| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Embryo; |
Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Chiba 277-8562, Japan
ABSTRACT
The gene expression pattern of differentiated oocytes is reprogrammed into that of totipotent preimplantation embryos before and/or after fertilization. To elucidate the mechanisms of genome reprogramming, we investigated histone H3 lysine 79 dimethylation (H3K79me2) and trimethylation (H3K79me3) in oocytes and preimplantation embryos via immunocytochemistry. In somatic cells and oocytes, H3K79me2 was observed throughout the genome, whereas H3K79me3 was localized in the pericentromeric heterochromatin regions in which there are no active genes. Because H3K79me2 is considered an active gene marker, H3K79 methylation seems to have differing functions depending on the number of methyl groups added on the same residues. Both H3K79me2 and H3K79me3 decreased soon after fertilization, and the hypomethylated state was maintained at interphase (before the blastocyst stage), except for a transient increase in H3K79me2 at mitosis (M phase). H3K79me3 was not detected throughout preimplantation, even at M phase. To investigate the involvement of H3K79me2 in genome reprogramming, somatic nuclei were transplanted into enucleated oocytes. H3K79me2 in these nuclei was demethylated following parthenogenetic activation. However, the nuclei that had been transplanted into the parthenogenetic embryos 7 h after activation were not demethylated. This suggests that the elimination of H3K79 methylation after fertilization is involved in genomic reprogramming.
early development, embryo, fertilization
Before and/or after fertilization, differentiated oocytes are transformed into totipotent preimplantation embryos. During this process, gene expression patterns are reprogrammed. Growing oocytes actively express genes, including oocyte-specific genes. However, oocytes cease gene expression before they are fully grown and remain in a transcriptionally silent state during meiotic maturation [1–3]. After fertilization, the zygote begins to express genes in an embryo-specific pattern; this pattern is dynamically altered during preimplantation development [4–7]. Although the reprogramming of gene expression is an important event in the creation of new life, the mechanisms regulating this event have not been elucidated.
Histone methylation plays an important role in regulating chromatin structure and gene expression. Several lysine residues in histones can be methylated, and each can be monomethylated, dimethylated, or trimethylated [8–11]. Unlike acetylation, which is generally associated with transcriptional activation, methylation is involved in both activation and repression, depending on which lysine residues are methylated [12, 13]. Furthermore, during transcriptional regulation, the number of methyl groups added to a lysine residue is also important [13]. For instance, histone H3 lysine 4 dimethylation (H3K4me2) and trimethylation (H3K4me3), which are known active markers of gene expression, are localized on euchromatin, a transcriptionally active chromosomal domain [14–18]. However, these different methylation states do not overlap completely. Whereas H3K4me2 is distributed across the whole body of active genes, H3K4me3 is localized specifically at the 5' end of these genes [10, 13, 19, 20]. Histone H3 lysine 9 dimethylation (H3K9me2) and histone H3 lysine 9 trimethylation (H3K9me3), which are known repressive markers, are localized on heterochromatin, a transcriptionally repressive chromosomal domain [15, 16, 21]. Although H3K9me2 has been detected on facultative heterochromatin (a changeable condensed chromatic domain) [13, 22], H3K9me3 has been observed on constitutive heterochromatin, which is composed of highly condensed pericentromeric heterochromatin [22–24]. In addition, heterochromatin protein 1 (HP1 or CBX5), which associates with methylated H3K9 to form heterochromatin, binds to H3K9me3 with twice the affinity that it binds to H3K9me2 [25]. Thus, in addition to the identity of methylated lysine residues, the number of methyl groups added to the lysine residue is also important in the regulation of chromatin structure and gene expression [10, 25, 26].
Of several methylated lysines, H3K79 methylation has many distinct characteristics. First, although most other methylated lysine residues are located near the N terminus, H3K79 is located in the globular domain [27]. Although the globular domain occurs at the center of the Histone H3 polypeptide chain, it is localized on the surface of the nucleosomal structure so that it can be accessed by transcriptional factors [28]. In addition, H3K79 methylation is catalyzed by the disrupter-of-telomere-silencing-like protein (DOT1L), which was originally identified as a gene whose mutation caused the disruption of telomere silencing in Saccharomyces cervisiae [27, 29]. DOT1L does not contain a set domain, which is a highly conserved characteristic motif in all other histone methyltransferases [30, 31]. Therefore, H3K79 methylation may be regulated by a pathway different from that of other histones.
H3K79 methylation may play a role as a marker of euchromatin [15, 16, 18, 32, 33]. In S. cervisiae, Sir2, which is involved in heterochromatin formation, is not recruited to the region of the chromosome containing methylated H3K79 [15, 27]. Furthermore, H3K79me2 does not occur in particular regions (e.g., telomere and mating-type loci) in which Sir2 association with chromatin leads to heterochromatin formation [15, 28, 33]. H3K79me2 is also suggested to be an active gene marker in mammalian cells [13, 16, 32]. It occurs on the promoter and 5' regions within the coding regions of transcriptionally active genes [34]. Following the first round of transcription, H3K79 is dimethylated. It is at this point that it is thought to play a role as an active gene marker [13, 15, 16, 18, 32, 33]. Similar to H3K79me2, H3K79me3 has also been suggested as an active gene marker in S. cervisiae [19]. However, virtually nothing is known concerning the function of H3K79me3 in mammalian cells.
We investigated H3K79 methylation states in oocytes and preimplantation embryos to elucidate the mechanisms that regulate the reprogramming of gene expression patterns during oogenesis and preimplantation development. We found that H3K79me2 and H3K79me3 occur in different regions of the chromosomes in oocytes; H3K79me2 was detected in the whole genome, whereas H3K79me3 was restricted to the pericentromeric region. This suggests that these two modifications have different functions; H3K79me2 may be an active marker of gene expression, whereas H3K79me3 may be a repressive marker. Interestingly, both of these modifications disappeared following fertilization, the point at which the embryo acquires totipotency. These results suggest that alterations in the H3K79 methylation state are involved in the reprogramming of gene expression in embryos.
Collection and Culture of Oocytes and Embryos
Oocytes at the germinal vesicle (GV) stage were collected from the ovaries of 3-wk-old BDF1 mice that had been injected with 5 IU eCG (Sankyo Co. Ltd., Tokyo, Japan). These cells then were incubated in K2+-modified simplex optimized medium (KSOM; [35]) containing 0.2 mM 3-isobutyl-1-methylxanthine (IBMX; Sigma Chemical Co., St. Louis, MO), as described previously [36]. The cells were subsequently cultured without IBMX to allow meiotic maturation in a humidified 5% CO2/95% air atmosphere at 38°C.
Unfertilized MII-stage oocytes were collected from the oviducts of 3-wk-old B6CDF1 mice (CLEA Japan Inc., Tokyo, Japan) that had been superovulated by the injection of 5 IU eCG followed 48 h later by 5 IU hCG (Sankyo) and were placed in KSOM. Spermatozoa were collected from the cauda epididymis of mature male ICR mice (SLC, Shizuoka, Japan) and placed in human tubal fluid medium [37] supplemented with 10 mg/ml BSA. Oocytes were inseminated with spermatozoa that had been incubated for 2 h at 38°C. The embryos were washed with KSOM containing 3 mg/ml BSA (KSOM/BSA) 5 h after insemination and then cultured in a humidified 5% CO2/95% air atmosphere at 38°C.
All of the procedures involving animals were reviewed and approved by the University of Tokyo Institutional Animal Care and Use Committee and were performed in accordance with the Guiding Principles for the Care and Use of Laboratory Animals.
At 6 h after insemination, the embryos were transferred into KSOM/BSA containing 3 µg/ml aphidicolin.
Oocytes and embryos were washed in PBS containing 1 mg/ml BSA (PBS/BSA), fixed for several hours or overnight in 3.7% paraformaldehyde in PBS at 3°C, and permeabilized with 0.5% Triton X-100 in PBS for 15 min at room temperature. The cells then were incubated overnight in a 1:100 dilution of primary antibodies against H3K79me2 (catalogue number ab3594; Abcam, Lake Placid, NY), H3K79me3 (catalogue number 2621; Abcam), centromere proteins (catalogue number 15–235; CREST; Antibodies Inc., Davis, CA), or HP1β (catalogue number ab11164; Abcam). Primary antibodies that were bound to the cells were probed by incubation for 45 min with a 1:50 dilution of the secondary antibodies. Secondary antibodies were purchased from Jackson Immunoresearch (West Grove, PA) as follows: fluorescein isothiocyanate (FITC)-conjugated anti-rabbit immunoglobulin G (IgG; catalogue number 111-095-003), Cy5-conjugated anti-rabbit IgG (catalogue number 111-175-003), FITC-conjugated anti-goat IgG (catalogue number 705-095-147), Cy5-conjugated anti-goat IgG (catalogue number 705-175-147), Cy5-conjugated anti-antibody human IgG (catalogue number 109-175-003), TRITC-conjugated anti-human IgG (catalogue number 109-025-003), and TRITC-conjugated anti-human IgG (catalogue number 705-025-149). DNA was stained with 100 µg/ml propidium iodide (PI; Sigma-Aldrich, Inc., St. Louis, MO) under RNase conditions for 20 min, and cells then were mounted on a glass slide in Vectashield antibleaching solution (Vector Laboratories, Burlingame, CA). In some experiments, cells treated with RNase were mounted on a glass slide in Vectashield antibleaching solution containing 100 µg/ml PI. Fluorescence was detected using a Carl Zeiss 510 laser scanning confocal microscope.
For triple staining of the chromosomes of MII-stage oocytes with anti-H3K79me3 antibody, CREST, and PI, the samples were prepared according to the procedures described by Hodges and Hunt [38].
MII-stage oocytes were collected from oviducts and placed in KSOM containing 3 mg/ml bovine testicular hyaluronidase (Sigma Chemical). After they were cultured for 1 h in KSOM, the oocytes were treated with 10 mM Sr2+ for 4 h in Ca2+-free KSOM. To obtain diploid parthenogenetic embryos, cytochalasin B (1% dimethylsulfoxide) was added to the KSOM/Sr2+.
Nuclear Transfer into Unfertilized Oocytes
The enucleation of oocytes was performed as described previously [39]. The nuclei of NIH3T3 cells were introduced into enucleated oocytes by electrofusion using a DC pulse of 1500 V/cm for 20 µsec in 300 mM mannitol containing 0.1 mM MgSO4, 0.1 mg/ml polyvinyl alcohol, and 3 mg BSA. As the control, the nucleus was embedded in the perivitelline space of the oocyte but not subjected to electrofusion.
Nuclear Transfer into Parthenogenetically Activated Embryos
Parthenogenetically activated embryos were prepared as described above. At 7 h after activation, they were enucleated and transplanted with the nuclei of NIH3T3 cells by electrofusion, using a DC pulse of 1500 V/cm for 20 µsec in 300 mM mannitol that contained 0.1 mM MgSO4, 0.1 mg/ml polyvinyl alcohol, and 3 mg BSA. As the control, the nucleus was embedded in the perivitelline space of the embryo but not subjected to electrofusion.
Subnuclear Localization of H3K79 Methylation in Somatic Cells
Because the localization of H3K79me3 has not been reported in any type of mammalian cell, we first examined the subnuclear localization of H3K79me2 and H3K79me3 in mouse somatic cells; that is, NIH3T3 cells and fibroblasts from the tail tip, at interphase and mitosis (M phase). Immunocytochemistry with specific antibodies against H3K79me2 and H3K79me3 revealed that the pattern of localization was similar between these two types of cells. Intense fluorescence signals of H3K79me2 were observed over the entire genome in these cells at interphase (Fig. 1, A and B; images a and g) and mitosis (M phase; Fig. 1, A and B; images b, c, h, and i). However, antibodies against H3K79me3 were detected as several foci in the nucleus at interphase. These foci seemed to be constitutive heterochromatin (e.g., pericentromeres or telomeres) because they were completely superimposed with highly condensed regions of DNA in PI-stained images (Fig. 1, A and B; images d and j). At metaphase, H3K79me3 was observed in many small foci that were localized on the edges of chromosomes (Fig. 1, A and B; images e and k). These foci were gathered on the outside region of chromosomal bundles during the time when the two sets of chromosomes were separated at anaphase (Fig. 1A; images f and l). This suggests that H3K79me3 is localized on or near the centromere. Interestingly, the H3K79me3 signal was detected in only one third of NIH3T3 cells and fibroblasts at interphase, although it was detected in all cells at M phase. To clarify the localization of H3K79me3, NIH3T3 cells were double stained with anticentromere protein antibody (i.e., CREST) and anti-H3K79me3. The results showed that the signals of H3K79me3 and CREST were nonoverlapping but were always localized adjacently (Fig. 1C). This suggests that H3K79me3 is localized on pericentromeric heterochromatin. Although it has been suggested that H3K79me2 is a marker of euchromatin [15, 16], these results indicate that H3K79me3 is a marker of heterochromatin. Thus, lysine 79 methylation appears to have differing functions depending upon the number of methyl groups.
|
Dynamics of H3K79 Methylation During Meiotic Maturation and Preimplantation Development
Oocytes and preimplantation embryos showed essentially the same nuclear localization of H3K79 methylation as the somatic cells. Their methylation levels, however, changed dynamically during preimplantation development. Although H3K79me2 was always detected in the whole genome of oocytes at the GV, GV breakdown, and MII stages, it was no longer detected after fertilization (Fig. 2). After fertilization, H3K79me2 was either not detected or detected at marginal levels during interphase before the four-cell stage. In one-cell-stage embryos, no signal was detected in the male pronucleus, although a faint signal was detected in the female pronucleus. The levels increased slightly but still remained low at the four-cell and morula stages. However, methylation increased prominently at the blastocyst stage. At M phase, H3K79me2 was always detected during preimplantation development. Thus, H3K79me2 was not detected or was detected at low levels at interphase but appeared transiently at the M phase before the blastocyst stage. At the blastocyst stage, H3K79me2 was detected at both interphase and M phase. Similarly to H3K79me2, H3K79me3 was detected in the oocytes and was no longer detected after fertilization (Fig. 3). It remained undetectable, however, through preimplantation development, even at the M phase.
|
|
In MII-stage oocytes, the localization of H3K79me3 on the pericentromeric region was confirmed by double staining with CREST and antibodies against H3K79me3 (Fig. 4A). Furthermore, because it has been reported that HP1β (CBX1) is localized in the pericentromeric region of GV-stage oocytes [40], the oocytes were double stained with antibodies against H3K79me3 and HP1β to confirm the pericentromeric localization of H3K79me3 in GV-stage oocytes. The results showed that H3K79me3 and HP1β were colocalized on foci with dense DNA (Fig. 4B).
|
The specificity of the antibodies against H3K79me2 and H3K79me3 was examined in adsorption experiments using their respective antigen peptides (Supplemental Figure S1 and S2 available at www.biolreprod.org). When each antibody was preincubated with its corresponding antigen peptides, the signals were no longer detected in MII-stage oocytes or NIH3T3 cells. In addition, adsorption experiments were conducted using swapped peptides between H3K79me2 and H3K79me3. Although the signal of the anti-H3K79me3 antibody did not decrease with preincubation with the H3K79me2 peptide, the signal of the anti-H3K79me2 antibody decreased with the H3K79me3 peptide. These results indicate that although the anti-H3K79me3 antibody does not recognize H3K79me2, the anti-H3K79me2 antibody cross-reacts with H3K79me3. However, because H3K79me3 was detected by its specific antibody only in the pericentromeric region, the signal of the anti-H3K79me2 antibody, which was observed in the whole genome, would show H3K79me2 localization.
In somatic cells, H3K79me2 is regulated in a cell cycle-dependent manner in which H3K79me2 levels decrease during S phase, reach the lowest point at G2 phase, increase during M phase, and are maintained at high levels during gap1 (G1 phase) [31]. To examine whether decreases in H3K79 methylation also depend on DNA synthesis following fertilization, methylation levels were determined 4, 8, and 12 h after fertilization. These times correspond to the gap1 (G1), DNA synthesis (S), and gap2 (G2) phases, respectively [41]. H3K79me2 deceased to a marginal level and H3K79me3 was completely lost as early as 4 h after fertilization, at which point DNA synthesis had not yet occurred (Fig. 5A). Because H3K79me2 remained at a marginal level at 4 h and was completely lost at 12 h, the independence of H3K79me2 demethylation from DNA synthesis was confirmed in an experiment in which DNA synthesis was inhibited in one-cell embryos by treatment with aphidicolin. H3K79me2 demethylation occurred in these embryos (Fig. 5B). Because the signals for H3K79me2 and H3K79me3 had decreased to marginal levels at 4 h after fertilization, the kinetics of this decrease were examined very soon after fertilization. The intensity of the signal for H3K79me2 decreased very little in maternal chromosomes at 1 h after insemination (Fig. 6A). The signal decreased markedly at 2 h, although it was still clearly detectable. We then observed a further decrease to marginal levels at 4 h, at which point female pronuclei had formed. The signal for H3K79me3 also was detected at 1 h after insemination, but this signal decreased markedly or disappeared altogether at 2 h (Fig. 6B). Thus, the demethylation of maternal chromosomes began before pronulei formation. In contrast, neither H3K79me2 nor H3K79me3 was detected in paternal chromosomes at any time after insemination (Fig. 6). Thus, demethylation of H3K79 appears to occur soon after fertilization and independently of DNA synthesis. To address this hypothesis, unfertilized oocytes were activated parthenogenetically, which mimics signals activated during fertilization. In these oocytes, both H3K79me2 and H3K79me3 disappeared (Fig. 7), suggesting that the fertilization triggers the demethylation of H3K79.
|
|
|
Taken together, H3K79 demethylation after fertilization possesses several distinct characteristics that have not been previously observed in somatic cells. First, demethylation occurs independently of DNA synthesis. In addition, H3K79 is almost completely demethylated following fertilization. Although H3K79me2 is demethylated during the S phase in somatic cells, an appreciable level of H3K79me2 still remains at the G2 phase, at which point H3K79 levels are the lowest [31].
Involvement of H3K79 Demethylation in Genome Reprogramming in Transplanted Somatic Nuclei
The observation that rapid demethylation of H3K79 occurs following fertilization led us to formulate the hypothesis that H3K79 demethylation is involved in genome reprogramming. It has been suggested that the program of gene expression in the differentiated oocytes is transformed into that of the totipotent embryo before and/or after fertilization [42]. During this process, epigenetic modifications marking active and/or inactive genes would be deleted in the oocytes/embryos [43]. It is suggested that H3K79me2 is a marker of active genes in somatic cells. In addition, H3K79me2 marking on parental nucleosomes is retained during M phase and inherited by daughter cells following cell division [34]. Therefore, it is possible that H3K79 demethylation is required for genome reprogramming. To address this hypothesis, H3K79me2 was examined in somatic nuclei transplanted into enucleated unfertilized oocytes. A number of recent reports have demonstrated that somatic nuclei are successfully reprogrammed after transplantation into unfertilized oocytes [42, 44, 45]. These results suggest that epigenetic modifications marking active genes are lost in these nuclei. Our results show that H3K79me2 disappears in the transplanted nucleus from a NIH3T3 cell at 5 h after parthenogenetic activation (Fig. 8A and Table 1). However, it remained in the control nucleus, which was embedded between the plasma membrane and the zona pellucida of the oocyte. This area is not exposed to the cytoplasm of the oocyte. These results suggest that the cytoplasm of the early one-cell embryo possesses the ability to demethylate H3K79 and that H3K79 demethylation is involved in genome reprogramming after fertilization.
|
|
Because it has been reported that the ability of the embryo to reprogram the genome of a transplanted somatic nucleus is lost as early as 5 to 6 h after fertilization [46], we examined whether the ability to demethylate H3K79 also would be lost in mid-one-cell-stage embryos. The nucleus from an NIH3T3 cell was transplanted into an enucleated oocyte at 7 h after activation. We found that H3K79 methylation still remained at a level comparable to the embedded nucleus after incubation for 5 h (Fig. 8B and Table 1). This indicates that the ability to demethylate H3K79me2 is lost before the mid-one-cell stage.
H3K79me2, As an Active Marker of Gene Expression, Is Not Deleted in Two-Cell Embryos
It has been suggested that H3K79me2 is an active gene marker in somatic cells that persists through the M phase and is inherited by the daughter cells following cell division [34]. However, it also has been reported that the level of H3K79me2 increases at M phase, during which time genes are not activated. In addition, this increased level of H3K79 is maintained during the G1 phase and then decreases to original levels during the S phase [31]. Thus, it seems that when H3K79me2 is functioning as an active gene marker, it is maintained throughout the entire cell cycle; furthermore, additional H3K79me2 is produced at M phase and plays a role in some M-phase-specific events. In preimplantation embryos, however, this pool of M-phase H3K79me2 decreased to undetectable or very low levels during interphase before the blastocyst stage (Fig. 2). Therefore, we considered that H3K79me2 does not act as an active gene marker and that the H3K79me2 that appeared during the M phase is completely deleted soon after cleavage in the preimplantation embryos (the G1 phase is very short in these embryos [47–49]). To address this hypothesis, somatic nuclei containing H3K79me2 as an active gene marker were transplanted into late-stage one-cell embryos and were examined for changes in H3K79 levels at M phase and after cleavage into the two-cell stage. The nuclei from NIH3T3 cells were transplanted into the enucleated embryos 10 h after parthenogenetic activation and then were observed for the presence of H3K79me2 after 2, 5, and 12 h. Five hours after transplantation, the nuclei present in embryos that had proceeded into M phase had prominent increases in their H3K79me2 levels (Fig. 9A). However, the embryos that had not yet proceeded into M phase showed no increase in H3K79me2 from the original levels measured before transplantation (data not shown). In cleaved embryos 12 h after transplantation, H3K79me2 decreased but still remained at levels comparable to original values. Thus, H3K79me2, which would have been an active gene marker in interphase somatic nuclei, persisted in two-cell embryos. When pronuclei (no H3K79me2 observed) were transplanted into parthenogenetically activated embryos, H3K79me2 also increased at M phase, but it was completely lost after cleavage into the two-cell stage (Fig. 9B). Taken together, H3K79, which increases at M phase but not interphase, appears to decrease in two-cell embryos.
|
We examined the localization of H3K79me2 and H3K79me3 in somatic cells and oocytes. The subnuclear localization was different between these two modifications. H3K79me2 was observed in the entire genome, whereas H3K79me3 was localized in the pericentromeric regions (Figs. 1–4). Both of these modifications were lost in oocytes soon after fertilization and remained at low levels before the blastocyst stage. The one exception was H3K79me2 levels, which appeared to increase transiently at M phase (Figs. 2, 3, 5, and 6). The signal for H3K79 methylation was also lost in somatic nuclei transplanted into enucleated oocytes following parthenogenetic activation (Fig. 8).
The differences in subnuclear localization observed between H3K79me2 and H3K79me3 suggest that they have different functions. In general, as more methyl groups are added, lysine residues become increasingly hydrophobic. In H3K9 methylation, both H3K9me2 and H3K9me3 are localized in heterochromatin regions [15, 16, 21]. However, heterochromatin protein 1 (HP1), which associates with methylated H3K9 to form heterochromatin, binds to H3K9me3 with twice the affinity that it binds to H3K9me2 [25]. This may explain the differential localization of H3K9me2 and H3K9me3, which are observed on the facultative and constitutive heterochromatin, respectively [13, 22]. Although H3K79me2 is thought to be a marker of active genes, we found that H3K79me3 is localized on pericentromeric heterochromatin (Figs. 1 and 4), which does not contain active genes, [15, 16, 18, 32, 33]. It is possible that H3K79 methylation acts as a switch of gene expression that is regulated by the number of methyl groups. When a methyl group is added to H3K79me2 on a gene that has been active, this gene might be placed into a repressed state and recruited into the pericentromeric region. It has been reported that H3K79me2 is enriched at the β-globin gene when it is active [32] and that it is repressed by its recruitment onto pericentromeric heterochromatin [50]. Because both H3K79me2 and H3K79me3 are catalyzed by the same enzyme, DOT1L, it is unknown how these different modifications are generated in a space- and time-dependent manner during preimplantation development. A recent study reported that there are multiple splice variants of Dot1l [51]. H3K79me2 and H3K79me3 may be catalyzed by different DOT1L variants, which are expressed and localized in different manners during preimplantation development.
H3K79me3 was localized adjacent to centromere proteins and was colocalized with HP1β in mouse somatic cells and oocytes (Figs. 1 and 4). This suggests that H3K79me3, as well as H3K9me3 and H4K20me3 [23, 52, 53], is characteristic of pericentromeric heterochromatin. The histone code hypothesis suggests that various histone modifications interact to play a role in the formation of various chromatin domains [54, 55]. It was suggested that H4K20me3 participates with H3K9me3 in the formation of pericentromeric heterochromatin [52]. H3K79me3 may also be involved in this process by interacting with these modifications. In yeast, previous studies using chromatin immunoprecipitation (ChIp) detected H3K79me3 on active genes [19]. This suggests that the localization of H3K79me3 may differ between mammals and fungi. However, it cannot be excluded that H3K79me3 is localized in the pericentromeric region in yeast because previous experiments investigated active genes, but not the pericentromeric domain, in ChIp experiments [19].
H3K79me3 disappeared after fertilization and remained undetectable during preimplantation development (Fig. 3). The possibility can be excluded that the absence of H3K79me3 signals would be attributed to the loss of accessibility of the antibody to the target in the preimplantation embryos, because a number of studies have shown that the signal is detected for other antibodies against histone H3 modifications (e.g., H3K9ac, H3K14ac, H3K9me2, H3K9me3, H3K4me2, and H3K4me3) in preimplantation embryos [56–58], indicating that these antibodies are accessible to their targets in the preimplantation embryos. It has been shown that pericentromeric heterochromatin undergoes dynamic reorganization during preimplantation development [59]. Therefore, the absence of H3K79me3 may be involved in the reorganization of pericentromeric heterochromatin during this period.
H3K79me2 either was not detected or was detected at a marginal level in embryos before the blastocyst stage. However, it appeared transiently only at M phase, suggesting that it does not function at interphase and instead plays a specific role in M phase during preimplantation development (Fig. 2). In somatic cells, H3K79me2 is detected at interphase, and its levels increase at M phase [31]. Although this result was obtained by immunoblotting, in which cytoplasmic histone H3 protein and that incorporated on nucleosomes were included in the analysis, H3K79me2 levels were very likely to reflect those in the nucleosomal protein, because DOT1L only acts when histone H3 is packaged into a nucleosome [27]. H3K79me2 occurs in the promoters and coding regions of activated genes at interphase following the first round of transcription and then persists during M phase, at which time transcription is suppressed [13, 34]. Therefore, it has been suggested that H3K79me2 on active genes is inherited by daughter cells following mitosis and acts as an epigenetic marker of active gene expression [34]. However, H3K79me2 levels increase at M phase, during which time there is no transcription, suggesting that there is an M-phase-specific function that differs from that at interphase [31, 34]. In embryos, H3K79me2 levels also increased at M phase (Fig. 2), suggesting that it may also have a specific function in M phase. However, H3K79me2 levels were not detected at interphase before the blastocyst stage (Fig. 2), although genes are actively transcribed. In addition, H3K79me2 did not increase in somatic nuclei transplanted into the embryos during interphase (Fig. 9). These results suggest that H3K79me2 does not act as an epigenetic marker of active gene expression in the early preimplantation embryos. A schematic view of this hypothesis is shown in Figure 10. Before the blastocyst stage, the embryos are not differentiated and seem to possess pluripotency. The absence of epigenetic markers may be involved in the genomic plasticity required to maintain pluripotency in these embryos. It is interesting that the stages at which H3K79me2 and H3K79me3 reappeared differed; H3K79me2 reappeared at the four-cell stage at a low level and increased at the blastocyst stage (Fig. 2), whereas H3K79me3 remained absent at the blastocyst stage (Fig. 3). It was reported that macroH2A, a heterochromatin marker, was lost from the nucleus after fertilization and reappeared after the eight-cell stage [40]. H3K9me2 decreases until the four-cell stage and thereafter prominently increases [57]. Thus, the preimplantation embryos may gradually lose their genome plasticity and prepare for the differentiation by adding various histone modifications at different stages during preimplantation development.
|
H3K79me2 has been suggested to be associated with the mechanism for the repair of DNA double-strand breaks, as well as gene expression. In fact, 53BP1, which plays a pivotal role in the DNA repair mechanism, binds to H3K79me2, and the decrease in the level of H3K79me2 by suppressing DOT1L, which catalyzes the methylation of H3K79, reduces the recruitment of 53BP1 to the sites of double-strand breaks and induces DNA damage checkpoint defects [60, 61]. In this context, it is interesting that the H3K79me2 level is low before the blastocyst stage (Fig. 2), because preimplantation embryos are hypersensitive to exposure to 
-irradiation, which induces double-strand breaks [62]. In addition, recent studies showed that the levels of -H2AX, which plays an important role in DNA repair, and expression of 53BP1 were low before the blastocyst stage [63–65]. Thus, the low level of H3K79me2 before the blastocyst stage may be associated with the deficiency in the DNA repair mechanism.
The gene expression patterns of differentiated oocytes are reprogrammed when they are transformed into totipotent embryos following fertilization [42, 44, 45]. During this process, epigenetic markers that had previously sustained oocyte-specific gene expression patterns would be erased, and genome reprogramming would occur [43]. In somatic nuclei transplanted into enucleated oocytes, the epigenetic markers would also be deleted, because the gene expression patterns in the differentiated somatic cells should be reprogrammed. We showed that both H3K79me2 and H3K79me3, which are involved in active gene expression and gene silencing, respectively, disappeared from the oocytes soon after fertilization. H3K79me2 also disappeared from the nuclei transplanted into the enucleated oocytes after activation (Fig. 8). These results suggest that the disappearance of these modifications may be involved in genome reprogramming. Furthermore, our results suggest that a phenomenon involved in genome reprogramming occurs soon after fertilization. Although active transcription occurs in the growing oocytes, it ceases when the oocytes are fully grown. The oocytes remain in a transcriptionally silent state until zygotic genes are activated after fertilization; in mice, the expression of zygotic genes begins as early as the middle to late one-cell stage [4]. Therefore, the deletion of epigenetic markers is likely to occur during this transcriptionally silent period. However, it was previously unknown when this phenomenon occurs (e.g., at the GV stage, during meiotic maturation, or after fertilization). Our results suggest that the deletion of some epigenetic markers occurs soon after fertilization. These results are consistent with reports that transplanted somatic nuclei are efficiently reprogrammed when they are transplanted into enucleated bovine oocytes soon after the oocytes are activated [66, 67]. The ability to reprogram a transplanted somatic nucleus decreases with time after the activation of enucleated oocytes; the percentage of transplanted oocytes that develop into the blastocyst stage gradually decreases following the time at which the somatic nucleus is transplanted, and no embryos that have been subjected to nuclear transplantation 5 to 6 h after fertilization develop into the blastocyst stage [46]. Consistent with this report, our results show that the somatic nuclei were not demethylated on H3K79 when they were transplanted into the enucleated embryos 7 h after activation (Fig. 8B). Thus, we suggest that global H3K79 demethylation is involved in genome reprogramming, and the ability to delete H3K79 methylation seems to be lost during the one-cell stage. The mechanism by which H3K79 methylation is deleted is unknown. Although it is likely that an enzyme catalyzing the demethylation of H3K79 is activated soon after fertilization, such an enzyme has not yet been identified in any organism. Alternatively, it is possible that the histones methylated on K79 are replaced by new ones that have not been methylated or that one or more proteins bind to H3K79me and interfere with the access of the antibodies.
Although many histone modifications have been examined in embryos and shown to change dynamically during preimplantation development, none of them disappeared during the early one-cell stage (H3K9me2 and H3K9me3 [57]; H3K27me3 [21]; H3K4me2 and H3K4me3 [58]; H4K20me3 [53]; H4K5ac, H4K8ac, H4K12ac, and H4K16ac [56, 68, 69]; H3K9ac and H3K14ac [56]). However, macroH2A is lost from the female pronucleus by the late one-cell stage [40]. Although it is currently unknown whether the mechanisms regulating H3K79 methylation and the deposition of macroH2A in the nucleosomes are associated, a dramatic remodeling of the genome seems to occur soon after fertilization.
Correspondence: 1Fugaku Aoki, Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Room # 302, Seimei-Building, Kashiwa, Chiba 277-8562, Japan. FAX: 81 471 36 3698; e-mail: aokif{at}k.u-tokyo.ac.jp
Received: 14 June 2007.
First decision: 17 July 2007.
Accepted: 2 November 2007.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
