observed that in meiotically dividing spermatocytes, GFP-mob4 had a reticular localization accumulating in membranous fibers surrounding the meiotic spindle (Fig. 4c). During individualization, GFP-mob4 accumulated in the cystic bulge (Fig. 4b). In early canoe stage spermatids, GFP-mob4 strongly accumulated in individual puncta at the basal side of nuclei in the vicinity of the basal body (Fig. 4d). This punctate localization appeared to be transient as it was absent when the actin cone was forming in spermatids at the late canoe stage (Supplementary Fig. 2). This dynamic behavior of Mob4 at different stages of spermiogenesis suggests a functional requirement for Mob4 in the parafusorial membranes or structures derived from or associated with them, at the basal body or transition zone in the initiation of axoneme elongation, and in the cystic bulge during individualization per se.

Fig. 4.

GFP-mob4 localization throughout spermatogenesis. a) Testes from 2-day-old males expressing GFP-mob4 were stained to reveal DNA (DAPI stain, in magenta), actin cones (phalloidin-594, in red) and fluorescence of GFP-mob4 (green) (b) GFP-mob4 accumulates in the cystic bulge surrounding actin cones in individualizing spermatid bundles. c) GFP-mob4 has a reticular localization around the parafusorial membranes and associated with microtubules during the meiotic divisions. d) GFP-mob4 (d') localizes to the basal end of the nuclei (d'') in early canoe stage spermatids (see also Supplementary Fig. 2).

Strip and Cka are required for spermatid individualization

We affinity-purified GFP-mob4 from Drosophila syncytial embryos and analyzed the complexes by mass spectrometry; samples prepared from embryos expressing GFP alone were used as a negative control to exclude proteins that either bind to the GFP-tag or to the beads themselves. This approach identified STRIPAK core components as the interactants with highest Mascot scores (Supplementary Table 2). We therefore questioned whether knockdown of STRIPAK core components might replicate the mob4^RNAi phenotypes. We found that ubiquitous depletion of either Strip or Cka, via RNAi, induced late larval/pupal lethality (Fig. 5a) as previously described for mutant alleles of these loci (Chen et al. 2002; Sakuma et al. 2014). When we downregulated Strip and Cka in the Drosophila germline, we found that depletion of Strip resulted in 100% male sterility while depletion of Cka led to ∼50% of males being sterile (Fig. 5b). Curiously, only Strip^RNAi had any impact on female fertility, with progeny dropping to a third of control levels (Fig. 5c).

Fig. 5.

STRIPAK components Mob4, strip, and Cka are required for male fertility in Drosophila melanogaster. a) Viability of control (mCherry^RNAi), mob4^RNAi, Strip^RNAi, and Cka^RNAi. Data points represent percentage of viable flies in 6 independent experiments. The da-GAL4 driver was crossed either to UAS-mCherry^RNAi (n = 417), UAS-mob4^RNAi (n = 266), UAS-Strip^RNAi (n = 244), or UAS-Cka^RNAi (n = 144). Means ± SEM are shown. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level); ****P < 0.0001. b) Male fertility test of control, mob4^RNAi, Strip^RNAi, and Cka^RNAi. Germinal downregulation was achieved by crossing UAS-RNAi flies (as in a) to the nos-GAL4 driver line. Males of each genotype were individually mated with wild-type females. Data points represent numbers of progeny from individual males. Means ± SEM are shown for at least n = 8 males per genotype. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level). c) Female fertility test of control, mob4^RNAi, Strip^RNAi, and Cka^RNAi. Germinal downregulation was achieved as in b. Females of each genotype were individually mated with wild-type males. Data points represent numbers of progeny from individual females. Means ± SEM are shown for at least n = 9 females per genotype. P-values of 2-tailed unpaired t-tests are indicated (99% confidence level); ****P < 0.0001, ns, not significant. (d–i) Whole mount testes stained with phalloidin-594 to visualize the terminal epithelium (highlighted pink) and seminal vesicles (highlighted white) in 2-day-old (d-f) and 10-day-old (g-i) control, mob4^RNAi, and Strip^RNAi unmated flies. Pink star highlights enlargement of terminal epithelium. Strip^RNAi (j) and Cka^RNAi (l) testes from 2-day-old males were immunostained with antipolyglycylated tubulin (green) to visualize spermatid axonemes. Actin cones and spermatid bundles from Strip^RNAi (k) and Cka^RNAi (m) depleted testes visualized with phalloidin-594 and immunostained to reveal polyglycylated tubulin (green).

Testes from 2-day-old Strip^RNAi males revealed scattered actin cones and no mature sperm accumulating in the seminal vesicle (Fig. 5j-k). As control animals aged, their seminal vesicles increased in volume, due to continued sperm accumulation (Fig. 5d/g). However, the seminal vesicles of either 10-day-old mob4^RNAi (Fig. 5e/h) or Strip^RNAi (Fig. 5f/i) males failed to fill with sperm, and no increase in volume was observed. Instead, there was enlargement of the terminal epithelium suggestive of continuous production of aberrant sperm and abnormal sperm coiling. Testes from Cka^RNAi males showed both mature sperm in the seminal vesicle and abnormal enlargement of the terminal epithelium. Accordingly, we could detect some intact ICs, but the great majority of the actin cones were scattered (Fig. 5l-m). We conclude that, like Mob4, Strip and Cka are required for sperm individualization and male fertility.

Discussion

Here, we describe a previously unknown function of Mob4 in spermatid differentiation. Mob4 depleted testes undertake cyst elongation after meiosis, but spermatids fail to individualize and the seminal vesicles are devoid of mature sperm. Our results accord with the testes being the tissue with highest levels of Mob4 expression (FlyAtlas2) (Krause et al. 2022).

One primary defect of spermiogenesis in mob4-depleted testes is the loss of the structural integrity of the axoneme. Spermiogenesis requires intense cytoskeletal reorganization, not only of microtubules in axoneme elongation but also with actin having a major role in spermatid individualization (Lindsley et al. 1980; Fabrizio et al. 1998). Several mutants for microtubule-binding proteins have been shown to affect the axoneme structure and lead to individualization defects. For example, testes mutant for Bug22, a conserved protein that associates with basal bodies and cilia, shows defects in sperm individualization characterized by disrupted spermatid axonemes (Maia et al. 2014). Flies depleted for TTLL3B, an essential member of the tubulin tyrosine ligase-like family, are male sterile with either missing axonemes or axonemes composed of singlet microtubules and having sperm individualization defects (Rogowski et al. 2009). To our knowledge, however, “expanded” axonemes having all 9 outer doublets but with complete loss of linkage between doublets is a unique characteristic of mob4 downregulation. It is possible that the short-lived presence of GFP-Mob4 at the basal body may correlate with such defects, but this would require further studies.

Previous genetic studies have shown Mob4 to be a regulator of microtubule organization and axonal transport in Drosophila neurons (Schulte et al. 2010), and to play roles in the microtubule-based mitotic spindles in cultured cells (Baillat et al. 2001). Moreover, in zebrafish and nematodes, mob4 mutants have defective microtubule networks, and Mob4 interacts directly with the tubulin- and actin-folding TRiC/CCT chaperone complex (Khabirova et al. 2014; Berger et al. 2022). We also identified components of the TRiC/CCT complex as molecular partners of GFP-Mob4 (Supplementary Table 2). In the light of these findings, it is tempting to suggest that the spermatogenesis defects we now report when downregulating mob4 are due to impaired microtubule function.

Our finding that downregulation of either Strip or Cka results in similar failures of sperm individualization to Mob4 suggests that this process requires the STRIPAK complex. Strip and Cka were recently found to act within the somatic lineage of the Drosophila male gonad, to regulate germline lineage proliferation by acting as negative regulators of JNK signaling (La Marca et al. 2019). Our current findings point to an additional requirement for Strip and Cka at later stages in spermiogenesis, particularly during the postmeiotic spermatid individualization. As spermiogenesis defects may be a consequence of defects occurring earlier during spermatogenesis, the use of drivers for expression in late spermatocytes would allow to clarify whether Mob4, Strip, and Cka are required specifically at spermiogenesis.

Mitochondria are known to assist axonemal growth during spermatid elongation. TEM analysis revealed unexpected defects in the mitochondrial derivatives of elongating cysts in mob4^RNAi testes. At the onset of elongation, the 2 mitochondrial derivatives extend in parallel with the growing axoneme (Regan and Fuller 1990). One of them accumulates a dense body of paracrystalline material and becomes the major mitochondrial derivative, while the other reduces in size and volume until individualization is completed. In the elongating cysts of mob4^RNAi males, it was possible to observe accumulation of paracrystalline material in both mitochondrial derivatives as also described in the big bubble 8 mutant (Vedelek et al. 2016). Moreover, following mob4^RNAi, both mitochondrial derivatives often displayed irregular shapes and dimensions; a similar phenotype to that was reported for the wampa (wam) mutant (Bauerly et al. 2020). Wampa is a dynein subunit, and wam mutant spermatids lack the axonemal outer dynein arm, leading to loss of flagellar motility. It therefore seems likely that defects in axonemal organization could lead to the defects in mitochondria structure of the type we observe. However, as mitochondria provide structural and physical support for the developing axoneme, defects in mitochondrial elongation could further destabilize the developing axoneme as previously suggested (Hoyle and Raff 1990; Fuller 1993).

At present, we can only speculate about the molecular function of Mob4 in spermiogenesis. We note that clathrin heavy chain (Chc4) mutants are also defective in receptor-mediated endocytosis, neurotransmitter secretion, and spermatid individualization (Bazinet et al. 1993). This raises a tempting possibility of common functions for Chc4 and Mob4, particularly since Mob4 has a role in vesicular trafficking functions and shares sequence homology with the σ subunits of clathrin adaptor complexes (Baillat et al. 2001; Bailly and Castets 2007). Moreover, Strip has also been shown to regulate endosome trafficking and microtubule stabilization during axon elongation (Sakuma et al. 2014). Further studies of proteins that interact with the STRIPAK complex in the testes will be required to test these possibilities.

In summary, our present study has identified essential roles in spermatogenesis and male fertility for the Mob4, Strip, and Cka genes. Mob4 is required for proper axoneme structure, and mob4-depleted testes display a range of defects throughout multiple stages of sperm development. Understanding how Mob4 regulates the architecture of axonemal microtubules and how this affects other aspects of cellular function will be indispensable for understanding the pleiotropic manifestation of diseases and clinical conditions, such as sterility, that arise from the dysfunction of STRIPAK genes.

Data availability

The authors confirm that all relevant data supporting the findings of this study are available within the article and/or its supplementary information file.

Supplemental material available at GENETICS online .

Funding

This work was funded by the Algarve 2020 Program, grant number ALG-01-0145-FEDER-030014, cofinanced by FEDER Funds through the Operational Program for Competitiveness Factors—COMPETE 2020 and by national funds through FCT—Foundation for Science and Technology under the Project PTDC/BIA-CEL/30014/2017. IBS was funded by FCT fellowship SFRH/BD/141734/2018. AW acknowledges Cancer Research UK for a PhD studentship (1999-2003). GC was funded by the Ministero dell’Istruzione, dell’Università e della Ricerca (MIUR 2020CLZ5XW). DMG acknowledges past grants from CRUK and Wellcome and current support from NIH grant R01NS119614.