Exploring functional significance of asymmetric DNA methylation in early mammalian development
DNA methylation is an epigenetic modification important for the regulation of transcriptional activity in processes like genomic imprinting, retrotransposon silencing and centromeric stabilisation. It is also crucial for correct embryonic development and the differentiation of embryonic stem cells (ESCs) into diverse cell types. Although in mammals DNA methylation occurs predominantly in the symmetric CG context, it has been shown that certain cell types and tissues (ESCs, oocytes, primordial germ cells) have substantial amounts of methylation outside of the CG dinucleotide, which is asymmetric. Its presence in early developmental stages related to toti- or pluri-potency raises the intriguing possibility that non-CG context methylation may have a role in differential gene expression during those stages, contributing to their transcriptional plasticity. I investigated the distribution, dynamics and functional significance of non-CG methylation in early mouse development. Most methods for DNA methylation analysis are either targeted for the analysis of CG methylation or, in the case of bisulfite sequencing, suffer from potentially confounding issues. I have compared existing approaches to detect methylation outside of CG context and developed novel tools, namely the use of antibodies against non-CG methylation, to either analyse global levels of non-CG methylation, or validate its genomic distribution. With these, I have evaluated the role of components of the DNA methylation machinery and have followed the dynamic changes of non-CG context methylation throughout development. My analysis reveals that the highest levels of non-CG methylation in the mouse are present in the mature oocyte and the zygote. The enzymes responsible for establishing and maintaining its levels are the de novo Dnmts (3a and 3b), among which the activity of Dnmt3a2 towards CH seems regulated, suggesting a specific rather than an unspecific role. In ES cells, CH methylation correlates with active histone marks e.g. H3K4me3, and inversely with H3K27me3. The distribution of mCH, both in ESC naïve and primed pluripotency states, is very heterogeneous, while its nuclear distribution is very homogenous. mCH is physically recognised by a number of pluripotency factors such as Oct4 and Sox2, as well as by other DNA modification-sensitive proteins like MeCP2, Foxk1 and Foxk2. Moreover, mCH cannot be hydroxylated by the Tet family of enzymes, and repels proteins involved in the initiation of base-excision repair (AID and RPA), thus potentially escaping active demethylation in the zygote. In summary, my results show that mCH is a valid methylation mark, with a functional significance in early mouse development.