Spata6 is required for normal assembly of the sperm connecting piece and tight head–tail conjunction

Multialignment and phylogenetic analyses of SPATA6 orthologs in 13 vertebrate species revealed that SPATA6 is highly conserved during evolution (SI Appendix, Fig. S1). Human SPATA6 shares 88%, 50%, and 33% amino acid sequences with its orthologs in mice, chicken, and zebrafish, respectively. Consistent with a previous study (23), four isoforms of Spata6 transcripts are predicted in mice (SI Appendix, Fig. S2A); however, isoform 1 appears to be the sole dominant isoform, with its mRNA and protein detected exclusively in testis among all 10 organs analyzed (SI Appendix, Fig. S2B and Fig. 1A). Both Spata6 mRNA and protein were detectable in testes at postnatal day (P) 14 (P14), and the levels kept increasing from P21 onward, with the highest levels detected in adult testes (SI Appendix, Fig. S2C and Fig. 1B). This expression pattern suggests that Spata6 mRNA and protein are expressed in late pachytene spermatocytes and spermatids. Indeed, Spata6 mRNA were detected only in purified spermatocytes and spermatids (SI Appendix, Fig. S2D), and in situ hybridization analyses also localized Spata6 mRNA to spermatocytes and spermatids (SI Appendix, Fig. S2 E and F).

Immunofluorescent staining using rabbit anti-SPATA6 polyclonal antibodies revealed that SPATA6 is localized exclusively to the connecting piece of spermatozoa collected from the epididymis of WT male mice (Fig. 1C). Supporting this antibody specificity, the specific staining was completely absent in Spata6-null spermatozoa with the head still attached to the flagella (Fig. 1D), which accounted for <1% of the epididymal spermatozoa in Spata6 KO males (see below). Considering the high degree of amino acid similarities among all vertebrates (SI Appendix, Fig. S1), we used the same anti-SPATA6 antibodies to stain spermatozoa collected from trouts (Fig. 1E), rats (Fig. 1F), rabbits (Fig. 1G), and monkeys (Fig. 1H), as well as ejaculated human spermatozoa (Fig. 1I). Specific staining was detected exclusively in the connecting piece of spermatozoa in both fish (Fig. 1E) and four mammalian species (Fig. 1 C, D, F–I), suggesting that SPATA6 is a structural protein specifically localized to the connecting piece of spermatozoa in vertebrates.

To further determine the subcellular localization of SPATA6, we performed immunogold labeling followed by transmission electron microscopy (IG-TEM) on mouse and rabbit epididymal spermatozoa (Fig. 1 J–Q). Gold particles were detected mainly in the segmented columns and the capitulum of the connecting piece of spermatozoa from mice (Fig. 1 J–L) and rabbits (Fig. 1P). No specific gold particles were detected in either WT mouse and rabbit spermatozoa stained with preimmune serum (Fig. 1 M and Q) or Spata6-null mouse spermatozoa stained with the SPATA6 antibodies (Fig. 1 N and O). Taken together, our expression and localization data demonstrate that SPATA6 is an evolutionarily conserved protein localized exclusively to the segmented columns and the capitulum of the connecting piece of spermatozoa.

To define the physiological role of SPATA6, we generated Spata6 global KO mice using an embryonic stem (ES) cell line obtained from the Knockout Mouse Project (KOMP) repository (EPD0224_2_B05, ID: CSD33302), in which a ‘‘knockout first’’ gene trap cassette was inserted in between exons 2 and 3 of the Spata6 locus, leading to a “gene trap” allele (Fig. 2 A and B). By examining the levels of Spata6 mRNA (Fig. 2C) and protein (Fig. 2D) in WT and KO testes, we confirmed that homozygous (Spata6^frt/frt) mice were completely lacking Spata6 mRNAs and protein and thus were truly Spata6-null. The complete ablation of both Spata6 mRNA and protein is also supported by the data showing an absence of signals in the in situ hybridization (SI Appendix, Fig. S2E), immunofluorescence (Fig. 1D), and IM-TEM (Fig. 1N) analyses of the KO testes.

Spata6 KO mice were viable and displayed normal development. Although all of the adult Spata6 KO females were fertile, the adult male KO mice did not produce any pups after breeding with fertility-proven adult WT female mice for 6 months, suggesting that Spata6 KO male mice are sterile. To uncover the cause of the male sterility phenotype, we examined the Spata6 KO males at both gross and histological levels. No differences in testis size and total testis weight were observed between WT and KO mice (Fig. 2 E and F). Computer-assisted sperm analyses (CASA) revealed reductions in both sperm counts (Fig. 2G) and total motility (Fig. 2H). Histological examination did not reveal any severe disruptions in the seminiferous epithelia in KO testes (Fig. 2I and SI Appendix, Fig. S3); however, fewer sperm heads were aligned along the lumen at stage VII in the KO testes compared with the WT stage VII tubules (SI Appendix, Fig. S3). Moreover, spermatozoa within the WT epididymis appeared to be more heavily stained with hematoxylin compared with those in the KO epididymis (Fig. 2I and SI Appendix, Fig. S3). These results suggest the presence of fewer sperm heads in both KO testes and epididymides, and are consistent with the reduced number of sperm heads in both the testicular and epididymal sperm suspensions from the KO mice (see below).

Phase-contrast microscopy examination revealed that the WT epididymal spermatozoa were mostly intact, with the majority carrying the typical cytoplasmic droplets (CDs) (Fig. 2J, Top and Middle Left), whereas the Spata6 KO cauda epididymis contained mainly sperm flagella bearing enlarged CD-like structures at the mid-principal piece junction (Fig. 2J, Top and Middle Right). Through immunostaining of SPEM1, a marker protein for the CD (24), we confirmed that those enlarged cytoplasmic remnants on KO sperm flagella were indeed CDs (Fig. 2J, Bottom). Quantitative analyses revealed that intact spermatozoa accounted for only ∼7%, ∼4%, and ∼3% of the total sperm cells (intact spermatozoa plus sperm flagella) in the KO testes, KO caput, and caudal epididymides, respectively (SI Appendix, Fig. S4). In contrast, the majority (96–99%) of the WT testicular and epididymal spermatozoa were intact (SI Appendix, Fig. S4). Moreover, more sperm flagella (93–97%) than sperm heads (14–16%) were present in both KO testes and epididymides compared with the WT samples (SI Appendix, Fig. S4). These results suggest that separation of sperm heads from the flagella occurs during spermiogenesis before spermiation, and that the separated sperm heads most likely are engulfed and absorbed by Sertoli cells. Interestingly, ∼10% of the headless sperm flagella collected from the KO cauda epididymis displayed limited progressive motility, and the major waveforms appeared to be confined only to the end piece instead of the principal piece, as seen in WT epididymal spermatozoa (SI Appendix, Movies S1 and S2). Taken together, the foregoing findings demonstrate that inactivation of Spata6 leads to male sterility owing to acephalic spermatozoa with close to 100% penetrance.

The fact that almost all spermatozoa are acephalic and sperm heads are scarce in the KO epididymis suggests that the separation of the sperm head from the flagellum occurs within the seminiferous tubules. To identify the underlying structural defects, we examined the ultrastructure of step 15–16 elongated spermatids in WT and Spata6 KO testes using TEM (Fig. 3). TEM analyses of the longitudinal sections and cross-sections of WT and KO step 15–16 elongated spermatids revealed no discernable defects in either the nucleus or the acrosome; however, the connecting piece was either completely lacking or only partially developed in Spata6-null elongated spermatids (Fig. 3). In WT step 15–16 elongated spermatids, the connecting piece was fully developed and consisted of well-defined segmented columns that united at the anterior and formed the capitulum (Fig. 3A). The capitulum was tightly attached to the basal plate of the caudal portion of the nucleus at the implantation fossa. The mitochondria were well aligned to surround the ODFs of the developing mid-piece (Fig. 3A). In contrast, normal connecting piece development was never observed in step 15–16 KO elongated spermatids, and two major defects included a complete lack of the segmented columns and/or the capitulum (Fig. 3 B–D) and partially formed segmented columns (Fig. 3E). In either case, there were no signs of mitochondrial sheath formation (Fig. 3 B–G), and a lack of mitochondrial sheath was always associated with misplacement of the annulus, which normally forms a belt at the mid-principal piece junction in step 15–16 WT elongated spermatids and epididymal spermatozoa (Fig. 3 H and I).

Because of the lack of connecting piece formation, the developing flagella, derived from the distal centrioles, were often misaligned with the developing nuclei/heads (Fig. 3D). To determine whether the failure in the assembly of the segmented columns could cause disrupted flagellar development, we further analyzed cross-sections of the Spata6-null sperm flagellum. Indeed, we identified severe flagellar defects in Spata6-null spermatozoa (Fig. 3J). In the mid-piece, WT spermatozoa have a well-defined mitochondrial sheath, enveloping nine ODFs and the axoneme consisting of the typical “9 + 2” microtubules (Fig. 3J), whereas the Spata6-null spermatozoa displayed a complete lack of or partially formed mitochondrial sheaths, along with incomplete ODFs and axonemes with missing ODFs and axonemal microtubules (Fig. 3J). In the end piece, although the fibrous sheath was intact in both WT and Spata6 KO spermatozoa, disorganized axonemal microtubules were obvious in Spata6-null spermatozoa, showing the “8 + 2” or “7 + 2” microtubular compositions instead of the normal “9 + 2” organization in the WT axoneme (Fig. 3J). These severe flagellar defects may explain why Spata6-null spermatozoa displayed minimal motility (Fig. 2H and SI Appendix, Movies S1 and S2). Taken together, these ultrastructral analyses reveal that failure of assembly of the segmented columns and the capitulum in developing spermatids represents the primary defects of SPATA6 deficiency, consistent with the exclusive localization of SPATA6 to the segmented columns and the capitulum in the connecting piece of spermatozoa (Fig. 1 C–Q).

Previous ultrastructral studies have documented that flagellar development starts with a functional transformation of the distal centriole into the basal body that serves as the template for the assembly of axonemal doublet microtubules (20, 25). As the flagellar axoneme grows, the basal body migrates to the cell periphery, where distal centrioles dock perpendicularly to the plasma membrane as the axoneme sprouts toward the extracellular space. Meanwhile, the axoneme growth is also accompanied by complex modifications in the dense “pericentriolar materials,” from which new proteins or structural components arise and organize to form the nine longitudinal segmented columns of the connecting piece (20, 25). Segmented columns are nine cylindrical structures with periodic densities that fuse cranially to form the capitulum, a curved plate-like disk that links the connecting piece to the basal plate, a dense structure lining the outer nuclear membrane at the implantation fossa (Figs. 1 J–M and 3A). At the caudal end, each segmented column is continuous with one of the nine ODFs that associate with peripheral microtubular doublets of the growing axoneme (20, 25).

To delineate the developmental process of the connecting piece during late spermiogenesis, we analyzed the ultrastructure of elongating (steps 9–12) and elongated spermatids (steps 13–16), as well as epididymal spermatozoa in WT and KO mice (Fig. 4A). In both WT and KO testes, nuclear condensation, nuclear elongation, and flagellar development start in step 9 spermatids. In step 9 spermatids, the proximal centriole was present at the caudal side of the nucleus, and signs of segmented column formation could be seen in the step 10 WT spermatids, but not in the step 10 KO spermatids (Fig. 4A). From step 11 to step 12, the segmented columns developed further and became highly defined, and the capitulum started to form in WT spermatids (Fig. 4A); however, in KO step 11–12 spermatids, the segmented columns and the capitulum were often absent or only partially formed and severely disorganized (Fig. 4A). In step 13–16 WT spermatids, both the segmented columns and capitulum were fully developed, and the flagellum was attached to the basal plates at the implantation fossa (Fig. 4A). In contrast, in step 13–16 KO spermatids, the segmented columns and the capitulum were either completely absent or only partially formed, and the annulus was misplaced (Fig. 4A). In the KO epididymis, sperm heads accounted for 14–16% of the total sperm cell content (SI Appendix, Fig. S4), and the headless sperm flagella usually carried enlarged CDs stuffed with mitochondria (Fig. 4A), which were never observed in CDs of WT epididymal spermatozoa (Fig. 4A). Taken together, the step-by-step TEM analyses further confirm that in the absence of SPATA6, the segmented columns either do not form or form only partially at a slower pace (Fig. 4B), leading to instability in the neck region and, ultimately, complete disconnection between the sperm heads and the flagella within the seminiferous epithelia (Fig. 4B).

TEM analyses of the KO sperm heads revealed no discernable defects, suggesting that disruptions in the connecting piece formation may have no impact on sperm head formation. This idea is also supported by the fact that ∼40% of the sperm heads collected from KO cauda epididymides maintained plasma membrane integrity and thus were considered “live” (26, 27), as demonstrated by eosin-nigrosin staining (SI Appendix, Fig. S5). To determine whether the KO sperm heads are functionally competent, we evaluated their fertility by intracytoplasmic sperm injection (ICSI). No significant differences in preimplantation development (from two-cell to blastocyst stage) or the number of live-born pups were observed when WT or Spata6 KO sperm heads were injected into WT eggs (Table 1). Moreover, the Spata6^+/− offspring derived from ICSI using Spata6-null spermatozoa, both male and female, were all healthy and fertile when they reached adulthood. These data suggest that the Spata6 KO sperm heads, although lacking the connecting piece and separated from the flagella, are fully developed and competent for fertilization and production of normal offspring when injected into WT oocytes.

Because SPATA6 has no known homologs in somatic tissues, it is difficult to predict its molecular function. To gain insight into the molecular action of SPATA6, we adopted an unbiased approach, with immunoprecipitation (IP) followed by liquid chromatography–mass spectrometry (LC-MS), to identify protein partners that interact with SPATA6. Using our rabbit polyclonal anti-SPATA6 antibodies, we performed IP on both WT and KO testes and then subjected the immunoprecipitants to LC-MS analyses. The IP and LC-MS assays identified a total of 308 putative SPATA6-interacting proteins, all of which showed greater than twofold enrichment (WT vs. KO) and were represented by at least five peptide reads in the LC-MS results (SI Appendix, Table S1). Gene Ontology (GO) term analyses identified 29 biological processes that were significantly enriched (P < 0.05) among all SPATA6-interacting proteins (SI Appendix, Table S2). The three most-enriched biological processes (P < 0.001) included cytoskeleton-dependent intracellular transport, actin filament-based process, and actin filament-based movement (Table 2), and the proteins with the greatest fold enrichment were four myosin subunits, including myosin light-chain (MYL) polypeptide 6 (MYL6), myosin heavy-chain (MYH) polypeptide 10 (MYH10), MYL polypeptide 11 (MYH11), and MYH polypeptide 14 (MYH14) (Table 2).

We then chose to analyze MYL6 to validate the proteomic findings. Western blot analyses revealed similar levels of MYL6 expression in both WT and KO testes (Fig. 5A), suggesting that Spata6 inactivation does not affect MYL6 expression. We further performed reciprocal IP assays using antibodies specific for SPATA6 and MYL6. In the immunoprecipitants of SPATA6 antibodies, both SPATA6 and MYL6 were detected in WT testes, but not in KO testes (Fig. 5B). In contrast, in the immunoprecipitants of anti-MYL6 antibodies, both SPATA6 and MYL6 were detected in WT testes, whereas MYL6, but not SPATA6, was detected in KO testes (Fig. 5B). These results confirm that SPATA6 and MYL6 are indeed bona fide interacting partners in the testes. Further supporting their interactions, our immunofluorescence analyses revealed that both SPATA6 and MYL6 were localized to the manchettes of elongating spermatids (steps 9–12) in WT testes (Fig. 5C). Although SPATA6 was eventually confined to the segmented column and the capitulum in spermatozoa (Fig. 1), it was expressed mainly on the manchette in step 9–14 spermatids within the testis (Fig. 5C). This finding is common for most, if not all, of the proteins that ultimately become sperm flagellar structural proteins, because the manchette, a microtubule-enriched perinuclear structure present exclusively in step 9–14 spermatids, plays a pivotal role in orchestrating the final assembly of functional spermatozoa, not only by providing the physical forces for shaping the sperm head, but also by serving as a “conveyer belt” that mediates the efficient transportation and processing of proteins required for sperm flagellar assembly (28–33).

Myosin has been considered a molecular motor that converts chemical energy (i.e., ATP) to mechanical power to generate force and to achieve movement in muscle and nonmuscle cells (34). Myosin is composed of two myosin heavy chains, two regulatory light chains, and two essential light chains (35). As a major form of cytoskeleton-dependent intracellular transport, myosin is known to bind actin filaments and to function in the transport of protein cargos in the cytoplasm (36, 37). Interactions between SPATA6 and myosin light and heavy chain subunits strongly suggest that SPATA6 is involved in myosin-based microfilament transport, which may be responsible for transporting basic building blocks for the assembly of the segmented columns during connecting piece formation in developing spermatids (Fig. 5D).

The University of Nevada, Reno’s Institutional Animal Care and Use Committee approved all animal use protocols. Details are provided in SI Appendix, Materials and Methods.

Spata6 mutant mice were generated using a targeted ES cell clone (EPD0224_2_B05; ID: CSD33302) obtained from the KOMP Repository, in which a ‘‘knockout first’’ gene trap cassette was inserted in between exons 2 and 3 of the Spata6 gene (Fig. 2A). Details are provided in SI Appendix, Materials and Methods.

The fertility test was carried out by mating adult Spata6 KO (Spata6^−/−) male mice with WT adult female mice with proven fertility for at least 3 mo. A total of eight Spata6 KO males and eight of their male heterozygous littermates (Spata6^+/−), as the control group, were tested. The number and size of litters sired by each breeding pair were recorded.

Histological evaluation was performed on periodic acid-Schiff–stained, paraffin-embedded sections of the testes, and on hematoxylin and eosin-stained epididymal spermatozoa. Details are provided in SI Appendix, Materials and Methods.

Total sperm motility and number were analyzed using the CASA system (version 14.0; Hamilton-Thorne Bioscience). Details are provided in SI Appendix, Materials and Methods.

The total number of intact spermatozoa, sperm heads, and flagella were counted, and the proportions of each of the three types of sperm components were calculated based on the common denominator (i.e., intact spermatozoa plus flagella) (SI Appendix, Fig. S4). Details are provided in SI Appendix, Materials and Methods.

TEM studies were performed as descried previously with minor modifications (24, 49, 50). Details are provided in SI Appendix, Materials and Methods.

SYBR Green-based quantitative PCR (qPCR) analyses were performed on an Applied Biosciences 7900HT Real-Time qPCR system. Primers used for the qPCR analyses are listed in SI Appendix, Table S3. Gapdh served as an internal control. Details are provided in SI Appendix, Materials and Methods.

Spata6 mRNA in situ hybridization in testis was performed as described previously (51). Details are provided in SI Appendix, Materials and Methods.

A rabbit anti-SPATA6 antibody (1:1,000) and a rabbit anti-MYL6 antibody (1:500; Abcam catalog no. ab84349) were used for Western blot analyses. Details are provided in SI Appendix, Materials and Methods.

SPATA6 IP and LC-MS analyses were conducted as described previously with some modifications (52). Details are provided in SI Appendix, Materials and Methods.

ICSI was carried out as described previously (53, 54), with slight modifications. Details are provided in SI Appendix, Materials and Methods.

Data are reported as mean ± SEM. Statistical differences between datasets were assessed by one-way ANOVA or the t test using SPSS 16.0 software. P ≤ 0.05 was considered to indicate a significant difference and P ≤ 0.01 to indicate a very significant difference between two groups. ICSI data were analyzed using the χ² test compared with the WT group.