Jumonji domain-containing protein 3 regulates histone 3 lysine 27 methylation during bovine preimplantation development

Abstract

Understanding the mechanisms of epigenetic remodeling that follow fertilization is a fundamental step toward understanding the bases of early embryonic development and pluripotency. Extensive and dynamic chromatin remodeling is observed after fertilization, including DNA methylation and histone modifications. These changes underlie the transition from gametic to embryonic chromatin and are thought to facilitate embryonic genome activation. In particular, trimethylation of histone 3 lysine 27 (H3K27me3) is associated with gene-specific transcription repression. Global levels of this epigenetic mark are high in oocyte chromatin and decrease to minimal levels at the time of embryonic genome activation. We provide evidence that the decrease in H3K27me3 observed during early development is cell-cycle independent, suggesting an active mechanism for removal of this epigenetic mark. Among H3K27me3-specific demethylases, Jumonji domain-containing protein 3 (JMJD3), but not ubiquitously transcribed tetratricopeptide repeat X (UTX), present high transcript levels in oocytes. Soon after fertilization JMJD3 protein levels increase, concurrent with a decrease in mRNA levels. This pattern of expression suggests maternal inheritance of JMJD3. Knockdown of JMJD3 by siRNA injection in parthenogenetically activated metaphase II oocytes resulted in inhibition of the H3K27me3 decrease normally observed in preimplantation embryos. Moreover, knockdown of JMJD3 in oocytes reduced the rate of blastocyst development. Overall, these results indicate that JMJD3 is involved in active demethylation of H3K27me3 during early embryo development and that this mark plays an important role during the progression of embryos to blastocysts.

Fig. 1.

H3K27me3 demethylation is independent of cell-cycle progression. H3K27me3 levels in bovine preimplantation embryos treated with aphidicolin. (A) Representative images of immunostaining using H3K27me3 antibody in embryos treated with and without aphidicolin at different stages of development. (B) Quantification of H3K27me3 nuclear fluorescence intensity using MetaMorph software and normalized against background. DNA was stained with Hoechst. hpa, hours postactivation.

Fig. 2.

JMJD3 and UTX gene-expression patterns in bovine oocytes and preimplantation embryos. (A) UTX transcript levels determined by quantitative RT-PCR at different stages of preimplantation development. Negligible levels of UTX were observed before the morula stage. (B) JMJD3 transcript levels determined by quantitative RT-PCR at different stages of preimplantation development. High levels of JMJD3 are found in oocytes, decreasing in early embryos and increasing again in blastocysts. (C) Immunoblot of GV and MII oocytes and two- to four-cell–stage embryos using JMJD3 antibody. A clear reactive band is observed in two-cell embryos but not in oocytes. Actin levels in the same samples are shown as loading controls. (D) Immunoblot of two-cell and eight- to 16-cell–stage embryos after blocking (+) or not (−) with JMJD3 peptide. The disappearance of the JMJD3 band when the antibody is preincubated with the blocking peptide indicates the antibody specificity. Results also confirm the presence of JMJD3 in cleavage stage embryos. (E) Immunofluorescence staining using an anti-JMJD3 antibody at different stages of oocyte and embryo development. Protein expression of JMJD3 in bovine preimplantation embryos reveals an increase of protein synthesis from the PN stage onwards. Bl, blastocyst; 2C, two-cell; 4C, four-cell; 8C, eight-cell; 16C, 16-cell; Mo, morula.

Fig. 3.

JMJD3 is required for H3K27me3 demethylation during preimplantation development. (A) Schematic representation of predicted JMJD3 mRNA. Two siRNA species were designed to target different regions of JMJD3 near the 5′ (JMJD3-1077) and to 3′ (JMJD3-5979) ends. Arrow indicates RT-PCR amplification sites. (B) Experimental outline of siRNA injection in bovine MII oocytes just before parthenogenetic activation and developmental stages analyzed. (C) JMJD3 mRNA levels analyzed by real-time RT-PCR in four-cell embryos after injection of each species of siRNA (JMJD3-1077 and JMJD3-5979) in bovine MII oocytes. Control group oocytes were injected with control siRNA without specificity to any gene in the genome. The letters “a” and “b” indicate significantly different mRNA levels (P < 0.05). (D) H3K27me3 (Upper) and H3K9me3 (Lower) immunostaining of four- and eight-cell–stage bovine embryos. Control siRNA oocytes were injected with control siRNA (without specificity against any gene); JMJD3-siRNA oocytes were injected with JMJD3-5979 siRNA and the DNA was stained with Hoechst. JMJD3 siRNA injection results in increased H3K27me3 levels without affecting H3K9me3 abundance.

Fig. 4.

Knockdown of JMJD3 affects development to blastocyst. (A) Cleavage rate and (B) blastocyst rate in parthenos evaluated at 40 h and 7 d, respectively, after JMJD3 siRNA injection of siRNA (JMJD3-1077 and JMJD3-5979) in bovine MII oocytes. Control group oocytes were injected with control siRNA lacking specificity against any gene in the genome. The letters “a” and “b” indicate significantly different blastocyst rates (P < 0.05). (C) Blastocyst cell number determined by counting cells in Hoechst-stained embryos. The letters “y” and “z” indicate a tendency for decreased cell number in JMJD3-injected embryos (P < 0.1). (D) JMJD3 mRNA expression levels analyzed by real-time RT-PCR at the blastocyst stage in the aforementioned groups and quantified relative to RPS18, a gene not affected by JMJD3 siRNA injection, to account for the difference in cell number between treatment and control blastocysts. The letters “a” and “b” indicate significantly different JMJD3 mRNA levels (P < 0.05).