DNA methylation is the enzymatic addition of a methyl group to a base within DNA. It modifies the chemical properties of DNA without changing its underlying base sequence. In mammals, it predominantly produces 5-methylcytosine and contributes to the regulation of gene expression. Because methylation patterns can persist through cell division, they are an important component of epigenetics, although not every methylation mark is permanent or inherited between generations. Its effects depend on the modified base, genomic location, and biological context. (genome.gov)
Chemical basis and genomic distribution
DNA methylation is catalyzed by DNA methyltransferases, which transfer a methyl group from S-adenosylmethionine to a DNA base. In cytosine methylation, attachment at the fifth carbon produces 5-methylcytosine, commonly abbreviated 5mC. This is a covalent modification of the base rather than a replacement of one base by another. (nature.com)
In mammalian DNA, methylation occurs mainly at CpG sites: positions where cytosine is followed by guanine along the same strand. The “p” denotes the intervening phosphate. Non-CpG methylation also occurs; genome-wide mapping demonstrated substantial amounts in human embryonic stem cells and showed that these patterns change with differentiation. The complete distribution of methylated bases across a genome is called its methylome. (nature.com)
Methylation is unevenly distributed. Many repetitive sequences and gene bodies are methylated, whereas numerous CpG islands—regions with relatively high densities of CpG sites—remain unmethylated. Consequently, a methylation measurement has meaning only in relation to its location: methylation within a gene body is not equivalent to methylation at its transcriptional start region. (nature.com)
Establishment and maintenance
Two broadly distinguished processes generate methylation patterns. De novo methylation establishes marks on previously unmethylated DNA. In mammals, DNMT3A and DNMT3B are major enzymes responsible for this process. Experiments disrupting their genes in mice demonstrated essential roles in establishing methylation during development, with partly distinct genomic targets and developmental functions. (pubmed.ncbi.nlm.nih.gov)
Maintenance methylation preserves existing patterns after DNA replication. Replication initially creates DNA in which the parental strand may be methylated while the newly synthesized strand is not. DNMT1 helps restore methylation on the new strand. The protein UHRF1 participates in this machinery and connects DNA methylation maintenance with chromatin features, including methylation of histone H3. These interactions allow chemical information to accompany the copied DNA sequence. (pubmed.ncbi.nlm.nih.gov)
This copying mechanism helps specialized cells retain characteristic gene-regulatory states. It does not make their methylomes immutable: methylation can be established, maintained, or lost differently as cells change developmental state. (genome.gov)
Demethylation and chemical intermediates
Methylation can decline when maintenance does not fully replace marks during successive rounds of replication. It can also be reversed through enzymatic modification and repair. In mammals, TET enzymes oxidize 5mC, initially producing 5-hydroxymethylcytosine or 5hmC. TET1’s ability to catalyze this reaction was demonstrated in cultured cells and with purified biochemical systems. (pubmed.ncbi.nlm.nih.gov)
Further oxidation can generate 5-formylcytosine and 5-carboxylcytosine. Thymine DNA glycosylase can remove these modified bases, after which DNA repair machinery restores an unmodified cytosine. This provides a route for active demethylation without requiring the original methylated base to be diluted solely through replication. Importantly, 5hmC is also a measurable DNA modification in its own right; detecting it does not establish that demethylation has been completed. (pubmed.ncbi.nlm.nih.gov)
Gene regulation and development
DNA methylation influences transcription by changing interactions between DNA and regulatory proteins. It operates alongside histone modifications and other chromatin-associated mechanisms rather than acting as an isolated universal on–off switch. Human methylome maps show that its relationship with transcription differs across genomic regions and cell types. (genome.gov)
Methylation participates in genomic imprinting, in which the activity of certain genes depends on whether their copy was inherited from the mother or father. It also contributes to the regulation of the inactive X chromosome and the silencing of repetitive DNA. Experimental disruption of methyltransferases demonstrates that normal methylation patterns are important for mammalian development. (genome.gov)
Inheritance through ordinary cell division should be distinguished from inheritance across organismal generations. Much of the epigenome is reset during reproduction and development. Persistence of a mark in a cell lineage therefore does not, by itself, show that an acquired methylation pattern will be transmitted to descendants. (genome.gov)
Variation among organisms
In plants, cytosine methylation occurs in CG, CHG, and CHH sequence contexts, where H represents adenine, cytosine, or thymine. Different methyltransferase pathways contribute to these patterns. In Arabidopsis, small interfering RNAs participate in RNA-directed DNA methylation, directing silencing machinery toward particular genomic regions. Non-CG methylation is important in suppressing transposable elements. (nature.com)
In bacteria, methylation also has functions beyond the developmental regulation familiar from mammals. Methylated DNA can be distinguished from foreign DNA by restriction–modification systems, and methylation patterns can influence replication and gene expression. These roles illustrate why DNA methylation should not be defined exclusively as a mechanism of gene repression. (nig.ac.jp)
Measurement and research interpretation
Bisulfite sequencing combines chemical treatment with DNA sequencing to map cytosine modifications. Unmodified cytosine is converted to uracil and subsequently read as thymine, while 5mC remains read as cytosine. Conventional bisulfite sequencing, however, does not distinguish 5mC from 5hmC. Oxidative bisulfite sequencing addresses this limitation by selectively oxidizing 5hmC before treatment and comparing the resulting measurements. (nature.com)
Methylation studies require attention to cell identity and genomic context. Differences between samples may reflect different cell populations, developmental states, or regulatory processes. Altered methylation has also been documented in cancer, including distinct patterns at regulatory regions in breast tumors. Such associations identify biologically informative differences, but do not alone establish whether a methylation change initiated a disease process or followed another cellular alteration. (nature.com)