MYCBPAP is a central apparatus protein required for centrosome–nuclear envelope docking and sperm tail biogenesis in mice Free

According to the latest report from the World Health Organization, an estimated 17.5% of couples worldwide are expected to experience infertility during their lifetime (WHO, 2023, see https://iris.who.int/bitstream/handle/10665/366700/9789240068315-eng.pdf). Male factors contribute to ∼50% to 60% of all infertility cases (Leslie et al., 2023); however, despite significant advancements in our understanding of the etiology of male infertility, a substantial portion of cases remains unexplained (Bhattacharya et al., 2024; Kimmins et al., 2024). One of the most severe subsets of male infertility is known as multiple morphological abnormalities of the sperm flagella (MMAF), characterized by abnormal sperm flagella, including coiled, bent, short or absent tails, coupled with decreased sperm motility (Dong et al., 2018). As genetic factors are considered to be highly associated with MMAF (Wang et al., 2020), elucidating the role of the responsible genes in sperm flagellum formation is crucial.

Haploid round spermatids undergo transformation into spermatozoa during the final phase of spermatogenesis, known as spermiogenesis, and is divided into 16 distinct steps in mice. In early spermiogenesis (step 2–3), the centrosome, which is composed of proximal and distal centrioles, docks onto the plasma membrane (Dunleavy et al., 2019). The distal centriole is distributed nearly perpendicularly against the proximal centriole and functions as the base for axonemal elongation (Xie et al., 2023). The axoneme is composed of a central pair of microtubule singlets surrounded by nine peripheral microtubule doublets, forming a recognizable ‘9+2’ microtubule structure (Miyata et al., 2020a; Pereira and Sousa, 2023). While the axoneme is being formed, the axonemal dynein arms, which are molecular motors that are crucial for sperm motility, are simultaneously integrated into the axoneme (Wu et al., 2023). It is noteworthy that the central pair of microtubules forms an asymmetric complex with proteinous projections; this is involved in dynein regulation and is known as the central apparatus (Gui et al., 2022; Lechtreck et al., 2013). The architecture of the central apparatus is conserved between mammals and the unicellular organism Chlamydomonas (Leung et al., 2021).

In the middle of spermiogenesis, round spermatids undergo further morphological changes, leading to the formation of elongated spermatids. As the flagellum continues to assemble, the centrosome translocates and docks onto the nuclear envelope at step 6 (Pereira et al., 2019; Shimada and Ikawa, 2023). This centrosome–nuclear envelope docking then accelerates the formation of the head-to-tail coupling apparatus (HTCA), disruption of which leads to male infertility (Wu et al., 2020). As this centrosome–nuclear envelope docking is spermiogenesis specific and does not occur in ciliogenesis, a large part of this process remains unknown (Dunleavy et al., 2019).

The MYC-binding protein (MYCBP) associated protein (MYCBPAP) has been identified as being highly expressed in human testes (Yukitake et al., 2002). Additionally, the ortholog of MYCBPAP in Chlamydomonas (FAP147) is known to exhibit specific localization at the central apparatus (Han et al., 2022). However, the role of MYCBPAP in male fertility has not been extensively investigated. To address this gap, we generated a Mycbpap-knockout (KO) mouse model and analyzed the resulting phenotype to reveal the function of MYCBPAP during spermiogenesis.

To understand the expression pattern of Mycbpap, we first performed RT-PCR using various mouse tissues, and found that Mycbpap is predominantly expressed in mouse testes (Fig. 1A). Furthermore, RT-PCR using postnatal testes indicated that Mycbpap is expressed on postpartum day (dpp) 21 and shows high expression from 28 dpp (Fig. 1B), which corresponds to the early stages of spermiogenesis. Single-cell RNA sequencing (scRNA-seq) data (Ernst et al., 2019) also shows that expression of Mycbpap is increased during early spermiogenesis in mice (Fig. S1A). Furthermore, the expression data of MYCBPAP from scRNA-seq of human adult testes (Guo et al., 2018) displays similar expression trends to that seen in mice, and increases dramatically from late spermatocytes to round spermatids (Fig. S1B).

To study the function of MYCBPAP, we generated KO mice lacking a large region of the open reading frame (ORF) of Mycbpap (Mycbpap^em1/em1) by introducing CAS9 and a pair of guide RNA (gRNA) targeting exon 2 and exon 18 of Mycbpap to B6D2F1×B6D2F1 genetic background zygotes (Fig. 1C). Of the 41 zygotes that were electroporated, 33 two-cell embryos were transplanted into the oviducts of two pseudopregnant females. A total of 12 pups were born and 10 of them had large deletion regions in Mycbpap. Mycbpap^em1/em1 mice were then obtained by caging F0 mice with wild-type (WT) mice and subsequent matings between F1 mice. Genotyping of Mycbpap mutant mice was carried out by genomic PCR (Fig. 1D). Subsequently, Sanger sequencing of the Mycbpap em1 allele showed that 11,419 base pairs (bp) of the Mycbpap gene was deleted (Fig. 1E).

We then caged individual Mycbpap^em1/em1 males with WT females and checked for vaginal plugs as a sign of copulation. Mycbpap^em1/em1 males were sterile, as no pups were obtained despite copulation occurring (Fig. 1F). These data suggest that MYCBPAP is indispensable for male fertility in mice.

To elucidate the causes of infertility, we first observed the testes of Mycbpap^em1/em1 males. As shown in Fig. 2A, adult Mycbpap^em1/em1 males exhibited smaller testes compared to Mycbpap^wt/em1 males. Consistent with this, testis weights of Mycbpap^em1/em1 males were significantly lower than those of control Mycbpap^wt/em1 males (100.9±12.8 mg for Mycbpap^wt/em1 males and 59.0±8.3 mg for Mycbpap^em1/em1 males; mean±s.d.) (Fig. 2B). As small testes can be related to defective spermatogenesis in mice (Lee et al., 2011; Takahashi et al., 2002; Wang et al., 2016), we next performed histological analyses of testis sections. In stage VII seminiferous tubules, step 16 spermatids are located on the interior side of step 7 spermatids, with their tail lining in the center of the lumen (Meistrich and Hess, 2013). However, in Mycbpap^em1/em1 males, a few of the step 16 spermatids were intermixed with the step 7 spermatids, and their tails were not visible in the sections (Fig. 2C). Further analyses of other stages revealed that step 13–16 (in stage I–VIII) elongated spermatids were almost absent in seminiferous tubules of Mycbpap^em1/em1 males (Fig. S2A). Although the number of elongated spermatids decreased, images of individual spermatids from testis sections were digitally isolated and designated as belonging to steps 1–16, to further study the progression of spermiogenesis in Mycbpap^em1/em1 males (Fig. 2D). The morphology of spermatids from step 1–8 was comparable between Mycbpap^wt/em1 and Mycbpap^em1/em1 males, whereas spermatids of Mycbpap^em1/em1 males exhibited abnormal morphology after step 9, eventually leading to the abnormal elongated spermatids found in step 16 (Fig. 2D). Altogether, absence of MYCBPAP alters late spermiogenesis in male mice.

We then performed histological studies on epididymis sections. Spermatozoa were visible in the caput, corpus and cauda epididymis in control Mycbpap^wt/em1 males, whereas spermatozoa were nearly absent in Mycbpap^em1/em1 epididymis (Fig. 3A). Moreover, the number of cauda epididymal spermatozoa was significantly decreased in Mycbpap^em1/em1 males (Fig. S2B). Consistent with the abnormal spermatid heads observed in the testis, epididymal spermatozoa exhibited round shape heads in Mycbpap^em1/em1 males (Fig. 3B). Furthermore, the length of the sperm tail ranged from ∼20 to 120 μm in Mycbpap^em1/em1 males, which was significantly shorter than those of Mycbpap^wt/em1 males (Fig. 3C). In addition, Mycbpap KO mature spermatozoa showed no motility (Movies 1 and 2). These results indicate that Mycbpap^em1/em1 males are infertile due to the reduced number, abnormal morphology and impaired motility of spermatozoa.

The morphological change of the sperm head depends on the manchette, a microtubule-based skirt-like structure that surrounds the caudal region of the sperm head (Yogo, 2022). Therefore, we labeled the main component of manchette, α-tubulin, to visualize the morphological change of sperm head oriented by the manchette (Fig. 3D). We noticed an abnormally long manchette in Mycbpap KO spermatids, which subsequently caused a club-shaped morphology, consistent with the head shape observed in the testis sections (Fig. 2D). These results indicate that both the sperm tail and head are defective in Mycbpap^em1/em1 males, and the latter is likely caused by an abnormally long manchette.

To further analyze the process of spermiogenesis, we observed Mycbpap^em1/em1 testes with transmission electron microscopy (TEM) (Fig. 4A,B). Resembling the histology result found in the stage VII seminiferous tubules of testis sections (Fig. 2C), we rarely observed step 16 spermatids inside the Mycbpap^em1/em1 seminiferous tubules by TEM (Fig. 4A). Moreover, condensed cytoplasm and cytoplasmic vacuoles were found in the same seminiferous tubules, which are signs of apoptosis (Taatjes et al., 2008). We then conducted terminal transferase dUTP nick end labeling (TUNEL) staining in testis sections, and found that some of the spermatids in Mycbpap^em1/em1 males were TUNEL positive (Fig. 4C), and the percentage of seminiferous tubules carrying TUNEL-positive spermatids significantly increased in Mycbpap^em1/em1 males (Fig. S3A), indicating that ablation of MYCBPAP caused apoptosis of spermatids during spermiogenesis, which might have led to the reduced number of mature spermatozoa.

Moreover, for step 4 spermatids, axonemes of both WT and Mycbpap^em1/em1 males were able to elongate (Fig. 4B). However, in those spermatids that progressed to step 9 in Mycbpap^em1/em1 males, the centrosome did not dock correctly onto the nuclear envelope (Fig. 4B). Centrosome–nuclear envelope docking should occur in step 6 spermatids (Shimada and Ikawa, 2023). Consistent with these results, in sections of Mycbpap^em1/em1 testis, sperm tails were almost absent in stage II–III and stage VII–VIII seminiferous tubules (Fig. S3B). In addition, in stage VII seminiferous tubules of Mycbpap^wt/em1 males, tails were observed intact on the opposite side of acrosomal vesicles in step 7 round spermatids; these tails normally grow towards the center of the lumen. However, in the counterpart Mycbpap^em1/em1 males, acetylated tubulin was accumulated around the round spermatids but not in the correct position (Fig. S3B). Taken together, these results suggest that the centrosome–nuclear envelope docking and tail biogenesis are abnormal in Mycbpap^em1/em1 spermatids.

To elucidate the localization and the binding partner of MYCBPAP, we generated transgenic (Tg) mice (Fig. S4A). The transgene possessed Mycbapap cDNA and a 3×FLAG tag in its C-terminal, and was driven by a testis-specific Clgn promoter (Ikawa et al., 2001; Watanabe et al., 1995). The success of Mycbapap Tg mouse generation was validated with western blotting, as clear bands were observed in the testis and sperm lysates of Mycbapap Tg mice by means of anti-FLAG antibody staining (Fig. S4B). We then caged Mycbpap^em1/em1 males that carried the transgene (Mycbpap^em1/em1 Tg) with WT females. Mycbpap^em1/em1 Tg males were fertile and exhibited comparable litter size to that from WT males (Fig. 5A).

We attempted to visualize the localization of MYCBPAP by performing immunocytochemistry (ICC) on spermatids and epididymal spermatozoa. The MYCBPAP–FLAG signal was observed along the sperm tail of Mycbpap^em1/em1 Tg males (Fig. S4C,D), indicating that MYCBPAP is a sperm flagellar protein. Although the fluorescence signal was also observed in the manchette of Mycbpap^em1/em1 Tg males, we could not exclude the possibility that the signal was background because spermatids of Mycbpap^wt/em1 males also exhibited a fluorescence signal in the manchette as well (Fig. S4C). Furthermore, for the cauda epididymal spermatozoa of Mycbpap^em1/em1 Tg males, we separated sperm heads and tails, and performed western blotting analyses. Consistent with the above results, western blotting indicated that MYCBPAP was localized in the sperm tails (Fig. 5B). Moreover, mature spermatozoa from Mycbpap^em1/em1 Tg males were fractionated using three different lysis buffers and the MYCBPAP–FLAG band was observed in the SDS-soluble fraction that contains axonemal proteins (Fig. 5C) (Castaneda et al., 2017). Taken together, these results indicate that MYCBPAP is an axoneme-associated flagellar protein.

To explore the binding partners of MYCBPAP, we next performed immunoprecipitation (IP) using testis lysates obtained from Mycbpap^em1/em1 Tg males. Mass spectrometry analyses after IP using anti-FLAG antibody showed that MYCBPAP–FLAG interacts with 30 proteins (Fig. 5D; Table S1), including five proteins whose orthologs are known to be localized at the central apparatus of Chlamydomonas [cilia- and flagella-associated protein 65 (CFAP65), kinesin family member 9 (KIF9), CFAP70, sperm associated antigen 16 (SPAG16) and MYCBP] (Han et al., 2022; Samsel et al., 2021). We verified the binding of MYCBPAP with CFAP70 and CFAP65 in testes by performing western blotting after IP (Fig. 5E), which suggested that MYCBPAP is localized in the central apparatus.

To further analyze whether MYCBPAP is localized in the central apparatus, we frayed axonemal microtubules (Fig. S5). By inducing microtubule sliding with ATP after demembranation of mature spermatozoa (Kinukawa et al., 2004), we separated bundles of microtubule doublets and their associated proteins from the axoneme. We confirmed that dynein regulatory complex subunit 3 (DRC3) was localized in several microtubule doublet bundles, but KIF9 was localized in only one microtubule doublet bundle, which was expected as DRC3 has been previously shown to localize on all nine axonemal microtubule doublets but KIF9 has been shown to localize in the central apparatus (Miyata et al., 2020b; Walton et al., 2023). We found that MYCBPAP–FLAG was also localized in only one microtubule doublet bundle, further supporting that MYCBPAP is a central apparatus protein in mouse sperm flagella (Fig. S5).

Previous cryo-electron microscopy (Cryo-EM) analyses of the Chlamydomonas central apparatus have shown that the orthologs of MYCBPAP, CFAP65, CFAP70, SPAG16, MYCBP are localized in the C2a projection, whereas the ortholog of KIF9 is localized in the C2c projection that is close to the C2a projection (Gui et al., 2022; Han et al., 2022), suggesting that MYCBPAP is possibly a component of C2a projection in mice as well. Intriguingly, CCP110, a centrosome protein, also interacts with MYCBPAP (Fig. 5D,E; Table S1) (Chen et al., 2021). We compared the amounts of CCP110 between WT and Mycbpap^em1/em1 testes by western blotting and found no differences (Fig. S4E). Furthermore, we applied gene ontology (GO) annotation on the interactome of MYCBPAP (Fig. 5F). GO analysis demonstrated that MYCBPAP-interacting proteins participate in not only the cilium movement but also the assembly of dynein. Moreover, analysis of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways revealed that MYCBPAP-interacting proteins were enriched in proteins annotated with the ‘motor proteins’ criteria.

MYCBPAP was previously identified as a testis-specific binding partner of MYCBP (Yukitake et al., 2002), yet the function of MYCBPAP during spermatogenesis remained largely unknown. Consistent with the previous report, in this study, we showed that Mycbpap is predominantly expressed in the mouse testis (Fig. 1A) from 21 days postnatal (Fig. 1B). Moreover, we generated Mycbpap^em1/em1 mice and showed that Mycbpap^em1/em1 males lose their fertility completely (Fig. 1F). Furthermore, we demonstrate that the male infertility of Mycbpap^em1/em1 males is caused by multiple abnormalities, such as a reduced number of spermatozoa in the epididymis (Fig. 3A and Fig. S2B), the immobility of their spermatozoa (Movies 1 and 2), short sperm tails (Fig. 3B,C), and abnormal sperm heads (Fig. 3B,D). We also show that spermatozoa from Mycbpap^em1/em1 males are unable to undergo correct docking of the centrosome to the nuclear envelope (Fig. 4B).

The interactome of MYCBPAP contains five central apparatus proteins, CFAP65, KIF9, CFAP70, SPAG16 and MYCBP (Fig. 5D; Table S1), and they are all conserved in Chlamydomonas as FAP65, KLP1, FAP70, PF20 and FAP174, respectively. Moreover, MYCBPAP is conserved in Chlamydomonas as FAP147, and the localization of FAP147 has been well studied by biochemical and cryo-EM analyses (Gui et al., 2022; Han et al., 2022; Hou et al., 2021). Strikingly, FAP147 localizes at the C2a projection with FAP65 and FAP70 in the fully assembled central apparatus, and FAP147 acts as a rachis that recruits FAP65 and FAP70 to C2a. These results largely resemble our interactome data, suggesting that the localization of MYCBPAP in the C2a projection might be highly conserved from Chlamydomonas to mice. In mice, KO of Cfap70 causes fewer spermatozoa to be observed in the cauda epididymis (Jin et al., 2023), and depletion of Cfap65 results in aberrant head morphogenesis (Wang et al., 2021). Moreover, both KOs of Cfap65 and Cfap70 lead to abnormal sperm tail biogenesis, similar to what we found in Mycbpap^em1/em1 males (Fig. 3B,C). Furthermore, we demonstrated that MYCBPAP–FLAG was localized in only one separated microtubule doublet bundle (Fig. S5). Collectively, these results suggest that MYCBPAP is a central apparatus component with CFAP65 and CFAP70, and contributes to sperm tail biogenesis. Abnormal sperm tail formation in Mycbpap^em1/em1 mice might lead to disrupted centrosome–nuclear envelop docking and subsequent abnormal manchette morphology.

Our TEM analyses showed that MYCBPAP is required for centrosome–nucleus docking during spermiogenesis. We could not conclude whether MYCBPAP functions at the centrosome directly or not, because our antibody is not specific enough to show its localization in the centrosome. Nevertheless, the interactome data of MYCBPAP show possible functional partners in respect to centrosome docking, namely CFAP65, KIF9 and CCP110. CCDC108 (an ortholog of CFAP65 in Xenopus) is reported to function in the apical trafficking of the centrioles during the assembly of ciliary cells (Zhao et al., 2022). Furthermore, KIF9 is localized at axonemes and basal bodies in multiciliated cells in human airways. Knockdown or KO of KIF9 results in impaired motility of cilia in humans and sperm flagella in mice, respectively (Konjikusic et al., 2023; Miyata et al., 2020b). Moreover, CCP110 localizes near the centrioles of human spermatozoa (Turner et al., 2023) and is crucial for controlling the maturation of centrioles, while also acting as a suppressor of ciliogenesis (Turner et al., 2023; Yadav et al., 2016), depletion of which causes abnormal outgrowth of cilia in mice (Spektor et al., 2007; Yadav et al., 2016). CCP110 needs to be regulated spatiotemporally to fulfill its correct function (Otto and Hoyer-Fender, 2023), and the MYCBPAP–CCP110 interaction we identified might participate in the spatial part of it, as the temporal levels of CCP110 largely depend on the ubiquitin-proteasome system (Dangiolella et al., 2010; Li et al., 2013). Overall, MYCBPAP might achieve the purpose of centrosome–nuclear envelope docking together with the centrosome-associated proteins.

In summary, our work identifies that MYCBPAP is required for centrosome–nuclear envelope docking and sperm tail biogenesis, with its dysfunction ultimately resulting in male infertility in mice (Fig. 5G). Multiple genomic variants on MYCBPAP have been predicted as loss-of-function in the Genome Aggregation Database (gnomAD; Karczewski et al., 2020). Among them, a stop-gain variant (P.Cys58Ter) was found with an allele frequency of 4.94e-4. Therefore, this study might lead to a better comprehension of male infertility in humans. Further studies on MYCBPAP could promote the understanding of centrosome–nuclear envelope docking and sperm tail biogenesis during spermiogenesis.

Mice were purchased from CLEA Japan (Tokyo, Japan) or Japan SLC (Shizuoka, Japan). The mice were housed in a specific-pathogen-free environment with unrestricted access to food, following an artificial 12-h light and 12-h dark cycle. All experimental procedures involving mice were approved by the Animal Care and Use Committee at the Research Institute for Microbial Diseases, Osaka University (Approval number: #Biken-AP-H30-01 and #Biken-AP-R03-01). All gene-modified mice generated in this study will be available through either the RIKEN BioResource Research Center or the Center for Animal Resources and Development (CARD), Kumamoto University.

Generation of KO and Tg mice was performed as described previously (Abbasi et al., 2018; Miyata et al., 2021). Female B6D2F1 mice were superovulated by intraperitoneal injection with CARD HyperOva (Kyudo, Saga, Japan) and human chorionic gonadotropin (hCG) (ASKA Pharmaceutica, Tokyo, Japan). Subsequently, they were caged with WT B6D2F1 males. The resulting two-pronuclear (2PN) zygotes were isolated from the female mice.

For generation of KO mice, guide RNAs with the sequence of 5′-TCCAGAAAAGAAGCGGGCGA-3′ and 5′-TAAAGTCTGCAAGCCGGGAC-3′ were designed and used to target exon 2 and exon 18 of Mycbpap, respectively. 2PN zygotes were electroporated with the ribonucleoprotein complex of CAS9 (#A36497, Thermo Fisher Scientific, Waltham, MA, USA), CRISPR RNA (crRNA) (Sigma-Aldrich, St Louis, MO, USA) and trans-activating crRNA (tracrRNA) (#TRACRRNA05N-5NMOL, Sigma-Aldrich).

For the generation of Tg mice, Mycbpap cDNA with a C-terminal 3×FLAG was cloned and inserted into a plasmid with a Clgn promoter (pClgn1.1, #173686 Addgene). The linearized DNA was then microinjected into the pronuclei of the 2PN zygotes.

For the generation of both KO and TG mice, zygotes were cultured in potassium simplex optimized medium (KSOM) (Ho et al., 1995) until the two-cell stage. Subsequently, the embryos were transplanted into the ampulla of oviducts of pseudopregnant ICR female mice. F0 pups were either naturally delivered or obtained through Cesarean section. Genomic PCR and Sanger sequencing were performed to verify the deleted or inserted sequences. PCR was performed using primers with the sequence 5′-ACCTGCACATGCTTGGTAGG-3′ and 5′-CTTCCTTGCCACTCTGTTGG-3′ for the KO allele, 5′-GTCACTTGCCCGTCCTATGG-3′ and 5′-TGACCTCGCGCTCCTCCTGG-3′ for the WT allele, and 5′-TTGAGCGGGCCGCTTGCGCACTGG-3′ and 5′-TCCTTTTTAATGGCCAAGGC-3′ for the Tg alelle.

RNA extraction was performed on multiple tissues and postnatal testes of C57BL/6J mice using Trizol (Thermo Fisher Scientific). Subsequently, the extracted RNA was reverse transcribed into cDNA using the SuperScript III First Strand Synthesis Kit (Thermo Fisher). The resulting cDNA was employed for gene expression analysis through PCR with KOD Fx Neo DNA Polymerase (Toyobo, Tokyo, Japan). Primers with the sequence 5′-CCTTCGCTGCTGTTCTGTGGG-3′ and 5′-AGTCTTCCTGGCGGTGTCCA-3′ were used for RT-PCR of Mycbpap, whereas 5′-CATCCGTAAAGACCTCTATGCCAAC-3′ and 5′-ATGGAGCCACCGATCCACA-3′ were used for Actb as a loading control.

Each sexually mature KO or WT male was housed with three 7-week-old females for 8 weeks. Throughout these 8 weeks, the presence of vaginal plugs was monitored as an indicator of successful copulation, and the number of resulting pups was recorded. Following these 8 weeks, the males were removed from the cages, and the number of pups was recorded for an additional 3 weeks, allowing for the delivery any final litters by the females.

Bouin's fluid (#16045-1, Polysciences, Inc., Warrington, PA, USA) was employed for the fixation of testes and epididymis. Subsequently, the tissues were embedded in paraffin-filled cassettes and sectioned into 5 μm slices. Periodic acid–Schiff (PAS) staining was carried out on these sections with 1% periodic acid (Nacalai Tesque, Kyoto, Japan) for 10 min and subsequent Schiff's reagent (Wako, Osaka, Japan) for 20 min. After washing with tap water, slides were counterstained with hematoxylin for 5 min, washed with tap water, and observed using an Olympus BX-53 microscope (Tokyo, Japan).

Spermatozoa from cauda epididymis were suspended in Toyoda, Yokoyama, Hoshi (TYH) medium (Muro et al., 2016). The number of spermatozoa was counted with a hemocytometer. For morphology and motility assessment, spermatozoa were placed on glass slides and observed using an Olympus BX-53 microscope.

Antibodies used were against: α-tubulin [#T6074; 1:1000 for immunocytochemistry (ICC)], FLAG [#F1804; 1:1000 for western blotting (WB), 1:200 for ICC, 2 μg for immunoprecipitation (IP)] from Sigma-Aldrich, acetylated tubulin [#5335; 1:1000 for WB, 1:500 for immunohistochemistry (IHC)] from Cell Signaling Technology (Danvers, MA, USA), FLAG (#PM020; 1:1000 for WB) from MBL (Nagoya, Aichi, Japan), AKAP4 (#611564; 1:1000 for WB) from BD Biosciences (San Jose, CA, USA), CFAP65 (#HPA055156; 1:100 for WB), CFAP70 (#HPA037582; 1:200 for WB), DRC3 (#HPA036040; 1:500 for ICC) and KIF9 (#HPA022033; 1:200 for ICC) from Atlas Antibodies (Bromma, Sweden), CP110 (CCP110) (#12780-1-AP; 1:500 for WB) from Proteintech (Rosemont, Illinois, USA), β-actin (#ab6276; 1:2000 for WB) from Abcam (Cambridge, UK), BASIGIN (#sc-46700; 1:500 for WB) from Santa Cruz Biotechnology (Dallas, Texas, USA). IZUMO1 antibody (#KS64-125; 1:1000 for WB) was described previously (Ikawa et al., 2011).

For immunohistochemistry (IHC) of testis, male mice were killed, and their testes were fixed using 4% paraformaldehyde (PFA, Thermo Fisher Scientific) in phosphate-buffered saline (PBS) and embedded in OCT compound (Sakura Finetek, Tokyo Japan). Subsequently, 10 μm sections were prepared using a cryostat (CryoStar NX70, Thermo Fisher Scientific).

For immunocytochemistry (ICC) of spermatids, spermatids were isolated from seminiferous tubules using fine forceps and air dried on glass slides. Glass slides were then treated with 4% PFA in PBS for 10 min.

Both slides for IHC and ICC were washed with PBS and blocked with a blocking buffer composed of 3% bovine serum albumin (BSA, Sigma-Aldrich) and 10% goat serum (Thermo Fisher Scientific) in PBS for 1 h at room temperature. Subsequently, the slides were incubated with the primary antibody in blocking buffer overnight at 4°C. After three washes with PBS, the samples were incubated with a fluorophore-conjugated secondary antibody [goat anti-mouse-IgG conjugated to Alexa Fluor 488; goat anti-rabbit IgG conjugated to Alexa Fluor 488 or 546 (Thermo Fisher Scientific)]. Following an additional three washes with PBS, a solution of 0.02% Hoechst 33342 (Thermo Fisher Scientific) in PBS was applied to the slides for 10 min to stain the nucleus. After three final washes, the slides were mounted using Immu-Mount (Thermo Fisher Scientific) before observation with an Olympus BX-53 microscope or a Nikon Eclipse Ti microscope connected to a C2 confocal module (Nikon, Tokyo, Japan).

TEM was performed as described previously (Shimada et al., 2019). Briefly, perfusion fixation with 4% PFA in PBS was performed on anesthetized males, and then, their testes were further fixed in 4% PFA. Testes were then fixed in 1% glutaraldehyde. For postfixation, testes were immersed in 1% osmium tetroxide (OsO₄) and 0.5% potassium ferrocyanide. After dehydration with ethanol, testes were incubated with 100% propylene oxide (PO) and then placed in a mixture of PO and epoxy resin. Finally, the mixture was substituted with pure epoxy resin. Ultrathin sections were cut and stained. The images of samples were taken under a JEM-1400 Plus electron microscope (JEOL, Tokyo, Japan) at 80 kV with a CCD Veleta 2K×2 K camera (Olympus).

The in situ Apoptosis Detection Kit (Takara Bio, Shiga, Japan) was employed for terminal transferase dUTP nick end labeling (TUNEL) staining, following the manufacturer's protocol. In brief, paraffin-embedded testes underwent dewaxing, followed by treatment with Protease K and hydrogen peroxide. The samples were then labeled with a FITC-labeling mixture and incubated with anti-FITC horseradish peroxidase (HRP) conjugate. Finally, the samples were subjected to a reaction with diaminobenzidine (DAB) and hydrogen peroxide, followed by counterstaining with hematoxylin. The prepared slides were observed using an Olympus BX-53 microscope.

Western blotting was performed as previously described with slight modifications (Miyata et al., 2022). Protein lysates were subjected to heat treatment at 95°C in SDS sample buffer (60 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 0.025% Bromophenol Blue) with 5% β-mercaptoethanol. These samples were loaded onto pre-made SDS-PAGE gels (ATTO, Osaka, Japan) and transferred onto PVDF membranes using the Trans-Blot Turbo system (BioRad, Munich, Germany). Then, membranes were blocked with blocking buffer with 5% skim milk in TBS-T (0.05 M Tris-HCl pH 7.4, 0.15 M NaCl and 0.05% Tween 20) for 1 h at room temperature. Following this, membranes were incubated overnight at 4°C with primary antibodies in the blocking buffer. After three consecutive washes with TBS-T, membranes underwent additional incubation with HRP-conjugated secondary antibodies in the blocking buffer for 1 h at room temperature. After three brief washes with TBS-T, membranes were treated with Amersham ECL Western Blotting Reagent (Cytiva, Tokyo, Japan) for 1 min, and subsequent band detection was achieved using the Amersham ImageQuant 800 system (Cytiva).

Sperm head and tail separation was performed as described previously (Miyata et al., 2015). Briefly, spermatozoa were released from the cauda epididymis into 1 ml of PBS. Subsequently, sonication was employed to dissociate sperm heads and tails. The separated sperm head and tail components were then isolated into two phases using 90% Percoll in PBS through centrifugation, followed by protein extraction after a brief washing step with PBS.

For protein fractionation, mature spermatozoa collected from the cauda epididymis underwent initial extraction with Triton lysis buffer (1% Triton X-100, 50 mM NaCl, 20 mM Tris-HCl), followed by subsequent extraction with 1% SDS solution (1% SDS, 75 mM NaCl and 24 mM EDTA), and finally extracted with SDS sample buffer. The resulting fractions were denoted as the Triton-soluble, SDS-soluble and SDS-resistant fractions, respectively (Castaneda et al., 2017).

Immunoprecipitation (IP) was performed as described previously (Morohoshi et al., 2023). The lysate derived from testes underwent IP using Invitrogen Dynabeads Magnetic Beads (Thermo Fisher Scientific), following the manufacturer's protocol.

Sperm demembranation and microtubule sliding were performed as previously described with slight modification (Kinukawa et al., 2004). Briefly, cauda epididymal spermatozoa were released in PBS, aliquoted to demembranation solution, and incubated at 37°C for 1 min. Then, a portion of the spermatozoa was transferred into reactivation medium containing ATP and 1 mM DTT to induce microtubule sliding at 37°C for 10 min. ICC was then performed on the spermatozoa as described above.

Proteins from IP were analyzed through nanocapillary reversed-phase liquid chromatography (LC)-tandem mass spectrometry (MS/MS). The chromatography employed a C18 column (IonOpticks, Victoria, Australia) integrated into a nanoLC system (Bruker Daltonics, Billerica, MA), connected to a timsTOF Pro mass spectrometer (Bruker Daltonics, Billerica, MA), and utilized the CaptiveSpray nano-electrospray ion source (Bruker Daltonics, Billerica, MA). Raw data were processed using DataAnalysis software (Bruker Daltonics, Billerica, MA), and protein identification was conducted through MASCOT (Matrix Science, Tokyo, Japan) by querying the UniProt database. Subsequently, quantitative values and fold changes were determined using Scaffold5 (Proteome Software, Portland, OR). Significantly enriched proteins identified through IP analysis were subjected to gene ontology (GO) and KEGG enrichment analyses using DAVID functional annotation (Sherman et al., 2022).

Statistical analyses were performed using a two-tailed unpaired Student's t-test. Data is displayed as the mean±s.d.

We would like to thank Dr Julio M. Castaneda and Ferheen Abbasi for the critical reading of the manuscript, Natsuki Furuta and Takako Ishikawa for technical assistance, and the members of the Department of Experimental Genome Research for experimental assistance and discussion. We also thank Hiroko Omori for ultrastructural analysis and Akinori Ninomiya and Fuminori Sugihara for mass spectrometry analysis (Core Instrumentation Facility, Research Institute for Microbial Diseases, Osaka University).

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.261962.reviewer-comments.pdf