SPACA1-deficient male mice are infertile with abnormally shaped sperm heads reminiscent of globozoospermia

The acrosome is a membrane-bound cap-like structure that covers the anterior portion of the mammalian sperm nucleus. Although the size and shape of the acrosome vary considerably from species to species, its basic structure is the same in all eutherian mammals (Yanagimachi, 1994). In the testis, acrosome formation is initiated by the production of numerous proacrosomal granules in trans-Golgi stacks, followed by fusion into a single large acrosomal granule that associates with the nuclear envelope (Abou-Haila and Tulsiani, 2000). Subsequently, the acrosome increases in size and flattens (thinning) over the surface of the nucleus and assumes its mature species-specific conformation.

Globozoospermia is a human infertility syndrome caused by defects manifested during spermatogenesis (Dam et al., 2007a). The characteristic feature of globozoospermia is the malformation or loss of the acrosome accompanied by an abnormal nuclear shape as well as an abnormal arrangement of the sperm mitochondria (Battaglia et al., 1997). Given the tendency for a familial appearance of globozoospermia, the disease has become a target for quantitative trait locus analysis in humans (Taskiran et al., 2006).

In the mouse, various genes were found to be associated with globozoospermia by gene disruption experiments, including Csnk2a2 (Xu et al., 1999), Hrb (Agfg1 – Mouse Genome Informatics) (Kang-Decker et al., 2001), Gopc (Yao et al., 2002), Gba2 (Yildiz et al., 2006), Zpbp1 (Zpbp – Mouse Genome Informatics) (Lin et al., 2007) and Pick1 (Xiao et al., 2009). Interestingly, except for Zpbp1, these disrupted genes are expressed in all body tissues, but the globozoospermia phenotype was found to arise exclusively in spermatogenesis. Although all six genes seem to function at different stages of Golgi transport and acrosomal biogenesis, these proteins must be closely related because the phenotypes of their defects are very similar. Elucidation of the relationships among these proteins is essential to understand the molecular mechanism that leads to globozoospermia (Matzuk and Lamb, 2008). We found, serendipitously, sperm acrosome associated 1 (Spaca1) to be a new member of the globozoospermia-related genes. By examining Gopc- and Zpbp1-disrupted mouse lines, we demonstrated that SPACA1 is missing in both. Evaluating the phenotypes of these mice will help elucidate the mechanism underlying globozoospermia.

The Gopc knockout mouse strain [RBRC01253 (Yao et al., 2002)] was provided by the RIKEN BioResource Center through the National Bio-Resource Project of MEXT, Japan. The Zpbp1 knockout mouse strain was provided by Dr Martin M. Matzuk [Baylor College of Medicine, Houston, TX, USA (Lin et al., 2007)]. Wild-type mice and rats were purchased from Japan SLC (Shizuoka, Japan). All animal experiments were approved by the Animal Care and Use Committee of the Research Institute for Microbial Diseases, Osaka University, Japan.

A targeting vector was constructed using pNT1.1 [http://www.ncbi.nlm.nih.gov/nuccore/JN935771 (Fujihara et al., 2010)]. A 1.5 kb NotI-SalI fragment as a short arm and a 1.2 kb EcoRI-AflII fragment including exons 3 and 4 as a middle arm and a 4.3 kb KpnI-AvrII fragment as a long arm were obtained by PCR amplification using genomic DNA derived from C57BL/6N mice as a template. The PCR primers used were (5′-3′): TGCGGCCGCTGCCTGGAATCGGCATGAT and CGAGTCGACGTAACTGCCATGGATCTTCCC for the short arm; CGAATTCGTGTTGGTATCAGAGAAGTAA and GCTTAAGTTGGTCTGGCTTTAGAAGC for the middle arm; CGGTACCGTGAGTATTCTGTGTATAAATAATCC and GCCTAGGCCTATCACAAAGGAACTCAC for the long arm. These three fragments were inserted into a pNT1.1 vector and the targeting construct was linearized with NotI. Embryonic stem (ES) cells were electroporated and colonies were screened as described below.

The mouse Spaca1 gene consists of seven exons and maps to chromosome 4. The targeting vector was designed to remove the third and fourth exons of Spaca1 (supplementary material Fig. S2A) and was electroporated into EGR-G01 [129S2 × (cag/acr-EGFP)C57BL/6N F1] ES cells after linearization. Potentially targeted ES cell clones were separated by positive/negative selection with G418 and ganciclovir. Correct targeting of the Spaca1 allele in ES cell clones and germline transmission were determined by PCR analysis. Screening primers were (5′-3′): CACAGGGAAGCTGCTCGTG and GCTTGCCGAATATCATGGTGGAAAATGGCC for the short arm; CAGCCTCTGAGCCCAGAAAGCG and AGGCTGCAAATAACCCAGATGC for the middle and long arms.

Spaca1-floxed mice obtained from germline transmission mouse chimeras had normal fertility. For the removal of loxP-flanked exons, mating between F1 mice and CAG-Cre transgenic mice [B6.Cg-Tg(CAG-Cre)CZ-MO2Osb] yielded offspring that carried a heterozygous deficiency of the Spaca1 gene (supplementary material Fig. S2A). Mating between heterozygous mutant mice yielded the expected Mendelian ratios: Spaca1^+/+, 22.7%; Spaca1^+/–, 50.0%; Spaca1^–/–, 27.3% of offspring (n=66). Genotyping primers were (5′-3′): GGGAGGGAAGATCCATGGCA and GCGCTAACACTCATTGATCTGGC. Both a 1.6 kb band for the wild-type allele and a 0.5 kb band for the mutant allele were amplified by PCR (supplementary material Fig. S2B). Western blot analysis showed that SPACA1 protein was undetectable in Spaca1^–/– testis (supplementary material Fig. S2C). Spaca1^–/– mice were healthy and showed no overt developmental abnormalities. Mice used in this study were of a B6;129 mixed background.

The Spaca1-disrupted mouse line has been submitted to the RIKEN BioResource Center (http://www.brc.riken.jp/inf/en/index.shtml) and is available to the scientific community.

Sexually mature male mice of genotypes Spaca1^+/+, Spaca1^+/– and Spaca1^–/– were caged with (C57BL/6N × DBA/2N)F1 (also known as B6D2F1) female mice (at least 2 months old; CLEA Japan, Tokyo, Japan) for 2 months and the number of pups in each cage was counted within a week of birth.

Rabbit anti-mouse SPACA1 polyclonal antiserum was produced by immunization with mouse SPACA1 polypeptide (QSPTDIPVHEDDALSEWNE). A polyclonal antibody against human SPACA1 (#BP5112, C-terminal domain: CGEDDALSEWNE) was purchased from Acris Antibodies (Herford, Germany). A rabbit anti-GOPC polyclonal antibody was a gift from Dr Tetsuo Noda [The Cancer Institute of the Japanese Foundation for Cancer Research, Tokyo, Japan (Yao et al., 2001)]. A rabbit anti-ZPBP1 antibody (G176) was a gift from Dr Martin M. Matzuk (Lin et al., 2007). The mouse monoclonal antibody MN9 was a gift from Dr Kiyotaka Toshimori [Chiba University, Chiba, Japan (Toshimori et al., 1992)]. A goat anti-basigin (CD147) polyclonal antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). A polyclonal antibody against mouse calnexin and the monoclonal antibody KS64-10 (for SLC2A3) were established previously in our laboratory according to standard methods (Ikawa et al., 2001; Yamaguchi et al., 2006).

Immunohistochemistry was performed as described previously for the testicular germ cells shown in Fig. 1D and supplementary material Fig. S1 (Kotaja et al., 2004) and for the epididymal spermatozoa shown in Fig. 1E (Yamaguchi et al., 2008) using a fluorescence microscope (Olympus BX50).

Scanning electron microscopy (SEM) (Fujihara et al., 2010) and transmission electron microscopy (TEM) (Inoue et al., 2011) were performed as described previously.

Immunoblot analyses (Inoue et al., 2008), extraction of human sperm proteins (Mandal et al., 1999) and phase separation of Triton X-114 extracts of sperm proteins (Yamaguchi et al., 2006) were performed as described. All samples were separated by SDS-PAGE under reducing conditions. Peptide-N-glycosidase F (PNGase F) (Roche Applied Science, Germany) treatment was performed as described (Inoue et al., 2008). Testis lysates were treated with 250 mU PNGase F for 16 hours at 37°C according to the manufacturer’s instructions. The treated lysates were subjected to SDS-PAGE followed by western blotting.

RT-PCR analysis using testicular cDNA was performed as described (Fujihara et al., 2010) using the following primers (5′-3′): GGCGGCAGTGAAATCCTTCT and TGCTCTAATGGTCTCTGGTTTCTC for Spaca1; CCACTCAGCTTCAGCTTCATGCC and AGCCTGGAGAACAGCTATGTGCC for Gopc; TTGGACATTTGGCTCGACTGCC and CCTTTTGGCCCATGCCATTGG for Zpbp1; ACATTCACGGAAGCGCTGGG and ACGGTGTTCTCAGCACAAGGC for Csnk2a2; AGTTTCCAGC AGCCTGCCTTC and GAGCTTCCTGTTGGAAGTTGTCCT for Hrb; CTTTGCAGCAAAGGAGCAGCTC and CGGCTGCCTGACGAAG ATGG for Spata16; GGCTCGAGCCCGCTTCTC and GCCTTGGT TAGGGCCATAGGC for Pick1; AAGTGTGACGTTGACATCCG and GATCCACATCTGCTGGAAGG for β-actin (Actb).

The amplification conditions were 1 minute at 96°C, followed by 40 cycles (for Fig. 1A) or 27 cycles (for Fig. 4A) of 96°C for 30 seconds, 65°C for 30 seconds, and 72°C for 30 seconds, with a final 1-minute extension at 72°C. The expression of Scp3 (Sycp3 – Mouse Genome Informatics) and Prm1, which were used as markers for the meiotic spermatocyte and the spermatid stages, respectively, had been established previously (Fujihara et al., 2010).

The expression of Spaca1 in various organs was examined by RT-PCR. Spaca1 was exclusively expressed in testis and the SPACA1 protein was found in spermatozoa as previously reported in human (Hao et al., 2002), boar, ram and bull (Jones et al., 2008) (Fig. 1A,B). Presence of SPACA1 was demonstrated by western blotting after proteins had been extracted from spermatozoa using Triton X-114 treatment followed by phase separation and SDS-PAGE. SPACA1 was detected in the detergent-enriched phase, indicating that SPACA1 is a transmembrane protein (Fig. 1C).

Immunostaining over a range of developmental stages (supplementary material Fig. S1) revealed that SPACA1 first appears at the step 2 stage of round spermatids and was localized exclusively on a peripheral part of the acrosomal membrane, forming a ring shape in step 7 spermatids (Fig. 1D). After the sperm head and acrosome had elongated, SPACA1 became localized on the equatorial segment of the acrosome. SPACA1, being a transmembrane protein, was not released from spermatozoa even after exocytosis of the acrosomal contents in the acrosome reaction, which is essential for spermatozoa to fertilize eggs (Fig. 1E).

Spaca1 was disrupted by homologous recombination (supplementary material Fig. S2A). RT-PCR analysis confirmed the successful elimination of the Spaca1 gene and western blot analysis demonstrated the absence of SPACA1 in Spaca1^–/– testes (supplementary material Fig. S2B,C). To examine the fertilizing ability of Spaca1-disrupted mouse line, three to five adult males of each genotype were bred with wild-type females for 2 months. Whereas Spaca1^+/– mice showed similar fertility to Spaca1^+/+ mice, the Spaca1^–/– males were completely sterile (Fig. 1F). The mean litter size in Spaca1^+/+ males was 9.6±1.6 (n=15) and in Spaca1^+/– males it was 8.7±2.5 (n=23). The testicular weights of Spaca1^–/– (109.5±12.9 mg; n=12) and Spaca1^+/+ (113.5±11.2 mg; n=12) mice were similar (supplementary material Fig. S2D). Spaca1^–/– females were fertile (Fig. 1F; mean litter size 8.7±2.6; n=23), reflecting the fact that Spaca1 is not expressed in female reproductive organs.

The swimming ability of Spaca1^–/– spermatozoa was significantly impaired (supplementary material Movie 1). Moreover, when the Spaca1^–/– spermatozoa were examined under a fluorescence microscope, all of the testicular spermatozoa showed a round-headed morphology (Fig. 2A), and spermatozoa from the cauda epididymidis showed a tail coiled around the head (Fig. 2B; supplementary material Fig. S2F). To analyze acrosomal integrity, Spaca1^–/– males were mated with Acr-EGFP transgenic mice (Nakanishi et al., 1999) in which acrosomes can be visualized by enhanced green fluorescence (EGFP). All of the round-headed spermatozoa from Spaca1^–/– mice had an acrosome of negligible size (Fig. 2A,B). A few of the Spaca1^–/– spermatozoa (<5%) had small acrosomes of aberrant shape.

These abnormalities were also examined by SEM and TEM. Disruption of Spaca1 caused hypoplasia of intermediate filament bundles, which normally form a characteristic dense structure on the nuclear membrane facing the inner acrosomal membrane (Fig. 2C,D) (Ito et al., 2004). Thinning of the acrosome was not observed in Spaca1^–/– spermatids. Mature spermatozoa had disorganized mitochondria trapped inside large cytoplasmic droplets, misshapen nuclei and ectopic localization of the flagella (Fig. 2E,F). The Spaca1^–/– spermatozoa were seen to be round-headed in SEM analysis, in contrast to the hook-shaped spermatozoa observed in wild type (Fig. 2F). These abnormalities are reminiscent of the defects seen in the human condition globozoospermia (Dam et al., 2007a). Similar to Gopc null (Suzuki-Toyota et al., 2007) and Zpbp1 null (Lin et al., 2007) mice, this morphological aberrance became more evident while the spermatozoa were passing through the epididymis. Although these globozoospermic spermatozoa are abnormal in appearance and exhibit no ability to fertilize eggs by themselves, when the sperm head is introduced into an egg by intracytoplasmic sperm injection (ICSI), healthy pups can be obtained (Dam et al., 2007a; Yao et al., 2002). In our studies, we demonstrated that Spaca1^–/– spermatids can produce healthy pups by performing round spermatid injection (ROSI) into eggs (supplementary material Fig. S3).

Abnormal spermatogenesis was not obvious in routine histology sections prepared from Spaca1^–/– testes stained with Hematoxylin and Eosin (supplementary material Fig. S2E). We crossed the Spaca1^–/– male mice with transgenic mice expressing EGFP in their acrosome so that acrosome formation during spermatogenesis could be examined easily by fluorescence microscopy (Nakanishi et al., 1999). We found that the acrosome was shaped normally until around step 3 of spermatogenesis (Fig. 3). The amount of EGFP (estimated from its brightness) transferred from the Golgi apparatus to the acrosome in Spaca1^–/– mice also seemed to be equivalent to that in wild-type spermatids, at least in the early steps of spermatogenesis (Fig. 3). However, the abnormality became evident in spermatids at around step 6. By this stage, accumulation of EGFP was observed in the acrosomal granule in wild-type spermatids together with a thinly extended peripheral part of the acrosome (Fig. 3, arrows). In Spaca1^–/– spermatids, by contrast, the accumulation of GFP in the acrosomal granule was not apparent and the acrosomal extension was insufficient to form a thin peripheral structure (Fig. 3, arrowheads). Formation of the acrosome did not progress further after this step and in most of the spermatids it degenerated and disappeared (Fig. 3). Differences in nuclear shape were also noted. The nuclei of Spaca1^–/– spermatids were smaller than those of wild-type spermatids, probably because they remained globular rather than increasing their apparent area by thinning as in wild-type spermatids (Fig. 2E,F).

To clarify the functional and temporal assembly of the acrosome with respect to the components expressed at different stages of acrosomal biogenesis, the onset of Spaca1 expression was compared by RT-PCR (Fig. 4A) and western blotting (Fig. 4B) with that of other globozoospermia-related genes: Csnk2a2 (Xu et al., 1999), Hrb (Kang-Decker et al., 2001), Gopc (Yao et al., 2002), Zpbp1 (Lin et al., 2007), Spata16 (Dam et al., 2007b) and Pick1 (Xiao et al., 2009). Spaca1 expression was the last to appear among all the strains examined (Fig. 4A). SPACA1 protein was detectable only after 4 weeks of age in these mice (Fig. 4B). Using PNGase F treatment, SPACA1 was shown to be glycosylated, as is the case for ZPBP1 (Lin et al., 2007) (Fig. 4C). It should be noted that the amounts of SPACA1 and ZPBP1 were decreased significantly in the Gopc-disrupted testis (Fig. 4C). By contrast, the amount of GOPC was unaffected by the disruption of either Zpbp1 or Spaca1 (Fig. 4D,E). Likewise, disruption of Zpbp1 caused the loss of SPACA1 (Fig. 4D), but ZPBP1 expression remained unchanged in the Spaca1-disrupted testis (Fig. 4E). These results indicate that SPACA1 must be functioning further downstream than any of the other key molecules associated with globozoospermia.

In eutherian mammals, the equatorial segment of the sperm head is the site where sperm-egg fusion initiates (Bedford and Cooper, 1978). To date, IZUMO1 is the only sperm factor reported to be essential for fusion, but it is localized over the entire head (Inoue et al., 2005). For this reason, we were actively seeking a fusion factor localized only in the equatorial segment. It was reported that SPACA1 (SAMP32) is localized in this region (Hao et al., 2002; Jones et al., 2008) and that an anti-SPACA1 antibody could inhibit the fusion of human spermatozoa with zona-free hamster eggs (Hao et al., 2002). Therefore, to investigate the role of SPACA1 in sperm-egg fusion, we produced a Spaca1-disrupted mouse line. The male mice were infertile, as we expected, but the major cause of this sterility was found to be an abnormal shaping of spermatozoa reminiscent of human globozoospermia.

Because the acrosome is a large Golgi-derived secretory vesicle containing a variety of enzymes, it was unsurprising to find that many of the factors (CSNK2A2, HRB, GOPC and PICK1) that are associated with globozoospermia function in protein transport to the Golgi apparatus and/or acrosome. However, SPACA1 is an integral acrosomal membrane protein, indicating that the major cause of globozoospermia is not aberrant protein transport in the Golgi or into the acrosome. Rather, the globozoospermia in Gopc- and Zpbp1-disrupted mouse lines is due to the loss of SPACA1 from their acrosomes (Fig. 4C,D).

Disruption of Spaca1 led to poor formation of intermediate filament bundles, which normally form a characteristic dense structure on the nuclear membrane facing the inner acrosomal membrane (Ito et al., 2004) (Fig. 2C,D). We presume that this leads both to the failure of thinning of the acrosome and to the impaired formation of the acroplaxome, an F-actin/keratin-containing plate that separates the inner acrosomal and nuclear membranes (supplementary material Fig. S4). At present, no apparent functional domain has been found in the SPACA1 protein. However, according to Jones et al. (Jones et al., 2008), SPACA1 possesses a potential Y268 tyrosine phosphorylation site that is conserved in all mammalian sequences. This might indicate that the function of SPACA1 is regulated by phosphorylation.

Kierszenbaum and Tres (Kierszenbaum and Tres, 2004) hypothesized that, during spermatogenesis, a marginal ring that forms at the edge of the acroplaxome reduces its diameter as it descends gradually along the elongating spermatid. Apparently, this process coincides with the extension of the acrosomal area attaching to the elongating spermatid nucleus. Our observations are in accordance with the above hypothesis, so impaired acrosome-acroplaxome-manchette formation might be involved in the cascades that lead to globular-headed spermatozoa (Fig. 5). This impaired complex formation did not seem to cause severe abnormalities in mitochondrial shape and sperm tail formation during spermatogenesis, but abnormal coiling of the tail around the sperm head together with misshapen mitochondria became prominent during sperm passage thorough the epididymis (Fig. 2A,B). The same phenomenon was also reported in both Gopc- and Zpbp1-disrupted spermatozoa (Lin et al., 2007; Suzuki-Toyota et al., 2007; Yao et al., 2002). The reasons for the progression of such abnormalities in the epididymis remain to be clarified.

SPACA1 is present in human spermatozoa and SPACA1 content is reported to be decreased in patients with globozoospermia as assessed by proteomic analysis (Liao et al., 2009). Further work will be required to clarify the mechanism of deficiency of SPACA1 leading to globozoospermia.

We thank Dr Tetsuo Noda at The Cancer Institute of the Japanese Foundation for Cancer Research for kindly providing the Gopc knockout mouse line and anti-GOPC polyclonal antibody; Dr Martin M. Matzuk at Baylor College of Medicine for kindly providing the Zpbp1 knockout mouse line and anti-ZPBP1 polyclonal antibody; and Akiko Kawai, Yumi Koreeda and Yoko Esaki for technical assistance in producing transgenic mouse lines.