Establishing, maintaining and modifying DNA methylation patterns in plants and animals
Julie A. Law,Steven E. Jacobsen
DOI: https://doi.org/10.1038/nrg2719
IF: 59.581
2010-02-09
Nature Reviews Genetics
Abstract:Key PointsRecent studies have shown that RNA-directed DNA methylation (RdDM) in Arabidopsis thaliana not only requires the production of 24-nucleotide small interfering RNAs (siRNAs) and the de novo DNA methyltransferase DOMAINS REARRANGED METHYLTRANSFERASE 2 (DRM2) but also requires the production of RNA polymerase V (Pol V)-dependent intergenic non-coding (IGN) transcripts. Two crucial RdDM components, ARGONAUTE 4 (AGO4) and SUPPRESSOR OF TY INSERTION 5-LIKE (SPT5L), interact with Pol V-dependent transcripts, which suggests that they serve as a scaffold for the recruitment of the RdDM machinery. This process ultimately leads to DNA methylation and silencing at loci that produce siRNAs and IGN transcripts.Analyses of transposon expression and DNA methylation patterns in pollen grains and in embryo and endosperm tissues, respectively, suggest that genome-wide decreases in DNA methylation occur during male and female gametogenesis in A. thaliana, which might facilitate enhanced RdDM and transposon silencing in the sperm and egg cells by an unknown mechanism.Biochemical purification of DNA methyltransferase 3-like (DNMT3L) revealed an interaction between DNMT3L and unmethylated histone 3 lysine 4 (H3K4) tails. As DNMT3L also interacts with the DNMT3A de novo methyltransferase and because H3K4 methylation is anticorrelated with DNA methylation, a model has been proposed in which DNMT3L interaction with unmethylated H3K4 tails targets de novo methylation.Several recent findings suggest that Piwi-interacting RNAs (piRNAs) target de novo DNA methylation at transposons and other repetitive elements of the genome during male gametogenesis in mammals. piRNA populations isolated early in development were found to be enriched for such sequence elements, and mutations in MILI — a Piwi-clade Ago protein that binds piRNAs — showed DNA methylation defects at the stage of development at which de novo methylation is observed.Characterization of the Piwi clade of Ago proteins revealed the presence of symmetrical dimethylarginine modifications on several family members from Xenopus laevis, Drosophila melanogaster and mice. Tudor domains are known to interact with this modification, and Tudor domain-containing 1 (TDRD1), a protein with several Tudor domains, interacts with MILI and is required for DNA methylation and transposon silencing.In mammals, DNA methylation is maintained during DNA replication through the activity of DNMT1, which catalyses the methylation of hemimethylated CG sites in newly synthesized DNA. This activity depends heavily on the presence of ubiquitin-like plant homeodomain and RING finger domain 1 (UHRF1), a protein that specifically recognizes hemimethylated DNA and is proposed to recruit DNMT1 to chromatin.In A. thaliana, 5-methylcytosine DNA glycosylases and the base excision repair pathway catalyse active DNA demethylation during female gametogenesis and in vegetative plant tissues. Demethylation during gametogenesis is required for imprinting, whereas demethylation in vegetative tissues is proposed to combat robust DNA methylation by the RdDM pathway.In zebrafish there is evidence for an active DNA demethylation pathway that also involves DNA glycosylase activity and the base excision repair pathway. However, unlike in A. thaliana, in which methylated cytosines are directly recognized and removed, in zebrafish the methylated cytosine is first deaminated by the activation-induced cytosine deaminase (Aid)/apolipoprotein B mRNA-editing enzyme (Apobec) family of proteins, generating a G/T mismatch. This base is then removed by a thymine DNA glycosylase in what seems to be a tightly coupled manner.In mammals, mechanisms for active DNA demethylation remain unclear. However, an early model proposed a mechanism similar to that recently demonstrated in zebrafish, and a recent study showing that AID is necessary for the reduced levels of DNA methylation normally observed in primordial germ cells also supports this hypothesis. The discovery of the 5-hydroxymethylcytosine modification in certain mammalian cell types led to speculation that this modification could be a substrate for active DNA methylation.
genetics & heredity