Demethylation of somatic and testis-specific histone H2A and H2B genes in F9 embryonal carcinoma cells.

In contrast to many other genes containing a CpG island, the testis-specific H2B (TH2B) histone gene exhibits tissue-specific methylation patterns in correlation with gene activity. Characterization of the methylation patterns within a 20-kb segment containing the TH2A and TH2B genes in comparison with that in a somatic histone cluster revealed that: (i) the germ cell-specific unmethylated domain of the TH2A and TH2B genes is defined as a small region surrounding the CpG islands of the TH2A and TH2B genes and (ii) somatic histone genes are unmethylated in both liver and germ cells, like other genes containing CpG islands, whereas flanking sequences are methylated. Transfection of in vitro-methylated TH2B, somatic H2B, and mouse metallothionein I constructs into F9 embryonal carcinoma cells revealed that the CpG islands of the TH2A and TH2B genes were demethylated like those of the somatic H2A and H2B genes and the metallothionein I gene. The demethylation of those CpG islands became significantly inefficient at a high number of integrated copies and a high density of methylated CpG dinucleotides. In contrast, three sites in the somatic histone cluster, of which two sites are located in the long terminal repeat of an endogenous retrovirus-like sequence, were efficiently demethylated even at a high copy number and a high density of methylated CpG dinucleotides. These results suggest two possible mechanisms for demethylation in F9 cells and methylation of CpG islands of the TH2A and TH2B genes at the postblastula stage during embryogenesis.     

The essential role of histone H3 Lys9 di-methylation and MeCP2 binding in MGMT silencing with poor DNA methylation of the promoter CpG island

Silencing of the O (6)-methylguanine-DNA methyltransferase (MGMT) gene, a key to DNA repair, is involved in carcinogenesis. Recent studies have focused on DNA hypermethylation of the promoter CpG island. However, cases showing silencing with DNA hypomethylation certainly exist, and the mechanism involved is not elucidated. To clarify this mechanism, we examined the dynamics of DNA methylation, histone acetylation, histone methylation, and binding of methyl-CpG binding proteins at the MGMT promoter region using four MGMT negative cell lines with various extents of DNA methylation. Histone H3K9 di-methylation (H3me2K9), not tri-methylation, and MeCP2 binding were commonly seen in all MGMT negative cell lines regardless of DNA methylation status. 5Aza-dC, but not TSA, restored gene expression, accompanied by a decrease in H3me2K9 and MeCP2 binding. In SaOS2 cells with the most hypomethylated CpG island, 5Aza-dC decreased H3me2K9 and MeCP2 binding with no effect on DNA methylation or histone acetylation. H3me2K9 and DNA methylation were restricted to in and around the island, indicating that epigenetic modification at the promoter CpG island is critical. We conclude that H3me2K9 and MeCP2 binding are common and more essential for MGMT silencing than DNA hypermethylation or histone deacetylation. The epigenetic mechanism leading to silent heterochromatin at the promoter CpG island may be the same in different types of cancer irrespective of the extent of DNA methylation.     

Mutagenesis in mice of nuclear hormone receptor binding sites in the Igf2/H19 imprinting control region

The H19/Igf2 imprinting control region (ICR) is a DNA methylation-dependent chromatin insulator in somatic cells. The hypomethylated maternally inherited ICR binds the insulator protein CTCF at four sites, and blocks activity of the proximal Igf2 promoter by insulating it from the shared distal enhancers. The hypermethylated paternally inherited ICR lacks CTCF binding and insulator activity, but induces methylation-silencing of the paternal H19 promoter. The paternal-specific methylation of the ICR is established in the male germ cells, while the ICR emerges from the female germ line in an unmethylated form. Despite several attempts to find cis-regulatory elements, it is still unknown what determines these male and female germ cell-specific epigenetic modifications. We recently proposed that five in vivo footprints spanning fifteen half nuclear hormone receptor (NHR) binding sites within the ICR might be involved, and here we report on the effects of mutagenizing all of these half sites in mice. No effect was obtained--in the female and male germ lines the mutant ICR remained hypomethylated and hypermethylated, respectively. The ICR imprinting mechanism remains undefined.