Mono-ubiquitylated H2A marks the transcriptionally silenced XY body during male meiotic prophase. Concomitant with H2AK119ub1, the ubiquitin-conjugating enzyme HR6B is also enriched on the XY body. We analyzed H2A and H2B ubiquitylation in Hr6b-knockout mouse spermatocytes, but no global changes were detected. Next, we analyzed phosphorylation of the threonine residues T120 and T119 that are adjacent to the K119 and K120 target sites for ubiquitylation in H2A and H2B, respectively. In wild-type cells, H2AT120ph and H2BT119ph mark meiotically unpaired and silenced chromatin, including the XY body. In Hr6b-knockout spermatocytes, the H2BT119ph signal was unchanged, but H2AT120ph was enhanced from late pachytene until metaphase I. Furthermore, we found increased H3K4 dimethylation on the X and Y chromosomes of diplotene Hr6b-knockout spermatocytes, persisting into postmeiotic round spermatids. In these cells, the X and Y chromosomes maintained an unchanged H3K9m2 level, even when this modification was lost from centromeric heterochromatin. Analysis of gene expression showed derepression of X chromosome genes in postmeiotic Hr6b-knockout spermatids. We conclude that HR6B exerts control over different histone modifications in spermatocytes and spermatids, and that this function contributes to the postmeiotic maintenance of X chromosome silencing.
Chromatin structure regulation requires the concerted actions of different histone-modifying enzymes and ATP-dependent chromatin remodeling complexes. The combined presence of specific histone modifications, such as acetylation, phosphorylation, methylation, ubiquitylation and sumoylation, on the four core histones (H2A, H2B, H3, H4) that constitute the nucleosome, has been termed the histone code which is `read' by regulatory effector proteins (Jenuwein and Allis, 2001).
Histone mono-ubiquitylation (Emre and Berger, 2006) and histone sumoylation (Nathan et al., 2006) involve addition of a peptide rather than a smaller organic group. Ubiquitin is a peptide with Mr 7000 that is attached to lysine residues of substrates through the subsequent action of ubiquitin-activating (E1), ubiquitin-conjugating (E2) and ubiquitin-ligating (E3) enzymes. Poly-ubiquitylation usually targets a substrate for degradation by the proteasome (Varshavsky, 2005) whereas mono-ubiquitylation is involved in various processes including DNA repair and regulation of gene expression (Pickart and Fushman, 2004).
The ubiquitin-conjugating enzymes HR6A and HR6B are two very similar mouse homologs of the S. cerevisiae RAD6 protein (Roest et al., 2004; Roest et al., 1996). In yeast, RAD6 is required for ubiquitylation of H2BK123, together with an E3 named BRE1 (Hwang et al., 2003; Robzyk et al., 2000; Wood et al., 2003). H2AK119 is not ubiquitylated in yeast, but it is a prominent ubiquitylation substrate in mammalian cells, because as much as 10% of H2A, compared with 0.1-2% of H2B, is ubiquitylated in most mammalian cell types investigated (West and Bonner, 1980). To study the possible role of HR6A and HR6B in histone ubiquitylation in mammalian cells, knockout or knockdown approaches are required. In mouse, the Hr6a gene is located on the X chromosome, and Hr6b localizes to chromosome 11. Knockout mice for each individual gene were generated, but Hr6a-Hr6b double-knockouts are not viable, indicating that the two encoded proteins perform essential redundant functions (Roest et al., 2004). In most cell types, HR6A and HR6B protein levels are approximately equal, but oocytes contain a relatively high HR6A dose, explained by the presence of two active X chromosomes (Roest et al., 2004). Conversely, male germline cells show a relatively low HR6A expression associated with inactivation of the X and Y chromosomes during male meiotic prophase (see below), whereas HR6B expression is maintained at a high level (Koken et al., 1996). In the female and male germlines, there is a specific requirement for Hr6a and Hr6b, respectively, as reflected by the infertility phenotypes of Hr6a-knockout females and Hr6b-knockout males (Roest et al., 2004; Roest et al., 1996).
During male meiotic prophase, the X and Y chromosomes remain largely unpaired, with the exception of the so-called pseudoautosomal regions. The XY body is formed when the X and Y chromosomes are transcriptionally inactivated, by a mechanism named `meiotic silencing of unsynapsed chromatin' (MSUC) (Schimenti, 2005). Proteins that are specifically associated with the XY body during meiotic prophase, may play a role in MSUC. In particular, recruitment of the checkpoint kinase ATR by the BRCA1 protein and subsequent phosphorylation of Ser139 of H2AX have been shown to be essential for MSUC (Fernandez-Capetillo et al., 2003; Turner et al., 2004; Turner et al., 2005). Interestingly, H2AK119ub1 is also enriched on the XY body, concomitant with accumulation of the HR6B enzyme (Baarends et al., 2005; van der Laan et al., 2004). However, HR6A- or HR6B-dependent histone modifications in mammalian cells have not been described.
Hr6b-knockout spermatocytes display an increased frequency of meiotic recombination, possibly related to disruption of the structural organization of the paired chromosomes as a consequence of dysregulation of histone modifications (Baarends et al., 2003). In yeast, RAD6-dependent H2BK123 ubiquitylation is required for H3K4 and H3K79 methylation in a trans-histone regulatory mechanism (Dover et al., 2002; Ng et al., 2002; Sun and Allis, 2002). In addition, neighboring amino acid residues can be modified by different posttranslational modifications, and this was proposed to function as a binary switch (Fischle et al., 2003). The existence of such a binary switch in H3 was recently shown to involve stable methylation on Lys9 and transient phosphorylation of Ser10 (Fischle et al., 2005). The H2A and H2B ubiquitin-accepting lysines together with their respective adjacent threonine residues may also form binary switches (Fischle et al., 2003). It is not known whether H2BT119 is phosphorylated, but H2AT120 is phosphorylated by nucleosomal histone kinase-1 (NHK-1) in Drosophila (Aihara et al., 2004). Based on the findings of trans-histone regulation in yeast and the binary switch model, we have analyzed histone ubiquitylation as well as H2AT120 and H2BT119 phosphorylation, and H3K4 and H3K9 methylation in wild-type and Hr6b-knockout animals. Finally, we have analyzed changes in X chromosomal gene expression in wild-type and Hr6b-knockout meiotic and postmeiotic spermatogenic cells.
Histone H2AK119 and H2BK120 ubiquitylation
By analogy to the function of RAD6 in yeast, ubiquitylation of H2BK120 is the most obvious candidate histone modification that could be affected in Hr6b-knockout mice. To study the effect of loss of HR6B on H2BK120 ubiquitylation in spermatogenesis we analyzed H2BK120 ubiquitylation in spermatocytes and spermatids isolated from wild-type and Hr6b-knockout testes. Antibodies that specifically recognize ubiquitylated H2BK120 are not available, and we used anti-H2B to detect both H2B and H2BK120ub1 (Fig. 1A). In addition, the use of antibodies against ubiquitin and H2AK119ub1 allowed us to compare the amounts of H2AK119ub1 and H2BK120ub1 (Fig. 1B). The results show that the global levels of H2AK119 and H2BK120 ubiquitylation are not affected in Hr6b-knockout spermatogenic cells (Fig. 1B).
Histone H2AT120 and H2BT119 phosphorylation
The binary switch model predicts that phosphorylation of the threonines that are immediately adjacent to Lys119 and Lys120 in H2A and H2B, respectively, would take part in a mechanism to regulate binding of effectors of the histone code (Fischle et spermatocytes, we performed immunocytochemical analysis using specific antibodies against phosphorylated H2AT120 and H2BT119. To identify the different substages of meiotic prophase, we co-stained with an antibody that recognizes SYCP3. This protein is a component of the axial elements (before synapsis) or lateral elements (during synapsis) of the synaptonemal complex (SC) that forms between the chromosomal axes of pairing homologous chromosomes during meiotic prophase (reviewed by Heyting, 1996). Round spermatids are recognized using DAPI staining of DNA, which also visualizes the chromocenter, a heterochromatic dense-staining round area containing the centromeric DNA in the center of the spermatid nucleus.
The highest level of H2BT119ph was observed along the unpaired axial elements of the X and Y chromosomes of pachytene and diplotene spermatocytes, and a lower level extended over the rest of the XY body chromatin (Fig. 2A). Interestingly, H2BT119ph was strongly reduced on a few foci along the unpaired axial elements of pachytene nuclei. All other spermatogenic cell nuclei were negative, except for metaphases of mitotic spermatogonia and meiotic spermatocytes, where H2BT119ph was mainly detected on centromeric DNA (Fig. 2B, and results not shown). In nuclei of Hr6b-knockout spermatocytes, the staining pattern and level of H2BT119ph were not different from the wild type (not shown). In cultured human somatic cells (HeLa), H2BT119 phosphorylation was restricted to centromeric DNA of metaphase nuclei (Fig. 2B). All observed signals could be competed by the phosphorylated peptide that was used to generate the antibody and not by similar amounts of non-phosphorylated peptide, indicating that the antibody selectively recognized H2BT119ph on nuclei (data not shown).
Next, we analyzed H2AT120ph in spread nuclei of testes from wild-type and Hr6b-knockout mice. In wild-type cells, H2AT120ph was found to be high in leptotene/zygotene spermatocyte nuclei, with the exception of regions containing centromeric DNA. Most H2AT120ph is lost at early pachytene, but it remains present on the XY body. During late pachytene and diplotene development, H2AT120 phosphorylation decreases further, and the staining disappears also from the XY body. Hereafter, H2AT120ph increases on centromeric DNA during meiotic divisions. Then, early round spermatids show a low level of H2AT120ph, but an increased level of H2AT120ph in the whole nucleus is detected at later stages of round spermatid development, with the highest signal in the chromocenter (Fig. 3A). Antibody specificity was again demonstrated through competition with phosphorylated and nonphosphorylated peptides (not shown). In Hr6b-knockout spermatogenic cells, we observed interesting alterations in the pattern of H2AT120 phosphorylation. In late pachytene and diplotene spermatocyte nuclei, phosphorylation of H2AT120 was increased compared with wild-type nuclei of the same stages (Fig. 3B). Most strikingly, we observed enhanced signals on the XY body. We also found much higher H2AT120ph on metaphase I chromosomes of Hr6b-knockout compared with wild-type cells. Following the meiotic divisions, H2AT120ph in Hr6b-knockout round spermatids was back to the wild-type level. To substantiate these findings, the H2AT120ph signal was quantified in wild-type and Hr6b-knockout diplotene spermatocytes (Fig. 3C). The results show a threefold increase in nuclear H2AT120ph of Hr6b-knockout diplotene spermatocytes compared to levels in the wild type, and a fivefold increase in XY-body-associated H2AT120ph.
Previously, we reported accumulation of H2AK119ub1 on the XY body of pachytene spermatocytes of wild-type and Hr6b-knockout mice (Baarends et al., 1999; Baarends et al., 2005). The staining for H2AK119ub1 was repeated, and quantification of the immunofluorescent signal confirmed that H2AK119ub1 levels were similar in wild-type and Hr6b-knockout spermatocytes (Fig. 3D). Next, we studied the developmental time course of H2AT120ph and H2AK119ub1 in wild-type and Hr6b-knockout meiotic spread nuclei using triple immunocytochemical staining. In wild-type early pachytene spermatocytes, H2AT120ph precedes H2AK119ub1 on the XY body. Then, H2AT120ph strongly decreases during mid-pachytene, as H2AK119ub1 increases to a high level that is reached during late pachytene (Fig. 4A). In Hr6b-knockout pachytene spermatocyte nuclei, H2AT120ph also precedes H2AK119ub1, but remains present until diplotene, which results in high signals for the two modifications within the same nucleus (Fig. 4B).
To investigate whether H2A molecules are present that contain both the phosphorylation and the ubiquitylation marks, we performed western blot analysis for basic nuclear protein extracts from testis. In addition to H2AT120ph, a faint but specific band with a size corresponding to H2AK119ub1 was identified, and this probably represents a small fraction of H2A that is both ubiquitylated and phosphorylated at the adjacent Lys119 and Thr120, respectively, in wild-type testis (Fig. 4C). Spermatocytes of Hr6b-knockout mice show a higher level of both H2AT120ph and H2AK119ub1T120ph compared with the wild type (Fig. 4D). The H2AT120ph signal is higher in spermatids compared with spermatocytes, but we observed no difference between wild-type and Hr6b-knockout spermatids. The pattern of H2AT120ph resembles the known pattern of H2AXS139 phosphorylation during meiotic prophase (Fernandez-Capetillo et al., 2003; Turner et al., 2004), but we detected no differences for the patterns of H2AXS139 phosphorylation in wild-type and Hr6b-knockout spermatocytes (not shown).
H2AK119ub1 is associated not only with the XY body, but also with other chromatin areas that remain unsynapsed during meiotic prophase (Baarends et al., 2005). These findings indicated that H2AK119 ubiquitylation could be important for MSUC, the silencing of this chromatin. To establish whether H2BT119ph and H2AT120ph are also associated with MSUC, we investigated these modifications in T/T' mice. These mice are double heterozygous for two very similar translocations between chromosomes 1 and 13, and the small 113 bivalent often shows regions of unsynapsed chromatin that are subject to MSUC (Baarends et al., 2005). The results show that both H2BT119ph and H2AT120ph localize not only to the XY body, but also to the 113 translocation bivalent (see supplementary material Fig. S1).
XY body nucleosome replacement
Histone variant H3.1 is deposited on DNA during DNA replication, whereas H3.3 is a replacement histone that can be incorporated into nucleosomes on DNA, independent of the cell cycle phase (Tagami et al., 2004). Using antibodies against these different histone variants, it was recently shown that all nucleosomes associated with the X and Y chromosomes are replaced during pachytene (van der Heijden et al., 2007). During this process, H3.1 gradually disappears from the XY body, concomitant with a gradual increase of the H3.3 level (van der Heijden et al., 2007). We analyzed XY body nucleosome replacement in Hr6b-knockout and wild-type spermatocytes, to assess whether the altered dynamics of H2AT120ph in Hr6b-knockout spermatocytes could be caused by disturbances in the general replacement of nucleosomes. Triple immunofluorescent analyses of SYCP3, H2AT120ph and H3.1 show that the disappearance of H3.1 from the XY body in wild-type pachytene spermatocytes coincides with the disappearance of H2AT120ph (see supplementary material Fig. S3A). However, in Hr6b-knockout late pachytene spermatocytes, the XY-body-associated H2AT120ph signal is still high when H3.1 has disappeared from the XY body (see supplementary material Fig. S3B). Thus, although H2AT120p dynamics have changed in Hr6b-knockout spermatocytes, H3.1 disappearance from the XY body follows the wild-type pattern.
Since H2BT119ph was not changed in Hr6b-knockout spermatocytes, we next analyzed H3K4 methylation in spread nuclei of spermatogenic cells as another possible indirect readout of dynamic changes in H2B ubiquitylation. We have used antibodies targeting H3K4m1, H3K4m2 and H3K4m3, and here we focus on results obtained with anti-H3K4m2. Similar results were obtained with anti-H3K4m1, but H3K4m3 levels were not consistently different between spermatogenic cells of wild-type and Hr6b-knockout mice (not shown). In agreement with the known general association of H3K4m2 with potentiated or transcriptionally active chromatin, we observed a low level of H3K4m2 in regions containing heterochromatic centromeric DNA (Fig. 5A). In meiotic prophase, the overall level of H3K4 dimethylation was highest on euchromatin of leptotene/zygotene spermatocytes. The overall H3K4m2 signal was very low in pachytene and diplotene spermatocyte nuclei, followed by an increase during the meiotic divisions and post-meiotic round spermatid development. The XY body shows an even lower level of H3K4m2 compared with the rest of the pachytene nucleus. During diplotene and subsequent stages, H3K4m2 on the XY body gradually increases to a level that exceeds the H3K4m2 signal on autosomal chromatin (Fig. 5A). In haploid round spermatids, either the X or the Y chromosome is located adjacent to the chromocenter. Localization of H3K4m2 on the X or Y chromosomes in round spermatids was verified using FISH (Fig. 5C). It should be noted that a significant fraction of round spermatids showed no increased H3K4m2 signal on X or Y, and this most likely reflects H3K4m2 loss from the sex chromosomes in round spermatid at steps 6-7 (just prior to spermatid elongation), as verified by immunohistochemical analysis (not shown).
When we compared developmental H3K4 dimethylation patterns in wild-type and Hr6b-knockout testis cell preparations, we observed increased H3K4m2 on the X and Y chromosomes from diplotene onwards (Fig. 5A-C, supplementary material Fig. S2). This ∼2.5-fold increase of H3K4m2 was verified by quantification of the fluorescent signal in nuclei and XY bodies of wild-type and Hr6b-knockout diplotene spermatocytes (Fig. 5B). The increased H3K4m2 signal persisted in metaphase I and metaphase/anaphase II spermatocytes (see supplementary material Fig. S2) and round spermatids (Fig. 5C).
Previously, we have found that the synaptonemal complexes (SCs) of late pachytene Hr6b-knockout spermatocytes are longer and thinner compared with the SCs of wild-type cells. In addition, we found a loss of SC components from near telomeric regions in Hr6b-knockout late pachytene nuclei (Baarends et al., 2003). These findings probably reflect a global change in chromatin structure in late pachytene and diplotene spermatocytes. To investigate this further, we analyzed additional histone modifications generally associated with active (H4K16 and H2AK119 acetylation) or inactive (H3K9m1, H3K9m2, H3K9m3, H3K27m2 and H3K27m3) chromatin during spermatogenesis in wild-type and Hr6b-knockout mice (not shown). Of these, only H3K9m2 appeared to be different between wild-type and Hr6b-knockout spermatocytes. H3K9m2 generally marks silent chromatin, and in wild-type diplotene spermatocytes the level is high on centromeric heterochromatin and the XY body. In Hr6b-knockout diplotene spermatocytes, H3K9m2 is much lower on centromeric heterochromatin compared with wild-type nuclei of the same stage, but a normal level is observed on the XY body (Fig. 6A,B). Also, in round spermatids, H3K9m2 is lower on the chromocenter, but not on either X or Y (Fig. 6A,B).
MacroH2A1 is also enriched on heterochromatin and the XY body in late pachytene and diplotene spermatocytes (Hoyer-Fender et al., 2000). We found a small increase in the overall level of macroH2A1 in Hr6b-knockout diplotene spermatocytes, compared with wild-type cells. This is a global change, with the same relative increase on autosomal and sex chromosomal chromatin (see supplementary material Fig. S3C,D). In round spermatids, macroH2A1 is gradually lost from autosomes and sex chromosomes, in a pattern that is not different between wild-type and Hr6b-knockout cells (not shown). Western blot analyses for basic nuclear proteins extracted from wild-type and Hr6b-knockout spermatocytes and spermatids, showed no effect of Hr6b mutation on macroH2A1 levels, and also not on H3K4m2 and H3K9m2 levels (see supplementary material Fig. S3E). As described in the legend to Figure S3E, the western blots represent a mixed cell population in which subnuclear changes that occur in germ cell substages that constitute a small fraction of the total will go undetected. The western blot results for these histone variants and modifications, therefore, confirm their presence, but cannot provide additional detail about the sub-nuclear and temporal control of these modifications.
Postmeiotic maintenance of X chromosome silencing
The relatively high level of H3K4m2 on the X and Y chromosomes in Hr6b-knockout spermatids may be related to changes in transcriptional activity. To investigate this further, we analyzed the expression of several X chromosomal and autosomal genes in cell preparations isolated from wild-type and Hr6b-knockout mice. Genes were selected based on data from Namekawa et al. (Namekawa et al., 2006) who showed that transcription from the X chromosome is largely repressed in meiotic and postmeiotic cells (Namekawa et al., 2006). We selected four genes that were reported to remain repressed in the postmeiotic spermatids (Chic1, Atp7a, Gla and Hprt) and two genes that showed postmeiotic reactivation (Ube1x and Pctk1) (Namekawa et al., 2006). In addition, we selected four autosomal genes that are expressed at meiotic (Spo11 and Sycp3) and postmeiotic (Tnp1 and Creb3l4) spermatogenic developmental steps. Real-time RT-PCR expression data were normalized to Actb (β-actin) mRNA, which showed equal expression in the different cell preparations (data not shown). For the autosomal genes, no significant differences between wild-type and Hr6b-knockout cells were detected (Fig. 7A). However, five out of the six X chromosomal genes that were tested show increased expression in Hr6b-knockout cells compared with the wild type (Fig. 7B). This effect was most clear for round spermatids. Pctk1 is expressed at a relatively high level compared with the other tested X-chromosomal genes, and no further upregulation of its postmeiotic expression was observed for the Hr6b-knockout cells (Fig. 7B). Taken together, the results indicate that postmeiotic maintenance of X-chromosomal gene silencing is compromised in Hr6b-knockout spermatids.
HR6A/B and histone ubiquitylation
RAD6 is a protein that is highly conserved in evolution. It is most well known for its pivotal function in replicative damage bypass, a pathway that allows replication to proceed in the presence of a damaged template (van der Laan et al., 2005). RAD6 is also essential for sporulation, and this involves H2B ubiquitylation by RAD6, together with BRE1 (Hwang et al., 2003; Robzyk et al., 2000; Wood et al., 2003). The present results indicate that HR6A and HR6B do not act as main determinants of global histone ubiquitylation in mammalian cells, since we observe no detectable defects in H2A or H2B ubiquitylation in spermatocytes from Hr6b-knockout males. HR6A is still expressed in these cells, at a low level (Baarends et al., 2003), and we cannot exclude the possibility that this small amount of HR6A is responsible for the observed maintenance of global histone ubiquitylation in Hr6b-knockout germline cells. In addition, redundancy with other E2 enzymes may prevent detection of some role of HR6B in dynamic control of histone ubiquitylation. Ideally, Hr6a-Hr6b double-knockout cells should be used to address the role of HR6A and HR6B in histone ubiquitylation in somatic and germline cells, but these cells are not obtained (Roest et al., 2004).
Recently, the human E2 enzyme UBCH6 has been shown to be able to ubiquitylate H2B in vitro, together with an E3 complex consisting of RNF20 and RNF40 proteins (Zhu et al., 2005). RNF20 and RNF40 are orthologs of BRE1, the yeast E3 that ubiquitylates H2B. In addition to the evolutionary conservation of these components, the pathway leading from H2B ubiquitylation to H3K4 methylation also appears to be conserved (Zhu et al., 2005). Possibly, the role of RAD6 in histone ubiquitylation has been taken over, at least in part, by UBCH6. Given the conservation of the trans-histone regulatory pathway, leading from H2BK120 ubiquitylation to H3K4 methylation, our observation that H3K4 methylation in Hr6b-knockout spermatogenic cells is not reduced, provides further evidence that H2B ubiquitylation is not affected in HR6B-deficient spermatocytes. HR6B localizes primarily to the XY body, and most likely is required to ubiquitylate certain XY body chromatin components, but not H2A or H2B. Somehow, lack of HR6B affects H2AT120ph, and this may subsequently lead to increased H3K4m1 and H3K4m2, specifically in the XY body. Previously, Khalil and Driscoll (Khalil and Driscoll, 2006), reported that H3K4m2 is upregulated on the silent XY body of wild-type diplotene spermatocytes. Our data confirm their findings, and we show that in Hr6b-knockout spermatocytes this modification is further upregulated on X and Y. We did not detect a consistent increase in H3K4m3 levels, but the anti-H3K4m3 antibodies crossreact to some extent with H3K4m2. Therefore, we cannot exclude that in addition to H3K4m2, H3K4m3 is also increased.
Global versus XY specific chromatin regulation
HR6A and HR6B probably exert multiple functions during spermatogenesis. During meiotic prophase, HR6B concentrates on the XY body, but is also present on autosomal chromatin. Previously, we have shown that Hr6b-knockout spermatocytes show a higher recombination frequency associated with some dysregulation of the structure of the synaptonemal complex (Baarends et al., 2003). Changes in the synaptonemal complex might follow after global changes in chromatin structure, and such global changes are indicated by increased macroH2A1 and decreased H3K9m2 levels of meiotic prophase chromatin. This provides a background for the observed differential effects of HR6B deficiency on autosomal versus XY associated histone modifications. Interestingly, however, the loss of H3K9m2 signal in Hr6b-knockout diplotene spermatocytes and round spermatids occurs on all heterochromatin, but not on XY chromatin. This adds to our observations on the marked XY-associated increase in H2AT120p and H3K4m2 signals.
H2AT120ph and regulation of chromatin organization
In Drosophila, nucleosomal histone kinase-1 (NHK-1) phosphorylates H2AT120 (Aihara et al., 2004). Female flies that carry a mutation in the gene encoding NHK-1 are infertile (Ivanovska et al., 2005). Loss of H2AT120ph was shown to be associated with a failure to disassemble the synaptonemal complex and with impaired loading of condensin (Ivanovska et al., 2005). H2AT120ph in mouse may also be associated with regulation at this level of chromatin organization. During meiosis, we observed a very high level of H2AT120ph, not only at metaphase, but also in leptotene-zygotene spermatocytes. This indicates that H2AT120ph may be relevant not only for disassembly of the synaptonemal complex (SC), but also for its assembly. Hr6b-knockout spermatocytes show SCs that are longer, with some loss of SC proteins from near centromeric regions (Baarends et al., 2003). Thus, in analogy to what is observed in Drosophila, the modified level and pattern of H2AT120ph in Hr6b-knockout spermatocytes may be related to a role of HR6B in the maintenance of a normal SC structure. Such a relationship may also appear from the observed loss of H3K9me2 from centromeric heterochromatin in Hr6b-knockout diplotene spermatocytes.
In wild-type spermatocytes going through meiotic prophase, H2AT120 phosphorylation decreases as H2AK119 ubiquitylation increases. Also, during spermatid elongation, H2AT120ph decreases again when H2AK119ub1 increases (our unpublished results).
This indicates that H2AT120 phosphorylation is downregulated prior to H2AK119 ubiquitylation. Loss of XY-body-associated H2AT120 phosphorylation occurs concomitantly with the exchange of all nucleosomes from the XY body, as visualized by the loss of H3.1. Therefore, no active dephosphorylation may be required at this stage. In Hr6b-knockout pachytene spermatocytes, H2AT120ph is not properly removed, but H2AK119 ubiquitylation increases as in the wild type. Apparently, H2AT120ph dephosphorylation is not a prerequisite for H2AK119 ubiquitylation.
In Hr6b-knockout pachytene and diplotene nuclei, H2AT120ph is increased throughout the nucleus, but the combined H2AK119ub1T120ph modification is present mainly on the XY body. This might lead to XY-body-restricted recruitment of complexes that recognize this combinatory code, and such a mechanism would provide an explanation for a subsequent increase of H3K4me2 only on the XY body of Hr6b-knockout diplotene spermatocytes.
H2AT120 and H2BT119 phosphorylation and meiotic silencing of unpaired chromatin (MSUC)
In meiotic prophase nuclei we detect H2BT119ph only on chromatin associated with the unpaired axial elements of the XY body. Then, when cells enter metaphase I, H2BT119ph localizes on centromeric DNA. On western blots of basic nuclear proteins isolated from testis, no specific signal could be obtained with anti-H2BT119ph antibody (results not shown). This may be due to the fact that the percentage of H2BT119ph is extremely low compared with the amount of unphosphorylated H2B. The localization of H2BT119ph on centromeric DNA of metaphase chromosomes of mitotic somatic cells is similar to the reported localization of H2AT120ph on metaphase chromosomes of mitotic cells from Drosophila (Aihara et al., 2004) and human (our unpublished observations). However, during spermatogenesis, the two modifications display different localization patterns and kinetics. H2AT120ph is first present on all chromatin during zygotene, and then persists on unpaired chromatin. By contrast, H2BT119ph is absent during leptotene-zygotene, and subsequently this phosphorylation is specifically induced on chromatin associated with unpaired axial elements. On the XY body, H2AT120ph covers all chromatin, whereas H2BT119ph is mainly concentrated on the axial elements that are unpaired, and it is excluded from pseudoautosomal regions, and a few unidentified small regions on the axial elements, which appear to show enhanced SYCP3 staining.
In spermatocytes from T/T' mice, phosphorylation of H2AT120 and H2BT119 is enhanced on the partially synapsed 113 bivalent. Taken together with the observations on these modifications for the unpaired and silenced XY chromatin, this indicates that H2AT120ph and H2BT119ph may be functionally relevant for MSUC.
Postmeiotic derepression of X-chromosomal gene expression in Hr6b-knockout spermatids
In Hr6b-knockout spermatocytes, accumulation of specific XY-body-associated histone modifications and nucleosome replacement all occur as in wild-type cells. The increase in H3K4m2 on X and Y chromatin becomes apparent in late meiotic prophase spermatocytes, and remains present in postmeiotic Hr6b-knockout cells. Therefore, we conclude that XY body formation in Hr6b-knockout spermatocytes is not affected; rather, late meiotic and postmeiotic regulation of XY chromatin appears to be disturbed. This is supported by our finding that the most obvious derepression of X-chromosomal genes is detected in postmeiotic spermatids.
H3K4 methylation is generally found in association with active genes, although recent data indicate that certain repressors of gene expression can also bind methylated H3K4 (reviewed by Becker, 2006). We find upregulation of five out of six tested X chromosomal genes in Hr6b-knockout spermatids, and no downregulation. This effect points to a link between H3K4 methylation and gene activation on the postmeiotic X chromosome. The autosomal genes tested show wild-type mRNA levels in the Hr6b-knockout cells. Although we do not exclude that autosomal gene expression is affected to some extent, the loss of HR6B activity seems to exert a more pronounced effect on transcriptional activity of the X chromosome.
The data presented herein provide evidence for a role of HR6B in the regulation of histone modifications in mammalian cells. This has been revealed in Hr6b-knockout spermatocytes, a cell type with a relatively low level of HR6A. HR6A and HR6B show 96% amino acid similarity, and at present we have no indications that HR6A and HR6B perform different activities, either in somatic or in germline cells. In spermatocytes, HR6B localizes mainly on the XY body, together with the putative partner ubiquitin ligase RAD18 (van der Laan et al., 2004). Although H2AK119ub1 is also enriched on the XY body, our findings indicate that HR6B activity is not responsible for this modification. Moreover, HR6B is not required for global H2BK120 ubiquitylation. Instead, we have established that loss of HR6B affects other aspects of histone modifications associated with the XY body, in particular exerting an effect on H2AT120ph and H3K4m2, in association with derepression of X-chromosomal genes in postmeiotic cells.
Materials and Methods
Isolation of different cell types from mouse testis
Spermatocytes and round spermatids were isolated from 4- to 5-week-old wild-type (FVB) and Hr6b-knockout mouse testes after collagenase and trypsin treatment, followed by sedimentation at unit gravity (StaPut procedure) (Grootegoed et al., 1986). This yielded a fraction containing approximately equal amounts of spermatocytes and round spermatids (spc/spt), with few other contaminating cells (<10%), and a fraction containing >90% pure spermatids (spt). These fractions were used for analysis of RNA. For protein analysis, cells were further purified (>90%) by density gradient centrifugation through Percoll (Grootegoed et al., 1986) resulting in preparations of spermatocytes (spc) and spermatids (spt). Cells were snap-frozen in liquid nitrogen and stored at –80°C.
Generation of antibodies
Antibody against peptide CPGGRKHSGKSGKPPL representing amino acids 22-37 of mouse SYCP3, with an added N-terminal cysteine, was generated at Eurogentec (Seraing, Belgium) according to their protocols. Antibodies against H2AT120ph, H2BT119ph, and H2AK119ac were generated in rabbits using peptides GTKAVT*KYTSS, LLPKKT*ESHH, and QAVLLPKK*TESH (asterisk indicates phosphorylated or acetylated residue), respectively, using standard protocols. Phosphorylated and non-phosphorylated and acetylated and non-acetylated peptides were used in competition experiments to test antibody specificity.
Isolation of acid-soluble nuclear proteins and western blotting
Nuclei and acid-soluble proteins were isolated from cell preparations or total testes according to Chen et al. (Chen et al., 1998). The isolated protein fraction was precipitated with 5% (w/v) trichloroacetic acid. 20 μg of protein per sample was separated on 12% SDS-polyacrylamide gels and the separated proteins were transferred to nitrocellulose membranes, using the Bio-Rad miniprotean III system and blot cells (Bio-Rad, Veenendaal, Netherlands). Membranes were stained with Ponceau S (Sigma-Aldrich, Zwijndrecht, Netherlands) according to the supplier's protocol.
Modified histones were detected with mouse monoclonal IgM anti-H2AK119ub1 (Upstate, Waltham, MA), rabbit polyclonal anti-ubiquitin (DakoCytomation, Glostrup, Denmark), rabbit polyclonal anti-H2B (Upstate) and rabbit polyclonal anti-H2AT120ph.
Meiotic spread nuclei preparations and immunocytochemistry
Testes were obtained from adult T(1;13)70H/T(1;13)1Wa (T/T') mice (Swiss random bred), and from 5-week-old wild-type and Hr6b-knockout mice (FVB background).
Testis tissues were processed to obtain spread nuclei for immunocytochemistry as described by Peters et al. (Peters et al., 1997), Spread nuclei of spermatocytes were stained with rat polyclonal anti-Sycp3, mouse monoclonal IgM anti-ubi-H2A (Upstate, Waltham, MA), rabbit polyclonal anti-H3K4m1 (Upstate), rabbit polyclonal anti-H3K4m2 (Upstate and Abcam, Cambridge, UK with similar results), rabbit polyclonal anti-H3K4m3 (Upstate and Abcam), rabbit polyclonal anti-H3K9m1 (Upstate), rabbit polyclonal anti-H3K9m2 (Upstate), mouse monoclonal anti-H3K9m2 (Abcam) (both anti-H3K9m2 antibodies yielded similar results), rabbit polyclonal anti-H3K27m2 (Upstate), mouse monoclonal anti-H3K27m3 (Abcam), rabbit polyclonal anti-H4K16ac (Abcam), rabbit polyclonal anti-macroH2A1 (Upstate), mouse monoclonal anti-H3.1 (van der Heijden et al., 2005), rabbit polyclonal anti-H2AT120ph, rabbit polyclonal anti-H2AK119ac, or rabbit polyclonal anti-H2BT119ph. Secondary antibodies were FITC-(Sigma, St Louis, MO), TRITC-(Sigma), Alexa Fluor 350, Alexa Fluor 594 or Alexa Fluor 488 (Molecular Probes)-labeled goat anti-rabbit, goat anti-mouse or goat anti-rat IgG antibodies; FITC-labeled goat anti-mouse IgM (Sigma) was used as secondary antibody for anti-H2AK119ub1 (IgM). Before incubation with antibodies, slides were washed in PBS (3×10 minutes), and non-specific sites were blocked with 0.5% w/v BSA and 0.5% w/v milk powder in PBS. First antibodies were diluted in 10% w/v BSA in PBS, and incubations were overnight at room temperature in a humid chamber. Subsequently, slides were washed (3×10 minutes) in PBS, blocked in 10% v/v normal goat serum (Sigma) in blocking buffer (supernatant of 5% w/v milk powder in PBS centrifuged at 14,000 rpm for 10 minutes), and incubated with secondary antibodies in 10% normal goat serum in blocking buffer at room temperature for 2 hours. Finally, slides were washed (3×10 minutes) in PBS (in the dark) and embedded in Vectashield containing DAPI to counterstain the DNA (Vector Laboratories, Burlingame, CA) or in Prolong Gold (Molecular Probes) when Alexa Fluor-labeled secondary antibodies were used. DAPI was omitted when triple immunostainings were performed. Fluorescent images from spread nuclei were observed using a fluorescent microscope (Axioplan 2; Carl Zeiss, Jena, Germany) equipped with a digital camera (Coolsnap-Pro, Photometrics, Waterloo, Canada). Digital images were processed using Adobe Photoshop software (Adobe Systems). For quantification of immunofluorescent signal, slides were analyzed on the same day. Fluorescent images were taken under identical conditions for all slides, and not further processed in Adobe. The signal present in the total nucleus and in the XY body was measured using ImageJ software analysis (National Institutes of Health).
Following immunocytochemistry, the position of selected nuclei on the slide was determined, and FISH was performed to detect the X chromosome using STAR*FISH mouse whole chromosome-specific paints (1200XmCy3; Cambio, Cambridge, UK) according to the manufacturer's protocol. If specific signal was not obtained, the procedure was performed for a second time, and this always resulted in a positive signal in most nuclei. Specificity of hybridization was confirmed using male meiotic spread nuclei preparations; positive signal colocalized with the XY body of pachytene spermatocytes. Digital images were obtained and processed as above. FISH images were combined with immunocytochemical images using Adobe Photoshop software (Adobe Systems).
For real-time RT-PCR, RNA was prepared from the isolated germ cell preparations by Trizol, DNase-treated and reverse transcribed using random hexamer primers and Superscript II reverse transcriptase (Invitrogen, Breda, The Netherlands). PCR was carried out with the iQ SYBR green PCR mastermix (Applied Biosystems) in the DNA engine Opticon 2 real-time PCR detection system (Bio-Rad). For Hprt, Pctk1 and Atp7a we used the following conditions: 3 minutes 95°C then 30 seconds 95°C, 30 seconds 62°C, 30 seconds 72°C for 45 cycles. For Chic1, we used 3 minutes 95°C then 30 seconds 95°C, 30 seconds 65°C, 30 seconds 72°C, 30 seconds 78°C for 45 cycles, For Gla we used 3 minutes 95°C then 30 seconds 95°C, 30 seconds 65°C, 30 seconds 72°C for 45 cycles and for Ube1x conditions were: 3 minutes 95°C then 30 seconds 95°C, 30 seconds 57°C, 30 seconds 72°C, 30 seconds 80°C for 45 cycles. For Spo11, Sycp2, Tnp1 and Creb3l4 we used 3 minutes 95°C then 30 seconds 95°C, 30 seconds 60°C, 30 seconds 72°C for 45 cycles. β-actin was included in each reaction and used to normalize the data. Two independent experiments were performed and each real-time PCR was performed in duplicate. All –RT reactions were negative. Forward and reverse primers for Hprt, Pctk1, Atp7a, Gla, Chic1 and Actb were as described (Namekawa et al., 2006). Forward and reverse primers (5′ to 3′): Ube1x: TGTCCACACCCACTTACT and GCACTCTGCAACTCC; Spo11: GCTCCTGGACGACAACTTCT and ATCTGCATCGACCAGTGTGA; Sycp2: TGGATGTGATGACAGCAAGA and TGGGTCTTGGTTGTCCTTTT; Tnp1: AAGAACCGAGCTCCTCACAA and CATCACAAGTGGGATCGGTA; Creb3l4: CCTCCGATTCGCATAGACAT and GCCAGCAGTTGCTTTTCTTC.
We would like to thank Wiggert A. van Cappellen, Jan H. J. Hoeijmakers and Henk P. Roest (Erasmus MC, Rotterdam, The Netherlands) for helpful discussions. We thank Peter de Boer and Johan van der Vlag (Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands) for providing T/T' mice and the H3.1 antibody, respectively. We thank Upstate Biotechnology and David Allis (The Rockefeller University, New York, USA) for their role in developing the H2AT120ph, H2BT119ph and H2AK119ac antibodies and Mahesh Chandrasekharan (Vanderbilt University School of Medicine, Nashville, USA) for the initial characterization of these antibodies. This work was supported by the Netherlands Organisation for Scientific Research (NWO) through ALW (VIDI 864.05.003). Z.-W.S. is supported by NIH (RO1CA109355).
Supplementary material available online at http://jcs.biologists.org/cgi/content/full/120/11/1841/DC1
- Accepted April 4, 2007.
- © The Company of Biologists Limited 2007