DNMT1 regulates the timing of DNA methylation by DNMT3 in an enzymatic activity-dependent manner in mouse embryonic stem cells

DNA methylation (DNAme; 5-methylcytosine, 5mC) plays an essential role in mammalian development, and the 5mC profile is regulated by a balance of opposing enzymatic activities: DNA methyltransferases (DNMTs) and Ten-eleven translocation dioxygenases (TETs). In mouse embryonic stem cells (ESCs), de novo DNAme by DNMT3 family enzymes, demethylation by the TET-mediated conversion of 5mC to 5-hydroxymethylation (5hmC), and maintenance of the remaining DNAme by DNMT1 are actively repeated throughout cell cycles, dynamically forming a constant 5mC profile. Nevertheless, the detailed mechanism and physiological significance of this active cyclic DNA modification in mouse ESCs remain unclear. Here by visualizing the localization of DNA modifications on metaphase chromosomes and comparing whole-genome methylation profiles before and after the mid-S phase in ESCs lacking Dnmt1 (1KO ESCs), we demonstrated that in 1KO ESCs, DNMT3-mediated remethylation was interrupted during and after DNA replication. This results in a marked asymmetry in the distribution of 5hmC between sister chromatids at mitosis, with one chromatid being almost no 5hmC. When introduced in 1KO ESCs, the catalytically inactive form of DNMT1 (DNMT1CI) induced an increase in DNAme in pericentric heterochromatin and the DNAme-independent repression of IAPEz, a retrotransposon family, in 1KO ESCs. However, DNMT1CI could not restore the ability of DNMT3 to methylate unmodified dsDNA de novo in S phase in 1KO ESCs. Furthermore, during in vitro differentiation into epiblasts, 1KO ESCs expressing DNMT1CI showed an even stronger tendency to differentiate into the primitive endoderm than 1KO ESCs and were readily reprogrammed into the primitive streak via an epiblast-like cell state, reconfirming the importance of DNMT1 enzymatic activity at the onset of epiblast differentiation. These results indicate a novel function of DNMT1, in which DNMT1 actively regulates the timing and genomic targets of de novo methylation by DNMT3 in an enzymatic activity-dependent and independent manner, respectively.

SPEN is required for Xist upregulation during initiation of X chromosome inactivation

At initiation of X chromosome inactivation (XCI), Xist is monoallelically upregulated from the future inactive X (Xi) chromosome, overcoming repression by its antisense transcript Tsix. Xist recruits various chromatin remodelers, amongst them SPEN, which are involved in silencing of X-linked genes in cis and establishment of the Xi. Here, we show that SPEN plays an important role in initiation of XCI. Spen null female mouse embryonic stem cells (ESCs) are defective in Xist upregulation upon differentiation. We find that Xist-mediated SPEN recruitment to the Xi chromosome happens very early in XCI, and that SPEN-mediated silencing of the Tsix promoter is required for Xist upregulation. Accordingly, failed Xist upregulation in Spen−/− ESCs can be rescued by concomitant removal of Tsix. These findings indicate that SPEN is not only required for the establishment of the Xi, but is also crucial in initiation of the XCI process.

The distinct effects of MEK and GSK3 inhibition upon the methylome and transcriptome of mouse embryonic stem cells

Mouse embryonic stem cells (mESCs) were first cultured in vitro in serum-containing medium with leukaemia inhibitory factor, in which they exhibit heterogeneous expression of both pluripotency and some early differentiation markers. Over the last decade, however, it has become commonplace to grow mESCs with inhibitors of MEK and GSK3 signalling, which together elicit a more homogeneously 'naive' state of pluripotency. Whilst 2i/L-cultured mESCs have been shown to be globally hypomethylated, a comprehensive understanding of the distinct effects of these signalling inhibitors upon the DNA methylome is still lacking. Here we carried out whole genome bisulphite and RNA sequencing of mESCs grown with MEK or GSK3 inhibition alone, including different time points and concentrations of MEK inhibitor treatment. This demonstrated that MEK inhibition causes a dose-dependent impairment of maintenance methylation via loss of UHRF1 protein, as well as rapid impairment of de novo methylation. In contrast, GSK3 inhibition triggers impairment of de novo methylation alone, and consequent hypomethylation is enriched at enhancers with a 2i/L-specific chromatin signature and coincides with upregulation of nearby genes. Our study provides detailed insights into the epigenetic and transcriptional impacts of inhibiting MEK or GSK3 signalling in mouse pluripotent cells.