Lashings of DNA methylation, forkfuls of chromatin remodeling.

Abstract

Epigenetic control of gene expression in animals and plants is often correlated with changes in cytosine methylation at specific chromosomal loci. This leads to the irreversible promoter silencing of many genes, transposons, and endogenous retroviruses (Bestor 2000; Martienssen and Colot 2001). How DNA methylation patterns are set up during development, especially in the context of chromatin, is not well understood. Cytosine methyltransferases catalyze the transfer of a methyl group onto the C-5 position of cytosine from the universal cofactor SAM (Bestor 2000). Two main types of methyltransferase activity exist in mammals: a de novo activity and a maintenance activity. Dnmt3a and Dnmt3b have been identified as de novo methyltransferases, which methylate cytosine at CpG dinucleotides on both strands with little sequence specificity (Okano et al. 1998; Lyko et al. 1999). Dnmt1 is predominantly a postreplicative maintenance methyltransferase, which recognizes hemimethylated substrates, and acts to restore methylated cytosines at CpGs on the newly duplicated strand (Lyko et al. 1999; Bestor 2000). Mutation studies in mice have shown that the methyltransferase genes are necessary for embryonic (Dnmt1 and Dnmt3b) and postnatal (Dnmt3a) development (Lei et al. 1996; Okano et al. 1999). All three enzymes have been shown to methylate naked DNA substrates in vitro. However, complicating the issue is the in vivo evidence suggesting that patterns of DNA methylation are established and maintained within a nucleosomal infrastructure (Bird and Wolffe 1999). As yet we know little about the mechanistic basis of Dnmt3a and Dnmt3b activity in vivo, but in the case of Dnmt1 it has been shown that this maintenance methyltransferase enzyme can be targeted to PCNA, an auxiliary component of the DNA replication complex (Chuang et al. 1997). It is probable that chromatin assembly precedes DNA methylation behind the replication fork. Assembly of DNA into nucleosomes can prevent access by methyltransferases to their in vivo substrates (Kladde and Simpson 1994). Yet in the genome, methylated CpGs are distributed with equal probability either in the DNA wrapped around the nucleosome or in the linker region (Reik et al. 2001).