Abstract
Spermatogenesis is a highly complex cell differentiation process that is governed by unique transcriptional regulation and massive chromatin alterations, which are required for meiosis and postmeiotic maturation. The underlying mechanisms involve alterations to the epigenetic layer, including histone modifications and incorporation of testis-specific nuclear proteins, such as histone variants and protamines. Histones can undergo methylation, acetylation, and phosphorylation among other modifications at their N-terminus, and these modifications can signal changes in chromatin structure. We have identified the temporal and spatial distributions of histone H3 mono-, di-, and trimethylation at lysine 4 (K4), and the lysine-specific histone demethylase AOF2 (amine oxidase flavin-containing domain 2, previously known as LSD1) during mammalian spermatogenesis. Our results reveal tightly regulated distributions of H3-K4 methylation and AOF2, and that H3-K4 methylation is very similar between the mouse and the marmoset. The AOF2 protein levels were found to be higher in the testes than in the somatic tissues. The distribution of AOF2 matched the cell- and stage-specific patterns of H3-K4 methylation. Interaction studies revealed unique epigenetic regulatory complexes associated with H3-K4 methylation in the testis, including the association of AOF2 and methyl-CpG-binding domain protein 2 (MBD2a/b) in a complex with histone deacetylase 1 (HDAC1). These studies enhance our understanding of epigenetic modifications and their roles in chromatin organization during male germ cell differentiation in both normal and pathologic states.
Introduction
Spermatogenesis is a highly complex cell differentiation process that is essential for sexual reproduction. While our knowledge of the molecular mechanisms that regulate and coordinate spermatogenesis remains incomplete, it is well established that pituitary gonadotropins and androgens are critical effectors [1]. Male germ cell development includes unique processes, such as meiosis, genetic recombination, haploid gene expression, formation of the acrosome and flagellum, and remodeling and condensation of the chromatin. These processes are intricate, highly ordered, and require novel gene products and a precise and stringently co-ordinated program of chromatin reorganization. As with any differentiating tissue in the body, the characteristics of the cell types that comprise the mammalian testis are dependent upon differential gene expression. This differential gene expression is achieved by unique chromatin remodeling, transcriptional regulation, posttranscriptional control of mRNAs, and the expression of testis-specific genes or isoforms [2–4]. Epigenetic events in the testis have just begun to be studied. New work on the function of specific histone modifications suggests that they serve a key role in the control of development [5–7].
The topic of epigenetics, which refers to heritable changes in gene expression that are not encoded in the DNA, includes covalent modification of histones and methylation of DNA. The epigenetic layer regulates chromatin higher order and controls access of the transcriptional and repair machinery to the DNA [8]. Importantly, altered male fertility and reproductive health have been linked with exposure to chemotherapeutic drugs and pesticides that disrupt the epigenetic program. This has been shown to be manifested as DNA damage and altered DNA methylation [9–11], while the role of histone modifications remains to be elucidated. Histones can undergo epigenetic modifications at their N-termini, including methylation, acetylation, and phosphorylation, among others. Indirect immunofluorescence studies of cultured cells have been highly informative regarding the functions of specific histone modifications that modulate cellular events, e.g., histone methylation at H3-K9 directs X-chromosome inactivation in mammals [12–15], and histone phosphorylation of H3-S10 regulates pericentric chromatin condensation [16–18]. Despite these advances in mammalian model systems and in development, a comprehensive map of the organization of histone modifications during male germ cell differentiation has not been established.
Histone methylation is catalyzed by histone methyltransferases (HMTs), which can be grouped into two major classes, those specific for arginine (R) methylation (PMRTs, protein arginine methyltransferases), and those specific for lysine (K) methylation (HKMTs, histone lysine methyltransferases). To date, five enzymes have been identified that methylate histone H3 at lysine 4 (H3-K4) [19]. Genomic analysis has revealed the H3-K4 di- and trimethylation are principally associated with active transcription [20]. The absolute requirement for histone methylation in spermatogenesis is demonstrated by the sterility of mice that bear defects in histone methyltransferases Suv39h1 and Suv39h2, which are specific for histone H3-K9 [15]. Importantly, at the earliest point in germ cell specification, histone methylation has been shown to be crucial. Disruption of Prdm1 (previously known as Blimp1), which is a potent transcriptional repressor of a histone methyltransferase subfamily, results in aberrant primordial germ-like cells with deregulated gene expression [21].
Histone methylation was previously considered to be a permanent marker, until the recent identification of the first histone demethylase, the lysine-specific histone demethylase AOF2 (amine oxidase flavin-containing domain 2, previously known as LSD1) [24]. AOF2 is specific for the removal of mono- and di-methylation of H3-K4 [22–25], and functions with histone deacetylases (HDAC1/2), REST corepressor 1 (RCOR1), and PHD finger protein 21A (PHF21A) as a transcriptional repressor [25, 26]. AOF2 has also been reported to participate in the demethylation of H3-K9 [27], by co-operating with the androgen receptor to activate androgen receptor target genes via H3-K9 demethylation.
This is the first comprehensive study to determine precisely the stage- and cell-specific distributions of H3-K4 methylation and AOF2 in mammalian male germ cells, and to identify novel chromatin regulatory complexes associated with AOF2. Importantly, our comparison of the temporal and spatial distributions of H3-K4 methylation in spermatogenesis between the mouse and the common marmoset, Callithrix jacchus, indicates that these epigenetic markers are likely to be involved in the highly conserved mechanisms of chromatin organization in male germ cell development. We found high levels of AOF2 in the mouse testis, in comparison with the levels in somatic tissues, and its directly overlapping distribution pattern with H3-K4 mono-and di-methylation suggests that it is a key regulator of methylation of histone H3-K4 in spermatogenesis. Our coimmunoprecipitation studies identified interactions between the epigenetic modifiers AOF2, HDAC1, and MBD2a,b in the mouse testis, which suggest that they play a role in chromatin reorganization. Epigenetic events in the testis have just begun to be studied, and this study on the distribution and functional protein partners associated with histone H3-K4 methylation indicates a fundamental role for H3-K4 methylation in germ cell development.
Materials and Methods
Mice
All tissues and cells were obtained from CD1 mice (Charles River), which were housed under normal light/dark conditions with free access to food and water. Mice were killed and the testes were collected during the first wave of spermatogenesis at postnatal days corresponding to the appearance of the spermatogenic cell types [28]: postnatal day (PND) 6, type A spermatogonia; PND 8, type A/B spermatogonia; PND 10, preleptotene/leptotene; PND 12, zygotene; PND 14, pachytene; PND 20, late pachytene; and adult, 8 weeks of age. All animal procedures were approved by the Animal Care and Use Committee of McGill University, Montreal, Canada.
Primates
All procedures were carried out according to German Animal Experimentation Law. Animals were housed according to standard German Primate Centre practice for the common marmoset (Callithrix jacchus) [29, 30]. Two adult animals were killed by an overdose of anaesthetic, and the testes were surgically removed and immediately processed for further analyses.
Histology and Immunohistochemistry
Tissues were fixed in Bouin solution, processed for embedding in paraffin, and sectioned using standard histological protocols. Immunohistochemical staining was performed on 5-μm-thick sections. Briefly, tissues were deparaffinized with Citrisolve and then rehydrated through three changes of alcohol. After washing in PBS-Brij (PBS plus 0.03% Brij 35) for 10 min, antigen retrieval was performed by incubating tissue sections in sodium citrate buffer, heating in the microwave until boiling, followed by cooling for 30 min at room temperature and washing with distilled water. The slides were then rinsed in PBS-Brij and endogenous peroxidase activity was blocked by incubation with 0.3% hydrogen peroxide in methanol for 30 min at room temperature. The sections were subsequently blocked in 5% BSA in PBS-Brij for 1 h, and then incubated with polyclonal rabbit anti-H3-K4 mono-methyl (ab8895, 1:500 dilution; Abcam), anti-H3-K4 di-methyl (Abcam, ab7766, 1/500), anti-H3-K4 trimethyl (ab8580, 1:500; Abcam), anti-AOF2 (otherwise known as LSD1, ab17721, 1:50; Abcam), or control rabbit IgG (Jackson Immunoresearch Laboratories) antibody at the same concentration as the primary antibody, with rocking overnight at 4°C. After washing, the sections were incubated with secondary horseradish peroxidase-conjugated anti-rabbit antibody (1:500; Jackson Immunoresearch Laboratories), for 1 h at room temperature, followed by washing. Immune complexes were revealed by diaminobenzidine (Sigma) and sections were counterstained with hematoxylin. Spermatogenic stages for mouse histology were determined according to Russell et al. [31], and for marmoset histology according to the criteria described by Holt and Moore [32] and Millar et al. [33].
Stage-Specific Cell Isolation and Immunofluorescence
For immunofluorescent colocalization of AOF2 and ZBTB16 (previously known as PLZF, spermatogonial marker for As, Apr, and Aal), and for ZBTB16 and H3-K4 mono-, di- and trimethylation, stage-specific populations of mouse germ cells were isolated using transillumination-assisted microdissection, followed by phase-contrast analysis, as described by Kotaja et al. [34]. Briefly, monolayers of stage-specific cell populations were snap-frozen in liquid nitrogen, fixed in 90% ethanol, followed by washing in PBS-T (PBS with 0.05% Triton-X) and blocking for 30 min at room temperature in 3% BSA and 10% normal goat serum in PBS-T. The cells were then incubated with polyclonal rabbit anti-AOF2 (1:300; Abcam), anti-H3-K4 mono-methylation (1:2000; Abcam), anti-H3-K4 di-methylation (1:1000; Abcam), anti-H3-K4 trimethylation (1:2000; Abcam) or control rabbit IgG (Jackson Immunoresearch Laboratories) antibody, for 2 h at room temperature. The cells were washed and incubated with anti-ZBTB16 antibody (anti-PLZF, OP128, 1:100; Calbiochem). Primary antibodies were detected using Alexa Fluor 594-conjugated anti-rabbit and Alexa Fluor 488-conjugated anti-mouse (Molecular Probes) antibodies used at dilutions of 1:1000. Cells were counterstained with 4',6-diamidino-2-phenylindole (DAPI).
Immunoprecipitation Assay
Total testis extracts were prepared in modified RIPA (50 mM Tris-HCl [pH 7.5], 300 mM NaCl, 1% NP-40, 50 mM NAF, 1 mM PMSF, 100 μM NaVO3, and a proteinase inhibitor cocktail [Sigma]) and kept on ice for 30 min. Soluble extracts were separated by centrifugation at 10 000 × g for 30 min. Cell extracts that contained 200 μg of total protein were precleared for 1 h on Protein A-agarose beads (Roche Diagnostics) and collected as supernatant fluids after centrifugation. The supernatants were incubated for 2 h at 4°C, with 2 μg of anti-HDAC1 (clone 2E10, Upstate Cell Signaling), anti-AOF2 (Abcam), anti-MBD2a (ab3754; Abcam) or anti-MBD2ab (kindly provided by Dr. M. Szyf, McGill University, Montreal) antibody, or the appropriate controls, which included nonimmune serum, mouse IgG, and rabbit IgG. Protein A-agarose beads were preadsorbed with 0.05% BSA prior to capturing the immune complex for 1 h at 4°C. The beads were then washed three times with lysis buffer and captured proteins were eluted in SDS-PAGE sample buffer for Western blot analysis.
Western Blotting
For Western blotting, total testis extracts were prepared in modified RIPA (50 mM Tris-HCl [pH 7.5], 170 mM NaCl, 1% NP-40, 50 mM NAF, 1 mM PMSF, 100 μM NaVO3, and a proteinase inhibitor cocktail), followed by quantitation using the Bradford assay. Equal amounts of protein were resolved by standard SDS-PAGE and electroblotted onto nitrocellulose membranes. The membranes were incubated overnight at 4°C in PBS that contained 5% low-fat milk, 0.05% Tween-20, and polyclonal rabbit anti-AOF2 (1:500), anti-H3-K4 mono-, di- or trimethylation (1:2000), anti-HDAC1 (1:5000), anti-MBD2a/b (1:2000) or anti-MBD2a (1:2000) antibody, or the loading controls, which included monoclonal anti-β-actin (Sigma, 1:10 000) and monoclonal anti-MAPK1 (ERK2 (C-14), 1:1000; Santa Cruz Biotechnology) antibodies. Donkey anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:5000; Jackson Immunoresearch Laboratories) or donkey anti-mouse antibody (1:20 000; Jackson Immunoresearch Laboratories) was diluted in 5% milk in PBS with 0.05% Tween-20 and labeling was detected using enhanced chemiluminescence (Pierce). Membranes were exposed to Kodak autoradiography BioMax film. Each experiment was replicated a minimum of three times.
Cell Isolation by Elutriation
Testes from 23- to 28-day-old CD1 mice (Charles River Italia, Calco, Italy) were used to obtain spermatocytes and round spermatids by centrifugal elutriation, as described previously [35]. Spermatogonia were obtained from 8-day-old mice, as described previously [36]. After elutriation, cell extracts were immunoblotted using standard techniques.
Results
Dynamic Modulation of H3-K4 Mono-, Di-, and Trimethylation During Mouse Spermatogenesis
Strikingly, H3-K4 mono-, di-, and trimethylation display tightly controlled temporal and spatial appearances during spermatogenesis. We used immunohistochemical analysis of testis sections obtained from adult males and prepubertal males collected at PNDs corresponding to the appearance of spermatogenic cell types, to determine the spermatogenic stage and cell distribution of H3-K4 methylation during spermatogenesis. Our analysis indicated that the levels of H3-K4 mono-, di-, and trimethylation were highly dynamic during germ cell development (Fig. 1 and Table 1). These epigenetic modifications were observed at the onset of spermatogenesis, with moderate to strong signals present in type A and B spermatogonia (Fig. 1). Positive staining for H3-K4 mono-, di-, and trimethylation was also detected in prepubertal testes at PND 6 and PND 10, corresponding to the appearance of type A and B spermatogonia (Supplemental Fig. 1, available online at www.biolreprod.org). At the onset of meiosis, in preleptotene spermatocytes, H3-K4 mono-, di-, and trimethylation were strongly up-regulated (Fig. 1 and Supplemental Fig. 1). In the mid- and late-meiosis stages (zygotene through pachytene), H3-K4 mono- and di-methylation were greatly reduced (Fig. 1, A and B), while H3-K4 trimethylation persisted in zygotene spermatocytes (Fig. 1C). Unlike mono- and di-methylation, which were low in round spermatids (Fig. 1, A and B), H3-K4 trimethylation was moderate in round spermatids from stages I to VIII and persisted in elongating spermatids from stages VIII to XI (Fig. 1C). At the onset of spermatid elongation, H3-K4 mono-, di-, and trimethylation temporarily increased (stages IX to XII) (Fig. 1). These cell-specific patterns of H3-K4 methylation observed in adult animals were confirmed by immunohistochemical analysis of developmentally staged testis collected at PND 6–20, corresponding to the appearance of the spermatogenic cell types (Supplemental Fig. 1) [28].
Spermatogenic cell distribution patterns of H3-K4-me1, H3-K4-me2, H3-K4-me3, and AOF2 in the mouse.
| Cell typesa | Staining intensityb | |||
|---|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | AOF2 | |
| As, Apr, Aal spermatogonia | Intermediate | Intermediate | Intermediate | Intermediate |
| In and B spermatogonia | Strong | Intermediate | Intermediate | Strong |
| Preleptotene spermatozytes | Strong | Strong | Strong | Strong |
| Leptotene spermatozytes | Strong | Strong | Intermediate | Strong |
| Zygotene spermatozytes | Weak | Weak | Strong | Weak |
| Pachytene spermatozytes | Weak | Weak | Weak | Weak |
| Round spermatids | Moderate | Moderate | Intermediate (stages I–VII) | Moderate |
| Elongating spermatids | Strong (stages IX–XI) | Strong (stages IX–XII) | Strong (stages IX–XI) | Strong (steps 10–16) |
| Cell typesa | Staining intensityb | |||
|---|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | AOF2 | |
| As, Apr, Aal spermatogonia | Intermediate | Intermediate | Intermediate | Intermediate |
| In and B spermatogonia | Strong | Intermediate | Intermediate | Strong |
| Preleptotene spermatozytes | Strong | Strong | Strong | Strong |
| Leptotene spermatozytes | Strong | Strong | Intermediate | Strong |
| Zygotene spermatozytes | Weak | Weak | Strong | Weak |
| Pachytene spermatozytes | Weak | Weak | Weak | Weak |
| Round spermatids | Moderate | Moderate | Intermediate (stages I–VII) | Moderate |
| Elongating spermatids | Strong (stages IX–XI) | Strong (stages IX–XII) | Strong (stages IX–XI) | Strong (steps 10–16) |
In, intermediate spermatogonia.
me1, Mono-methylated; me2, di-methylated; me3, tri-methylated.
Spermatogenic cell distribution patterns of H3-K4-me1, H3-K4-me2, H3-K4-me3, and AOF2 in the mouse.
| Cell typesa | Staining intensityb | |||
|---|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | AOF2 | |
| As, Apr, Aal spermatogonia | Intermediate | Intermediate | Intermediate | Intermediate |
| In and B spermatogonia | Strong | Intermediate | Intermediate | Strong |
| Preleptotene spermatozytes | Strong | Strong | Strong | Strong |
| Leptotene spermatozytes | Strong | Strong | Intermediate | Strong |
| Zygotene spermatozytes | Weak | Weak | Strong | Weak |
| Pachytene spermatozytes | Weak | Weak | Weak | Weak |
| Round spermatids | Moderate | Moderate | Intermediate (stages I–VII) | Moderate |
| Elongating spermatids | Strong (stages IX–XI) | Strong (stages IX–XII) | Strong (stages IX–XI) | Strong (steps 10–16) |
| Cell typesa | Staining intensityb | |||
|---|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | AOF2 | |
| As, Apr, Aal spermatogonia | Intermediate | Intermediate | Intermediate | Intermediate |
| In and B spermatogonia | Strong | Intermediate | Intermediate | Strong |
| Preleptotene spermatozytes | Strong | Strong | Strong | Strong |
| Leptotene spermatozytes | Strong | Strong | Intermediate | Strong |
| Zygotene spermatozytes | Weak | Weak | Strong | Weak |
| Pachytene spermatozytes | Weak | Weak | Weak | Weak |
| Round spermatids | Moderate | Moderate | Intermediate (stages I–VII) | Moderate |
| Elongating spermatids | Strong (stages IX–XI) | Strong (stages IX–XII) | Strong (stages IX–XI) | Strong (steps 10–16) |
In, intermediate spermatogonia.
me1, Mono-methylated; me2, di-methylated; me3, tri-methylated.
Histone H3 methylation at lysine 4 (H3-K4) during murine spermatogenesis. Immunohistochemical analysis of adult mouse testis reveals dynamic modulation of histone H3-K4 using antibodies specific for (A) mono-methylated H3-K4 (H3-K4-me1), (B) di-methylated H3-K4 (H3-K4-me2), and (C) trimethylated H3-K4 (H3-K4-me3). Brown staining indicates positive reactivity. Roman numerals indicate the stage of spermatogenesis. B, Spermatogonia type B; ES, elongating spermatids; M, metaphase spermatocyte; L, leptotene spermatocytes; P, pachytene spermatocytes; Pl, preleptotene spermatocytes; RS, round spermatids; SG, spermatogonia; SGA, spermatogonia type A; Z, zygotene spermatocytes. The sections are counterstained with hematoxylin.
Confocal analysis of spermatogenic stage-specific isolated populations of germ cells [34] revealed identical patterns of H3-K4 methylation to those identified in Bouin-fixed tissue sections (Figs. 1 and 2). Using this technique, we were able to visualize the chromatin localization patterns. As predicted for H3-K4 methylation based on its previously identified association with gene transcription [37–39], H3-K4 mono-, di, and trimethylation appeared to be preferentially associated with euchromatic regions (Fig. 2). Confocal analysis of immunofluorescently stained cells revealed preferential colocalization of H3-K4 mono-, di-, and trimethylation to the regions of cells with low levels of DAPI staining, which correspond to regions of less-condensed DNA (Fig. 2). An intriguing observation was the polarization of H3-K4 trimethylation in spermatids (Fig. 2D), which occurred just prior to the onset of elongation and preparation for replacement of the majority of the histones, initially with transition proteins and subsequently with protamines [4]. Confirming our immunohistochemical observations that H3-K4 mono-, di-, and trimethylation localize to spermatogonial cells was the observation that these epigenetic markers were present in ZBTB16 (PLZF)-positive spermatogonia (Fig. 2). ZBTB16 is a specific transcriptional repressor that has been shown to regulate self-renewal and maintenance of the stem cell pool [40].
Confocal laser scanning analyses of the H3-K4 methylation patterns of staged spermatogenic cells. Immunofluorescence using antibodies specific for (A) H3-K4 mono-methylation (H3-K4-me1, red), (B) H3-K4 di-methylation (H3-K4-me2, red), and (C) H3-K4 trimethylation (H3-K4-me3, red) shows preferential colocalization of methylated H3-K4 to regions of the cell nucleus with less condensed DNA, which are indicated by low-level DAPI staining (blue). Zinc finger- and BTB domain-containing 16 (ZBTB16, previously known as PLZF; green) is specifically expressed in type A spermatogonia. D) Polarization of H3-K4-me3 (green) in elongating spermatids. ES, Elongating spermatids; L, leptotene spermatocytes; P, pachytene spermatocytes; SGA, spermatogonia type A. All pictures are original magnification ×63.
Western Blot Analysis of Histone H3-K4 Mono-, Di-, and Trimethylation During the First Wave of Spermatogenesis in Prepubertal, Pubertal, and Adult Testes
As our immunohistochemical results showed distinct increases in the levels of H3-K4 methylation in early meiotic cells at the preleptotene stage, we confirmed these results by Western blotting of testicular extracts obtained from developmentally staged testicular extracts corresponding to the appearance of spermatogenic cell types [28]. The aforementioned approach was chosen because centrifugal elutriation and the STAPUT methods of germ cell separation do not permit the isolation of pure populations of early meiotic cells, such as preleptotene, leptotene, and zygotene cells, from adult mice. There were increases in H3-K4 mono- and di-methylation in testicular extracts obtained from mice at PND 10, corresponding to the appearance of preleptotene spermatocytes (Fig. 3, A and B), followed by decreases at PND 14 and the appearance of pachytene spermatocytes. Confirming our immunohistochemical observations, H3-K4 trimethylation increased at PND 12 to PND 14, at the time of zygotene to pachytene development, and then decreased at PND 20, at which point late pachytene spermatocytes appear (Fig. 3C). These patterns of H3-K4 methylation during the first wave of spermatogenesis were confirmed by immunohistochemical analysis of testicular cross-sections that were collected according to the appearance of spermatogenic cell types in prepubertal mice (Supplemental Fig. 1).
Western blot analyses of the methylation status of H3-K4 in total testis protein extracts obtained from prepubertal mice. Testes were collected during the first wave of spermatogenesis at postnatal days (PND) corresponding to the appearance of the spermatogenic cell types (PND 6, spermatogonia type A; PND 8, spermatogonia types A and B; PND 10, preleptotene and leptotene spermatocytes; PND 12, zygotene spermatocytes; PND 14, early pachytene spermatocytes; PND 20, late pachytene spermatocytes; ad, adult), and protein extracts were analyzed for (A) mono-methylated H3-K4 (H3-K4-me1), (B) di-methylated H3-K4 (H3-K4-me2), and (C) trimethylated H3-K4 (H3-K4-me3) in the developing mouse testis. The Western blots were reprobed with an anti-β-actin antibody as a loading control. Data are presented as the ratios of H3-K4-me to β-actin. Mr values are ×10−3.
Dynamic Modulation of H3-K4 Mono-, Di-, and Trimethylation During Marmoset (Callithrix jacchus) Spermatogenesis
Immunohistochemistry was used to compare the distributions of histone H3 mono-, di-, and trimethylation between mice and the well-established primate reproductive model of the common marmoset. H3-K4 mono-, di-, and trimethylation were detected in spermatogonia, as well as in preleptotene, leptotene, and zygotene spermatocytes (Fig. 4). As in the mouse, pachytene spermatocytes at all stages of spermatogenesis were negative for H3-K4 methylation (Fig. 4). Round spermatids that were transitioning to elongated spermatids were positive for di- and trimethylation at H3-K4, while mono-methylation was not detected in the round spermatids (Fig. 4). Similarly, elongated spermatids were positive for H3-K4 di- and trimethylation at stage VII (Fig. 4). These distribution patterns are summarized in Table 2.
Spermatogenic cell distribution patterns of H3-K4-me1, H3-K4-me2, and H3-K4-me3 in the common marmoset (Callithrix jacchus).
| Cell types | Staining intensitya | ||
|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | |
| Spermatogonia | Strong | Strong | Strong |
| Preleptotene spermatozytes | Strong | Strong | Moderate |
| Leptotene spermatozytes | Strong | Strong | Moderate |
| Zygotene spermatozytes | Strong | Intermediate | Strong |
| Pachytene spermatozytes | Absent | Absent | Absent |
| Round spermatids | Absent | Intermediate (step 5,6) | Strong (step 4–6) |
| Elongating spermatids | Absent | Intermediate (step 7) | Strong (step 7) |
| Cell types | Staining intensitya | ||
|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | |
| Spermatogonia | Strong | Strong | Strong |
| Preleptotene spermatozytes | Strong | Strong | Moderate |
| Leptotene spermatozytes | Strong | Strong | Moderate |
| Zygotene spermatozytes | Strong | Intermediate | Strong |
| Pachytene spermatozytes | Absent | Absent | Absent |
| Round spermatids | Absent | Intermediate (step 5,6) | Strong (step 4–6) |
| Elongating spermatids | Absent | Intermediate (step 7) | Strong (step 7) |
me1, Mono-methylated; me2, di-methylated; me3, tri-methylated.
Spermatogenic cell distribution patterns of H3-K4-me1, H3-K4-me2, and H3-K4-me3 in the common marmoset (Callithrix jacchus).
| Cell types | Staining intensitya | ||
|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | |
| Spermatogonia | Strong | Strong | Strong |
| Preleptotene spermatozytes | Strong | Strong | Moderate |
| Leptotene spermatozytes | Strong | Strong | Moderate |
| Zygotene spermatozytes | Strong | Intermediate | Strong |
| Pachytene spermatozytes | Absent | Absent | Absent |
| Round spermatids | Absent | Intermediate (step 5,6) | Strong (step 4–6) |
| Elongating spermatids | Absent | Intermediate (step 7) | Strong (step 7) |
| Cell types | Staining intensitya | ||
|---|---|---|---|
| H3-K4-me1 | H3-K4-me2 | H3-K4-me3 | |
| Spermatogonia | Strong | Strong | Strong |
| Preleptotene spermatozytes | Strong | Strong | Moderate |
| Leptotene spermatozytes | Strong | Strong | Moderate |
| Zygotene spermatozytes | Strong | Intermediate | Strong |
| Pachytene spermatozytes | Absent | Absent | Absent |
| Round spermatids | Absent | Intermediate (step 5,6) | Strong (step 4–6) |
| Elongating spermatids | Absent | Intermediate (step 7) | Strong (step 7) |
me1, Mono-methylated; me2, di-methylated; me3, tri-methylated.
Alterations of histone H3-K4 during spermatogenesis in Callithrix jacchus. Immunohistochemistry of adult Callithrix jacchus testis sections shows dynamic modulation of histone H3-K4 using antibodies specific for mono-methylated H3-K4 (H3-K4-me1), di-methylated H3-K4 (H3-K4-me2) and trimethylated H3-K4 (H3-K4-me3). Brown staining indicates positive reactivity. Roman numerals indicate the stage of the spermatogenic cycle. ES, Elongating spermatids; L, leptotene spermatocytes; P, pachytene spermatocytes; Pl, preleptotene spermatocytes; RS, round spermatids; SG, spermatogonia; SGA, spermatogonia type A; Z, zygotene spermatocytes. The sections are counterstained with hematoxylin.
AOF2 Is Preferentially Distributed in the Testis in a Stage-Specific Manner That Directly Corresponds to the H3-K4 Mono- and Di-Methylation Patterns
Histone H3-K4 mono- and di-methylation can be reversed in vitro by the demethylase activity mediated by AOF2 [24]. Strikingly, using Western blot analysis, we detected significantly more AOF2 in the testes than in the somatic tissues (P ≤ 0.05) (Fig. 5, A and B), and its distribution was tightly regulated during spermatogenesis (Figs. 5C and 6, and Supplemental Figs. 2 and 3, available at www.biolreprod.org). Highly pure populations of spermatogenic cells isolated by elutriation and analyzed by Western blotting revealed that the AOF2 levels were highest in spermatocytes, in comparison to the respective levels in spermatogonia and spermatids (P ≤ 0.05) (Fig. 7H). These levels detected by Western blot analysis in purified cell populations directly matched the levels obtained in our immunohistochemical analysis of both adult tissues and prepubertal testes (Fig. 6 and Supplemental Fig. 2). Importantly, immunolocalization revealed that, similar to H3-K4 mono- and di-methylation, AOF2 was strongly expressed in spermatogonia, increased in preleptotene spermatocytes, and then dramatically declined in pachytene spermatocytes. Again, as for H3-K4 mono- and di-methylation, AOF2 transiently reappeared in haploid round spermatids at the onset of elongation and was restricted to stage IX and X spermatids (Fig. 6 and Table 1). Immunofluorescence analysis of staged spermatogenic cells revealed that AOF2 was found in a punctuate distribution in euchromatic regions (Supplemental Fig. 3). The overlapping distribution patterns of AOF2 and H3-K4 mono and di-methylation suggest tight gene-specific regulation governed by H3-K4 methylation and AOF2 interaction at methylated domains in H3-K4.
Protein levels of amine oxidase (flavin-containing) domain 2 (AOF2, previously known as LSD1) in mouse tissues. A) AOF2 protein levels are high in the mouse testis in comparison to somatic tissues. Western blot analyses of AOF2 protein levels in total protein extracts obtained from different mouse tissues. T, Testis; E, epididymis; U, uterus; St, stomach; C, colon; I, small intestine; B, brain; L, lung; Li, liver; H, heart; S, spleen; K, kidney. The Western blots were reprobed with an anti-β-actin antibody as a loading control. Each experiment was repeated three times. Mr values are ×10−3. B) Data in the graph indicate the ratios of AOF2 to β-actin. The stacked column represents the contribution of each experiment. C) Western blot analysis of AOF2 levels in total testis protein extracts obtained from mice at different postnatal developmental stages. Testes were collected during the first wave of spermatogenesis at postnatal days (PND) corresponding to the appearance of the spermatogenic cell types (PND 6, spermatogonia type A; PND 8, spermatogonia types A and B; PND 10, preleptotene and leptotene spermatocytes; PND 12, zygotene spermatocytes; PND 14, early pachytene spermatocytes; PND 20, late pachytene spermatocytes; ad, adult) and protein extracts were analyzed for the levels of AOF2 in the developing mouse testis. The Western blots were reprobed with an anti-β-actin antibody as a loading control. Each experiment was repeated three times. The graphs show the ratios of AOF2 to β-actin. Stacking lines indicate the trend of changes in testicular AOF2 protein expression during the first wave of spermatogenesis. Mr values are ×10−3. *P < 0.05.
Immunolocalization of amine oxidase (flavin-containing) domain 2 (AOF2, previously known as LSD1) in the mouse testis. Immunohistochemical analysis of adult mouse testis sections shows AOF2 expression in a stage-specific pattern. Roman numerals indicate the stage of the spermatogenic cycle. Brown staining indicates positive reactivity. B, Spermatogonia type B; ES, elongating spermatids; In, intermediate spermatogonia; L, leptotene spermatocytes; Pl, preleptotene spermatocytes; RS, round spermatids; SC, Sertoli cell; (SGA), spermatogonia type A; Z, zygotene spermatocytes. The sections are counterstained with hematoxylin.
Interaction studies and protein analyses of known epigenetic modifiers in the mouse testis. A) Western blot analysis of amine oxidase (flavin-containing) domain 2 (AOF2, previously known as LSD1) coimmunoprecipitated with histone deacetylase 1 (HDAC1) and vice versa (C). B) Western blot analysis of AOF2, which coimmunoprecipitates with methyl binding domain proteins 2a and 2b (MBD2a and MBD2b), and in turn, MBD2a (E) and MBD2b (F) coimmunoprecipitate with AOF2. D) Western blot analysis of histone deacetylase 1 (HDAC1) coimmunoprecipitated with MBD2b and vice versa (G). H) Western blot analyses of HDAC1, AOF2, and MBD2a in protein extracts from isolated murine testicular germ cells. The Western blots were reprobed with an antibody against mitogen-activated protein kinase 1 (MAPK1) as a loading control. Each experiment was repeated three times. The data in the graph indicate the ratios of AOF2 or HDAC1 or MBD2a to MAPK1. Mr values are ×10−3. IP, Immunoprecipitation; MIgG, mouse IgG; NRS, normal rabbit serum; RIgG, rabbit IgG; RS, round spermatids; SG, spermatogonia; Spc, spermatocytes.
AOF2 and HDAC1 Complex with Each Other and with Methyl-CpG-Binding Proteins (MBD) in the Mouse Testis
Epigenetic regulation of histone H3 methylated at K4 involves interactions with multiple proteins that act together to target gene-specific regions and reorganize chromatin. In previous studies, HDAC1 has been shown to be functionally associated with AOF2, and they act in concert to enhance the deacetylation of histone H3 and to remove the methyl groups on histone H3 at lysine 4 [22]. In addition, the methyl-CpG-binding proteins (MBD) have been associated with a corepressor complex comprised of HDAC1 [41]. Using Western blot analysis of purified germ cell populations, we determined that AOF2, HDAC1, and MBD2 were present in the same spermatogenic cell types (Fig. 7H). In other cell systems, HDAC1 has been shown to interact with AOF2 and MBD2. Do HDAC1 and MBD2a/b interact with AOF2 in male germ cells? We answered this question by in vivo coimmunoprecipitation experiments to determine which chromatin remodeling complex governs dynamic histone H3 methylation in spermatogenesis. Immunoprecipitated AOF2 was found to coimmunoprecipitate with HDAC1 (Fig. 7A), MBD2a, and MBD2b (Fig. 7B). In turn, HDAC1 interacted with AOF2 (Fig. 7C) and with MBD2b (Fig. 7D). Conversely, MBD2a (Fig. 7E) and MBD2b (Fig. 7F) coimmunoprecipitated with AOF2, and MBD2b coimmunoprecipitated with HDAC1, respectively (Fig. 7G). Immunolocalization of MBD2a or MBD2b in either isolated germ cells or in testicular cross-sections was not possible due to the limited reactivities of the available antibodies for immunocytochemistry. In order to confirm that these proteins have overlapping distributions in germ cells, we proceeded to analyze by Western blotting pure populations of spermatogonia, spermatocytes, and round spermatids. Indeed, the concomitant presence of these proteins in purified populations of spermatogonia, spermatocytes, and round spermatids was detected, confirming their potential to form chromatin-modifying complexes in germ cells (Fig. 7H). It is noteworthy that the protein level of AOF2 was highest in spermatocytes, while the HDAC1 level was highest in round spermatids (Fig. 7H). Protein loading was normalized to the levels of MAPK1 (Fig. 7H).
Discussion
We have shown that mono-, di-, and trimethylation of histone H3 at the fourth lysine is tightly regulated during male germ cell development. In the mouse, there is striking up-regulation of H3-K4 methylation in preleptotene spermatocytes, which is followed by down-regulation in pachytene spermatocytes. H3-K4 methylation reappeared transiently in the haploid spermatids. Transcriptional activity has been correlated with specific histone modifications of N-terminal tails [42, 43]. Genome-wide data correlate the H3-K4 methylation state, gene locations, and corresponding gene expression levels. For example, trimethylation of histone H3-K4 at the 5'-end of the transcribed region is a hallmark of gene expression [38, 44, 45], as is H3-K4 di-methylation [39, 46]. Mono-methylation of H3-K4 differs from di- and trimethylation, as in yeast it has been localized to the 3'-region and is associated with genes that are poised for expression [47]. In other eukaryotes, H3-K4 methylation has been linked to transcriptional activation, so it seems plausible that it has a similar role in transcriptional control in germ cells. Microarray studies based on purified germ cell populations have shown increased gene expression in early meiotic cells and in postmeiotic cells [48]. These waves of gene expression follow the H3-K4 methylation patterns described herein.
In addition to unique transcriptional regulation, spermatogenesis also involves dramatic chromatin alterations during meiosis and through the postmeiotic histone to protamine exchange. Meiosis requires specialized chromatin remodeling to facilitate transcriptional regulation, homologous chromosome pairing, alignment, and genetic recombination. Little is known about the epigenetic mechanisms that regulate these steps, and our data highlight the possibility that H3-K4 methylation is involved in the massive chromatin reorganization required for successful completion of meiosis. The strong distribution of H3-K4 methylation in preleptotene spermatocytes coincides with condensation of the chromatin in preparation for chromosome synapsis, which occurs in zygonema.
During postmeiotic male germ cell development, preceding histone replacement, there is massive chromatin remodeling, which is underpinned by alterations to the epigenetic layer. Global histone H4 acetylation coincides with the initial steps of chromatin restructuring in elongating spermatids [49], and this process is conserved across mammals, fish, and birds as a prerequisite to histone displacement [50, 51]. An interesting temporal feature of H3-K4 methylation is that it is strongly detected at the onset of spermatid elongation at stage IX. Importantly, this methylation at H3-K4 is retained during the massive chromatin remodeling process that begins at stage X with the replacement of 90% of the histones, initially with transition proteins and subsequently with protamines [4, 52, 53]. At the later stages of spermatid maturation (from steps 13 to 16, corresponding to the deposition of protamines), H3-K4 methylation is not detected. The timing of H3-K4 methylation in elongating spermatids hints at a possible involvement in the chromatin remodeling process at the histone-to-transition protein exchange.
Genomic methylation patterns are established in the germline and accurate setting of methylation markers is required for successful passage through meiosis [54]. Importantly, H3-K4 di-methylation has recently been implicated in the differential marking of imprinting control regions during mouse spermatogenesis, being associated with the unmethylated allele [55]. Underscoring the requirement for histone H3-K4 trimethylation in spermatogenesis is the observation that mice in which the Prdm9 gene (previously known as Meisetz, which encodes the histone-lysine methyltransferase with catalytic activity for H3-K4 trimethylation) is disrupted are completely sterile [56]. These mice display reduced H3-K4 trimethylation and spermatogenic arrest at the pachytene stage. Extended microarray analysis of cDNA isolated from PND-14 Prdm9 null mice revealed reduced transcription of 109 genes in comparison to wild-type controls [37]. Based on our in-depth stage- and cell-specific analyses of histone H3 trimethylation, its effects do not appear to be limited to meiotic events, as this marker is strongly evident in round spermatids.
The common marmoset, Callithrix jacchus, has been used intensively for many years as a nonhuman primate model to describe the physiology of testicular development and function [33, 57–60]. Importantly, the marmoset shows striking similarities to the organization of the human seminiferous epithelium [61]. We wondered if the dynamic histone H3 methylation patterns at lysine 4 represented a global epigenetic mechanism that is conserved across species. The temporal- and cell-specific patterns we observed by immunohistochemical comparisons of the mouse and marmoset were highly similar. In both the marmoset and mouse, at the onset of the meiotic prophase, preleptotene and leptotene spermatocytes show strong evidence of mono-, di-, and trimethylated H3-K4. Interestingly, in zygotene spermatocytes, in which alignment of homologous chromosomes is initiated and the formation of the synaptonemal complex commences, strong H3-K4 methylation persists in the marmoset testis but is lost in the mouse. This difference between species may represent chromatin reorganization events that persist slightly longer in marmoset meiotic cells. For both species, H3-K4 methylation is lost in pachytene spermatocytes, coinciding with the transcriptional silencing that accompanies chromatin condensation mediated in part by aurora kinase B phosphorylation of histone H3 at serine 10 [62, 63], full chromosome synapsis, and genetic recombination. As in the mouse, in the marmoset there is a transient reappearance of H3-K4 concomitant with the massive chromatin remodeling associated with spermatid morphological transformation and transition protein-to-protamine deposition. As further confirmation of the involvement of global and conserved epigenetic mechanisms in spermatogenesis, Rathke and coauthors [64] have reported dynamic levels of H3-K4 methylation during spermatogenesis in Drosophila melanogaster, and similar to our present findings, they observed high levels of H3-K4 methylation prior to histone removal.
Histone-lysine methylation is a reversible epigenetic modification with important functions in both gene repression and gene activation [8, 27, 65]. How lysine methylation affects chromatin organization and function is dependent upon several factors, including the cellular and gene contexts, the site of lysine methylation, and importantly, the recruitment of proteins that recognize and bind the methylated lysine. We report that AOF2 is preferentially expressed in the mouse testis in comparison to somatic tissues, and that its distribution closely matches those of H3-K4 mono- and di-methylation. The coexistence of AOF2 and H3-K4 methylation in the same cells suggest that H3-K4 methylation is regulated by AOF2 during germ cell development. The preferential expression of AOF2 in the testis may reflect the massive alterations in the epigenetic layer that are marked by erasure of epigenetic markers and their resetting [6, 7, 37, 66]. Notably, regions of DNA methylation have been shown to exclude H3-K4 di- and trimethylation as a means of protecting DNA-mediated gene silencing [67].
The translational output and the downstream effects of a particular histone modification on chromatin status are determined by the recruitment of opposing enzyme pairs, such as histone methyltransferases and histone demethylases. AOF2 has been shown to act both as a corepressor, when associated with the RCOR1 complex and HDAC1 [22, 24, 26], and as a coactivator in association with androgen receptor target genes and demethylation of H3-K9 [27]. The simultaneous detection of HDAC1, AOF2, and MBD2 in male germ cells prompted us to investigate whether these proteins could complex to form a potential chromatin-reorganizing complex in the testis. Coimmunoprecipitation experiments revealed strong associations between these proteins, and suggest that this complex is important in chromatin organization by promoting the removal of H3-K4 methylation. Methyl-CpG-binding proteins provide a link between DNA methylation and chromatin remodeling and several mechanisms have been proposed to couple epigenetic modifications of histones and DNA to the silencing of gene expression [68, 69]. Based on the association reported herein between AOF2 and HDAC1, we postulate that these proteins form a chromatin regulatory complex. Zhang et al. [70] have reported that MBD2 recruits a transcriptional repressor complex, termed NuRD, to methylated DNA to promote gene silencing. Similarly, in the mouse testis, MBD2a,b may tether these interaction partners (AOF2 and HDAC1) to methylated DNA. Once in place, the complex may serve to recruit MBD proteins to invoke a heterochromatic state similar to that reported by Kaji et al. [71]. In embryonic stem cells, MBD3a or MBD3b gene targeting has been proposed to recruit the NuRD transcriptional repressor complex to invoke gene silencing associated with differentiation of embryonic stem cells [71]. A similar mechanism may operate in male germ cells to remove H3-K4 methylation and to promote the chromatin reorganization required for meiotic progression. The association we observed between HDAC1 and AOF2 and the timing of disappearance of H3-K4 methylation in elongating spermatids imply that AOF2 and HDAC1 participate in a co-ordinated effort to alter epigenetic marks prior to the histone-to-protamine exchange and the incorporation of histone variants [4]. As previously mentioned, in elongating spermatids there is global hyperacetylation of histone H4 [50, 51], which coincides with the methylation of H3-K4.
Elucidating the molecular pathways that control male germ cell chromatin organization is important in understanding testicular pathologies, such as cancer and infertility. The dynamic H3-K4 methylation patterns we have revealed as occurring during mouse and marmoset spermatogenesis implicate these epigenetic modifications in the highly specialized chromatin reorganization that governs spermatogenesis. Further methylation of H3-K4 during spermatogenesis seems to be a global mechanism that is important for the proper differentiation of germ cells of different species. The sensitivity of methyl markers to environmental disruption through various factors, such as toxicants [10], and drugs [72, 73], places upon these epigenetic markers a pivotal role for understanding the effects of the environment on the epigenome of developing germ cells and for future studies on transgenerational inheritance [10, 74, 75].
Acknowledgments
We are grateful to Bernard Robaire, Jacquetta Trasler, Dan Bernard, and Barbara Hales for helpful comments and discussions.
References
Author notes
Supported by grants from NSERC (to S.K.), the Dr. L.J. Johnson Foundation (to S.K.), and the Lance Armstrong Foundation (to C.S.). M.G. is supported by a fellowship from the McGill Faculty of Agriculture and Environmental Sciences, McGill University, Montreal, Canada.
