Genomic patterns of DNA methylation: targets and function of an epigenetic mark

Methylation of cytosines can mediate epigenetic gene silencing and is the only known DNA modification in eukaryotes. Recent efforts to map DNA methylation across mammalian genomes revealed limited DNA methylation at regulatory regions but widespread methylation in intergenic regions and repeats. This is consistent with the idea that hypermethylation is the default epigenetic state and serves in maintaining genome integrity. DNA methylation patterns at regulatory regions are generally stable, but a minor subset of regulatory regions show variable DNA methylation between cell types, suggesting an additional dynamic component. Such promoter de novo methylation might be involved in the maintenance rather than the initiation of silencing of defined genes during development. How frequently such dynamic methylation occurs, its biological relevance and the pathways involved deserve investigation.

Introduction

In prokaryotes, DNA can be methylated on both cytosines and adenines and this methylation is involved in various processes, including DNA repair and defense against foreign DNA [1]. Eukaryotes show methylation almost exclusively at cytosines: in mammals it occurs only in the context of CpG dinucleotides (CpGs), while in fungi (e.g. Neurospora crassa) and plants (e.g. Arabidopsis thaliana) methylation is seen in various symmetrical and asymmetrical sequence contexts (reviewed in [2]). However, cytosine methylation does not occur in all eukaryotes, as it is absent in Saccharomyces cerevisiae and in many invertebrates, like the nematode Caenorhabditis elegans. Regarding insects, low levels of cytosine methylation have been reported in Drosophila melanogaster and substantial methylation has been found in the honey bee Apis mellifera [3].

Four DNA methyltransferases (DNMTs) sharing a conserved DNMT domain have been identified in mammals. The founding member, DNMT1, maintains DNA methylation during replication by copying the DNA methylation of the old DNA strand onto the newly synthesized strand [4]. DNMT3a and DNMT3b are responsible for de novo methylation, as they are able to target unmethylated CpG sites [5]. They also cooperate with DNMT1 to propagate methylation patterns during cell division [6]. DNMT2 has only weak DNA methyltransferase activity in vitro and has recently been shown to efficiently methylate a tRNA [7^•].

DNA methylation is generally associated with a repressed chromatin state and inhibition of promoter activity. Two models of repression have been proposed: first, cytosine methylation can prevent the binding of some transcription factors, and second, DNA methylation can affect chromatin states indirectly through the recruitment of methyl-CpG-binding proteins (MBPs) [8]. DNA methylation is essential for mammalian development, as shown by the lethality of various DNMT deficiencies in mice [5, 9]. DNA-methylation-mediated repression has been directly implicated in X-chromosome inactivation and genomic imprinting (see review by Edwards and Ferguson-Smith in this issue); however, other functions of DNA methylation in developmentally regulated gene expression remain less definite.

Mammalian genomes are globally depleted for CpGs, except at short DNA stretches called CpG islands, which are frequently associated with gene promoters. This unequal distribution of CpGs needs to be considered when interpreting global maps of DNA methylation, because the amount of methylated cytosines at a given region depends both on the degree of methylation and on the density of CpGs.

Here we review how recent advances in determining the sites of DNA methylation on a genome-wide scale have given new insights into the biological function of DNA methylation in maintaining genome integrity and cell identity. Due to space limitations we will not focus on aberrant DNA methylation in cancer, for which we refer the reader to recent summary articles [10, 11].