Dynamic nucleosome remodeling mediated by YY1 underlies early mouse development.

Nucleosome positioning can alter the accessibility of DNA-binding proteins to their cognate DNA elements, and thus its precise control is essential for cell identity and function. Mammalian preimplantation embryos undergo temporal changes in gene expression and cell potency, suggesting the involvement of dynamic epigenetic control during this developmental phase. However, the dynamics of nucleosome organization during early development are poorly understood. In this study, using a low-input MNase-seq method, we show that nucleosome positioning is globally obscure in zygotes but becomes well defined during subsequent development. Down-regulation of the chromatin assembly in embryonic stem cells can partially reverse nucleosome organization into a zygote-like pattern, suggesting a possible link between the chromatin assembly pathway and fuzzy nucleosomes in zygotes. We also reveal that YY1, a zinc finger-containing transcription factor expressed upon zygotic genome activation, regulates the de novo formation of well-positioned nucleosome arrays at the regulatory elements through identifying YY1-binding sites in eight-cell embryos. The YY1-binding regions acquire H3K27ac enrichment around the eight-cell and morula stages, and YY1 depletion impairs the morula-to-blastocyst transition. Thus, our study delineates the remodeling of nucleosome organization and its underlying mechanism during early mouse development.

Histone H3K36me2 and H3K36me3 form a chromatin platform essential for DNMT3A-dependent DNA methylation in mouse oocytes.

Establishment of the DNA methylation landscape of mammalian oocytes, mediated by the DNMT3A-DNMT3L complex, is crucial for reproduction and development. In mouse oocytes, high levels of DNA methylation occur exclusively in the transcriptionally active regions, with moderate to low levels of methylation in other regions. Histone H3K36me3 mediates the high levels of methylation in the transcribed regions; however, it is unknown which histone mark guides the methylation in the other regions. Here, we show that, in mouse oocytes, H3K36me2 is highly enriched in the X chromosome and is broadly distributed across all autosomes. Upon H3K36me2 depletion, DNA methylation in moderately methylated regions is selectively affected, and a methylation pattern unique to the X chromosome is switched to an autosome-like pattern. Furthermore, we find that simultaneous depletion of H3K36me2 and H3K36me3 results in global hypomethylation, comparable to that of DNMT3A depletion. Therefore, the two histone marks jointly provide the chromatin platform essential for guiding DNMT3A-dependent DNA methylation in mouse oocytes.

TFB2M and POLRMT are essential for mammalian mitochondrial DNA replication.

Two classes of replication intermediates have been observed from mitochondrial DNA (mtDNA) in many mammalian tissue and cells with two-dimensional agarose gel electrophoresis. One is assigned to leading-strand synthesis in the absence of synchronous lagging-strand synthesis (strand-asynchronous replication), and the other has properties of coupled leading- and lagging-strand synthesis (strand-coupled replication). While strand-asynchronous replication is primed by long noncoding RNA synthesized from a defined transcription initiation site, little is known about the commencement of strand-coupled replication. To investigate it, we attempted to abolish strand-asynchronous replication in cultured human cybrid cells by knocking out the components of the transcription initiation complexes, mitochondrial transcription factor B2 (TFB2M/mtTFB2) and mitochondrial RNA polymerase (POLRMT/mtRNAP). Unexpectedly, removal of either protein resulted in complete mtDNA loss, demonstrating for the first time that TFB2M and POLRMT are indispensable for the maintenance of human mtDNA. Moreover, a lack of TFB2M could not be compensated for by mitochondrial transcription factor B1 (TFB1M/mtTFB1). These findings indicate that TFB2M and POLRMT are crucial for the priming of not only strand-asynchronous but also strand-coupled replication, providing deeper insights into the molecular basis of mtDNA replication initiation.

A histone H3.3K36M mutation in mice causes an imbalance of histone modifications and defects in chondrocyte differentiation.

Histone lysine-to-methionine (K-to-M) mutations have been identified as driver mutations in human cancers. Interestingly, these 'oncohistone' mutations inhibit the activity of histone methyltransferases. Therefore, they can potentially be used as versatile tools to investigate the roles of histone modifications. In this study, we generated a genetically engineered mouse line in which an H3.3K36M mutation could be induced in the endogenous H3f3b gene. Since H3.3K36M has been identified as a causative mutation of human chondroblastoma, we induced this mutation in the chondrocyte lineage in mouse embryonic limbs. We found that H3.3K36M causes a global reduction in H3K36me2 and defects in chondrocyte differentiation. Importantly, the reduction of H3K36me2 was accompanied by a collapse of normal H3K27me3 distribution. Furthermore, the changes in H3K27me3, especially the loss of H3K27me3 at gene regulatory elements, were associated with the mis-regulated expression of a set of genes important for limb development, including HoxA cluster genes. Thus, through the in vivo induction of the H3.3K36M mutation, we reveal the importance of maintaining the balance between H3K36me2 and H3K27me3 during chondrocyte differentiation and limb development.

Reprogramming of the histone H3.3 landscape in the early mouse embryo.

Epigenetic reprogramming of the zygote involves dynamic incorporation of histone variant H3.3. However, the genome-wide distribution and dynamics of H3.3 during early development remain unknown. Here, we delineate the H3.3 landscapes in mouse oocytes and early embryos. We unexpectedly identify a non-canonical H3.3 pattern in mature oocytes and zygotes, in which local enrichment of H3.3 at active chromatin is suppressed and H3.3 is relatively evenly distributed across the genome. Interestingly, although the non-canonical H3.3 pattern forms gradually during oogenesis, it quickly switches to a canonical pattern at the two-cell stage in a transcription-independent and replication-dependent manner. We find that incorporation of H3.1/H3.2 mediated by chromatin assembly factor CAF-1 is a key process for the de novo establishment of the canonical pattern. Our data suggest that the presence of the non-canonical pattern and its timely transition toward a canonical pattern support the developmental program of early embryos.

Zfp281 Shapes the Transcriptome of Trophoblast Stem Cells and Is Essential for Placental Development.

Placental development is a key event in mammalian reproduction and embryogenesis. However, the molecular basis underlying placental development is not fully understood. Here, we conduct a forward genetic screen to identify regulators for extraembryonic development and identify Zfp281 as a key factor. Zfp281 overexpression in mouse embryonic stem cells facilitates the induction of trophoblast stem-like cells. Zfp281 is preferentially expressed in the undifferentiated trophoblast stem cell population in an FGF-dependent manner, and disruption of Zfp281 in mice causes severe defects in early placental development. Consistently, Zfp281-depleted trophoblast stem cells exhibit defects in maintaining the transcriptome and differentiation capacity. Mechanistically, Zfp281 interacts with MLL or COMPASS subunits and occupies the promoters of its target genes. Importantly, ZNF281, the human ortholog of this factor, is required to stabilize the undifferentiated status of human trophoblast stem cells. Thus, we identify Zfp281 as a conserved factor for the maintenance of trophoblast stem cell plasticity.