DNA methylation: evolution of a bacterial immune function into a regulator of gene expression and genome structure in higher eukaryotes

The amino acid sequence of mammalian DNA methyltransferase has been deduced from the nucleotide sequence of a cloned cDNA. It appears that the mammalian enzyme arose during evolution via fusion of a prokaryotic restriction methyltransferase gene and a second gene of unknown function. Mammalian DNA methyltransferase currently comprises an N-terminal domain of about 1000 amino acids that may have a regulatory role and a C-terminal 570 amino acid domain that retains similarities to bacterial restriction methyltransferases. The sequence similarities among mammalian and bacterial DNA cytosine methyltransferases suggest a common evolutionary origin. DNA methylation is uncommon among those eukaryotes having genomes of less than 10(8) base pairs, but nearly universal among large-genome eukaryotes. This and other considerations make it likely that sequence inactivation by DNA methylation has evolved to compensate for the expansion of the genome that has accompanied the development of higher plants and animals. As methylated sequences are usually propagated in the repressed, nuclease-insensitive state, it is likely that DNA methylation compartmentalizes the genome to facilitate gene regulation by reducing the total amount of DNA sequence that must be scanned by DNA-binding regulatory proteins. DNA methylation is involved in immune recognition in bacteria but appears to regulate the structure and expression of the genome in complex higher eukaryotes. I suggest that the DNA-methylating system of mammals was derived from that of bacteria by way of a hypothetical intermediate that carried out selective de novo methylation of exogenous DNA and propagated the methylated DNA in the repressed state within its own genome.(ABSTRACT TRUNCATED AT 250 WORDS)     

Association of ATRX with pericentric heterochromatin and the Y chromosome of neonatal mouse spermatogonia

Background  

Establishment of chromosomal cytosine methylation and histone methylation patterns are critical epigenetic modifications required for heterochromatin formation in the mammalian genome. However, the nature of the primary signal(s) targeting DNA methylation at specific genomic regions is not clear. Notably, whether histone methylation and/or chromatin remodeling proteins play a role in the establishment of DNA methylation during gametogenesis is not known. The chromosomes of mouse neonatal spermatogonia display a unique pattern of 5-methyl cytosine staining whereby centromeric heterochromatin is hypo-methylated whereas chromatids are strongly methylated. Thus, in order to gain some insight into the relationship between global DNA and histone methylation in the germ line we have used neonatal spermatogonia as a model to determine whether these unique chromosomal DNA methylation patterns are also reflected by concomitant changes in histone methylation.     

Viral genome methylation as an epigenetic defense against geminiviruses

Geminiviruses encapsidate single-stranded DNA genomes that replicate in plant cell nuclei through double-stranded DNA intermediates that associate with cellular histone proteins to form minichromosomes. Like most plant viruses, geminiviruses are targeted by RNA silencing and encode suppressor proteins such as AL2 and L2 to counter this defense. These related proteins can suppress silencing by multiple mechanisms, one of which involves interacting with and inhibiting adenosine kinase (ADK), a cellular enzyme associated with the methyl cycle that generates S-adenosyl-methionine, an essential methyltransferase cofactor. Thus, we hypothesized that the viral genome is targeted by small-RNA-directed methylation. Here, we show that Arabidopsis plants with mutations in genes encoding cytosine or histone H3 lysine 9 (H3K9) methyltransferases, RNA-directed methylation pathway components, or ADK are hypersensitive to geminivirus infection. We also demonstrate that viral DNA and associated histone H3 are methylated in infected plants and that cytosine methylation levels are significantly reduced in viral DNA isolated from methylation-deficient mutants. Finally, we demonstrate that Beet curly top virus L2- mutant DNA present in tissues that have recovered from infection is hypermethylated and that host recovery requires AGO4, a component of the RNA-directed methylation pathway. We propose that plants use chromatin methylation as a defense against DNA viruses, which geminiviruses counter by inhibiting global methylation. In addition, our results establish that geminiviruses can be useful models for genome methylation in plants and suggest that there are redundant pathways leading to cytosine methylation.     

RNA-directed DNA methylation mediated by DRD1 and Pol IVb: a versatile pathway for transcriptional gene silencing in plants

RNA-directed DNA methylation, which is one of several RNAi-mediated pathways in the nucleus, has been highly elaborated in the plant kingdom. RNA-directed DNA methylation requires for the most part conventional DNA methyltransferases, histone modifying enzymes and RNAi proteins; however, several novel, plant-specific proteins that are essential for this process have been identified recently. DRD1 (defective in RNA-directed DNA methylation) is a putative SWI2/SNF2-like chromatin remodelling protein; DRD2 and DRD3 (renamed NRPD2a and NRPD1b, respectively) are subunits of Pol IVb, a putative RNA polymerase found only in plants. Interestingly, DRD1 and Pol IVb appear to be required not only for RNA-directed de novo methylation, but also for full erasure of methylation when the RNA trigger is withdrawn. These proteins thus have the potential to facilitate dynamic regulation of DNA methylation. Prominent targets of RNA-directed DNA methylation in the Arabidopsis thaliana genome include retrotransposon long terminal repeats (LTRs), which have bidirectional promoter/enhancer activities, and other types of intergenic transposons and repeats. Intergenic solitary LTRs that are targeted for reversible methylation by the DRD1/Pol IVb pathway can potentially act as switches or rheostats for neighboring plant genes. The resulting alterations in gene expression patterns may promote physiological flexibility and adaptation to the environment.     

AtMBD6, a methyl CpG binding domain protein,maintains gene silencing in <Emphasis Type="Italic">Arabidopsis</Emphasis> by interacting with RNA binding proteins

DNA methylation, mediated by double-stranded RNA, is a conserved epigenetic phenomenon that protects a genome from transposons, silences unwanted genes and has a paramount function in plant or animal development. Methyl CpG binding domain proteins are members of a class of proteins that bind to methylated DNA. The Arabidopsis thaliana genome encodes 13 methyl CpG binding domain (MBD) proteins, but the molecular/biological functions of most of these proteins are still not clear. In the present study, we identified four proteins that interact with AtMBD6. Interestingly, three of them contain RNA binding domains and are co-localized with AtMBD6 in the nucleus. The interacting partners includes AtRPS2C (a 40S ribosomal protein), AtNTF2 (nuclear transport factor 2) and AtAGO4 (Argonoute 4). The fourth protein that physically interacts with AtMBD6 is a histone-modifying enzyme, histone deacetylase 6 (AtHDA6), which is a known component of the RNA-mediated gene silencing system. Analysis of genomic DNA methylation in the atmbd6, atrps2c and atntf2 mutants, using methylation-sensitive PCR detected decreased DNA methylation at miRNA/siRNA producing loci, pseudogenes and other targets of RNA-directed DNA methylation. Our results indicate that AtMBD6 is involved in RNA-mediated gene silencing and it binds to RNA binding proteins like AtRPS2C, AtAGO4 and AtNTF2. AtMBD6 also interacts with histone deacetylase AtHDA6 that might have a role in chromatin condensation at the targets of RdDM.     

DNA methylation in health, disease, and cancer

The spatial arrangement and three-dimensional structure of DNA in the nucleus is controlled through the interdigitation of DNA binding proteins such as histones and their modifiers, the Polycomb-Trithorax proteins, and the DNA methyltransferase enzymes. DNA methylation forms the foundation of chromatin and is crucial to epigenetic gene regulation in mammals. Disease pathogenesis mediated through infectious agents, inflammation, aging, or genetic damage often involves changes in gene expression. In particular, cellular transformation coincides with multiple changes in chromatin architecture, many of which appear to affect genome integrity and gene expression. Infectious agents, such as viruses directly affect genome structure and induce methylation of particular sequences to suppress host immune responses. Hyperproliferative tissues such as those in the gastrointestinal tract and colon have been shown to gradually acquire aberrant promoter hypermethylation. Here we review recent findings on altered DNA methylation in human disease, with particular focus on cancer and the increasingly large number of genes subject to tumor-specific promoter hypermethylation and the possible role of aberrant methylation in tumor development.     

Arabidopsis DEMETER-LIKE proteins DML2 and DML3 are required for appropriate distribution of DNA methylation marks

Cytosine DNA methylation is a stable epigenetic mark for maintenance of gene silencing across cellular divisions, but it is a reversible modification. Genetic and biochemical studies have revealed that the Arabidopsis DNA glycosylase domain-containing proteins ROS1 (REPRESSOR OF SILENCING 1) and DME (DEMETER) initiate erasure of 5-methylcytosine through a base excision repair process. The Arabidopsis genome encodes two paralogs of ROS1 and DME, referred to as DEMETER-LIKE proteins DML2 and DML3. We have found that DML2 and DML3 are 5-methylcytosine DNA glycosylases that are expressed in a wide range of plant organs. We analyzed the distribution of methylation marks at two methylated loci in wild-type and dml mutant plants. Mutations in DML2 and/or DML3 lead to hypermethylation of cytosine residues that are unmethylated or weakly methylated in wild-type plants. In contrast, sites that are heavily methylated in wild-type plants are hypomethylated in mutants. These results suggest that DML2 and DML3 are required not only for removing DNA methylation marks from improperly-methylated cytosines, but also for maintenance of high methylation levels in properly targeted sites.     

MSAP markers and global cytosine methylation in plants: a literature survey and comparative analysis for a wild‐growing species

Methylation of DNA cytosines affects whether transposons are silenced and genes are expressed, and is a major epigenetic mechanism whereby plants respond to environmental change. Analyses of methylation‐sensitive amplification polymorphism (MS‐AFLP or MSAP) have been often used to assess methyl‐cytosine changes in response to stress treatments and, more recently, in ecological studies of wild plant populations. MSAP technique does not require a sequenced reference genome and provides many anonymous loci randomly distributed over the genome for which the methylation status can be ascertained. Scoring of MSAP data, however, is not straightforward, and efforts are still required to standardize this step to make use of the potential to distinguish between methylation at different nucleotide contexts. Furthermore, it is not known how accurately MSAP infers genome‐wide cytosine methylation levels in plants. Here, we analyse the relationship between MSAP results and the percentage of global cytosine methylation in genomic DNA obtained by HPLC analysis. A screening of literature revealed that methylation of cytosines at cleavage sites assayed by MSAP was greater than genome‐wide estimates obtained by HPLC, and percentages of methylation at different nucleotide contexts varied within and across species. Concurrent HPLC and MSAP analyses of DNA from 200 individuals of the perennial herb Helleborus foetidus confirmed that methyl‐cytosine was more frequent in CCGG contexts than in the genome as a whole. In this species, global methylation was unrelated to methylation at the inner CG site. We suggest that global HPLC and context‐specific MSAP methylation estimates provide complementary information whose combination can improve our current understanding of methylation‐based epigenetic processes in nonmodel plants.     

On how mammalian transcription factors recognize methylated DNA

DNA methylation is an epigenetic mark that is essential for the development of mammals; it is frequently altered in diseases ranging from cancer to psychiatric disorders. The presence of DNA methylation attracts specialized methyl-DNA binding factors that can then recruit chromatin modifiers. These methyl-CpG binding proteins (MBPs) have key biological roles and can be classified into three structural families: methyl-CpG binding domain (MBD), zinc finger, and SET and RING finger-associated (SRA) domain. The structures of MBD and SRA proteins bound to methylated DNA have been previously determined and shown to exhibit two very different modes of methylated DNA recognition. The last piece of the puzzle has been recently revealed by the structural resolution of two different zinc finger proteins, Kaiso and ZFP57, in complex with methylated DNA. These structures show that the two methyl-CpG binding zinc finger proteins adopt differential methyl-CpG binding modes. Nonetheless, there are similarities with the MBD proteins suggesting some commonalities in methyl-CpG recognition across the various MBP domains. These fresh insights have consequences for the analysis of the many other zinc finger proteins present in the genome, and for the biology of methyl-CpG binding zinc finger proteins.     

A Dnmt2-like protein mediates DNA methylation in Drosophila

The methylation status of Drosophila DNA has been discussed controversially over a long time. Recent evidence has provided strong support for the existence of 5-methylcytosine in DNA preparations from embryonic stages of fly development. The Drosophila genome contains a single candidate DNA methyltransferase gene that has been termed Dnmt2. This gene belongs to a widely conserved family of putative DNA methyltransferases. However, no catalytic activity has been demonstrated for any Dnmt2-like protein yet. We have now established a protocol for the immunological detection of methylated cytosine in fly embryos. Confocal analysis of immunostained embryos provided direct evidence for the methylation of embryonic DNA. In order to analyse the function of Dnmt2 in DNA methylation, we depleted the protein by RNA interference. Depletion of Dnmt2 had no detectable effect on embryonic development and resulted in a complete loss of DNA methylation. Consistently, overexpression of Dnmt2 from an inducible transgene resulted in significant genomic hypermethylation at CpT and CpA dinucleotides. These results demonstrate that Dnmt2 is both necessary and sufficient for DNA methylation in Drosophila and suggest a novel CpT/A-specific DNA methyltransferase activity for Dnmt2 proteins.     

Involvement of putative SNF2 chromatin remodeling protein DRD1 in RNA-directed DNA methylation

In plants, the mechanism by which RNA can induce de novo cytosine methylation of homologous DNA is poorly understood. Cytosines in all sequence contexts become modified in response to RNA signals. Recent work has implicated the de novo DNA methyltransferases (DMTases), DRM1 and DRM2, in establishing RNA-directed methylation of the constitutive nopaline synthase promoter, as well as the DMTase MET1 and the putative histone deacetylase HDA6 in maintaining or enhancing CpG methylation induced by RNA. Despite the identification of enzymes that catalyze epigenetic modifications in response to RNA signals, it is unclear how RNA targets DNA for methylation. A screen for mutants defective in RNA-directed DNA methylation identified a novel putative chromatin-remodeling protein, DRD1. This protein belongs to a previously undefined, plant-specific subfamily of SWI2/SNF2-like proteins most similar to the RAD54/ATRX subfamily. In drd1 mutants, RNA-induced non-CpG methylation is almost eliminated at a target promoter, resulting in reactivation, whereas methylation of centromeric and rDNA repeats is unaffected. Thus, unlike the SNF2-like proteins DDM1/Lsh1 and ATRX, which regulate methylation of repetitive sequences, DRD1 is not a global regulator of cytosine methylation. DRD1 is the first SNF2-like protein implicated in an RNA-guided, epigenetic modification of the genome.