Analysis of active chromatin modifications in early mammalian embryos reveals uncoupling of H2A.Z acetylation and H3K36 trimethylation from embryonic genome activation

Abstract

Early embryonic development is characterized by dramatic changes in cell potency and chromatin organization. The role of histone variants in the context of chromatin remodeling during embryogenesis remains under investigated. In particular, the nuclear distribution of the histone variant H2A.Z and its modifications have not been examined. Here we investigated the dynamics of acetylation of H2A.Z and two other active chromatin marks, H3K9ac and H3K36me3, throughout murine and bovine pre-implantation development. We show that H2A.Z distribution is dynamic during the earliest stages of mouse development, with protein levels significantly varying across stages and lowest at the 2-cell stage. When present, H2A.Z localizes preferentially to euchromatin at all stages analyzed. H2A.Z is acetylated in pre-implantation blastomeres and is preferentially localized to euchromatin, in line with the known role of H2A.Zac in transcriptional activation. Interestingly, however, H2A.Zac is undetectable in mouse embryos at the 2-cell stage, the time of major embryonic genome activation (EGA). Similarly, H3K36me3 is present exclusively in the maternal chromatin immediately after fertilization but becomes undetectable in interphase nuclei at the 2-cell stage, suggesting uncoupling of these active marks with global embryonic transcription activation. In bovine embryos, which undergo EGA at the 8-cell stage, H2A.Zac can be detected in zygotes, 4-, 8- and 16-cell stage embryos as well as in blastocysts, indicating that the dynamics of H2A.Zac is not conserved in mammals. In contrast, H3K36me3 displays mostly undetectable and heterogeneous localization pattern throughout bovine pre-implantation development. Thus, our results suggest that 'canonical' active chromatin marks exhibit a dynamic behavior in embryonic nuclei, which is both stage- and species-specific. We hypothesize that chromatin of early embryonic nuclei is subject to fine-tuning through differential acquisition of histone marks, allowing for proper chromatin remodeling and developmental progression in a species-specific fashion.

Keywords: :

• epigenetics
• Histone Modifications
• histone variants
• mammalian embryo

Introduction

Pre-implantation development of mammals takes place from the fertilization of the oocyte by the sperm head until the hatching of the blastocyst and its implantation into the uterine wall of the mother. Upon fertilization, both parental chromatins are reorganized to form a novel chromatin configuration in the embryo necessary for a new developmental program. The events that orchestrate such reprogramming and chromatin assembly upon zygote formation are not fully understood. Oocytes and, subsequently, zygotes contain a large pool of proteins and untranslated mRNAs, which support the development of the new organism before the embryonic genome becomes activated. The onset of embryonic transcription marks the transition from maternal to embryonic control of development.

The timing of the onset of genome activation differs between species. In mice, the first major wave of embryonic genome activation (EGA) occurs at the 2-cell stage of development,^1,2 while bovine embryos start transcribing their genome later, at the 8-cell stage.³ Although the exact mechanisms regulating EGA are not fully known, underlying chromatin-based processes are expected to be in place to ensure a tight control of EGA. These include changes in transcriptionally repressive environment as well as the action and availability of enhancer elements.^4-6 The first round of DNA replication has also been suggested to play a regulatory role in this process.^7,8

In a eukaryotic cell, DNA is packaged into a nucleoprotein complex, called chromatin. The basic unit of chromatin is the nucleosome, in which core histones (two copies of H2A, H2B, H3 and H4) form an octamer around which ~146 bp of DNA is wrapped.⁹ Chromatin is further organized into higher-order structures, in part due to the action of the linker H1 histone and non-histone proteins. The structure of chromatin, which can render DNA more or less accessible to factors acting upon it, can and must be modified in response to the vast variety of signals received from the environment. Changes in chromatin structure occur through the action of ATP-dependent chromatin remodeling complexes, post-translational modifications of histones or through the incorporation of specific chromatin proteins like histone variants.^10,11 Incorporation of such “replacement” histones is one of the emerging layers of chromatin regulation.¹¹ Of the core histones, H2A has the largest number of variants. Arguably, the most extensively studied of them is H2A.Z. H2A.Z, a highly conserved protein,¹² was discovered in the 1980s, and was first implicated in gene activation, as it was exclusively found in the transcriptionally active macronucleus of Tetrahymena thermophilla.¹³ Subsequent investigations uncovered a vast array of H2A.Z functions, including maintenance of telomeric heterochromatin boundaries, suppression of antisense RNAs and epigenetic memory.^14-18 It is now thought that, despite the high conservation of the primary sequence of H2A.Z throughout evolution, the function of the protein has adapted to play specific roles in different organisms.

Post-translational modifications of histones have been extensively studied. For example, H3K9me3 and H3K27me3 are considered as repressive marks, whereas H3K4me3 is linked to transcriptional activity.¹⁰ Like canonical histones, H2A.Z can be posttranslationally modified. The N-terminal lysines (K4, 7, 11, 13 and 15 in mice and human) are subject to acetylation, which modulates the essential charge patch of H2A.Z.^19-21 Acetylation of H2A.Z (H2A.Zac) has been linked to active genes in chicken erythrocytes and in the maintenance of chromosome architecture in fission yeast.^22,23 Recently, it was shown that H2A.Zac might be involved in gene deregulation occurring in cancer cells.²⁴

H2A.Z is the first histone variant shown to be essential in mammals.²⁵ Mice lacking H2A.Z die at the time of implantation, presumably due to defects in inner cell mass (ICM) proliferation and differentiation.²⁵ Detailed analysis of H2A.Z localization at the blastocyst stage reports that H2A.Z first becomes detectable in trophectoderm cells, where it is targeted to pericentromeric heterochromatin and subsequently to other parts of the nucleus.²⁶ Similarly, H2A.Z becomes detectable in ICM derivatives upon differentiation but is not detected initially in the undifferentiated ICM. A study investigating the global pattern of H2A variants in early mouse embryogenesis reported an absence of H2A.Z from the 2-cell stage until the late 8-cell stage.²⁷ However, the H2A.Z antibody used in this study cannot discriminate between modified and unmodified H2A.Z and does not recognize the ubiquitylated form of H2A.Z.²⁷ The nuclear localization of H2A.Z in early embryos or whether it exists in a modified form has not been addressed.

H3K9ac, H3K36me3 and H3K4me3 are marks of actively transcribed chromatin. H3K36me3 is associated with transcriptional elongation and is enriched in the coding region of transcribed genes.^28,29 Studies in yeast and higher eukaryotes have shown that the H3K36 methyltransferase SET2 interacts with the elongating RNA polymerase II through the CTD of its largest Rbp1 subunit.^30,31 Acetylation of H3K9 is an active chromatin mark, directly antagonizing the occurrence of trimethylation on the same residue.¹⁰ H3K9 is acetylated by Gcn5 and CBP^10,32 and is enriched at the transcription start site of active genes in all systems analyzed to date. Moreover, similarly to H3K4me3, the enrichment of H3K9ac at a promoter correlates directly to the level of expression of a gene.

So far, most of the analyses of histone modifications in early embryos have been focused on heterochromatic marks,^33-39 but the contribution of active histone modifications to the 'marking' of the newly established embryonic chromatin has received less attention.^40-42

Here, we have investigated the dynamics of several active histone modifications in the nuclei of cleavage stage mouse and bovine embryos. We show that H2A.Z is acetylated in early mammalian embryos. H2A.Zac exhibits a dynamic temporal pattern and localizes to euchromatin. Three other euchromatin marks, H3K9ac, H3K4me3 and H3K36me3, were detected in mouse embryonic chromatin, but exhibited different spatiotemporal patterns. We show that H3K36me3 defines an epigenetic asymmetry in mouse zygotes. In contrast, in bovine embryos a parental asymmetry is defined by H2A.Zac, which is enriched in the maternal pronucleus. Notably, H2A.Zac and H3K36me3 are undetectable in mouse embryos during the major wave of zygotic genome activation. The combined analysis of bovine and mouse embryos suggests that H3K36me3 is globally uncoupled from EGA and highlights species-specific temporal dynamics of H2A.Z acetylation during embryogenesis.

Results

We first set out to determine the presence of the histone variant H2A.Z in the nuclei of cleavage-stage mouse embryos. Immediately after fertilization H2A.Z is only barely detectable in the two pronuclei (Fig. 1A and B). However, H2A.Z is readily detected in the maternal chromatin of the polar body. This suggests that H2A.Z might be either differentially segregated upon meiosis or 'removed' from the maternal chromatin after fertilization. H2A.Z remains undetectable in embryonic chromatin at the early 2-cell stage and starts to be incorporated again—albeit at very low levels—only at the late 2-cell stage (Fig. 1C). From the 4-cell stage onwards, H2A.Z is abundant and uniformly distributed in all embryonic nuclei but excluded from the DAPI rich regions. Furthermore, H2A.Z was readily detected in both the trophectoderm and the ICM at the blastocyst stage, although H2A.Z levels were slightly higher in trophectoderm cells (not shown), which is in line with previously published data.²⁶ Immunostaining with the H2A.Z specific antibody showed a strong H2A.Z signal in mitotic chromosomes (Fig. 1A, blastocyst), suggesting that H2A.Z remains incorporated in chromatin during cell division. H2A.Z was largely excluded from the DAPI-rich regions, suggesting a preferentially euchromatic localization throughout embryogenesis (Fig. 1B, arrows). Thus, H2A.Z shows a stage-dependent, dynamic distribution, with lowest levels in the zygote and 2-cell stage embryos.

Figure 1. Dynamics of H2A.Z during mouse pre-implantation development. Freshly collected embryos were fixed and stained with an a-H2A.Z antibody (red). DNA is shown in blue. (A) Top and middle panels show full projections of Z-sections taken every 1 µm on a confocal microscope. The bottom panel shows a merge of the corresponding middle sections (1 µm) for the blue (DNA) and red (H2A.Z) channels. The arrowheads point to the polar body; male and female pronuclei are indicated; in the blastocyst image the arrow points to mitotic chromosomes. Images were acquired using the same confocal parameters and on the same slide, therefore the fluorescence levels are directly comparable. At least 8 embryos per stage were analyzed in at least two independent experiments. Scale bar = 20 µm. (B) H2A.Z levels decrease between the zygote and the 2-cell stage, after which they increase from the 4-cell stage onwards. Shown are higher magnification of pronuclei (at zygotic stage) and individual nuclei at indicated stages of development stained with a-H2A.Z antibody. Top panels show a single confocal section of H2A.Z staining (in gray scale). The bottom panels show a merge image of the corresponding Z-section of DNA (blue) and H2A.Z (red). Scale bar = 10 µm.(C) H2A.Z is absent from embryonic chromatin in the period that follows fertilization and until the second replication site at the late 2-cell stage. Early and late 2-cell stage embryos were stained with the H2A.Z antibody and analyzed under confocal microscopy under the same conditions and in parallel.

Because acetylation of the N-terminal lysines of H2A.Z has been shown to exert a positive role in transcriptional regulation,^22,24,43 we next asked whether H2A.Z exists in its acetylated form (H2A.Zac) during early embryogenesis. Using a specific antibody that recognizes acetylated K4, K7 and K11 of H2A.Z, we performed immunostaining on freshly collected cleavage stage embryos. We detected acetylation of H2A.Z at early pronuclear stages (PN0-PN3) at very low levels in both pronuclei (Fig. 2A and B). In agreement with the absence of H2A.Z from embryonic chromatin at the 2-cell stage, we did not detect acetylation of H2A.Z in 2-cell stage embryos. However, H2A.Z was highly acetylated from the 4-cell stage onwards, with high acetylation levels persisting to the blastocyst stage. We did not detect any obvious difference between acetylation of H2A.Z in the trophectoderm vs. the ICM (Fig. 2A). H2A.Zac was enriched in euchromatin in all stages analyzed and, contrary to the unmodified H2A.Z, was undetectable on mitotic chromosomes (Fig. 2C), in agreement with previous reports.²² Thus, H2A.Zac was detected in all stages of mouse pre-implantation development, apart from the 2-cell stage.

Figure 2. H2A.Z is acetylated during pre-implantation development. (A) Embryos were fixed at the indicated stages and processed for immunostaining with an antibody specific for acetylated H2A.Z (green). DNA is shown in blue. Images were acquired using confocal microscopy. Top and middle panels show full projections of Z-sections taken every 1 µm of representative embryos stained with the H2A.Zac antibody and DAPI, respectively. The bottom panel shows the merge of corresponding middle sections in the blue (DNA) and green (H2A.Zac) channels. Changes of fluorescence between cleavage stages are comparable as all embryos shown were processed in parallel and using identical acquisition settings. Scale bar = 20 µm. (B) Levels of acetylation of H2A.Z change during the course of pre-implantation development. H2A.Z acetylation is very low in both pronuclei in the zygote and becomes undetectable at the early 2-cell stage. H2A.Zac is present in euchromatic regions of morula and blastocyst stage nuclei. Full projections of nuclei stained with the H2A.Zac antibody (top) and a middle, merge section (bottom) are shown. Scale bar = 10 µm. (C) H2A.Z acetylation is not detected on mitotic chromosomes. Shown is a full projection of a morula stage blastomere stained with the H2A.Zac antibody. DNA is shown in blue.

The major wave of EGA in mouse embryos takes place at the 2-cell stage. Surprisingly, at this stage we observed an absence of H2A.Zac from embryonic chromatin, an active chromatin mark with well-documented roles in transcriptional activation. This prompted us to investigate the presence and dynamics of other known marks of actively transcribed chromatin, and their relation to EGA at the 2-cell stage. We chose to look at H3K9 acetylation, H3K36 trimethylation and H3K4 trimethylation, all of which are well-established euchromatic marks. In addition, to our knowledge, their presence has not been investigated thus far in the context of early mammalian embryogenesis, and only few reports document the analysis of H3K9ac and H3K4me3 in mouse zygotes.^33,40-46

We found that H3K9ac is abundant in all stages of mouse pre-implantation development, with high levels of H3K9ac detectable from the zygote till the blastocyst stage (Fig. 3A). H3K9ac is present in the zygote from the earliest pronuclear stages (PN1) and is equally abundant in the male and the female pronucleus (Fig. 3A and B). Thus, H3K9ac is present in the embryonic chromatin prior to EGA. Because chromatin in early developmental stages is considered to be more plastic or 'open', we suggest that high levels of H3K9ac might contribute to such an open chromatin state. This could be, for example, through rendering H3K9 pools unavailable for methylation, a known mark of heterochromatin.

Figure 3. H3K9ac dynamics throughout early mouse embryogenesis. (A) Acetylation of H3K9 occurs throughout pre-implantation development. Mouse embryos at the indicated stages were stained with a H3K9ac antibody (green) and with DAPI (blue). Full projections of Z-sections acquired every 1 μm (top and middle panels) and middle, merge section (bottom panel) are shown. (B) Distribution of H3K9ac in male and female pronuclei. Representative male (right) and female (left) pronuclei (PN) in zygotes at indicated pronuclear stages were stained with an H3K9ac antibody. Note that H3K9ac accumulates uniformly in both pronuclei. Scale bar = 10 µm.

H3K36me3, on the other hand, showed a markedly different pattern of localization in mouse embryonic nuclei. We detected high levels of H3K36me3 in the zygote, but exclusively in the maternal pronucleus (Fig. 4A and B). Importantly, we did not detect H3K36me3 staining in the paternal pronucleus, indicating that during pronuclear formation and zygotic development paternal chromatin does not acquire trimethylation on H3K36 (Fig. 4A and B). Therefore, H3K36me3 contributes to the established epigenetic asymmetry of both pronuclei. At the 2-cell stage, the nuclear H3K36me3 signal decreases substantially to almost undetectable levels in interphase nuclei (Fig. 4A), presumably through active demethylation, from the maternal chromatin. A strong signal remains readily detectable in the polar body, which also serves as a positive control for immunostaining. H3K36me3 levels become detectable, albeit at low levels, throughout the 4-cell stage (Fig. 4A). H3K36me3 is more apparent at the 8-cell stage, albeit at lower signal intensity compared with the zygote, which remain low even at the blastocyst stage. Interestingly, not all nuclei within the blastocyst are positive for H3K36me3, which shows therefore a heterogeneous distribution (Fig. 4A). H3K36me3, a mark of transcriptional elongation, seems therefore to be absent in interphase nuclei at the 2-cell stage, a time of intensive transcription of the embryonic genome.

Figure 4. H3K36me3 shows a heterogeneous pattern in early mouse embryos. (A) H3K36me3 shows an epigenetic asymmetry between the maternal and paternal pronucleus immediately after fertilization and is undetectable at the 2-cell stage. Embryos at the indicated stages were fixed for immunostaining with an H3K36me3 antibody. Shown are full projections (top and middle panels) or single merge sections of representative embryos. (B) Distribution of H3K36me3 in parental pronuclei at zygotic stage. Male and female pronuclei at indicated stages of zygotic development were stained with an H3K36me3 antibody. H3K36me3 is readily detectable in maternal chromatin and the polar body, but undetectable in the male pronuclei. Clearly, H3K36me3 defines an epigenetic parental asymmetry and is absent from the paternal chromatin. Scale bar = 10 µm.

The apparent uncoupling of H3K36me3 and H2A.Zac from EGA in the mouse embryo prompted us to investigate the distribution of H3K4me3 at these stages. H3K4me3 displayed an asymmetry immediately after fertilization, with the maternal pronucleus displaying higher levels of H3K4me3 compared with the paternal one (Fig. 5A, B and refs. 40,47). This is consistent with previous reports that have suggested a dichotomy between the genome-wide localization of H3.3 in the male pronucleus and undetectable levels of H3K4me3.⁴⁷ Subsequently, H3K4me3 was detected in all stages of pre-implantation development with clear euchromatic localization (Fig. 5A and B). Thus, H3K36me3 and H2A.Zac are uncoupled from EGA in mouse embryos, but H3K4me3 and H3K9ac are globally not.

Figure 5. H3K4me3 is present throughout mouse pre-implantation development. (A) Freshly collected mouse embryos from natural matings at the indicated stages were stained with a H3K4me3 antibody (green) and with DAPI (blue). Full projections of Z-sections acquired every 1 μm (top and middle panels) and middle, merge section (bottom panel) are shown. Note that similar results were obtained with embryos derived from superovulation experiments (not shown). (B) H3K4me3 displays euchromatic localization in embryonic nuclei. Shown are higher magnifications of pronuclei (at zygotic stage) and individual nuclei at indicated stages of development stained with an H3K4me3 antibody. Top panels show a single confocal section of H2A.Z staining (in gray scale). The bottom panels show a merge image of the corresponding Z-section of DNA (blue) and H3K4me3 (green). Scale bar = 10 µm.

In bovine embryos, the major wave of EGA occurs later than in mice, at the 8-cell stage. To determine if uncoupling of these two active marks from EGA is conserved in another mammalian species, we performed immunostaining with the H2A.Zac and H3K36me3 antibodies on bovine embryos at various developmental stages. We also analyzed H3K9ac. We found that H2A.Zac is present in all stages of bovine embryogenesis analyzed (Fig. 6A). Interestingly, in zygotes (1-cell stage PN2/PN3 embryos), H2A.Zac is detected almost exclusively in one of the pronuclei (Fig. 6A), with only 18% of the zygotes analyzed (n = 32) displaying a very faint staining in the second pronucleus. Co-staining with an H3K9me3 antibody—which marks maternal chromatin⁴⁸—revealed that H2A.Zac is strongly enriched on the maternal chromatin of bovine zygotes (Fig. 6B). Subsequently, H2A.Zac shows a uniform distribution in all nuclei from the 2/4-cell stage onwards and is abundant both in the trophectoderm and ICM of day 6 blastocysts (Fig. 6A).

Figure 6. Developmental dynamics of H2A.Zac, H3K36me3 and H3K9ac in bovine embryos. (A) Analysis of H2A.Z acetylation in cleavage stage bovine embryos. H2A.Z is acetylated in the zygote and throughout pre-implantation development. Note that in bovine embryos H2A.Zac defines a parental epigenetic asymmetry as only the female PN contains acetylated H2A.Z. The number of embryos analyzed was: 32 zygotes, 24 8-cell stage embryos and 21 blastocysts. Scale bar is 5µm. (B) H2A.Z marks exclusively the maternal chromatin in bovine zygotes. Co-staining of H2A.Zac and H3K9me3 in bovine zygotes. Zygotes (n = 32) were analyzed by confocal microscopy as above. (C) H3K36me3 shows a heterogeneous distribution in bovine pre-implantation embryos. H3K36me3 is undetectable in bovine zygotes after fertilization. H3K36 methylation occurs only in few nuclei in 4-cell stage embryos and blastocysts. A total of 45 zygotes, 21 2-cell stage embryos, 16 4-cell stage embryos and 18 blastocysts were analyzed. Scale bar is 5µm. (D) H3K36me3 marks the female pronucleus exclusively at the earliest stages of bovine development immediately after fertilization. Early zygotes (17h post-fertilization) were processed for immunostaining with the H2A.Zac and the H3K36me3 antibody. Out of 20 embryos analyzed, 8 zygotes had H3K36me3 staining, which co-localized with H2AZac on the female PN. (E) H3K9ac is present throughout bovine pre-implantation development. Embryos were analyzed as above with an H3K9ac antibody. H3K9ac marks equally well both pronuclei. Note that the 2-cell stage is a very transient stage and therefore it is not included in our figure.

In contrast to H2A.Zac, H3K36me3 was mostly undetectable in bovine zygotes (Fig. 6C). Of the 45 zygotes analyzed, only 11 zygotes showed a faint staining in only one of the pronucleus (Fig. 6C). A more detailed analysis revealed that the zygotes showing a faint staining in one of the pronuclei correspond to earlier stage zygotes (17 h post-fertilization). Co-staining with the H2A.Zac antibody revealed that, when labeled, the H3K36me3-labeled pronucleus is the maternal one (Fig. 6D), suggesting that the epigenetic asymmetry demarcated by H3K36me3 might be conserved in mice and cattle. Later stage zygotes (20 h post-fertilization) displayed mostly undetectable H3K36me3 levels in both pronuclei. We first detected H3K36me3 staining again at the transition between the 4- and 8-cell stage, but not all nuclei are marked with H3K36me3 (Fig. 6C). This heterogeneous pattern of H3K36me3 persisted until the blastocyst stage, where some nuclei are H3K36me3 positive while others are devoid of H3K36me3 (Fig. 6C).

Contrary to the heterogeneous pattern of H3K36me3, we detected H3K9ac throughout bovine pre-implantation embryogenesis (Fig. 6E). Moreover, both pronuclei appeared equally stained for H3K9ac (Fig. 6E) and H3K9ac persisted in 2-cell, 4-cell and 8-cell embryos and in the blastocyst (Fig. 6E and data not shown).

Discussion

The period that follows mammalian fertilization is a time of intense chromatin remodeling. It has been suggested that the epigenetic asymmetries that arise on the parental chromatins are a key aspect of the reprogramming process that underlies early development. The precise constitution of embryonic chromatin and the contribution of histone variants and their modifications and histone modifications in general to this remodeling program are not well understood. Here, we have documented the temporal dynamics of H2A.Z and its acetylated form as well as H3K36me3, H3K4me3 and H3K9ac, three other euchromatic marks, in early mammalian embryos.

Our data indicate that H2A.Zac is not present in mouse embryonic chromatin at the time of global embryonic transcriptional activation, which occurs at the 2-cell stage. This is somewhat surprising because within in a heterotypic nucleosome (H2A.Z/H2A), H2A.Z acetylation contributes to the overall destabilization of the nucleosome core particle, acting in concert with acetylation of other core histones, which presumably facilitates transcription.⁴⁹ Indeed, H2A.Zac has been ascribed a role in transcriptional activation in chicken erythrocytes and enrichment of H2A.Z at TSS in ES cells correlates with transcriptional activity.^22,29,50 Therefore, our data suggest that regulation of embryonic transcription might be orchestrated by different chromatin signatures to that in somatic cells or at later stages of development.

The dynamics of H2A.Z in mouse embryos were recently reported in another study.²⁷ Our temporal dynamics differ to those presented in that study, in which they report the absence of H2A.Z from the 2-cell stage until the late morula and blastocyst stage, while we detect H2A.Z in embryonic chromatin from the late 2-cell stage. Indeed, the presence of H2A.Z from the 4-cell stage is also confirmed by the use of the H2A.Zac antibody, as we readily detect H2A.Zac from this stage onwards. The discrepancies are likely due to differences in experimental procedures. Note that, in contrast to the previous analysis in which in vitro fertilized embryos were used, we used freshly collected embryos fertilized in vivo. We think this is of particular importance when analyzing chromatin modifications given the potential influence of culture and in vitro fertilization on the stability of some chromatin modifications and DNA methylation.^51,52 It is important to note that the H2A.Z antibody used in both studies is directed against the C-terminal part of the protein, which is sensitive to monoubiquitylation that occurs at K120 and K121 of H2A.Z.⁵³ There are no H2A.Zub antibodies currently available, and therefore it remains to be determined whether H2A.Z is ubiquitylated during embryogenesis. Indeed, ubiquitylation of H2A.Z could be involved in silencing of heterochromatin in mammalian cells, as it has been proposed for other model systems. In line with this, it has been suggested that H2A.Z localizes to heterochromatin of trophectoderm-like cells derived from blastocyst outgrowths.²⁶

H3K36me3, a mark of transcriptional elongation, is also undetectable from interphase nuclei of mouse embryos at the 2-cell stage, a time of intensive transcription of the embryonic genome, suggesting that during embryonic development, regulation of elongation might be independent of H3K36me3 and/or that H3K36me3 might have other functions. Interestingly, H3K36me3 marks exclusively the maternal chromatin after fertilization and is absent from the paternal pronucleus. In somatic cells, H3K36me3 is highly enriched in the histone variant H3.3⁵⁴ and therefore the absence of H3K36me3 in the paternal pronucleus is in sharp dichotomy with the exclusive presence of H3.3 in the paternal pronucleus at this stage.^47,55-57 Moreover, H3K36me3 appears to be an exception, as it is so far the only euchromatic histone modification that is exclusive to the maternal chromatin. Indeed, it is generally believed that the epigenetic asymmetry of the two pronuclei after fertilization is mainly due to the absence of heterochromatic marks in the paternal pronucleus that would be permissive for epigenetic reprogramming.^33,36 It is therefore also surprising to find H2A.Zac only in the maternal pronuclei of bovine zygotes while other histone acetylation marks can be observed in both pronuclei.⁴¹ All together, this suggests that only some specific active marks could be required to reprogam the paternal pronucleus.

Similarly to mouse embryogenesis, we observed very low to undetectable levels of H3K36me3 in bovine embryos. This is in contrast to H3K9ac, which is present throughout bovine pre-implantation development (our work and ref. 45). Remarkably, H3K36me3 is absent from interphase nuclei at the time of EGA, i.e., the 8-cell stage, suggesting that regulation of transcriptional elongation might be independent of H3K36me3 as in mouse embryos.

The lack of H3K36me3 is intriguing as the mRNAs of the H3K36 methyltransferases NSD1 and Whsc1l1 are present in oocytes and 2-cell stage embryos in the mouse (our unpublished observations). The rapid disappearance of H3K36me3 from the maternal chromatin suggests that this might occur through an active mechanism of demethylation. Interestingly, although initially present in maternal chromatin, H2A.Z is also undetectable from embryonic chromatin within a few hours of fertilization.²⁷ Nashun and colleagues suggested that the timely disappearance of H2A.Z (and the canonical H2A) from the maternal embryonic chromatin is essential for normal progression through development in mouse. Lack of detection of H3.1 in the maternal chromatin immediately after fertilization has also been documented.⁵⁷ This phenomena seems to be conserved in the model systems analyzed so far, as the maternal chromatin of the plant Arabidopsis thaliana is equally subject to intensive remodeling, including the eviction of maternal histones.⁵⁸ The mechanisms and significance of these findings in mammals remain to be investigated, but it is possible that the maternal chromatin is also remodeled after fertilization to a greater extent than originally thought and that this process might have been overseen.

In the future, it will be important to investigate the effects of specific histone amino acids and/or their modifications on early mouse development through, for example, introduction of histone mutant mRNA as it has been done for H3 variants.⁵⁶

In summary, we have documented the distribution of four euchromatic marks during mouse pre-implantation development. Our data indicate that H3K36me3 does not seem to be a global mark for embryonic transcriptional activation in the two mammalian species that we analyzed. We show that H3K36me3 establishes an epigenetic asymmetry in the mouse zygote and is exclusively present in the maternal chromatin. Likewise, H3K4me3 shows a very early, initial asymmetry but then is detected in embryonic chromatin throughout mouse pre-implantation development. Furthermore, H2A.Z is acetylated during early embryogenesis, but in mice, this modification is undetectable at the stage of major EGA. The analysis of acetylated H2A.Z highlights species-specific temporal dynamics of H2A.Z acetylation during embryogenesis. Our work contributes to the understanding of the reprogramming process by defining the temporal dynamics and the components of the embryonic chromatin and their modifications during early mammalian development.

Materials and Methods

Collection of mouse embryos

Mouse embryos were collected from natural matings of CD1 crosses bred in a 12 h light cycle as previously described.⁵⁹ Zygotes, 2-cell stage, 4-cell stage embryos and morulas were obtained by puncturing the oviduct 12, 34, 46 and 58 h post-coitum (hpc), respectively. Blastocysts were collected by flushing the uterus with M2 medium (Sigma) 72 hpc. Experiments with animals were performed according to valid legislation in France.

Oocyte collection, in vitro maturation, fertilization and development of bovine embryos

Cattle ovaries were obtained from a slaughterhouse and transported to the laboratory in sterile PBS at 35°C within 3 h. Cumulus oocyte complexes (COCs) were aspirated in TCM 199 (Sigma) from 3 to 6 mm diameter follicles.⁶⁰ Compact COCs with several dense layers of cumulus cells were selected and washed three times in TCM 199. Maturation was performed in TCM 199 supplemented with 10% fetal calf serum (FCS; Life Technologies), 10 mg/ml porcine follicle-stimulating hormone/luteinizing hormone (Stimufol, Merial), and 1 mg/ml estradiol 17-β, for 22 h at 39°C in a humidified atmosphere of 5% CO2 in air. At the end of the maturation period, in vitro matured oocytes were co-incubated with heparin-capacitated frozen–thawed spermatozoa in Tyrode’s albumin lactate pyruvate medium for 18 h.⁶¹ After IVF, presumptive zygotes were cultured in SOF medium with 1% estrus cow serum and 1% BSA and incubated for 7 d at 39°C under 5% CO2 and 5% 02. Cleavage was assessed at day 2 after insemination and blastocyst formation was evaluated at day 7. For immunofluorescence analyses zygotes and embryos were analyzed at 18 h, 30 h, 40 h, 96 h and 6 d, which corresponded to the pronucleus (PN2/3), 2-cell, 4-cell, 8/16-cell and young blastocysts stages respectively.

Immunostaining

Mouse embryos were fixed immediately after collection. After removal of the zona pellucida with acid Tyrode solution (Sigma), embryos were washed three times in PBS and fixed as described.⁴⁷ Fixation was performed at 37° to ensure preservation of nuclear architecture. After permeabilization with 0.5% Triton, embryos were washed three times in PBS-T (0.1% Tween in PBS), blocked and incubated with the primary antibodies in 3% BSA 0.1% Tween in PBS for ~12 h at 4°C. Embryos were then washed twice in PBS-T, blocked for 30 min and incubated for 2 h at 25°C with the corresponding secondary antibodies in 3% BSA in PBS-T. After washing, embryos were mounted in Vectashield (Vector Laboratories) containing 4'-6-Diamidino-2-phenylindole (DAPI) for visualizing DNA. Bovine embryos were fixed following the same protocol, except that the zona pellucida was not removed. Primary antibodies used were: anti-H2A.Z (Upstate 07–594); anti-acetylated H2A.Z (abcam ab18262); anti-H3K9acetylated (abcam ab4441); anti-H3K36me3, (abcam ab9050); anti-H3K4me3 (Diagenode pAb-MEHAHS-024). Dilution of all primary antibodies was 1:250. Secondary antibodies used were: A488-conjugated goat anti-rabbit IgG, A555-conjugated goat anti-rabbit IgG, or A488-conjugated donkey anti-sheep IgG (Molecular probes).

Confocal analysis and microscopy

Confocal microscopy for mouse embryos was performed using a 63x oil objective in a Leica SP2 AOBS MP inverted microscope. Z-sections were taken every 1 μm. All the stainings were repeated independently at least two times with at least 8 embryos analyzed per stage. Bovine embryos were analyzed on a Zeiss Axiovert 200M inverted microscope equipped with a fluorescence module, ApoTome and B/W digital camera for imaging (MIMA2 platform: Microscopie et Imagerie des Microorganismes, Animaux et Aliments). The images were captured and processed using the AxioVision software (Zeiss).

Acknowledgments

We thank Y. Miyanari and A. Burton for critical reading of the manuscript, the IGBMC-ICS imaging facility for support and Pierre Adenot for providing access to the MIMA2 platform (Microscopie et Imagerie des Microorganismes, Animaux et Aliments) at INRA in Jouy-en-Josas. M.E.T.-P. acknowledges funding from ANR-09-Blanc-0114, EpiGeneSys NoE, ERC-Stg ‘NuclearPotency’ and the FP7 Marie-Curie Actions ITN Nucleosome4D. N.B. acknowledges funding from EU FP7 PLURISYS project (HEALTH-F4–2009–223485) and from the REVIVE consortium, an ANR “Laboratoire d’Excellence” program.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.