Loss of perinuclear theca ACTRT1 causes acrosome detachment and severe male subfertility in mice

Spermiogenesis refers to the complex process of differentiation and maturation of spermatids, which includes dramatic changes in cell shape. The perinuclear theca (PT) is a special class of cytoskeleton and closely surrounds the nuclei of spermatids and sperm. The PT, although structurally continuous, can be distinguished in different segments: (1) the subacrosomal region (SAR), intercalated between the inner acrosomal membrane (IAM) and the nuclear envelope (NE); (2) PT overlying the equatorial segment (ES), between the plasma membrane and outer acrosomal membrane (OAM); and (3) PT of the postacrosomal sheath (PAS). SAR-PT and ES-PT give rise to the subacrosomal layer (SAL) of the PT (Oko and Sutovsky, 2009). The PT is characterized by resistance to high ionic strength and detergents and by a special (testis-specific) protein composition, including calicin (Longo et al., 1987), cylicin 1 (Hess et al., 1993) and 2 (Hess et al., 1995), the actin-capping proteins CAPZA3 (Geyer et al., 2009) and CPβ3 (CAPZB) (von Bulow et al., 1997), the WW domain-binding proteins WBP2 (Hamilton et al., 2018) and PAWP (also known as WBP2NL) (Wu et al., 2007), the histone variant H2BL1 (Aul and Oko, 2002), signal transducer and activator of transcription STAT4 (Herrada and Wolgemuth, 1997), phospholipase C PLCζ (Hachem et al., 2017) and the glutathione S-transferase GSTO2 (Hamilton et al., 2017). SAL-PT and PAS-PT are different in both protein constituents and timing of formation. It has been suggested that SAL proteins originate from acrosomal vesicles during acrosome biogenesis, whereas PAR proteins are translated in the cytoplasmic lobe and then transported via the manchette, concomitantly with spermatid elongation (Oko and Sutovsky, 2009). It has been suggested that SAL-PT is involved in acrosome assembly whereas PAS-PT is involved in sperm-egg interactions at fertilization (Oko and Sutovsky, 2009). However, the physiological roles of PT proteins in sperm are not fully understood.

Actin-related proteins (ARPs) share significant amino acid sequence and basal structure similarities with conventional actins. Several ARPs, such as ACTRT1 (Heid et al., 2002), ACTRT2 (Heid et al., 2002), ACTRT3 (Hara et al., 2008), ACTL7A (Boëda et al., 2011; Xin et al., 2020), ACTL7B (Tanaka et al., 2003), ACTL9 (Dai et al., 2021) and ACTL11 (rodent-specific) (Oh et al., 2013), have been reported to be sperm specific and selectively localized to the PT structure of sperm heads. ACTL7A interacts with ACTL9 to form a multimeric protein complex (Dai et al., 2021). Despite their expression pattern being well understood, the physiological roles of these ARPs in sperm biology are far from clear. Two recent interesting studies suggest that disruption of ACTL7A (Xin et al., 2020) or ACTL9 (Dai et al., 2021) causes total fertilization failure and male infertility in humans and mice. Transmission electron microscopy (TEM) analysis further revealed a phenotype of acrosome detachment in ACTL7A- and ACTL9-mutated sperm (Xin et al., 2020; Dai et al., 2021).

In this study, we generated an Actrt1-KO mouse model and found that Actrt1-KO males are severely subfertile as a result of detached acrosomes and a deficiency in fertilization. Intriguingly, the phenotype of Actrt1-KO sperm shows a high similarity to that of Actl7a- and Actl9-mutated sperm, exhibiting acrosome detachment. Mechanistically, ACTRT1 could form a large complex with ACTRT2, ACTL7A and ACTL9. ACTRT1 further interacted with the IAM protein SPACA1 and the NE proteins PARP11 and SPATA46. Collectively, these data suggest a crucial role of the sperm-specific ARP complex ‘ACTRT1-ACTRT2-ACTL7A-ACTL9’ in anchoring the developing acrosome to the nucleus via an IAM-acroplaxome-NE connection.

Western blot analyses of different subcellular components from ACTRT1-Flag-transfected HEK293T cells showed that ACTRT1 was mostly located in the cytoskeleton, with some cytoplasm and membrane expression (Fig. 1A). This observation is consistent with the concept that ACTRT1 is a protein of sperm cytoskeletal PT (Heid et al., 2002). Likewise, ACTRT2, ACTL7A and ACTL9 were all mainly distributed in the cytoskeletal components (Fig. 1A). To establish the cellular localization of ACTRT1, we performed immunofluorescence staining (Fig. 1B). In round and elongated spermatids, ACTRT1 signals were detected in the SAR, and a certain distance existed between ACTRT1 signals and PNA-FITC (an acrosome dye) signals. The ACTRT1 signals in wild-type spermatids were specific because no staining was present using Actrt1-KO spermatids (Fig. S1A). Although colocalization of ACTRT1 and ACTL7A could not be assessed because both are rabbit antibodies, ACTL7A also exhibited SAR distribution in spermatids (Fig. 1B). Intriguingly, ACTRT1 and ACTL7A were translocated to the PAS of mature sperm in the epididymis (Fig. S1B). Although PAS proteins are proposed to be synthesized in the cytoplasmic lobe of elongating spermatids and transported up the manchette for final deposition and assembly in the postacrosomal region (Oko and Sutovsky, 2009), some PAS proteins, such as calicin (Longo et al., 1987; Lécuyer et al., 2000), CAPZA3 (Geyer et al., 2009), ACTRT1 (Heid et al., 2002; this study) and ACTL7A (Boëda et al., 2011; this study), are present in the acroplaxome of spermatids and then translocated to the PAS of spermatozoa. Importantly, Myc-tagged ACTRT1 was immunoprecipitated with Flag-tagged ACTRT2, ACTL7A and ACTL9 in HEK293T cells (Fig. 1C). The interaction between ACTL7A and ACTRT1/2 was further confirmed by endogenous co-immunoprecipitation (co-IP) experiments using mouse testis extracts (Fig. 1D). The assembly of such sperm-specific ARPs into a protein complex suggests that they may play coordinated roles in spermiogenesis.

To examine the physiological role of ACTRT1 in vivo, we generated Actrt1-KO mice using the CRISPR/Cas9 system. The Actrt1 gene has one transcript (ENSMUST00000059466.2) and is located on chromosome X. Exon 1 of Actrt1 was selected as the knockout region because this region contains the entire coding sequence. After genotyping PCR, Actrt1^+/− females were mated with wild-type males to obtain Actrt1^−/Y (Actrt1-KO) mice (Fig. 2A). Specific and complete absence of Actrt1 mRNA in the testis was confirmed by reverse transcription-polymerase chain reaction (RT-PCR) analysis using specific primers for Actrt1. The deletion of ACTRT1 protein in Actrt1-KO testis was confirmed by western blot using an anti-ACTRT1 antibody (Fig. 2A). In general, Actrt1-KO mice were viable and showed no overt abnormalities. To examine the reproductive ability of Actrt1-KO mice, we performed a 2-month fertility test using Actrt1-KO mice and their littermate wild-type males (n=3 each group). Both the pregnancy rate (21.57±5.06%) and number of pups (2.33±0.42) that were born from Actrt1-KO males were significantly reduced compared with those from wild-type males (pregnancy rate: 90.23±5.29%; litter size: 8.56±0.30) (Fig. 2B,C). The severely subfertile phenotype was not due to abnormal mating behaviour because vaginal plugs were detected in all females mated with Actrt1-KO mice. We next performed in vitro fertilization using sperm collected from the cauda epididymis of Actrt1-KO and wild-type male mice. The percentage of two-cell embryos using sperm from Actrt1-KO mice (18.38±1.50%) was significantly decreased compared with that using wild-type sperm (82.87±1.96%) (Fig. 2D). Accordingly, we suggest that Actrt1-KO mice were severely subfertile as a result of a deficiency in fertilization.

Next, we performed sperm analysis of Actrt1-KO mice. Sperm count (1.50×10⁶±0.03×10⁶) and sperm motility (73.33±3.28%) in Actrt1-KO mice showed no significant differences compared with wild-type males (sperm count: 1.63×10⁶±0.05×10⁶; sperm motility: 80.02±2.31%) (Fig. 2E,F). Histological examination of the testis and epididymis revealed that the components in the seminiferous epithelium and sperm number in the caudal epididymis of Actrt1-KO mice were all similar to those of wild-type males (Fig. 2G). It is notable that deformed heads of elongated spermatids (in testis) and sperm (in epididymis) could be observed under high magnification. Accordingly, we examined the morphology of sperm in detail using Papanicolaou staining (Fig. 3A; Fig. S2). In contrast to sperm from wild-type males, which appeared normal, a higher proportion of sperm from Actrt1-KO mice showed deformed heads (58.71±4.35% versus 13.04±0.79%) (Fig. 3B). Likewise, significantly more sperm with abnormal nuclear morphology were observed in Actrt1-KO mice (60.50±2.16%) than in wild-type males (15.15±0.63%), as observed by DAPI staining (Fig. S3A,B). Double staining of sperm using PNA-FITC (an acrosome dye) and MitoTracker Red (a mitochondria dye) further revealed a disturbance of the typical crescent moon shape of the acrosome in Actrt1-KO mice (Fig. 3C). Notably, the proportion of acrosomes that were detached from the nucleus in Actrt1-KO sperm (39.74±2.16%) was significantly higher than that in wild-type sperm (9.13±1.44%) (Fig. 3D). We explored the phenomenon of acrosome detachment from the Actrt1-KO sperm nucleus further by TEM analysis (Fig. 3E; Fig. S4). The number of sperm with detached acrosomes was significantly increased in Actrt1-KO males (76.10±2.78%) compared with wild-type mice (13.16±0.84%) (Fig. 3F). The number of detached acrosomes in Actrt1-KO sperm as observed by TEM was higher than that identified by PNA-FITC staining, owing to the high resolution of TEM. The acroplaxome, an F-actin-containing plate between the acrosome and nucleus, is thought to be involved in acrosome-nuclear connection (Oko and Sutovsky, 2009). The distribution of the acroplaxome was disturbed, and sometimes it broke away from the nucleus in Actrt1-KO sperm (Fig. 3G). Moreover, the manchette of Actrt1-KO spermatids was largely disorganized (Fig. S3C). The manchette anomaly observed in Actrt1-KO spermatids may be a secondary effect of the acrosome anomaly after the loss of ACTRT1 because manchette assembly depends on the stability of acrosome structure (Kierszenbaum and Tres, 2004).

TEM analysis of the testis further revealed a loosened acroplaxome structure in Actrt1-KO spermatids at the cap/acrosome phase of acrosome formation (Fig. 4A). The percentage of spermatids with a loosened acroplaxome structure in Actrt1-KO spermatids (75.59±1.82%) was much higher than that in wild-type spermatids (7.08±1.07%) (Fig. 4B). The acroplaxome thickness that was measured in TEM images was significantly larger in Actrt1-KO spermatids (53.89±2.23 nm) than in wild-type spermatids (23.11±1.30 nm) (Fig. 4C). Moreover, the contents of PT-enriched proteins were analysed in sperm samples from wild-type and Actrt1-KO mice by western blot. We found that the protein levels of ACTL7A and PLCζ were significantly lower in Actrt1-KO sperm than in wild-type sperm. The protein contents of ACTRT2 and lamin A/C were significantly increased, whereas CAPZA3 and CAPZB expression was unchanged in sperm samples after ACTRT1 deletion (Fig. 4D). PLCζ is a well-known sperm-borne oocyte activation factor (Hachem et al., 2017), and disruption of ACTL7A causes sperm fertilization failure owing to reduced expression and abnormal localization of PLCζ (Xin et al., 2020). The reduced ACTL7A and PLCζ that we observed in Actrt1-KO sperm may partially explain the fertilization deficiency of Actrt1-KO male mice.

ACTRT1 may interact with IAM and NE proteins to anchor the developing acrosome to the nucleus. As immunoprecipitation-mass spectrometry could not be performed because of the inability of commercial ACTRT1 antibodies to immunoprecipitate endogenous ACTRT1 protein, we chose a candidate-based approach by searching the literature for IAM and NE proteins for which knockout mice exhibit a deficiency of the IAM-acroplaxome-NE structure. Based on the similar phenotype with Actrt1-KO sperm, SPACA1, DPY19L2, FAM209, PARP11 and SPATA46 are candidate partners of ACTRT1. No close association of the IAM with the NE is formed in Spaca1-deficient spermatids (Fujihara et al., 2012), and, importantly, the IAM protein SPACA1 has been shown to interact with ACTL7A to anchor the acrosome to the acroplaxome (Chen et al., 2021). DPY19L2 is the first NE protein shown to be essential for anchoring the acrosome to the nucleus (Pierre et al., 2012). FAM209 colocalizes and interacts with DPY19L2 at the NE to maintain the developing acrosome (Castaneda et al., 2021). Both PARP11 and SPATA46 are NE proteins, and deficient spermatids exhibit structural defects in the NE associated with an abnormal nuclear shape (Meyer-Ficca et al., 2015; Chen et al., 2016). We found that Flag-tagged ACTRT1 was immunoprecipitated with Myc-tagged IAM protein SPACA1 in HEK293T cells (Fig. 5A). Myc-tagged ACTRT1 could be immunoprecipitated with the Flag-tagged NE proteins PARP11 and SPATA46, whereas no interactions existed between ACTRT1 and the NE proteins DPY19L2 and FAM209 in HEK293T cells (Fig. 5A). The interaction between the IAM protein SPACA1 and ACTRT1 was further confirmed by endogenous co-IP using mouse testis extracts (Fig. 5B). An endogenous interaction of ACTRT1 with the NE proteins PARP11 and SPATA46 was not detected because their commercial antibodies could not be used for IP experiments. Moreover, loss of ACTRT1 did not affect the ACTL7A-ACTRT2 interaction but reduced the connection between ACTL7A and the IAM protein SPACA1 (Fig. 5C). The ACTL7A-SPACA1 interaction is known to anchor the acrosome to the acroplaxome (Chen et al., 2021). Collectively, we suggest that ACTRT1 within the sperm PT-specific ARP complex (ACTRT1-ACTRT2-ACTL7A-ACTL9) anchors developing acrosomes to the nucleus by connecting with the IAM protein SPACA1 and the NE proteins PARP11 and SPATA46 (Fig. 6).

One of the causes of male infertility is abnormal development of spermatids into mature sperm. It is well known that the acrosome and flagella are responsible for oocyte fertilization and sperm motility, respectively. Correspondingly, globozoospermia (round-headed sperm without acrosome) and multiple morphological anomalies of the sperm flagella (MMAF) are two types of asthenoteratozoospermia (Jiao et al., 2021). Our understanding of acrosome biogenesis and flagella formation originates from a large number of gene-deficient animal studies. In contrast, although the protein constituents of PT have been well studied during the last 20 years (Aul and Oko, 2002; Boëda et al., 2011; Hamilton et al., 2017; Heid et al., 2002; Herrada and Wolgemuth, 1997; Hess et al., 1993; Longo et al., 1987; Müjica et al., 2003; Oko and Morales, 1994; Olson and Winfrey, 1988; von Bulow et al., 1997), the physiological roles of the PT remain largely unclear owing to a lack of knockout/mutated mouse models. Among PT-enriched proteins, PLCζ, PAWP, CAPZA3, ACTL7A and ACTL9 have been investigated by loss-of-function studies in mice. Deletion of the Plcz1 gene abolishes the ability of sperm to induce Ca²⁺ oscillations in eggs (Hachem et al., 2017), whereas no abnormalities are observed in Pawp-KO mice (Satouh et al., 2015). Capza3-mutated mice show low sperm concentration, poor motility and abnormally shaped sperm heads (Geyer et al., 2009). Disruption of Actl7a or Actl9 causes acrosomal ultrastructural defects in sperm (Xin et al., 2020; Dai et al., 2021). In this study, we generated a novel mouse line lacking ACTRT1 and found a high proportion of detached acrosomes and partial fertilization failure in Actrt1-KO sperm. Interestingly, the phenotype of Actrt1-KO males is similar to that of Actl7a- and Actl9-mutated mice, which exhibit detached acrosomes in sperm. The study of Actl7a-, Actl9- and Actrt1-deficient sperm indicates a physiological role of the PT in anchoring the developing acrosome to the nucleus. More knockout/mutated mouse models are still needed for a comprehensive understanding of the physiological roles of PT proteins in sperm biology.

Actin is one of the major components of the eukaryotic cytoskeleton. In addition to actin, eukaryotes also contain a conserved family of ARPs that have 17-45% sequence identity to actin. Several sperm PT-specific ARPs have been reported, including ACTRT1 (Heid et al., 2002), ACTRT2 (Heid et al., 2002), ACTRT3 (Hara et al., 2008), ACTL7A (Tanaka et al., 2003), ACTL7B (Chadwick et al., 1999) and ACTL9 (Dai et al., 2021). Two recent studies suggest that ACTL7A and ACTL9 form complexes and participate in acrosomal anchoring and egg activation (Dai et al., 2021; Xin et al., 2020). Chen et al. (2021) further suggested an interaction between ACTL7A and the IAM protein SPACA1, explaining how ACTL7A anchors the acrosome to the acroplaxome. However, the mechanism underlying the anchoring of developing acrosomes to the NE is still largely unclear. ACTRT1 interacts with ACTRT2, ACTL7A and ACTL9, as revealed by our co-IP analysis, suggesting the formation of an ARP complex in the sperm PT layer. More importantly, ACTRT1 could interact with the IAM protein SPACA1 and the NE proteins PARP11 and SPATA46 to mediate the connection between the acrosome and the nucleus by the acroplaxome. However, whether the protein-protein interaction is direct or indirect needs further investigations. The phenotypes of Actrt1-KO, Actl7a-mutated and Actl9-mutated mice show a high degree of similarity, including detachment of the acrosome, loosened acroplaxome structure and sperm-derived fertilization failure. However, the defect in proacrosomal vesicle fusion shown in Actl7a- and Actl9-mutated spermatids was not observed in Actrt1-KO spermatids. It is unclear whether the fusion of proacrosomal vesicles into the acrosome is dependent on proper acroplaxome formation. Furthermore, Actrt1-KO mice showed partial fertilization failure whereas Actl7a- and Actl9-mutated mice exhibited total fertilization failure. We suggest a partial functional redundancy between ACTRT1 and ACTRT2 because (1) these two proteins share over 70% similarity in nucleotide sequence; (2) ACTRT2 expression was increased in sperm samples from Actrt1-KO mice compared with those from control mice; and (3) ACTRT2 could also interact with ACTL7A, ACTL9, SPACA1 and PARP11 (Fig. S5). Generation and characterization of Actrt1/2 double knockout mice will be helpful to examine whether functional redundancy truly exists between these two ARPs. The partial fertilization failure of Actrt1-KO mice could be explained by the reduced protein content of PLCζ and ACTL7A in Actrt1-KO sperm because PLCζ is a well-known sperm PT protein that induces Ca²⁺ oscillations after entering oocytes (Escoffier et al., 2016; Hachem et al., 2017), and Actl7a-mutated sperm also fail to fertilize via intracytoplasmic sperm injection (Xin et al., 2020). Although reduced expression and/or abnormal localization of PLCζ are observed in Actl7a-, Actl9- and Actrt1-deficient sperm, the mechanisms underlying the regulation of PLCζ by ACTL7A, ACTL9 and ACTRT1 are still unknown. These ARPs may influence the content and localization of PLCζ in an indirect way.

While we were preparing our manuscript, Sha et al. reported two point mutations (c.95G>A and c.662A>G) of the ACTRT1 gene in two ASS patients, and approximately 60% of sperm from Actrt1-KO mice were headless (Sha et al., 2021). Surprisingly, the phenotype observed in our Actrt1-KO mice was inconsistent with that in the Actrt1-KO mice in the Sha study. There was no significant difference in the deletion strategy between these two KO mouse models because both used the C57BL/6 mouse strain and targeted the whole Actrt1 gene (this gene has only one exon). However, we did not observe any ASS phenomenon in our Actrt1-KO mice; instead, detachment of the acrosome from the nucleus was identified in our mouse model. Headless sperm are actually ‘pinhead sperm’ because the unremoved cytoplasm and misarranged mitochondria attached to the top of the flagella are often regarded as heads of reduced size (Baccetti et al., 1989; Le Lannou, 1979; Perotti and Gioria, 1981). SPATA6 (Yuan et al., 2015), SUN5 (Shang et al., 2017), PMFBP1 (Zhu et al., 2018) and CNTLN (Zhang et al., 2021) are well-known ASS-related genes; indeed, residual cytoplasm and disordered mitochondria can be clearly observed at the fore-end of the flagella in their corresponding KO mice (Shang et al., 2017; Yuan et al., 2015; Zhang et al., 2021; Zhu et al., 2018). In Sha's Actrt1-KO sperm, the separated heads and tails are morphologically normal, and, notably, with no cytoplasm or misarranged mitochondria attached to the top of the tails. The sperm head-to-tail coupling apparatus (HTCA) ensures sperm head-tail integrity, and defective HTCA is the basic cause of acephalic spermatozoa. ACTRT1 is a well-known PT protein surrounding the nuclei of sperm (Heid et al., 2002), and there is no evidence to indicate that ACTRT1 can regulate the HTCA structure. Data from independent laboratories (Chen et al., 2021; Dai et al., 2021), including ours (Figs 1 and 5), suggest that ACTRT1, ACTL7A and ACTL9 form a large ARP complex and interact with the IAM protein SPACA1. The phenotype of our Actrt1-KO males (Figs 3 and 4) was identical to that of Actl7a- and Actl9-mutated mice (Dai et al., 2021; Xin et al., 2020), showing a phenotype of acrosome detachment but not ASS. Many studies from independent laboratories support the suggestion that mutations in SUN5 and PMFBP1 are the most common aetiologies of human ASS (Elkhatib et al., 2017; Liu et al., 2020; Shang et al., 2017; Zhu et al., 2018; Zhu et al., 2016). In contrast, the clinical evidence to link ACTRT1 to human ASS is not strong enough because the consequences of a single-point mutation on ACTRT1 function have not been proven by any experiments. Accordingly, we suggest that there is currently not enough evidence to indicate that ACTRT1 is an ASS-related gene.

Animal experiments were approved by the Animal Care and Use Committee of the College of Life Sciences, Beijing Normal University. The mouse Actrt1 gene has one transcript (ENSMUST00000059466.2) and is located on chromosome X. The Actrt1 gene has only one exon, and the whole Actrt1 gene was selected as the knockout region. In brief, we selected two sgRNA primers with high scores (CRISPR finder: https://www.sanger.ac.uk/htgt/wge/find_crisprs) to generate a deletion of the whole Actrt1 gene in mice (GRCm38). The wild-type genomic sequence is provided in Table S1, in which the Actrt1 gene and the position of gRNAs are marked in red and blue, respectively. The sgRNAs (5′-CCATTGGTTGCTCAGTTCAA-3′ and 5′-CTGGATAAGTAAAGTAACTC-3′) were synthesized by Sangon Biotech (Shanghai, China). The two complementary DNA oligos of each sgRNA target were annealed (95°C for 5 min and then naturally cooled to room temperature) and ligated to the pUC57-sgRNA plasmid (Addgene plasmid #51132) for cloning. The recombinant plasmid was transformed into DH5α competent cells, and the positive clone was screened based on kanamycin resistance and sequencing. The recombinant plasmid was linearized and purified by phenol chloroform extraction. Transcriptions of the sgRNAs in vitro were performed using the MEGAshortscript Kit (AM1354, Ambion) and purified using the MEGAclear Kit (AM1908, Ambion). Cas9 mRNA was purchased from TriLink BioTechnologies (L-7206). Mouse zygotes were co-injected with an RNA mixture of Cas9 mRNA (50 ng/μl) and sgRNA (30 ng/μl). The injected zygotes were transferred into pseudopregnant recipients to obtain the F0 generation. DNA was extracted from tail tissues from 7-day-old offspring, and PCR amplification was carried out with genotyping primers (Table S2) using the Mouse Tissue Direct PCR Kit (Tiangen Biotech) under the following conditions: 94°C for 5 min; 35 cycles of 94°C for 30 s, 61°C for 30 s and 72°C for 30 s; and a final step of 72°C for 5 min. PCR products were run on a 1% agarose gel in 1×TBE buffer and then subjected to sequencing (Sangon Biotech). A stable F1 generation (heterozygous mice) was obtained by mating positive F0 generation mice with wild-type C57BL/6JGpt mice.

Total RNA was extracted from mouse testes using an RNA Easy Fast Tissue/Cell Kit (Tiangen Biotech, DP451). First-strand cDNA synthesis and RT-PCR were performed using a FastKing One-Step RT-PCR Kit (Tiangen Biotech, KR123) according to the manufacturer's instructions. The amplification conditions for the RT-PCR were 5 min at 94°C, followed by 35 cycles of 94°C for 30 s, 62°C for 30 s and 72°C for 30 s, with a final 5-min extension at 72°C. The primers for the specific recognition of Actrt1 and Actrt2 mRNA are listed in Table S3.

Mouse Actrt1 cDNA was chemically synthesized by GenScript Biotech Corporation (Suzhou, China) and inserted into Flag- or Myc-tagged pCMV vectors (D2632 and D2672, Beyotime, Shanghai, China). Full-length cDNA encoding ACTRT2, ACTL7A, ACTL9, SPACA1, DPY19L2, FAM209, SPATA46 and PARP11 was amplified by PCR using mouse testis cDNA as the template and cloned into Flag- or Myc-tagged pCMV vectors. The primers for plasmid construction are listed in Table S4. The construction of expression plasmids in this study was confirmed by sequencing (Sangon Biotech). HEK293T cells (ATCC, CRL-11268) were cultured at 37°C in a 5% CO₂ incubator (Panasonic/Sanyo CO₂ incubator MCO-18AIC) with Dulbecco's modified Eagle's medium (Thermo Fisher Scientific, Gibco™, 10569-044) with 10% fetal bovine serum (HyClone, 10099-141) and 1% penicillin-streptomycin (Thermo Fisher Scientific, Gibco™, 15140-163). The transient transfection of HEK293T cells was performed using Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific, Invitrogen™, 11668-019) following the manufacturer's protocol (Thermo Fisher Scientific). Cells were then harvested 48 h after transfection.

To confirm the fertility of Actrt1-KO mice, natural mating tests were conducted. Briefly, three Actrt1-KO and three littermate control sexually mature male mice (8-12 weeks old) were paired with two 6- to 8-week-old C57BL/6J females (each male was mated with two female mice) for 2 months. The vaginal plugs of the mice were examined every morning. Then, the female mice with vaginal plugs were separately fed, and the number of pups per litter was recorded.

Testes and epididymis from adult male mice were dissected and fixed with 4% paraformaldehyde (PFA) (Beyotime, P0099) overnight at 4°C. Fixed tissues were embedded in paraffin, sectioned (5 μm thick), dewaxed and rehydrated. The sections were stained with Haematoxylin & Eosin solution (Solarbio Life Sciences, G1120) before imaging using a DM500 optical microscope (Leica Microsystems, Germany).

The cauda epididymis was recovered, immersed in prewarmed HTF medium (Nanjing Aibei Biotechnology, M1130), and cut into several pieces to let sperm swim out at 37°C in a 5% CO₂ humidified incubator for 15 min. Sperm counts were determined using a fertility counting chamber (Makler, Haifa, Israel) under a light microscope, and sperm mobility was assessed using a computer-assisted sperm analysis (CASA) system.

Sperm were collected from the cauda epididymis and washed three times in PBS buffer. The sperm suspension was mounted on a glass slide, air-dried, and fixed with 4% PFA for 10 min at room temperature. The slides were stained with Papanicolaou solution (Solarbio Life Sciences, G2571) and observed using a DM500 optical microscope (Leica Microsystems).

Eight-week-old C57BL/6J female mice were superovulated by injecting 5 IU (0.1 ml) of pregnant mare serum gonadotropin (Nanjing Aibei Biotechnology, M2620), followed by 5 IU (0.1 ml) of human chorionic gonadotropin (hCG) (Nanjing Aibei Biotechnology, M2520) 48 h later. The sperm was released from the cauda epididymis of 10-week-old male mice, and sperm capacitation was performed for 50 min using TYH solution (Nanjing Aibei Biotechnology, M2030). Cumulus-oocyte complexes (COCs) were obtained from the ampulla of the uterine tube at 14 h after hCG injection. The ampulla was torn with a syringe needle, and the COCs were gently squeezed onto the liquid drops of HTF medium. COCs were then incubated with 5-10 μl sperm suspension (sperm concentration: 1∼5×10⁶) in HTF liquid drops at 37°C under 5% CO₂. After 6 h, the eggs were washed several times using HTF medium to remove the cumulus cells and then transferred to liquid drops of KSOM medium (Nanjing Aibei Biotechnology, M1430). Two-cell embryos were counted at 24 h postfertilization. All lipid drops were covered with mineral oil (Nanjing Aibei Biotechnology, ART-4008P) and equilibrated overnight at 37°C under 5% CO₂.

Spermatogenic cells were isolated from adult mouse testes by digestion with 1 mg/ml collagenase IV (Solarbio Life Sciences, G8160), 1 mg/ml hyaluronidase (Solarbio Life Sciences, H8030), 1 mg/ml trypsin (Solarbio Life Sciences, P7340) and 0.5 mg/ml DNase I (Solarbio Life Sciences, D8071) for 20 min on a shaker. A cell suspension containing spermatids was mounted on a glass slide, air-dried, and fixed with 4% PFA for 10 min at room temperature. After permeabilization with 1% Triton X-100 (Solarbio Life Sciences, P1080) for 30 min, the slides of sperm and spermatids were blocked with 5% goat serum (Solarbio Life Sciences, SL050) for 45 min. Anti-ACTRT1 antibody or anti-ACTL7A antibody was added to the slide and incubated overnight at 4°C. After washing three times with PBS, slides were incubated with Alexa Fluor 555-labelled donkey anti-rabbit IgG for 1 h at room temperature. For the staining of the acrosome, acroplaxome and manchette, PNA-FITC dye (Sigma-Aldrich, L7381), Actin-Tracker Red-555 dye (Beyotime, C2203) and Tubulin-Tracker Red dye (Beyotime, C1050) were used, respectively. For mitochondrial sheath staining, live sperm (before 4% PFA fixation) were stained with Mito Tracker Red (Beyotime, C1035). The nuclei were counterstained with DAPI dye (Beyotime, C1005).

Precipitation of mouse sperm and testis tissues (∼1 mm³) were fixed with 2.5% (vol/vol) glutaraldehyde in 0.1 M phosphate buffer (PB) (pH 7.4) for 24 h at 4°C. The samples were washed four times in PB and first immersed in 1% (wt/vol) OsO₄ and 1.5% (wt/vol) potassium ferricyanide aqueous solution at 4°C for 2 h. After washing, the samples were dehydrated through graded alcohol (30%, 50%, 70%, 80%, 90%, 100%, 100%, 10 min each) into pure acetone (10 min twice). Samples were infiltrated in a graded mixture (3:1, 1:1, 1:3) of acetone and SPI-PON812 resin (21 ml SPO-PON812, 13 ml DDSA and 11 ml NMA), and then the pure resin was changed. The specimens were embedded in pure resin with 1.5% BDMA and polymerized for 12 h at 45°C, 48 h at 60°C, cut into ultrathin sections (70 nm thick), and then stained with uranyl acetate and lead citrate for subsequent observation and photography with a Tecnai G2 Spirit 120 kV (FEI) electron microscope. All reagents were purchased from Zhongjingkeyi Technology (Beijing, China).

Forty-eight hours after transfection, HEK293T cells were lysed with Pierce™ IP Lysis Buffer (Thermo Fisher, 87787) with protease inhibitor cocktail (MedChemExpress, HY-K0010) for 30 min at 4°C and then centrifuged at 12,000 g for 10 min. Pierce IP Lysis Buffer was composed of 25 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40 and 5% glycerol. To prepare input samples, 30 μl protein lysates (∼25 μg) were collected and boiled for 5 min in 1.2×SDS loading buffer (Beyotime, P0015). The lysates were precleared with 10 μl Pierce™ Protein A/G-conjugated Agarose (Thermo Fisher, 20422) for 1 h at 4°C. Precleared lysates were incubated overnight with 2 μg anti-Myc antibody (Abmart, M20002) or anti-Flag antibody (Abmart, M20018) at 4°C. The lysates were then incubated with 20 μl Pierce™ Protein A/G-conjugated Agarose for 2 h at 4°C. The agarose beads were washed four times with Pierce™ IP Lysis Buffer and boiled for 5 min in 30 μl 1.2×SDS loading buffer. Input (∼25 μg) and IP (∼30 μl) samples were analysed by western blotting using anti-Flag or anti-Myc antibodies. For endogenous co-IP, adult mouse testis tissues were lysed with Pierce™ IP Lysis Buffer. Precleared lysates were separated into two groups: one group was treated with 2 μg anti-ACTL7A antibody (Proteintech, 17355-1-AP) or anti-SPACA1 antibody (Abcam, ab191843), and another group (negative control) was treated with 2 μg rabbit IgG (Beyotime, A7016). Other endogenous co-IP procedures were similar to the co-IP assay in HEK293T cells.

Protein extraction from different subcellular structures was performed according to the manufacturer's instructions (Sangon Biotech, C500073). Proteins were divided into four fractions containing cytosolic proteins, membrane and organelle proteins, nuclear proteins, or cytoskeletal filaments. Briefly, Extraction Buffer 1 was added to the cells and oscillated for ∼10 min to lyse the cells. After centrifugation at 800 g for 10 min, the supernatant was collected as the cytosolic fraction. The pellet was resuspended in Extraction Buffer 2 and oscillated for ∼30 min to solubilize membrane proteins. We collected the supernatant as a membranous fraction after another centrifugation step at 5000 g for 10 min. Extraction Buffer 3 was used for the solubilization of nuclear proteins under oscillation for ∼30 min. Finally, the cytoskeletal proteins were pelleted at 6780 g for 10 min and completely dissolved in Extraction Buffer 4. Markers of cytosolic proteins (HSP90), membrane and organelle proteins (calnexin), nuclear proteins (MLH1) and cytoskeletal filaments (vimentin) were used to assess protein loading and fraction purity.

Proteins from HEK293T cells were extracted using RIPA lysis buffer (Applygen, C1053) containing 1 mM phenylmethylsulphonyl fluoride and protease inhibitors (Applygen, C1055) on ice. For sperm samples, 1% SDS (Beyotime, ST628) was added to the RIPA lysis buffer. The supernatants were collected following centrifugation at 12,000 g for 20 min. We loaded 30 μg protein per lane. Proteins were electrophoresed in 10% SDS-PAGE gels (Beyotime, P0052A) and transferred to nitrocellulose membranes (GE Healthcare). The blots were blocked in 5% milk (Beyotime, P0216) and incubated with primary antibodies overnight at 4°C, followed by incubation with anti-rabbit or mouse IgG H&L (HRP) (Abmart) at a 1/10,000 dilution for 1 h. The signals were evaluated using the Super ECL Plus Western Blotting Substrate (Applygen, P1050) and a Tanon-5200Multi chemiluminescence imaging system (Tanon, Shanghai, China). The antibodies used in this study are listed in Table S5.

Data were compared for statistical significance using GraphPad Prism version 5.01 (GraphPad Software). Unpaired, one-tailed Student's t-test was used for the statistical analyses. The data are presented as the mean±s.e.m, and statistically significant differences are represented as *P<0.05, **P<0.01 and ***P<0.001.

We would like to thank Dr Liu Jin and Dr Xi Chao from the Experimental Technology Center for Life Sciences, Beijing Normal University for technical assistance. We also acknowledge Dr Li Xi-Xia and Lv Zhong-Shuang from the Center for Biological Imaging (CBI), Institute of Biophysics, Chinese Academy of Sciences for TEM analysis.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200489.