TSSK6 is required for γH2AX formation and the histone-to-protamine transition during spermiogenesis Free

During mammalian spermatogenesis, spermatogonia undergo several rounds of mitotic divisions to yield spermatocytes, which then complete two meiotic divisions and give rise to haploid round spermatids (Hess and Renato de Franca, 2008). The round spermatids develop into mature sperm by undergoing a lengthy and complex process of cell differentiation called spermiogenesis, which includes elongation and condensation of nuclei, formation of the acrosome and flagellum, and removal of excess cytoplasm (Cheng and Mruk, 2002; Griswold, 1995). This process includes an extensive chromatin reorganization, often referred to as the histone-to-protamine transition, wherein histones are replaced first by transition proteins (TP) and subsequently by protamines (Prms), leading to a more compact packaging of DNA in sperm (Braun, 2001; Govin et al., 2004; Rathke et al., 2014; Ward and Coffey, 1991). Genetic ablation of Prm1 or Prm2 leads to sterility in male mice (Cho et al., 2001), and abnormal chromatin condensation in sperm is closely linked to infertility in men (Evenson et al., 1999; Hammoud et al., 2009). The molecular basis of nuclear condensation is poorly understood but involves key events such as expression of testis-specific histone variants, post-translational modifications of histones, transient DNA strand breaks and transcriptional silencing. Several post-translational modifications of histones, such as acetylation, phosphorylation, ubiquitination and methylation, all facilitate the displacement of histones from chromatin and gene regulation in spermatids (Govin et al., 2004; Li et al., 2014; Rathke et al., 2014; Zhuang et al., 2014). Modified histones are recognized by ‘reader’ proteins, which are often components of chromatin-remodeling complexes that are involved in the histone-to-protamine transition. For example, bromo-domain-containing proteins, such as BRDT and BRWD1, bind to hyperacetylated histones before histone removal from the chromatin, and mice lacking BRWD1 or expressing mutant BRDT are sterile due to impaired spermiogenesis (Pattabiraman et al., 2015; Shang et al., 2007).

In somatic cells, one of the first responses to DNA breaks is phosphorylation of histone H2AX (also known as H2AFX) at residue Ser139 (hereafter referred to as γH2AX) by PI3K-like protein kinases, such as ATM, DNA-PKcs and ATR (Bellani et al., 2005; Burma et al., 2001; Stiff et al., 2004; Ward and Chen, 2001), in order to recruit the repair machinery to these sites (Rogakou et al., 1998; Rossetto et al., 2012). γH2AX has also been implicated in chromatin remodeling that takes place during several other biological processes, such as male sex chromosome inactivation in germ cells, X chromosome inactivation in somatic cells, asymmetric sister chromosome segregation in stem cells and cellular senescence maintenance in fibroblasts (Turinetto and Giachino, 2015). Targeted deletion of histone H2AX in mice results in male, but not female, infertility due to failure of sex-body formation and meiotic sex chromosome inactivation (MSCI) in spermatocytes (Celeste et al., 2002; Fernandez-Capetillo et al., 2003). In leptotene spermatocytes, γH2AX is formed in response to recombination-associated DNA double-strand breaks (Blanco-Rodríguez, 2009; Carofiglio et al., 2013; Celeste et al., 2002). In pachytene spermatocytes and in the absence of DNA breaks, γH2AX is detected in the sex body, and thought to be produced by ATR (Carofiglio et al., 2013; Fernandez-Capetillo et al., 2003; Mahadevaiah et al., 2001; Turner et al., 2004). In elongating spermatids, γH2AX foci and transitory DNA breaks occur simultaneously (Agnieszka, 2014; Govin et al., 2004; Leduc et al., 2008b; Marcon and Boissonneault, 2004; Rathke et al., 2014; Wojtczak et al., 2008), but the kinase responsible for the H2AX phosphorylation and the role of γH2AX in spermiogenesis is not clearly understood.

The family of testis-specific serine/threonine kinases (TSSKs) comprises six members that are present only in spermatids and sperm (Li et al., 2011). Little is known about the factors that drive expression of the TSSK genes during spermatogenesis, with the exception of TSSK6 (also known as small serine/threonine kinase, SSTK), which is highly dependent on BRWD1 (Pattabiraman et al., 2015). Heat shock protein 90 (HSP90) is essential for the stability of all TSSKs and for catalytic activation of TSSK2 and TSSK4 (Jha et al., 2013, 2010). Tssk6-knockout (KO) and Tssk1/Tssk2 double KO mice exhibit male infertility (Shang et al., 2010; Spiridonov et al., 2005; Xu et al., 2008), whereas the targeted disruption of Tssk4 causes subfertility in male mice (Wang et al., 2015). Electron microscopy analysis of testis sections from Tssk6-KO mice indicates DNA condensation defects, and KO sperm have abnormal morphology, highly reduced motility and are incapable of fusing with zona pellucida-free eggs (Sosnik et al., 2009). Furthermore, variations in the TSSK6 gene are associated with impaired spermatogenesis and infertility in men (Su et al., 2010).

While numerous hallmarks of spermiogenesis have been well documented, there remain unanswered questions and an incomplete understanding of the elaborate mechanisms required to transform round spermatids into sperm. The close connection between Tssk6 expression and nuclear condensation during spermatid development, genetic ablation of Tssk6 and male infertility, and our previous observation that TSSK6 can phosphorylate histones such as H2AX in vitro (Spiridonov et al., 2005) led us to perform a systematic investigation of cellular and biochemical events during spermiogenesis in wild-type (WT) and Tssk6-KO mice. We have found that TSSK6 mediates γH2AX production and the histone-to-protamine transition required for proper nuclear condensation, and for the first time, we directly link γH2AX formation in maturing spermatids to male fertility. Based on our findings, we are now able to position TSSK6 within our current understanding of spermiogenesis and provide a model for the role of TSSK6 in male fertility.

As spermatogenesis progresses, germ cells move from the basal membrane towards the lumen of the seminiferous tubule, and the cycle of spermatogenesis in mice is divided into 12 stages representing distinct cellular associations in serial cross-sections of seminiferous tubules (Ahmed and de Rooij, 2009; Oakberg, 1956). Spermiogenesis in the mouse is further subdivided into 16 steps based on the morphology of the nucleus and acrosome of maturing spermatids. To thoroughly characterize the expression of Tssk6 mRNA and protein during spermatogenesis, we performed in situ hybridization (ISH) and immunofluorescence (IF) staining on testis sections. Using periodic acid Schiff (PAS) and hematoxylin staining, we evaluated the morphology of developing germ cells in WT and Tssk6-KO seminiferous tubules but observed no obvious differences when various stages and steps of spermatogenesis were compared (Fig. 1A,B and data not shown). Chromogenic ISH detected Tssk6 transcripts in both round and elongating spermatids, but not in spermatocytes (Fig. S1) or in condensed spermatids (data not shown). IF analysis demonstrated intense TSSK6 protein staining in step 11-12 spermatids in stage XI-XII tubules, as characterized by thin and compact spermatid nuclei (Fig. 1C,E,F). Notably, TSSK6 protein was confined to the nuclei of the elongating spermatids as the TSSK6 staining overlapped with that of DAPI, and no IF positivity was observed in sections from Tssk6-KO mice (Fig. 1D). Post-meiotic expression and nuclear localization of TSSK6 suggested that TSSK6 may have an important role in the nuclear condensation and/or regulation of gene expression in developing spermatids.

To determine whether TSSK6 influenced spermatid gene expression, we performed RNA sequencing (RNA-Seq) and compared the mRNA profiles of purified spermatids from WT and Tssk6-KO mice. RNA-Seq analyses were performed on three sets of purified spermatids from age-matched adult WT and KO mice, wherein each set represented a pool of spermatids from 4-6 mice. Enriched spermatid populations with >90% purity were obtained as determined by microscopic evaluation of size and morphological criteria described previously (Bellvé, 1993). Out of 23,235 genes present in the reference genome (Mus musculus_UCSC_mm10), transcripts from a total of 19,885 genes were detected in WT or Tssk6-KO spermatids, and the top 500 most abundant transcripts were ranked by fragments per kilobases per million mapped reads (FPKM) values (Table S1). Transcripts for spermatid-specific genes such as Prm1 and Prm2, Tnp1 and Tnp2 (TPs), Smcp, Akap4, Ldhc, Odf and those encoding TSSKs etc. were included in this list and were not significantly affected by the targeted deletion of Tssk6 (P>0.05). Interestingly, transcripts for Tssk6 and its activators Tsacc and Hsp90aa1 (Jha et al., 2013, 2010) were the 242nd, 180th and 106th most abundant transcripts, respectively, underscoring the importance of TSSK6 in spermatid development. A very small number of low-level transcripts (mean FPKM of 0.5–15 in either WT or KO) were either increased (34) or decreased (5) by >50% in KO versus WT cells, and exhibited a differential expression at P≤0.05 (Table S2). However, none of the proteins expressed by these transcripts are known to be important for spermiogenesis. Table S3 lists genes with less-abundant transcripts than those listed in Table S1 but that are relevant to spermiogenesis, such as chromatin remodeling mediators (Chd5, Ctcf), bromo-domain containing proteins (Brd genes), histones (Hist genes), DNA damage response proteins (Brca genes, H2afx), PI3K-like protein kinases (Atm, Atr, Prkdc) or other TSSK genes. The levels of these transcripts were also found to be comparable in WT and KO spermatids (P>0.05). Thus, genetic ablation of Tssk6 did not cause significant alteration in gene expression in spermatids as the amount of transcripts for the vast majority of genes were similar in WT and KO spermatids.

Nuclear abnormalities observed previously in electron micrographs of Tssk6-KO spermatids indicate defects in chromatin condensation (Spiridonov et al., 2005). Based on that observation and nuclear localization of TSSK6 in elongating spermatids in the present study, we investigated the relationship between TSSK6 protein expression and γH2AX generation during spermiogenesis. WT testis sections were co-labeled with antibodies against γH2AX and TSSK6, and immunoreactivity was detected using fluorescence (Fig. 2). Strong γH2AX staining was detected in the nuclei of spermatids at step 10-12 that then drastically reduced in the later steps (Fig. 2). As expected, γH2AX was also detected in leptotene and pachytene spermatocytes (Blanco-Rodríguez, 2009; Hamer et al., 2003). IF staining for TSSK6 was detected in step 10-12 elongating spermatids and was retained in step 13-16 condensing/condensed spermatids in the lumen of seminiferous tubules at stage I-VI (Fig. 2B,E,H). Importantly, γH2AX and TSSK6 colocalized in the nuclei of developing spermatids as demonstrated by the merged images of γH2AX, TSSK6 and DAPI (Fig. 2C,F,I). These results demonstrated that TSSK6 expression is coincident with γH2AX formation in nuclei of elongating spermatids, suggesting a relationship between TSSK6 and γH2AX formation during spermiogenesis.

To assess whether TSSK6 is involved in γH2AX formation, we performed IF studies on testis sections from WT and KO mice. Remarkably, no staining was observed in spermatids from KO mice (Fig. 3Ag-Al), but in parallel experiments γH2AX was readily detected in the nuclei of spermatids in WT sections (Fig. 3Aa-Af). These results demonstrated that TSSK6 is essential for γH2AX formation in spermatids. Conversely and as expected, γH2AX staining was indistinguishable in spermatocytes from WT and KO testis sections, demonstrating that genetic ablation of Tssk6 does not affect γH2AX formation that is associated with either homologous recombination or sex-body formation during meiosis (Fig. 3A). A drastically reduced amount of γH2AX in the lysate from purified KO spermatids was seen by western blotting, when compared to WT lysate (Fig. 3B). In the western blotting analyses, some γH2AX was detected due to unavoidable spermatocyte contamination of the spermatid preparation (Bellvé, 1993). To confirm that the absence of γH2AX in KO spermatids was due to a lack of Ser139 phosphorylation and not to reduced H2AX protein, we performed IF and western blotting experiments on WT and KO samples. IF staining of testis sections showed the presence of H2AX protein in the nuclei of spermatocytes and elongating spermatids, and no differences were observed in WT versus KO testis sections (Fig. S2) consistent with the RNA-Seq analysis (H2afx in Table S3). Finally, western blotting demonstrated that similar amounts of H2AX protein were present in the lysates of purified spermatids from WT and KO mice (Fig. 3B). Histone H4 hyperacetylation is another early molecular event that is believed to be essential for chromatin condensation (Bao and Bedford, 2016; Rathke et al., 2014). Histone H4 hyperacetylation was detected in step 8-12 spermatids in stage VIII-XII tubules, but unlike γH2AX, no differences in the staining pattern were observed in WT and KO testis sections (Fig. S3 and data not shown).

In our previous work, we have observed that both histone H2A and H2AX are phosphorylated to similar extents by TSSK6 in vitro (Spiridonov et al., 2005). Because H2A and H2AX have almost identical amino acid sequences, with the exception of the extended C-terminus in H2AX that contains Ser139, it is unlikely that H2AX Ser139 is a good phosphorylation site for TSSK6. Nevertheless, a temporal and spatial colocalization of TSSK6 and γH2AX in elongating spermatids led us to directly test whether Ser139 could be phosphorylated by TSSK6. We expressed Myc-tagged TSSK6 in Cos-7 cells and performed in vitro immunokinase reactions using peptides corresponding to the H2AX C-terminal region containing the wild-type sequence or Ser to Ala changes (Table S4). Histone H2AX peptides were synthesized and characterized as described previously (Fan et al., 2004). No kinase activity was detected for TSSK6 with any of these H2AX peptides, whereas kinase activity was easily detected with histone H2A protein as substrate. As a positive control, we confirmed that commercially available purified ATM kinase phosphorylated the peptides containing Ser139 (H2AX-P1 and H2AX-P2), but phosphorylation was dramatically reduced when Ser139 was changed to Ala (H2AX-P3) (Table S4).

Among the kinases that are known to phosphorylate Ser139 in histone H2AX, ATR transcripts were most abundant in our RNA-Seq analysis of spermatids (Table S3), and ATR is also reported to be responsible for H2AX Ser139 phosphorylation in spermatocytes (Turner et al., 2004). To test whether TSSK6 was involved in the protein expression and/or localization of ATR in spermatids, we performed western blotting and IF experiments to measure ATR on WT and Tssk6-KO testis samples. Western blotting demonstrated that similar amounts of ATR were present in the lysates of purified spermatids from WT and KO mice (Fig. 3B). In IF analyses, ATR was detected in the nuclei of spermatids at steps 9-12 and, importantly, the staining pattern was very similar in WT and KO spermatids (Fig. 3C). Taken together, these results demonstrate that TSSK6 is essential for γH2AX generation during spermiogenesis, but is most likely not the kinase that phosphorylates H2AX at Ser139 in spermatids.

Due to the lack of γH2AX formation in the KO spermatids, we performed a sensitive fluorescent terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay to evaluate DNA breaks on testis sections. As shown in Fig. 4, TUNEL labeling in both WT and Tssk6-KO testis was very similar, and detected as a homogeneous nuclear stain in spermatocytes and step 9-12 elongating spermatids (stages IX-XII). Reduced staining in both WT and KO spermatids was observed at step 13 (stage I), with no detection in condensed spermatids (stages VII-VIII). These results demonstrate that the spermiogenesis-associated transient DNA strand breaks occur in the absence of γH2AX formation, and the extent and timing are the same in WT and Tssk6-KO mice.

The final packaged DNA in sperm nuclei contains bound residual histones and protamines. To characterize the histone profile of KO sperm, we performed western blotting of sperm lysates with antibodies against histones H1, H2A, H2B, H3 and H4. As shown in Fig. 5A,B, loss of Tssk6 resulted in significant increases in histone H3 and H4 in sperm. Densitometry analysis revealed that histone H3 and H4 proteins were ∼3.7- and ∼2.9-fold higher in KO compared to WT sperm, respectively (Fig. 5B). Histones H1, H2A and H2B were undetectable in western blots of KO and WT sperm lysates (data not shown). Western blotting of purified spermatid lysates did not reveal any difference in the amounts of histones H1, H2A, H2B, H3 or H4, or of TP1 and TP2 proteins in WT and KO spermatids (Fig. S4A), indicating that the basic nuclear proteins profile is largely unaltered during early spermiogenesis in Tssk6-KO mice.

Unlike histones, protamines are not readily extracted from sperm, and the analysis of protamine levels requires a protocol that results in quantitative extraction of proteins from the compacted chromatin followed by protein fractionation in acid-urea-polyacrylamide gels (Balhorn et al., 1977; Yu et al., 2000; Zhao et al., 2001). Proteins extracted from the chromatin of WT and KO sperm were separated in these gels and either stained with Coomassie Brilliant Blue (Fig. 5C, left) or analyzed by western blotting for Prm1 (middle) and Prm2 (right). No differences were noted in KO and WT samples for Prm1 and mature Prm2, whereas strikingly increased amounts of both Prm2 precursor and intermediate(s) were observed in KO sperm compared to in WT.

Consistent with the observation in sperm, the acid-urea-polyacrylamide gel analysis also revealed an enhanced presence of the Prm2 precursor and intermediate(s) in spermatids from the KO mice (Fig. S4B). No TSSK6 protein was detected by western blotting in these spermatid extracts from WT or KO mice (data not shown), indicating that TSSK6, unlike protamines, is not tightly bound to chromatin. Prm1 and Prm2 were detected in condensing/condensed spermatid nuclei by IF, however, no difference in staining was observed in WT and KO sections (data not shown). These results demonstrated that TSSK6 is required for murine sperm to possess normal quantities of histone H3, H4 and Prm2 precursor and intermediate(s).

TSSK6 is a testis-specific protein kinase that is highly conserved in mammals and essential for spermatogenesis. To determine the biological role of TSSK6, here we have performed a comparative analysis of molecular events during spermiogenesis in WT and Tssk6-KO mice. Based on our study, the relationship of TSSK6 to key events that occur during chromatin condensation can be better understood. Fig. 6 highlights TSSK6-dependent and -independent events as round spermatids mature into condensed spermatids. Tssk6 mRNA is expressed in round spermatids, and the transcripts are still detected in elongating spermatids, whereas TSSK6 protein first appears in elongating spermatids and is retained in mature sperm. Genetic ablation of Tssk6 did not have a significant impact on the transcriptome, indicating that TSSK6 is not involved in gene expression. Histone H4 hyperacetylation, transient DNA breaks, chromatin displacement of histones H1, H2A and H2B, and the TP exchange all appeared to be normal in Tssk6-KO spermatids. Further, Prm1 and mature Prm2 levels were found to be comparable in WT and KO chromatin-bound protein extracts from spermatids and sperm. Conversely, TSSK6 was found to be essential for the production of γH2AX, removal of histones (H3 and H4) and effective processing of the Prm2 precursor and intermediate(s). Most of the core nucleosomal histones are displaced from the chromatin during steps 9-12 of spermiogenesis in the mouse, and condensed spermatids or sperm retain only small residual amounts of histones H3 and H4 (Li et al., 2014; Meistrich et al., 2003). Increased retention of histone H3 and H4 in spermatids/sperm has been closely linked to male infertility in many gene-KO mouse models (Bell et al., 2014; Li et al., 2014; Lu et al., 2010; Zhuang et al., 2014). Prm2 is synthesized as a precursor protein that binds to the chromatin and then undergoes proteolytic processing giving rise to several intermediates and finally mature Prm2 (Balhorn, 2007; Wu et al., 2000). Thus, incomplete Prm2 processing of the chromatin-associated precursor and intermediate(s) occurs in the absence of TSSK6. Numerous studies have implicated altered Prm1 or Prm2 amounts, and aberrant processing of Prm2 precursor and intermediate(s) in sterility in mice (Bao and Bedford, 2016; Hernández-Hernández et al., 2016; Li et al., 2014; Lu et al., 2010; Rathke et al., 2014; Wu et al., 2000; Yuen et al., 2014) and humans (Carrell et al., 2007; Oliva, 2006; Torregrosa et al., 2006). We interpret the defects in the histone-to-protamine transition, as evidenced by altered histone and protamine homeostasis in the chromatin of Tssk6-KO spermatids and sperm, as the cause for infertility in the null mice.

γH2AX marks sites of DNA breakage, and γH2AX foci are detected in spermatocytes during meiotic recombination (Blanco-Rodríguez, 2009; Carofiglio et al., 2013; Celeste et al., 2002). However, γH2AX can also be formed in a DNA break-independent manner, as in the case of sex-body-associated γH2AX in pachytene spermatocytes (Carofiglio et al., 2013; Fernandez-Capetillo et al., 2003; Mahadevaiah et al., 2001). As expected, and because expression of TSSK6 is limited to spermatids and sperm, genetic ablation of TSSK6 did not cause any defect in γH2AX formation that is associated with homologous recombination or sex-body formation in spermatocytes. Detection of γH2AX foci in step 9-12 elongating spermatids is coincident with DNA strand breaks (Govin et al., 2004; Leduc et al., 2008a; Rathke et al., 2014), and here we have demonstrated that TSSK6 is essential for γH2AX production during spermiogenesis. The timing and level of the DNA breaks were very similar in WT and Tssk6-KO spermatids, and can be interpreted two ways. Either DNA breaks are not the driver of the H2AX phosphorylation in spermiogenesis or TSSK6 is required for γH2AX to be generated in response to the breaks. Previous studies have suggested that precise control of γH2AX may be important during spermiogenesis since prolonged or increased amounts have been observed in KO mouse models with a sterile phenotype (Li et al., 2014; Wang et al., 2016). To the best of our knowledge, our findings regarding TSSK6 are the first that link γH2AX formation during spermiogenesis to the production of functional sperm and fertility. Based on the high abundance of Atr transcripts in spermatids, nuclear spermatid localization of ATR, a recognized role for ATR in γH2AX generation in pachytene spermatocytes (Turner et al., 2004) and the inability of TSSK6 to phosphorylate H2AX at Ser139 in vitro, it is more likely that ATR, and not TSSK6, is the kinase that phosphorylates Ser139 of H2AX during spermiogenesis. However, it is important to note that genetic ablation of Tssk6 had no impact on ATR expression and nuclear localization in spermatids and thus, there is no direct evidence that ATR phosphorylates H2AX in this instance. Nevertheless, the requirement of TSSK6 for γH2AX generation demonstrates that TSSK6 is an upstream kinase responsible for the proper localization and/or activation of the kinase that directly produces γH2AX during spermatid nuclear condensation.

IF studies demonstrated that TSSK6 is confined to the nuclei, but it was not detected in the chromatin-bound protein extracts derived from spermatids. Accordingly, these findings indicate that while TSSK6 is in the nucleus as chromatin is condensing, it is not tightly bound to DNA. Based on the insights gained from the present study, we postulate that TSSK6 mediates γH2AX generation to govern a chromatin-associated remodeling complex that is essential in the histone-to-protamine transition. Further studies are warranted to identify the TSSK6 protein substrates that are required for γH2AX formation, the histone-to-protamine transition and the production of normal sperm in mammals.

Antibodies against histones H3 (96C10), H4 (L64C1), H2A (catalogue number 2578), H2AX (catalogue number 2595) and γH2AX (20E3) were purchased from Cell Signaling Technology Inc; against histones H1 (AE-4), H2B (FL-126) and TP2 (K-18) were from Santa Cruz Biotechnology Inc; against TNP1 (catalogue number 17178-1-AP) were from ProteinTech Group Inc (Chicago, IL); and against Prm1 (Hup1N) and Prm2 (Hup2B) were from Briar Patch Biosciences (Livermore, CA). Recombinant human ATM kinase and antibody against hyperacetylated histone H4 (Penta) were from EMD Millipore (Billerica, MA), and terminal deoxynucleotidyl transferase (TdT), TdT buffer and biotin-16-dUTP were from Sigma (St Louis, MO). Monoclonal antibody against mouse TSSK6 was generated in our lab and has been described previously (Jha et al., 2010). All reagents used in this study were of analytical grade.

The Tssk6-KO mouse model was generated in our lab and has been described previously (Spiridonov et al., 2005). Mice were handled and killed according to the guidelines of the Animal Care and Use Committee (Center for Biologics and Evaluation Research, U.S. Food and Drug Administration). Adult male Tssk6-KO mice and WT littermates were used in this study.

RNA ISH was performed using the RNAScope ISH technology (Advanced Cell Diagnostics, Hayward, CA) (Wang et al., 2012). Briefly, testes were harvested and fixed in 10% neutral buffered formalin for 24 h at room temperature, dehydrated and embedded in paraffin. Tissue sections were cut at 5 μm thickness, air-dried at room temperature and processed for RNA in situ detection by using the RNAscope 2.5 HD Reagent Kit-RED (catalogue number 322360). To ensure RNA integrity and assay procedure, pre-treatment conditions were optimized to 15 min heat-pre-treatment, followed by 30 min protease digestion. Custom RNA ISH probes were prepared by Advanced Cell Diagnostics to detect mouse Tssk6, and a probe to bacterial mRNA DapB served as a negative control. Slides were scanned using Panoramic MIDI slide scanner (3DHISTECH) to acquire images of ISH.

Testes were fixed either in 4% paraformaldehyde or Carnoy solution (ethanol, chloroform, acetic acid in the ratio 6:3:1, respectively) and embedded in paraffin, and 5-μm thick cross-sections were mounted on glass slides. Sections were deparaffinized with xylene, rehydrated through a graded ethanol series and washed briefly in water. Slides were then processed for antigen retrieval in 10 mM citrate buffer (pH 6.0) for 20 min at 100°C, washed with PBS and blocked with 1% bovine serum albumin (BSA) in PBS containing 0.1% Tween-20 (PBS-T). Sections were incubated with primary antibody or isotype control IgG overnight, and the antibody dilution used in IF studies was 1:100 with the exception of γH2AX (20E3) antibody, which was used at 1:400 dilution. Following primary antibody incubation, the slides were washed with PBS-T and incubated with Alexa-Fluor-conjugated secondary antibody (1:200 dilution) (Molecular Probes, Eugene, OR) for 1 h. Slides were washed with PBS-T, counterstained with DAPI and mounted with Slow-Fade light reagent (Molecular Probes). Sections were examined under a confocal microscope (LSM 710, Zeiss), and images were captured. Established morphological criteria were used to determine specific stages of the mouse seminiferous epithelium cycle (Ahmed and de Rooij, 2009; Meistrich and Hess, 2013). Experiments were repeated with testes sections from at least three different sets of WT and Tssk6-KO mice.

A modified TUNEL assay was performed to detect DNA breaks in testis sections as previously reported (Marcon and Boissonneault, 2004). Briefly, Carnoy-fixed testis sections were deparaffinized with xylene and rehydrated through a graded ethanol series. Sections were washed with PBS and equilibrated with TdT buffer for 30 min at 37°C. TdT buffer was removed and a freshly prepared terminal transferase reaction mix (50 µl TdT buffer containing 25 U terminal transferase enzyme and 0.5 nmol Biotin-16-dUTP) was added. End-labeling was performed for 1 h at 37°C, followed by washing with PBS and incubation with fluorescein–avidin (1:100 dilution in PBS) for 1 h at room temperature. Sections were washed with PBS, counterstained with DAPI, mounted with Slow-Fade light reagent, and images were captured using a confocal microscope. A negative control with no TdT enzyme was included in all experiments.

Image acquisition from IF and TUNEL staining was performed at room temperature. Data analysis was performed using a LSM 710 confocal laser scanning microscope (Zeiss) equipped with EC Plan-Neofluar 20×/0.8 NA, 40×/1.3 NA oil or Plan-Apochromat 63×/1.4 NA oil objective lens. Images were captured with AxioCam camera (Zeiss) using Zen imaging software (Zeiss) and processed using Photoshop CS4 (Adobe). For ISH imaging, slides were scanned using Panoramic MIDI slide scanner (3DHISTECH) equipped with a Plan-Apochromat 20×/0.8 NA objective lens (Zeiss) and a VCC-FC60FR19CL/4MP camera (CIS). Image analysis was performed using Panoramic Viewer software (3DHISTECH), and images were processed using Photoshop CS4. No gamma correction was made in IF, TUNEL or ISH images, and scale bars were added.

An enriched population of mouse spermatids was purified following the STA-PUT method as described previously (Jha et al., 2013). Briefly, testes from adult mice were enzymatically dissociated, and the cells were separated by sedimentation at unit gravity in 2-4% BSA gradient in a STA-PUT apparatus. Fractions were examined under a microscope, and germ cell types were determined based on the size and morphology (Bellvé, 1993). Fractions enriched with spermatids were pooled to obtain a population of spermatids with >90% purity.

Spermatids were purified from three sets of age-matched adult WT and Tssk6-KO mice, and total RNA was isolated using RNeasy kit (Qiagen) with DNase I treatment. RNA-Seq libraries were prepared using Illumina TruSeq Stranded mRNA sample preparation kit and sequenced on Illumina HiSeq 2500. The quality of raw data was assessed and passed by FastQC, and reads were mapped to the reference genome (Mus musculus_UCSC_mm10). Cufflinks software (v2.2.1) was used to estimate transcript levels represented by the FPKM, and differential expression between samples was determined using Cuffdiff software (v.2.2.1).

Proteins were fractionated in 4-20% SDS-PAGE gels. Western blotting and immunoprecipitation experiments were performed as described previously (Jha et al., 2013, 2010), and the primary antibody dilution used in western blotting was 1:1000.

Myc-tagged Tssk6 was expressed in Cos-7 cells (American Type Culture Collection, Manassas, Virginia) and immunoprecipitated with an antibody against Myc, and in vitro kinase reactions were performed (Jha et al., 2013). Briefly, the reaction was performed at room temperature for 30 min in a reaction buffer containing 25 mM HEPES (pH 7.4), 10 mM MgCl₂, 10 mM MnCl₂, 2 mM EGTA, 30 µM ATP, 10 µCi of [γ-³²P]ATP, and synthetic histone H2AX peptides or histone H2A protein. The reaction was terminated by acidification, and phosphorylation of the H2AX peptides or H2A protein was quantified by measuring incorporated radioactive ³²P (Fan et al., 2004). Values are represented as mean kinase activity (fmol/min)±s.e.m. (n=3).

Sperm from mouse cauda epididymis were collected in modified Krebs–Ringer medium (Whitten's-HEPES buffered medium) containing 100 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 5.5 mM glucose, 1 mM pyruvic acid, 4.8 mM L(+) lactic acid hemicalcium salt in 20 mM HEPES (pH 7.3), as described previously (Jha et al., 2006, 2008).

Chromatin-bound proteins from spermatids and sperm were extracted and analyzed following the methods described previously (Balhorn et al., 1977; Yu et al., 2000; Zhao et al., 2001). Testicular germ cells were isolated by the enzymatic dissociation of WT and Tssk6-KO testes, and subjected to sonication followed by centrifugation through a layer of 5% sucrose in 0.2× concentration of MP buffer (5 mM MgCl₂ and 5 mM sodium phosphate, pH 6.5) containing 0.25% Triton X-100 and protease inhibitors. The pellet fraction, which comprises sonication-resistant nuclei of condensing/condensed spermatids, was washed in MP buffer containing 0.22 M NaCl and incubated at 37°C for 10 min in 10 mM DTT and protease inhibitors. DNA was precipitated with 0.5 M HCl on ice for 1 h, and the supernatant was subjected to protein precipitation with 25% trichloroacetic acid (TCA), followed by washing with acidified acetone and dissolving the protein precipitate in acid-urea gel sample buffer (5 M urea, 5% acetic acid, 1% β-mercaptoethanol, and Methyl Green dye). Extraction from sperm was performed as above with the exception that the nuclei pellet isolated after sonication was incubated in 5 M guanidium chloride solution (pH 8.0) for 30 min on ice, and then dissolved in a solution containing 3 M urea, 0.5 M β-mercaptoethanol and 2 M NaCl. Supernatant collected after DNA precipitation was subjected to buffer exchange with 0.01 N HCl by centrifugal filtration (3 kDa cut-off) before the proteins were precipitated with 25% TCA. Chromatin-bound proteins that had been isolated from spermatids or sperm were resolved in acid-urea-polyacrylamide gels and subjected to either Coomassie Brilliant Blue staining or western blotting. Experiments were performed on three sets of sperm from WT and KO mice, wherein each set comprised sperm from three to five mice.

Western blots were scanned, and protein band intensities were quantified using ImageJ software (National Institutes of Health, Bethesda, USA). Densitometry values of protein bands were normalized to those of β-tubulin, and then the normalized protein level in KO relative to WT samples was calculated. Western blots from three independent experiments were used for statistical analysis, and results were expressed as mean±s.e.m. Student's t-test was performed to calculate the P-values.

We are grateful to Dr Wells Wu, Core facility, Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, for his assistance with RNA-Seq analysis.