Dynamic replacement of histone H3 variants reprograms epigenetic marks in early mouse embryos

Abstract

Upon fertilization, reprogramming of gene expression is required for embryo development. This step is marked by DNA demethylation and changes in histone variant composition. However, little is known about the molecular mechanisms causing these changes and their impact on histone modifications. We examined the global deposition of the DNA replication-dependent histone H3.1 and H3.2 variants and the DNA replication-independent H3.3 variant after fertilization in mice. We showed that H3.3, a euchromatic marker of gene activity, transiently disappears from the maternal genome, suggesting erasure of the oocyte-specific modifications carried by H3.3. After fertilization, H3.2 is incorporated into the transcriptionally silent heterochromatin, whereas H3.1 and H3.3 occupy unusual heterochromatic and euchromatin locations, respectively. After the two-cell stage, H3.1 and H3.3 variants resume their usual respective locations on heterochromatin and euchromatin. Preventing the incorporation of H3.1 and H3.2 by knockdown of the histone chaperone CAF-1 induces a reciprocal increase in H3.3 deposition and impairs heterochromatin formation. We propose that the deposition of different H3 variants influences the functional organization of chromatin. Taken together, these findings suggest that dynamic changes in the deposition of H3 variants are critical for chromatin reorganization during epigenetic reprogramming.

Figure 1. H3.3 disappears from the oocyte genome immediately after fertilization.

(A) Deposition of Flag-H3 variants in growing oocytes. Each Flag-H3 variant mRNA was microinjected into the cytoplasm of growing oocytes, and the oocytes were immunostained 24 h later with anti-Flag antibody. DNA was counterstained with PI. Scale bar, 20 µm. (B) Analysis of Flag-H3.3 dynamics before and after fertilization. Fully grown oocytes (GV) were microinjected with Flag-H3.3 mRNA in the presence of IBMX, which inhibits meiotic maturation. Five hours later, Flag-H3.3 was detected in GV-stage oocytes, and 2 h and 18 h after the removal of IBMX it was still present in MI- and MII-stage oocytes, respectively. Merged images are shown in the inset, where the arrows show little or no signal for Flag-H3.3 in the heterochromatic regions. Four hours after fertilization in vitro (1-cell G1), Flag-H3.3 was not detected in the whole maternal pronucleus (m, arrowheads), whereas it was detected in the paternal pronucleus (p). Scale bar, 20 µm. (C) Analysis of H3.3 dynamics in the maternal genome during the early one-cell stage. The embryos were collected at 0, 1, 2, 3, and 4 h after insemination and examined for Flag-H3.3 incorporation into the maternal genome. The polar body genome from each embryo is shown by arrows (2 h) or in the inset (3 and 4 h). Scale bar, 20 µm. (D) Analysis of Flag-H4 dynamics before and after fertilization. The nuclear deposition of Flag-H4 was examined using the same procedure as for Flag-H3.3 in (B). Flag-H4 was not detected in the maternal pronucleus (m, arrowheads), but was detected in the paternal pronucleus. (E) The nuclei of growing oocytes (GO) and GV-stage oocytes, the chromosomes of MI- and MII-stage oocytes, and the female pronuclei of one-cell G1 embryos (1-cell G1) from a Zp3-Flag-H3.3 transgenic mouse were analyzed by immunostaining with anti-Flag antibody. Scale bar, 10 µm.

Figure 2. Incorporation of H3 variants and H4 into nuclei at an early stage after fertilization.

MII-stage oocytes were microinjected with mRNA encoding a Flag-H3 variant or Flag-H4, and then fertilized 2 h later. Four hours after fertilization, one-cell embryos were stained with anti-Flag antibodies; DNA was stained with DAPI. (A) No signal for any Flag-H3 variant was detected in the maternal pronuclei of early one-cell embryos. (B) Flag-H4 was also undetectable in the maternal pronucleus, although it was clearly detected in the paternal pronucleus. m, maternal pronucleus; p, paternal pronucleus.

Figure 3. Incorporation of Flag-H3 variants into nuclei in preimplantation embryos.

(A) Immunofluorescence analysis of Flag-H3 variants in one- and two-cell embryos. MII-stage oocytes were microinjected with each Flag-H3 variant mRNA and then fertilized 2 h later. Staining with anti-Flag antibody was observed in the one- and two-cell embryos at 12 h and 28 h post-fertilization, respectively. m, maternal genome; p, paternal genome. The DNA was stained with PI. Scale bar, 20 µm. (B) The fluorescence intensity of Flag-H3.1 and Flag-H3.2 in the two-cell embryo was determined as described in the Materials and Methods section. The numbers of nuclei examined were 80 and 64 for the Flag-H3.1- and Flag-H3.2-injected embryos, respectively. (C) Quantitative RT-PCR analysis for H3.1 or H3.2 mRNA in two-cell embryos injected with or without Flag-H3.1 or Flag-H3.2 mRNA prior to fertilization. RT-PCR was performed using the primer pairs for the H3.1 or H3.2 open reading frames, which measured the total amounts of endogenous and injected mRNAs. The target mRNA level was normalized to the amount of exogenous rabbit α-globin mRNA in the same sample. The average value of duplicate experiments is shown. (D) Analysis of Flag-H3.1 and Flag-H3.2 mRNA and protein level in two-cell embryos by RT-PCR and immunoblotting (WB). RT-PCR was performed using a common primer pair recognizing exogenous Flag-H3.1 and Flag-H3.2 mRNA, but not endogenous H3.1 or H3.2. Endogenous cyclin A2 mRNA served as a control. Flag-H3.1 and Flag-H3.2 proteins were detected by immunoblotting with anti-Flag antibody. The relative values of the band densities of H3.1 vs. H3.2 were 0.63 vs. 1.0. Antibody against α-tubulin was used for the loading control. The experiments were conducted twice with similar results. (E) Immunofluorescence of Flag-H3 variants in late preimplantation embryos. Each Flag-H3 variant mRNA was microinjected into one blastomere of two-cell embryos. Therefore, signals for the Flag-H3 variants were detectable in half of the blastomeres. Deposition of each H3 variant in embryos at the four-cell, eight-cell, and blastocyst stages was examined at 44 h, 56 h, and 80 h post-fertilization, respectively. Scale bar, 20 µm.

Figure 4. Intranuclear distribution of H3 variants changes dynamically during preimplantation development.

(A) Immunodetection of Flag-H3 variants at the two-cell and blastocyst stages. Heterochromatin domains, which are characterized by DNA-dense foci, are confined to the peripheries of nucleoli at the two-cell stage, whereas they localize at discrete foci in the nucleoplasm at the blastocyst stage. Arrowheads indicate the typical heterochromatin foci. Scale bar, 10 µm (two-cell stage), 5 µm (Blastocyst). (B) Fluorescence intensity profiles of Flag-H3 variants and DNA in the nucleus of two-cell and blastocyst-stage embryos. Lines were drawn on the images of nuclei (see merged images in (A)), and the pixel intensities for Flag-H3 variant staining (green) and DNA staining (magenta) were quantified along the lines. The horizontal axis represents the distance from the starting (farthest left) point of the analysis on the line.

Figure 5. Knockdown of CAF-1 p150 impairs distribution of H3 variants and leads to the inhibition of heterochromatin formation.

One-cell embryos were microinjected with siRNA targeting CAF-1 p150 (sip150#1 or sip150#2) or with siRNA targeting EGFP (siEGFP) as a control. (A) The morphology of embryos treated with siRNA at 96 h post-fertilization and the percentage of treated embryos that successfully progressed to the blastocyst stage are shown; n indicates the number of embryos examined. Scale bar, 100 µm. (B) The siRNA-treated embryos were microinjected with each Flag-H3 variant mRNA at the two-cell stage. The mRNAs were microinjected into one blastomere of two-cell embryos. Signals for the Flag-H3 variants were detectable in half of the blastomeres. Embryos that developed to the morula stage were immunostained with anti-Flag antibody, and the DNA was stained with DAPI. Scale bar, 20 µm. (C) Quantitative analysis of Flag-H3.3 staining intensity in the siRNA-treated morulae. The numbers of nuclei examined were 69 and 50 for the siEGFP- and sip150-treated embryos, respectively. (D) Fluorescence intensity analysis of DNA and Flag-H3.3 in siRNA-treated morulae. Arrowheads indicate the typical chromatin regions in which DNA is densely stained. Scale bar, 10 µm. (E) Immuno-DNA-FISH analysis for localizing major satellites and Flag-H3.3 in the nucleus of sip150-treated embryos. Arrows indicate the typical major satellites. Scale bar, 10 µm.

Figure 6. Distribution of H3 variants affects the pattern of histone modifications at late preimplantation stages.

(A, B) Quantitative analysis of the fluorescence intensity of H3K9me2 (A) and H3K9ac (B) in entire nuclei of sip150-treated embryos. Embryos that developed to the two-, four-, eight-cell, and morula stage were immunostained with anti-H3K9me2 and anti-H3K9ac antibodies, and the fluorescence intensities were determined as described in the Materials and Methods section. More than 45 nuclei were examined for each group. (C, D) Immunostaining analysis for H3K9me2 (C) and H3K9ac (D) in the nuclei of siEGFP- and sip150-treated morulae. Representative images of the embryos are shown. Scale bar, 10 µm.